United States Air and Radiation EPA420-R-00-026
Environmental Protection December 2000
Agency
&EPA Regulatory Impact
Analysis:
Heavy-Duty Engine and
Vehicle Standards and
Highway Diesel Fuel
Sulfur Control
Requirements
> Printed on Recycled Paper
-------
THIS PAGE INTENTIONALLY BLANK
-------
Executive Summary
Executive Summary
This Regulatory Impact Analysis assesses the feasibility, costs, benefits, cost-
effectiveness, and other issues associated with the Environmental Protection Agency's finalized
program that sets new federal emission standards for heavy-duty vehicles and places limits on the
level of sulfur in diesel fuel. A complete discussion of the details of the program can be found in
the preamble to the regulations published in the Federal Register. The key results of this
Regulatory Impact Analysis are discussed below.
Health and Welfare Concerns
When revising emissions standards for heavy-duty vehicles, the Agency considers the
effects of air pollutants emitted from heavy-duty vehicles on public health and welfare. As
discussed in more detail below, the outdoor, or ambient, air quality in many areas of the country
is expected to violate federal health-based ambient air quality standards for ground level ozone
and particulate matter during the time when this rule would take effect. In addition, some studies
have found public health and welfare effects from ozone and PM at concentrations that do not
constitute a violation of their respective NAAQS. Other studies have associated diesel exhaust
with a variety of cancer and noncancer health effects. Of particular concern is human
epidemiological evidence linking diesel exhaust to an increased risk of lung cancer. Emissions
from heavy-duty vehicles also contribute to a variety of environmental and public welfare effects
such as impairment of visibility/ regional haze, acid deposition, eutrophication/nitrification, and
POM deposition. The standards finalized in this rule will result in a significant improvement in
ambient air quality and public health and welfare.
Feasibility of Emission Standards
During the past 15 years advancements have continued to be made in the development of
diesel exhaust emission control devices. Several emission control devices have emerged to
control harmful diesel particulate matter constituents, including the diesel oxidation catalyst and
the many forms of particulate filters or traps. Diesel oxidation catalysts have been shown to be
durable in-use, but they control only a small fraction of the total particulate matter and
consequently do not address our concerns sufficiently. The same is true of un-catalyzed diesel
particulate filters. Catalyzed diesel particulate filters have the potential to provide major
reductions in diesel particulate matter emissions and provide the durability and dependability
required for diesel applications. Precious metal catalyzed particulate filters, in conjunction with
low sulfur diesel fuel, have been shown to be more than 90 percent effective over the federal test
procedure and the not-to-exceed zone, a level of efficiency that demonstrates a capability of
meeting the applicable standards. Therefore, we believe the catalyzed diesel particulate filter
will be the control technology of choice for future control of diesel particulate matter emissions.
in
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
However, these devices cannot be brought to market on diesel applications unless low sulfur
diesel fuel is available.
Several exhaust emission control devices have also been developed to control diesel NOx
emissions. Today's lean NOx catalyst is capable at best of steady-state NOx reductions of less
than 10 percent, eliminating it from serious consideration as a tool for meeting the future
emission standards. Both selective catalytic reduction systems and NOx adsorbers have the
potential to provide significant emission reductions, although we believe that the NOx adsorber
is the most likely candidate to be used to meet future low diesel exhaust emission standards that
apply to the heavy-duty diesel market. However, the NOx adsorber technology cannot be
brought to market on diesel engines and vehicles unless low sulfur diesel fuel is available.
These developments make the widespread commercial use of diesel exhaust emission
controls feasible. Through the use of these devices, emissions control similar to that attained by
gasoline applications will be possible with diesel applications. However, without low sulfur
diesel fuel, these technologies cannot be implemented on heavy-duty diesel applications. Low
sulfur diesel fuel will at the same time allow these technologies to be implemented on light-duty
diesel applications.
Improvements also continue to be made to technologies for controlling emissions from
gasoline engines and vehicles. This includes improvements to catalyst designs in the form of
improved washcoats and improved precious metal dispersion. Significant effort has also been
put into improving cold start strategies that allow for more rapid light-off of the catalyst. These
strategies include retarding the spark timing to increase the temperature of the exhaust gases and
using air-gap manifolds, exhaust pipes, and catalytic converter shells to decrease heat loss from
the system. These improvements to gasoline emission controls will be made in response to recent
regulations from California and the EPA that established more stringent emission standards for
the light-duty sector. These improvements should transfer well to the heavy-duty gasoline
segment of the fleet. With the optimization of these and additional existing technologies for the
heavy-duty gasoline sector, we believe that significant reductions in emissions from heavy-duty
gasoline engines and vehicles can be realized, thus allowing vehicles to meet more stringent
emission standards. The sulfur content of the fuel is a critical ingredient for gasoline engines as
well. The Tier 2 gasoline sulfur reduction that requires sulfur levels to be reduced to a 30 parts
per million average with an 80 parts per million cap will enable the technology needed to meet
the heavy-duty standards in the same way that it enables compliance with the Tier 2 standards.
Fuel Standard Feasibility
In order to meet the 15 parts per million sulfur cap, refiners are likely to further hydrotreat
their highway diesel fuel in much the same way as it is being done today to meet the current
federal sulfur limits. Improvements to current hydrotreaters can be used to reduce diesel fuel
IV
-------
Executive Summary
sulfur beyond that being done to meet the current requirements. However, these improvements
alone do not appear to be sufficient to provide compliance with the 15 parts per million cap.
Based on past commercial experience, it is very possible to incorporate current distillate
hydrotreaters into designs which provide compliance with the proposed 15 parts per million cap.
Thus, the equipment added to meet the current requirements in the early 1990's will continue to
be very useful in meeting a more stringent standard.
The primary changes to refiners' current distillate hydrotreating systems are likely to be:
1) the use of a second reactor to increase residence time, possibly incorporating
counter-current flow characteristics, or the addition of a completely new second
stage hydrotreater;
2) the use of more active catalysts, including those specially designed to desulfurize
sterically hindered sulfur containing material;
3) greater hydrogen purity and less hydrogen sulfide in the recycle gas; and,
4) possible use of higher pressure in the reactor.
Existing commercial hydrotreaters are already producing distillate with average sulfur
levels below 10 parts per million, which should be more than sufficient to meet the new
requirements. Therefore, the 15 parts per million cap appears to be quite feasible given today's
distillate processing technology. Advances continue to be made in catalyst technology, with
greater amounts of sulfur being able to be removed at the same reactor size, temperature and
pressure. Therefore, it is reasonable to expect that distillate hydrotreaters put into service in the
2006 timeframe will utilize even more active catalysts than those available today.
Other existing methods may help to reduce diesel fuel sulfur levels, but will generally not
be sufficient to provide compliance with a 15 parts per million cap. However, we expect that a
number of refiners will utilize these techniques to reduce the severity of their distillate
hydrotreaters and reduce hydrogen consumption (particularly by avoiding aromatic saturation).
Some of these techniques would tend to increase the supply of highway diesel fuel while others
would tend to decrease it.
Biodesulfurization technology holds promise to reduce distillate sulfur without the high
temperatures and pressures involved in hydrotreating. Efforts are underway to demonstrate that
this technology can achieve 50 parts per million sulfur or less in the next few years. However, it
is not clear whether this technology would be sufficient to meet a 15 parts per million cap.
In addition, despite the heightened challenge to the distribution industry caused by our
sulfur program, it will be feasible to distribute 15 parts per million highway diesel fuel with
relatively minor modifications to existing systems to limit contamination from higher sulfur
products. These modifications can be accomplished at modest additional costs.
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Economic Impact: Diesel Engines
The technologies we expect to be used to meet the new requirements represent significant
technological advancements for controlling emissions, but also make clear that much effort
remains to develop and optimize these new technologies for maximum emission-control
effectiveness with minimum negative impacts on engine performance, durability, and fuel
consumption. On the other hand, it has become clear that manufacturers have a great potential to
advance beyond the current state of understanding by identifying aspects of the key technologies
that contribute most to hardware or operational costs or other drawbacks and pursuing
improvements, simplifications, or alternatives to limit those burdens. To reflect this investment
in long-term cost savings potential, the cost analysis includes an estimated $385 million in R&D
outlays for heavy-duty engine designs and $220 million in R&D for catalysts systems giving a
total R&D outlay for improved emission control of more than $600 million. The cost and
technical feasibility analyses accordingly reflect substantial improvements on the current state of
technology due to these future developments.
Estimated costs are broken into additional hardware costs and life-cycle operating costs.
The incremental hardware costs for new engines are comprised of variable costs (for hardware
and assembly time) and fixed costs (for R&D, retooling, and certification). Total operating costs
include the estimated incremental cost for low-sulfur diesel fuel, any expected increases in
maintenance cost or fuel consumption costs along with any decreases in operating cost expected
due to low-sulfur fuel. Cost estimates based on these projected technology packages represent an
expected incremental cost of engines in the 2007 model year. Costs in subsequent years will be
reduced by several factors, as described below. Separate projected costs were derived for engines
used in three service classes of heavy-duty diesel engines. All costs are presented in 1999
dollars.
The costs of these new technologies for meeting the 2007 model year standards are
itemized in the RIA and summarized in Table V.A-1. For light heavy-duty vehicles, the cost of
an engine is estimated to increase by $1,990 in the early years of the program reducing to $1,170
in later years and operating costs over a full life-cycle to increase by approximately $600. For
medium heavy-duty vehicles the cost of a new engine is estimated to increase by $2,560 initially
decreasing to $1,410 in later years with life-cycle operating costs increasing by approximately
$1,200. Similarly, for heavy heavy-duty engines, the vehicle cost in the first year is expected to
increase by $3,230 decreasing to $1,870 in later years. Estimated additional life-cycle operating
costs for heavy heavy-duty engines are approximately $4,600. The higher incremental increase
in operating costs for the heavy heavy-duty vehicles is due to the larger number of miles driven
over their lifetime (714,000 miles on average) and their correspondingly high lifetime fuel usage.
Emission reductions are also proportional to VMT and so are significantly higher for heavy
heavy-duty vehicles.
VI
-------
Executive Summary
We also believe there are factors that will cause cost impacts to decrease over time,
making it appropriate to distinguish between near-term and long term costs. Our analysis
incorporates the effects of this learning curve by projecting that the variable costs of producing
the low-emitting engines decrease by 20 percent starting with the third year of production (2009
model year) and by reducing variable costs again by 20 percent starting with the fifth year of
production. Additionally, since fixed costs are assumed to be recovered over a five-year period,
these costs are not included in the analysis after the first five model years. Finally, manufacturers
are expected to apply ongoing research to make emission controls more effective and to have
lower operating cost over time. However, because of the uncertainty involved in forecasting the
results of this research, we have conservatively not accounted for it in this analysis.
Table ES-1 lists the projected costs for each category of vehicle in the near- and long-
term. For the purposes of this analysis, "near-term" costs are those calculated for the 2007 model
year and "long term" costs are those calculated for 2012 and later model years.
vn
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table ES-1. Projected Incremental System Cost and Life Cycle Operating Cost
for Heavy-Duty Diesel Vehicles
(Net Present Values in the year of sale, 1999 dollars)
Vehicle Class
Light
heavy-duty
Medium
heavy-duty
Heavy
heavy-duty
Model Year
near term
long term
near term
long term
near term
long term
Hardware
Cost
$1,990
$1,170
$2,560
$1,410
$3,230
$1,870
Life-cycle
Operating
CostA
$627
$543
$1,165
$1,007
$4,626
$4,030
A Incremental life-cycle operating costs include the incremental costs to refine and
distribute low sulfur diesel fuel, the service cost of closed crankcase filtration
systems, the maintenance cost for PM filters and the lower maintenance costs
realized through the use of low sulfur diesel fuel (see discussion in Section V.C).
Economic Impact: Gasoline Vehicles
To perform a cost analysis for the final gasoline standards, we first determined a package
of likely technologies that manufacturers could use to meet the standards and then determined the
costs of those technologies. In making our estimates, we have relied on our own technology
assessment which included publicly available information such as that developed by California,
confidential information supplied by individual manufacturers, and the results of our own in-
house testing.
In general, we expect that heavy-duty gasoline vehicles would (like Tier 2 light duty
vehicles) be able to meet these standards through refinements of current emissions control
components and systems rather than through the widespread use of new technology. More
specifically, we anticipate a combination of technology upgrades such as the following:
Vlll
-------
Executive Summary
Improvements to the catalyst system design, structure, and formulation, plus an increase
in average catalyst size and loading.
Air and fuel system modifications including changes such as improved oxygen sensors,
and calibration changes including improved precision fuel control and individual cylinder
fuel control.
• Exhaust system modifications, possibly including air gapped components, insulation, leak
free exhaust systems, and thin wall exhaust pipes.
• Increased use of fully electronic exhaust gas recirculation (EGR).
Increased use of secondary air inj ection.
• Use of ignition spark retard on engine start-up to improve upon cold start emission
control.
• Use of low permeability materials and minor improvements to designs, such as the use of
low-loss connectors, in evaporative emission control systems.
We expect that the technologies needed to meet the heavy-duty gasoline standards will be
very similar to those required to meet the Tier 2 standards for vehicles over 8,500 pounds
GVWR. Few heavy-duty gasoline vehicles currently rely on technologies such as close coupled
catalysts and secondary air injection, but we expect they would to meet the new standards.
For each group we developed estimates of both variable costs (for hardware and
assembly time) and fixed costs (for R&D, retooling, and certification). Cost estimates based on
the current projected costs for our estimated technology packages represent an expected
incremental cost of vehicles in the near-term. For the longer term, we have identified factors that
would cause cost impacts to decrease over time. First, since fixed costs are assumed to be
recovered over a five-year period, these costs disappear from the analysis after the fifth model
year of production. Second, the analysis incorporates the expectation that manufacturers and
suppliers would apply ongoing research and manufacturing innovation to making emission
controls more effective and less costly over time. Our analysis incorporates the effects of this
"learning curve" by projecting that a portion of the variable costs of producing the new vehicles
decreases by 20 percent starting with the third year of production.
We have prepared our cost estimates for meeting the new heavy-duty gasoline standards
using a baseline of current technologies for heavy-duty gasoline vehicles and engines. Finally,
we have incorporated what we believe to be a conservatively high level of R&D spending at
$2,500,000 per engine family where no California counterpart exists. We have included this
large R&D effort because calibration and system optimization is likely to be a critical part of the
effort to meet the standards. However, we believe that the R&D costs may be generous because
the projection probably underestimates the carryover of knowledge from the development
required to meet the light-duty Tier 2 and CARB LEV-II standards.
IX
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table ES-2 provides our estimates of the per vehicle cost for heavy-duty gasoline vehicles
and engines. The near-term cost estimates are for the first years that vehicles meeting the
standards are sold, prior to cost reductions due to lower productions costs and the retirement of
fixed costs. The long-term projections take these cost reductions into account. In the absence of
changes to gasoline specifications and with no decrease in fuel economy expected, we do not
expect any increase in vehicle operating costs.
Table ES-2. Projected Incremental System Cost and Life Cycle Operating Cost
for Heavy-Duty Gasoline Vehicles
(Net Present Values in the year of sale, 1999 dollars)
Vehicle Class
Heavy-Duty
Gasoline
Model Year
near term
long term
Incremental
System
Cost
$198
$167
Life-cycle
Operating
Cost
$0
$0
Economic Impact: Fuel Sulfur Requirements
We estimate that the overall net cost associated with producing and distributing 15 ppm
diesel fuel, when those costs are allocated to all gallons of highway diesel fuel, will be
approximately 5.0 cents per gallon in the long term, or an annual cost of roughly $2.2 billion per
year once the program is fully effective starting in 2010. During the initial years under temporary
compliance option, the overall net cost is projected to be 4.5 cents per gallon, or an annual cost
of roughly $1.7 billion per year.
This cost consists of a number of components associated with refining and distributing
the new fuel. The majority of the cost is related to refining. From 2006-2010, refining costs are
estimated to be approximately 3.3 cents per gallon of highway diesel fuel, increasing to 4.3 cents
per gallon once the program is fully in place. In annual terms, the 2006-2010 refining costs are
expected to be about $1.4 billion per year, increasing to about $1.8 billion in 2010. These figures
include the cost of producing slightly more volume of diesel fuel because: 1) desulfurization
decreases the energy density of the fuel and 2) slightly more highway diesel fuel is expected to be
downgraded to nonroad diesel fuel in the distribution system.
-------
Executive Summary
A small cost of 0.2 cents per gallon is associated with an anticipated increase in the use of
additives to maintain fuel lubricity. Also, distribution costs are projected to increase by 1.0 cents
per gallon during the initial years under the temporary compliance option, including the cost of
distributing slightly greater volumes of fuel. Together, these two cost components only amount
to about $0.5 billion per year beginning in 2006. These costs drop to only about $0.3 billion in
2011 when the temporary compliance option and hardship provisions are over.
Operation with 15 parts per million sulfur diesel fuel is expected to reduce average
vehicle maintenance costs by approximately 1 cent on a per gallon basis. Beginning in 2011, this
reduction in maintenance costs will total roughly $400 million per year.
Economic Impact: Aggregate Costs
Using current data for the size and characteristics of the heavy-duty vehicle fleet and
making projections for the future, the diesel per-engine, gasoline per-vehicle, and per-gallon fuel
costs described above can be used to estimate the total cost to the nation for the emission
standards in any year. Figure ES-1 portrays the results of these projections.
5.0
4.0
3.0
in
O
= 2.0
m
1.0
0.0
Diesel
engine
Gasoline
vehicle
Diesel
fuel
Total
2005
2010
2015 2020 2025 2030
Figure ES-1. Total Annualized Costs
XI
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
As can be seen from the figure, the annual costs start out at less than 1.0 billion dollars in
year 2006 and increase during the initial years to about $3.6 billion in 2010. Thereafter, total
annualized costs are projected to continue increasing due to the effects of projected growth in
engine sales and fuel consumption.
Future consumption of 15 parts per million diesel fuel may be influenced by a potential
influx of diesel-powered cars and light trucks into the light-duty fleet. The possibility exists 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. A
sensitivity analysis of diesel penetration into the light-duty vehicle fleet results in the expectation
that the effect of increased penetration of desiels in the light-duty fleet will likely have little or no
impact on the aggregate costs estimated for the standards being finalized in today's action.
Cost-Effectiveness
We have calculated the cost-effectiveness of our diesel engine/gasoline vehicle/diesel
sulfur 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 a single vehicle meeting our
standards. This per-vehicle approach focuses on the cost-effectiveness of the program from the
point of view of the vehicles and engines which will be used to meet the new requirements.
However, the per-vehicle approach does not capture all of the costs or emission reductions from
our diesel engine/gasoline vehicle/diesel sulfur program since it does not account for the use of
15 parts per million diesel fuel in current diesel engines. Therefore, we have also calculated a
30-year net present value cost-effectiveness using the net present value of costs and emission
reductions for all in-use vehicles over a 30-year time frame. The baseline or point of comparison
for this evaluation is the previous set of engine, vehicle, and diesel sulfur standards (in other
words, the applicable 2006 model year standards).
The cost of complying with the new standards will decline over time as manufacturing
costs are reduced and amortized capital investments are recovered. To show the effect of
declining cost in the per-vehicle cost-effectiveness analysis, we have developed both near term
and long term cost-effectiveness values. More specifically, these correspond to vehicles sold in
years one and six of the vehicle and fuel programs.
The 30-year net present value approach to calculating the cost-effectiveness of our
program involves the net present value of all nationwide emission reductions and costs for a 30
year period beginning with the start of the diesel fuel sulfur program and introduction of model
year 2007 vehicles and engines in year 2006. This 30-year timeframe captures both the early
period of the program when very few vehicles that meet our standards will be in the fleet, and the
later period when essentially all vehicles in the fleet will meet the new standards. We have
xn
-------
Executive Summary
calculated the 30-year net present value cost-effectiveness using the net present value of the
nationwide emission reductions and costs for each calender year.
Our per-vehicle and 30-year net present value cost-effectiveness values are given in the
following tables. The tables summarize the net present value lifetime costs, NMHC + NOx and
PM emission reductions, and resulting cost-effectiveness results for our diesel engine/gasoline
vehicle/diesel sulfur standards using sales weighted averages of the costs (both near term and
long term) and emission reductions of the various vehicle and engine classes affected for the two
different approaches. Diesel fuel costs applicable to diesel engines have been divided equally
between the adsorber and trap, since 15 parts per million diesel fuel is intended to enable all
technologies to meet our standards. In addition, since the trap produces reductions in both PM
and hydrocarbons, we have divided the total trap costs equally between compliance with the PM
standard and compliance with the NMHC standard.
The tables also display cost-effectiveness values based on two approaches to account for
the reductions in SO2 emissions associated with the reduction in diesel fuel sulfur. While these
reductions are not central to the program and are therefore not displayed with their own cost-
effectiveness, they do represent real emission reductions due to our program. The first set of
cost-effectiveness numbers in the tables simply ignores these reductions and bases the cost-
effectiveness on only the NOx, NMHC, and PM emission reductions from our program. The
second set accounts for these ancillary reductions by crediting some of the cost of the program to
SO2. The amount of cost allocated to SO2 is based on the cost-effectiveness of SO2 emission
reductions that could be obtained from alternative, potential future EPA programs. The SO2
credit was applied only to the PM calculation, since SO2 reductions are primarily a means to
reduce ambient PM concentrations.
Xlll
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table ES-3. Per-EngineA Cost Effectiveness of the Standards for 2007 and Later MY
Vehicles
Pollutants
Near-term costs
NOx + NMHC
PM
Long-term costs
NOx + NMHC
PM
Discounted
lifetime
vehicle &
fuel costs
$1937
$1055
$1346
$755
Discounted
lifetime emission
reductions (tons)
0.8421
0.0672
0.8421
0.0672
Discounted
lifetime cost
effectiveness
per ton
$2,300
$15,697
$1,599
$11,243
Discounted lifetime
cost effectiveness
per ton with SO2
creditB
$2,300
$9,058
$1,599
$4,604
A As described above, per-engine cost effectiveness does not include any costs or benefits from the existing, pre-
control, fleet of vehicles that would use the 15 parts per million diesel fuel.
$446 credited to SO2 (at $4800/ton) for PM cost effectiveness
Table ES-4. 30-year Net Present ValueA Cost Effectiveness of the Standards
NOx + NMHC
PM
30-year n.p.v.
engine,
vehicle, & fuel
costs
$34.7 billion
$10.2 billion
30-year
n.p.v.
reduction
(tons)
16.2 million
0.8 million
30-year n.p.v.
cost
effectiveness
per ton
$2,137
$13,598
30-year n.p.v.
cost effectiveness
per ton with SO2
creditB
$2,137
$4,383
A This cost effectiveness methodology reflects the total fuel costs incurred in the early years of the program
when the fleet is transitioning from pre-control to post-control diesel vehicles. In 2007 <10% of highway diesel
fuel is anticipated to be consumed by 2007 MY vehicles. By 2012 this increases to >50% for 2007 and later MY
vehicles.
B $6.9 billion credited to SO2 (at $4800/ton).
Cost-Benefit Analysis
We also made an assessment of the monetary value of the health and general welfare
benefits that are expected from the HD Engine/Diesel Fuel rule in 2030. We estimate that this
xiv
-------
Executive Summary
rule would, in the long term, result in substantial benefits, such as the yearly avoidance of:
approximately 8,300 premature deaths, approximately 5,500 cases of chronic bronchitis, roughly
361,400 asthma attacks, and significant numbers of hospital visits, lost work days, and multiple
respiratory ailments (including those that affect children). Our standards will also produce
welfare benefits related to the reduction of agricultural crop damage, impacts on forest
productivity, visibility, and nitrogen deposition in rivers and lakes.
Total monetized benefits of the HD Engine/Diesel Fuel rule in 2030 are expected to be
approximately $70.4 billion. Total monetized benefits, however, are driven primarily by the
value placed on the reductions in premature deaths. In the primary estimate, these represent close
to 89 percent of total monetized benefits. 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. Since the publication of the Tier 2/Gasoline Sulfur standards earlier this year, EPA has
obtained additional advice from its Science Advisory Board (SAB) on the proper characterization
of this value and alternatives to EPA's primary estimate of mortality benefits. Following the
advice of the SAB, EPA currently uses the VSL approach in calculating the primary estimate of
mortality benefits, because the method reflects the direct application of what EPA and the SAB
consider to be the most reasonable estimates for valuation of premature mortality available in the
current economics literature.
However, the economics literature concerning the appropriate method for valuing
reductions in premature mortality risk is still developing. There is general agreement that the
value to an individual of a reduction in mortality risk tends to 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 individual's attitudes towards risk, and the health status of the individual. While the
limited empirical basis for adjusting the VSL used by EPA for many of these factors does not
meet the SAB's standards of reliability at this time, a thorough discussion of these factors is
contained in the benefits TSD for this RIA (Abt Associates, 2000). Age in particular may be an
important difference between populations affected by air pollution mortality risks and
populations affected by workplace risks. Premature mortality risks from air pollution tend to
affect the very old more than the working age population. As such, any adjustments to VSL for
age differences may have a large impact on total benefits. EPA recognizes the need for further
research to improve estimates of the value of premature mortality risk reduction, including
potential adjustments to VSL for age and other factors mentioned above.
Based on recent advice from the SAB, our benefits estimates account for expected growth
in real income. Economic theory argues that a person's willingness to pay for most goods (such
as environmental protection) will increase as real incomes increase. There is substantial
empirical evidence in the economics literature for this idea, although there is uncertainty about its
exact value. Based on a review of the available literature, we adjust the valuation of human
xv
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
health benefits and visibility improvements upward to account for projected growth in real U.S.
income to 2030.
As summarized in Table ES-5, our primary estimate of monetary benefits realized in
2030 will be approximately $70.4 billion dollars ($1999), including an adjustment for growth in
real income as described above. Comparing this estimate of the economic benefits with the
adjusted cost estimate indicates that in 2030 the net economic benefits of the HD Engine/Diesel
Fuel rule to society are approximately $66.2 billion dollars ($1999). Due to the uncertainties
associated with this estimate of net benefits, it should be considered along with other components
of this RIA, such as: reductions in adverse health and environmental outcomes, total cost, cost-
effectiveness, and other benefits and costs that could not be monetized.
Table ES-5. 2030 Annual Monetized Costs, Benefits,
and Net Benefits for the Final HD Engine/Diesel Fuel RuleA
Annual compliance costs
Monetized PM-related benefits8
Monetized Ozone-related benefits8'0
NMHC-related benefits
CO-related benefits
Total annual benefits
Monetized net benefits13
Billions of 1999$
$4.2
$69.0 + BpM
$1.4 + B0zone
not monetized (BNMHC)
not monetized (Bco)
$70.4 +BPM + B0zone + BNMHC + Bco
$66.2 + B
A For this section, all costs and benefits are rounded to the nearest 100 million. Thus, figures presented in this chapter may not exactly equal
benefit and cost numbers presented in earlier sections of the chapter.
B Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table VII-1. Unmonetized PM- and ozone-related benefits are indicated by BPM. And B0zone, respectively.
c Ozone-related benefits are only calculated for the Eastern U.S. due to unavailability of reliable modeled ozone concentrations in the Western
U.S. This results in an underestimate of national ozone-related benefits. See US EPA (2000a) for a detailed discussion of the UAM-V ozone
model and model performance issues.
D B is equal to the sum of all unmonetized benefits, including those associated with PM, ozone, CO, and NMHC.
Table ES-6 shows the impact of alternative assumptions about key inputs to the benefits
analysis, including the concentration-response function relating particulate matter and premature
mortality and the dollar value of reductions in the risk of premature mortality. These calculations
are based on specific, plausible alternatives to the inputs used in deriving our primary estimate in
Table ES-5. See Chapter Vn of the RIA for a complete discussion of these and other important
alternative calculations and their associated uncertainties.
xvi
-------
Executive Summary
Table ES-6. Key Alternative Benefits Calculations
for the HD Engine/Diesel Fuel Rule in 2030A
Description of Alternative
Avoided Incidences
Impact on Primary Benefits
Estimate Adjusted for
Growth in Real Income
(billion 1999$)
Alternative Concentration-Response Functions for PM-related Premature Mortality
1
2
3
Krewski/ACS Study Regional Adjustment
Model8
Pope/ACS Study0
Krewski/Harvard Six-city StudyD
9,400
9,900
24,200
+$7.4 (+11%)
+12.8 (+18%)
+$118.5 (+169%)
Alternative Methods for Valuing Reductions in Incidences of PM-related Premature Mortality
Value
mortal]
specifi
of avoided premature
ty incidences based on age-
c VSL.E
Jones-Lee
(1989)
Jones-Lee
(1993)
8,300
8,300
-$28.5 (-41%)
-$6.8 (-10%)
A Please refer to Section 7.F of the RIA for complete information about the estimates in this table.
B This C-R function is included as a reasonable specification to explore the impact of adjustments for broad regional correlations, which have
been identified as important factors in correctly specifying the PM mortality C-R function..
c The Pope et al. C-R function was used to estimate reductions in premature mortality for the Tier 2/Gasoline Sulfur benefits analysis. It is
included here to provide a comparable estimate for the HD Engine/Diesel Fuel rule.
D The Krewski et al. "Harvard Six-cities Study" estimate is included because the Harvard Six-cities Study featured 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. and as such
provides a reasonable alternative to the primary estimate.
E Jones-Lee (1989) provides an estimate of age-adjusted VSL based on a finding that older people place a much lower value on mortality risk
reductions than middle-age people. Jones-Lee (1993) provides an estimate of age-adjusted VSL based on a finding that older people value
mortality risk reductions only somewhat less than middle-aged people.
Regulatory Flexibility Act
Our Regulatory Flexibility Analysis evaluates the impacts of the heavy-duty engine
standards and diesel fuel sulfur standards on small businesses. Prior to issuing our proposal we
analyzed the potential impacts of our program on small businesses. We convened a Small
Business Advocacy Review Panel, as required under the Regulatory Flexibility Act (RFA) as
amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA). The
small business provisions of today's action reflect revisions to the proposed program based upon
updated analysis as well as comments heard at the public hearings on the rulemaking and those
submitted in writing during the public comment period. The RFA requires us to determine, to
the extent feasible, our rule's economic impact on small entities, explore regulatory options for
xvn
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
reducing any significant economic impact on a substantial number of such entities, and explain
our ultimate choice of regulatory approach.
In developing this rule, we concluded that the heavy-duty engine and diesel fuel sulfur
standards would likely have a significant impact on a substantial number of small entities. We
identified several categories of small entities associated with diesel fuel production or
distribution. To our knowledge, no manufacturers of heavy-duty engines meet the Small
Business Administration definition of a small business. We have determined that the only small
entities that may be significantly affected by today's rule are small refiners, since they will have
to invest in desulfurization technology to produce low sulfur highway diesel fuel. We quantified
the economic impacts on the identified small entities. We determined the refinery costs for
average size refineries and small refiners to produce low sulfur diesel fuel. We also estimated
diesel distribution costs for the entire distribution system, including pipeline and tank wagon
deliveries.
For today's action, we have structured a selection of temporary flexibilities for qualifying
small refiners, both domestic and foreign. Generally, we structured these provisions to address
small refiner hardship while achieving air quality benefits expeditiously and ensuring that the
reductions needed in diesel sulfur coincide with the introduction of 2007 model year diesel
vehicles.
All refiners producing highway diesel fuel are able to take a advantage of the temporary
compliance option offered in the final regulations. Diesel producers that also market gasoline in
the GPA may receive additional flexibility under today's rule. Refiners that seek and are granted
small refiner status may choose from the following three options under the diesel sulfur program.
These three options have evolved from concepts on which we requested and received comment
in the proposal.
500 ppm Option. A small refiner may continue to produce and sell diesel fuel meeting the
current 500 ppm sulfur standard for four additional years, until June 1, 2010, provided
that it reasonably ensures the existence of sufficient volumes of 15 ppm fuel in the
marketing area(s) that it serves.
Small Refiner Credit Option. A small refiner that chooses to produce 15 ppm fuel prior
to June 1, 2010 may generate and sell credits under the broader temporary compliance
option. Since a small refiner has no requirement to produce 15 ppm fuel under this
option, any fuel it produced at or below 15 ppm sulfur will qualify for generating credits.
Diesel/Gasoline Compliance Date Option. For small refiners that are also subject to the
Tier 2/Gasoline sulfur program (40 CFR Part 80), the refiner may choose to extend by
xvin
-------
Executive Summary
three years the duration of its applicable interim gasoline standards, provided that it also
produces all its highway diesel fuel at 15 ppm sulfur beginning June 1, 2006.
xix
-------
Table of Contents
Table of Contents
Executive Summary iii
Table of Contents T-l
List of Tables T-ll
List of Figures T-19
List of Acronyms T-21
Chapter I: Introduction 1-1
Chapter II: Health and Welfare Concerns and Emissions Benefits D-l
A. Health and Welfare Concerns D-l
1. Health and Welfare Concerns Raised During Public Hearings n-1
2. Ozone H-3
a. Health and Welfare Effects of Ozone and its Precursors n-4
b. Photochemical Ozone Modeling D-7
c. Results of Photochemical Ozone Modeling n-11
d. Ozone Modeling and Analysis in 1-Hour State Implementation
Plan Submittals and Other Local Ozone Modeling n-15
e. Public Health and Welfare Concerns from Prolonged and Repeated
Exposures to Ozone n-40
3. Particulate Matter H-55
a. Health and Welfare Effects of Ambient Particulate Matter . . . 11-55
b. Public Health and Welfare Concerns from Exposure to
Fine PM H-64
4. Diesel Exhaust H-74
a. Cancer and Noncancer Effects of Diesel Exhaust n-74
b. The Link Between Diesel Exhaust and Diesel
Particulate Matter H-77
c. Ambient Concentrations and Exposure to Diesel Exhaust . . . n-79
d. Potential for Cancer Risk H-90
5. Gaseous Air Toxics n-95
a. Health Effects H-96
b. Assessment of Exposure n-101
6. Visibility/Regional Haze n-105
7. Acid Deposition n-107
T-l
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
8. Eutrophication and Nitrification n-108
9. POM Deposition II-l 10
10. Carbon Monoxide II-l 10
B. Heavy-Duty Diesel Inventory Impacts D-l 11
1. Description of Calculation Method n-111
a. Baseline Emissions Inventory II-l 11
b. Controlled Emissions Inventory (Air Quality Analysis Case) n-115
c. Controlled Emissions Inventory (Updated Control Case) . . . 11-117
2. HDDE Emission Reductions II-l 19
a. Anticipated Reductions due to the New HDDE Standards . . II-l 19
b. Additional Reductions due to the New HDDE Standards . . . n-127
3. HDGV Emission Reductions n-129
a. NOx Reductions H-130
b. Exhaust NMHC Reductions n-131
c. Evaporative Emission Reductions n-133
d. Air Toxics Reductions n-135
4. Total Emission Reductions n-135
5. Differences from NPRM Inventory n-139
a. Heavy-Duty Diesel Engines n-140
b. Heavy-Duty Gasoline Vehicles D-141
6. Sensitivity Analysis for In-Use PM Deterioration n-142
a. Methodology n-142
b. Results H-143
7. Contribution of HDVs to National Inventory n-144
Chapter n. References D-153
Chapter HI: Emissions Standards Feasibility ffi-1
A. Feasibility of the Heavy-Duty Diesel Standards ffi-2
1. Engine Out Improvements ffi-2
2. Meeting the PM Standard ffi-4
a. Catalyzed Diesel Particulate Filters ffi-5
b. Control of Ultra-Fine PM ffi-14
3. Meeting the NOx Standard ffi-15
a. Lean NOx Catalysts ffi-15
b. NOx Adsorbers ffi-16
c. Selective Catalytic Reduction (SCR) ffi-72
d. Non-Thermal Plasma Assisted Catalysts ffi-74
4. Meeting the NMHC Standard ffi-74
5. Meeting the Crankcase Emissions Requirements ffi-76
6. The Complete System ffi-77
T-2
-------
Table of Contents
7. The Need for Low Sulfur Diesel Fuel IH-78
a. Catalyzed Diesel Particulate Filters and the Need for Low Sulfur
Fuel IH-79
b. Diesel NOx Catalysts and the Need for Low Sulfur Fuel . . . IH-87
c. Contribution of Sulfur from Engine Lubricating Oils 111-91
B. Feasibility of the 2008 Standards for Heavy-Duty Gasoline
Vehicles & Engines 111-93
1. Gasoline Exhaust Emission Control Technology Descriptions 111-95
a. Base Engine Improvements 111-96
b. Improvements in Air-Fuel Ratio Control 111-99
c. Improvements in Fuel Atomization ID-102
d. Improvements to Exhaust and Exhaust Emission
Control Systems ffl-103
e. Improvements in Engine Calibration Techniques ID-106
2. The 2008 Heavy-Duty Gasoline Exhaust Emission Standards .... III-107
3. Current Exhaust Emission Certification Levels for Heavy-Duty Gasoline
Vehicles & Engines ffl-109
4. Technological Feasibility of the 2008 Heavy-Duty Gasoline Exhaust
Emission Standards ffi-112
5. The 2008 Heavy-Duty Gasoline Evaporative Emission Standards . ffi-119
6. Technological Feasibility of the 2008 Heavy-Duty Gasoline Evaporative
Emission Standards ID-121
Chapter HI. References IH-124
Chapter IV: Fuel Standard Feasibility IV-1
A. Feasibility of Removing Sulfur from Highway Diesel Fuel IV-1
1. Sources of Diesel Fuel Sulfur IV-1
2. Current Levels of Sulfur in Highway Diesel Fuel IV-6
3. Current Levels of Other Fuel Parameters in Highway Diesel Fuel . . . IV-8
4. Overview of Diesel Fuel Sulfur Control IV-12
5. Hydrotreating and Other Hydrogen-Based Processes Which Remove
Sulfur IV-14
a. Fundamentals of Distillate Hydrotreating IV-15
b. Meeting a 15 ppm Cap with Distillate Hydrotreating IV-20
c. Low Sulfur Performance of Distillate Hydrotreating IV-26
d. Undercutting Cracked Stocks IV-29
6. Other Desulfurization Technologies IV-30
a. Biodesulfurization IV-30
b. Chemical Oxidation and Extraction IV-31
c. Sulfur Adsorption IV-31
T-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
d. FCC Feed Hydrotreating IV-32
7. Will There Be Enough Supply of Highway Diesel Fuel? IV-33
a. Required Investment per Refinery IV-33
b. Historic Refining Profit Margins IV-34
c. Variation in Compliance Costs Faced by Refiners IV-35
d. Other Markets for Highway Diesel Fuel IV-38
e. Uncertainly in Requisite Desulfurization Technology IV-39
f. Likely Price and Import Response to the New Standard .... IV-40
g. Impact of Desulfurization Processes on Fuel Volume IV-41
h. Impact of Fuel Transport on Supply IV-42
i. Charles River Associates and Baker and O'Brien Study .... IV-42
8. Conclusions IV-44
9. Fuel Availability in 2006 IV-45
a. Summary IV-45
b. Diesel Fuel Refining Under the Temporary
Compliance Option IV-46
B. Interaction with Other Programs IV-56
1. Design and Construction Services IV-56
C. The Need for Lubricity Additives IV-63
1. What Impacts Will the Sulfur Change Have on Lubricity? IV-64
2. How Can One Determine Whether the Lubricity of a Fuel Is
Adequate? IV-66
3. What Experience Has There Been with Low-sulfur Fuels? IV-68
4. What Can Be Done About Poor Lubricity Fuels? IV-72
5. Today's Action on Lubricity: A Voluntary Approach IV-75
6. Are There Concerns Regarding the Impact of Diesel Desulfurization on
Other Fuel Properties? IV-76
D. Feasibility of Maintaining Off highway Fuel in the Distribution System ... IV-77
1. Overview IV-77
2. Feasibility of Limiting Sulfur Contamination in the Pipeline System IV-89
a. Interface Handling Practices IV-90
b. Identifying the Location of the Interface Between Fuel
Batches IV-95
c. Dead Legs IV-96
d. Line-Fill IV-97
e. Leaking Valves IV-97
f. Surface Accumulation of Sulfur-Containing Substances .... IV-98
3. Limiting Sulfur Contamination at Stationary Storage Facilities .... IV-99
a. Quality Control Testing IV-99
b. Product Switching in Stationary Storage Tanks IV-100
c. Tank Manifolds IV-102
T-4
-------
Table of Contents
4. Limiting Sulfur Contamination During Transport by Surface
Vehicles IV-102
5. Limiting Sulfur Contamination During Transport by Marine VesselsIV-105
6. Limiting Sulfur Contamination from Diesel Fuel Additives IV-106
7. Handling Batches of Highway Diesel Found to Exceed the Sulfur Standard
Downstream of the Refinery IV-108
E. Misfueling IV-109
1. Introduction IV-109
2. What Provision Are We Adopting to Ensure 2007 and Later Heavy-Duty
Diesel Vehicles Use 15 ppm Sulfur Fuel? IV-110
3. Are Additional Requirements Necessary to Address Deliberate
Misfueling? IV-111
4. Are Additional Requirements Necessary to Address Accidental
Misfueling? IV-114
Chapter V: Economic Impact V-l
A. Economic Impact of the 2007 Model Year Heavy-Duty Diesel Standards ... V-l
1. Methodology for Estimating Costs V-l
2. Heavy-Duty Diesel Technologies for Compliance with the Standards V-3
3. Technology/Hardware Costs for Diesel Vehicles and Engines V-8
a. NOx Adsorber Catalyst Costs V-8
b. Catalyzed Diesel Paniculate Filter Costs V-l2
c. Diesel Oxidation Catalyst (HC and H2S "Clean-Up" Catalyst) V-15
d. Closed Crankcase Filtration Systems V-l6
4. Fixed Costs V-17
a. Research and Development V-17
b. Tooling Costs V-21
c. Certification Costs V-24
d. Summary of Fixed Costs V-26
5. Operating Costs V-27
a. Low Sulfur Diesel Fuel V-28
b. Maintenance Costs for Closed Crankcase Ventilation Systems V-28
c. Maintenance Costs for Catalyzed Diesel Paniculate Filters . V-29
d. Maintenance Savings due to Low Sulfur Diesel Fuel V-30
e. Fuel Economy Impacts V-30
6. Summary of Near and Long Term Costs V-32
7. Total Incremental Nationwide Costs for 2007 Heavy-Duty
Diesel Engines V-37
B. Economic Impact of the 2008 Model Year Heavy-Duty Gasoline Standards V-41
1. Methodology for Estimating Heavy-Duty Gasoline Costs V-41
T-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
2. Technology Packages for Compliance with the 2008 Model Year Heavy-
Duty Gasoline Standards V-42
3. Technology/Hardware Costs for Gasoline Vehicles and Engines ... V-44
a. Improved Catalysts and Catalyst Systems V-44
b. Oxygen Sensors V-50
c. Exhaust Gas Recirculation (EGR) V-50
d. Secondary Air Injection with Closed Loop Control V-50
e. Exhaust Systems V-50
f. Evaporative Emission Control Systems V-51
g. Summary of Technology/Hardware Costs V-51
4. Heavy-Duty Gasoline Fixed Costs V-53
a. R&D and Tooling Costs V-54
b. Certification Costs V-55
5. Summary of Heavy-Duty Gasoline Costs V-55
6. Total Nationwide Costs for 2008 Heavy-Duty Gasoline Vehicles . . V-56
C. Diesel Fuel Costs V-60
1. Methodology V-60
a. Overview V-60
b. Derivation of the Fraction of LCO and other Cracked Blendstocks
in Highway Diesel Fuel for Each Refinery V-62
c. Technology and Cost Inputs from Vendors V-66
d. Development of Diesel Desulfurizati on Cost Projections ... V-74
e. Development of Desulfurizati on Cost Factors for Individual Diesel
Blendstocks
V-80
f. Future Diesel Fuel Volumes V-96
2. Projected Refinery Costs of Meeting the 15 ppm Sulfur Cap V-97
a. Other Cost Estimates for Desulfurizing Highway Diesel Fuel V-113
3. The Added Cost of Distributing Low-Sulfur Fuel V-l 19
a. Summary V-l 19
b. Cost of Distributing the Additional Volume of Highway Diesel
Fuel Needed to Compensate for a Reduction in
Energy Density V-123
c. Cost of Downgrading an Increased Volume of Highway Diesel
Fuel to a Lower Value Product During Shipment by PipelineV-124
d. Increased Cost of Downgrading the Current Interface Volumes
Associated with Pipeline Shipments of Highway Diesel Fuel V-l26
e. Increased Cost of Downgrading the Interface Between Pipeline
Shipments of Highway Diesel Fuel and Jet Fuel or Kerosene V-l 27
f. Cost of Additional Quality Control Testing at
Petroleum Terminals V-129
T-6
-------
Table of Contents
g. Cost of Downgrading the Additional Pipeline Interface Volumes
Associated with the Shipment of Highway Diesel Fuel that Meets a
500 ppm Sulfur Cap During the Initial Years of Our Sulfur
Program V-129
h. Cost of Optimizing the Distribution System to Distribute 15 ppm
Highway Diesel Fuel V-130
i. Additional Measures by Tank Truck, Tank Wagon, and Rail Car
Operators to Limit Contamination V-131
j. Potential Costs Associated with the Voluntary Phase Out of High
Sulfur Diesel Additives V-131
k. Costs During the Initial Years of Our Program Due to the Need for
Additional Storage Tanks to Handle Two Grades of Highway
Diesel Fuel V-131
4. What is the Cost of Lubricity Additives? V-133
5. Benefits of 15 ppm Diesel Fuel for the New and Existing
Diesel Fleet V-135
a. Methodology V-136
b. Extended Oil Change Intervals V-137
c. Extended EGR System Life V-138
d. Extended Exhaust System Life V-141
e. Extended Rebuild Intervals and Engine Life V-142
f. Aggregate Cost Savings for the New and Existing Diesel Fleet
Realized from Low Sulfur Diesel Fuel V-142
6. Per-Engine Life-Cycle Fuel Costs V-145
D. Combined Total Annual Nationwide Costs V-147
Chapter V. References V-151
Chapter VI: Cost-Effectiveness VI-1
A. Overview of the Per-vehicle Analysis VI-1
1. Temporal and Geographic Applicability VI-2
2. Baselines VI-3
B. Diesel Costs VI-4
1. Near and Long-Term Cost Accounting VI-4
2. Diesel Engine and Fuel Costs VI-5
3. Methodology for assigning costs to NOx, NMHC, and PM VI-5
4. Cost Crediting for SO2 VI-7
C. Emission Reductions from Diesel Engines VI-8
1. NOx, NMHC, and PM VI-8
2. Sulfur Dioxide VI-11
D. Costs and Emission Reductions for Heavy-duty Gasoline Vehicles VI-12
T-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1. Gasoline Vehicle Costs VI-12
2. Emission Reductions from Gasoline Vehicles VI-13
E. 30-year Net Present Value Cost-Effectiveness VI-15
F. Results VI-16
APPENDIX VI - A: Factors Used in Diesel Engine Calculations for Cost-effectiveness . . . VI-22
APPENDIX VI - B: Costs used in 30-year Net Present Value Cost Effectiveness Analysis
(Smillions) VI-23
APPENDIX VI - C: Emission Reductions Used in 30-year Net Present Value Cost Effectiveness
Analysis (thousand tons) VI-24
Chapter VI. References VI-25
Chapter VH: Benefit-Cost Analysis VH-1
A. Emissions Inventory Implications VH-10
B. Air Quality Impacts VH-11
1. Ozone Air Quality Estimates VH-11
a. Modeling Domain VH-12
b. Simulation Periods VH-13
c. Converting UAM-V Outputs to Full-Season Profiles for Benefits
Analysis VII-14
d. Ozone Air Quality Results VII-14
2. PM Air Quality Estimates VH-17
a. Modeling Domain VH-18
b. Simulation Periods VH-18
c. Model Inputs VH-18
d. Converting REMSAD Outputs to Benefits Inputs and Model
Performance VH-21
e. PM Air Quality Results VH-22
3. Visibility Degradation Estimates VH-25
a. Residential Visibility Improvements VH-26
b. Recreational Visibility Improvements VH-27
4. Nitrogen Deposition Estimates VH-29
C. Benefit Analysis VH-30
1. Methods for Estimating Benefits from Air Quality Improvements . VII-30
2. Methods for Describing Uncertainty VH-35
D. Assessment of Human Health Benefits VH-39
1. Estimating Baseline Incidences for Health Effects VH-39
2. Accounting for Potential Health Effect Thresholds VII-40
T-8
-------
Table of Contents
3. Quantifying and Valuing Individual Health Endpoints VII-40
a. Premature Mortality: Quantification VII-45
b. Premature Mortality: Valuation VII-49
c. Chronic Bronchitis: Quantification VII-53
d. Chronic Bronchitis: Valuation VII-54
e. Hospital and Emergency Room Admissions: Quantification VII-55
f. Hospital Admissions: Valuation VII-56
g. Asthma Attacks: Quantification VII-57
h. Asthma Attacks: Valuation VH-58
i. Other Health Effects: Quantification VH-59
j. Other Health Effects: Valuation VH-61
k. Lost Worker Productivity: Quantification and Valuation . . VII-62
1. Estimated Reductions in Incidences of Health Endpoints and
Associated Monetary Values VII-62
E. Assessment of Human Welfare Benefits VD-65
1. Visibility Benefits VH-65
2. Agricultural and Forestry Benefits VII-68
a. Agricultural Benefits VII-68
b. Forestry Benefits VH-69
c. Other Effects VH-71
3. Benefits from Reductions in Materials Damage VD-72
4. Benefits from Reduced Ecosystem Damage VD-72
5. Estimated Values for Welfare Endpoints VD-74
F. Total Benefits VH-76
G. Comparison of Costs to Benefits VD-85
Appendix VII-A: Supplementary Benefit Estimates and Sensitivity Analyses of Key Parameters
in the Benefits Analysis VII-98
A. Introduction and Overview VII-98
B. Supplementary Benefit Estimates VII-98
C. Sensitivity Analyses VD-lOl
1. Alternative Lag Structures VII-101
2. PM Health Effect Threshold VH-103
3. Income Elasticity of Willingness to Pay VII-105
Chapter Vn. Appendix A References VII-106
Chapter VIII: Regulatory Flexibility Analysis Vffi-1
A. Regulatory Flexibility Analysis Vffi-1
B. Need for and Objectives of the Rule Vffi-2
C. Summary of Significant Public Comments on the IRFA Vffi-3
D. Types and Number of Small Entities To Which The Rule Will Apply Vffi-3
T-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1. Small Refiners Vffi-5
2. Small Distributors/Marketers of Highway Diesel Fuel VIH-5
E. Projected Costs of the Diesel Sulfur Standards Vffi-6
F. Projected Reporting, Recordkeeping, and Other Compliance Requirements of the
Rule VIH-7
1. Registration Reports VIE-?
2. Pre-Compliance Reports VIII-7
3. Annual Compliance Reports VIII-8
4. Product Transfer Documents (PTDs) VIII-8
5. Recordkeeping Requirements VIII-8
6. Diesel Fuel Pump Labeling Vffi-9
G. Regulatory Alternatives Vffi-9
1. 500 ppm Option Vffi-10
2. Small Refiner Credit Option Vffi-11
3. Diesel/Gasoline Compliance Date Option Vffi-11
4. Relationship of the Options to Each Other Vffi-12
H. Other Relevant Federal Rules Which May Duplicate, Overlap, or Conflict with the
Low Sulfur Diesel Fuel Rule Vffi-13
Chapter IX: Sulfur Control in Alaska & Territories IX-1
A. What is the Authority For Exemptions? IX-1
B. Alaska Exemption From the 500 ppm Sulfur Standard IX-2
1. Why Are We Considering an Exemption for Alaska? IX-2
2. Who Commented on the 1998 Proposal for a Permanent Exemption? IX-3
3. What are the Relevant Factors Unique to Alaska? IX-3
a. Geography, meteorology, and fuel production,
distribution, usage IX-4
b. Environmental and Health Factors IX-13
c. Engine and emission control system factors IX-17
d. Are there alternatives to granting or denying Alaska's Petition for
permanent exemption? IX-21
e. What Flexibility Are We Offering Alaska? IX-22
C. American Samoa, Guam, and Commonwealth of Northern Mariana Islands
(CNMI) IX-24
1. Why Are We Considering an Exemption for American Samoa, Guam, and
CNMI? IX-24
2. What are the Relevant Factors? IX-24
a. American Samoa IX-24
b. Guam IX-25
c. Commonwealth of Northern Mariana Islands (CNMI) IX-26
3. What Are the Options for the Territories? IX-27
T-10
-------
Table of Contents
4. What Flexibility are we Offering the Territories? IX-28
Appendix A: Legal Authority for Diesel Fuel Sulfur Control A-l
A. EPA's Current Regulatory Requirements for Diesel A-l
B. How the Proposed Diesel Sulfur Control Program Meets the CAA Section 21 l(c)
Criteria
A-l
1. Health and Welfare Concerns of Air Pollution Caused by Sulfur in Diesel
Fuel A-2
2. Impact of Diesel Sulfur Emission Products on Emission
Control Systems A-3
3. Sulfur Levels that Exhaust Aftertreatment for Heavy-Duty Vehicles Can
Tolerate A-4
4. Sulfur Sensitivity of Other Catalysts A-5
5. Effect of Diesel Sulfur Control on the Use of Other Fuels or Fuel
Additives A-7
APPENDIX B: Vehicle Miles Traveled by HDDE Class for Split by Pre-2007 and 2007+ Model
Years (MY) B-l
T-ll
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
T-12
-------
List of Tables
List of Tables
Table ES-1. Projected Incremental System Cost and Life Cycle Operating Cost for Heavy-Duty
Diesel Vehicles viii
Table ES-2. Projected Incremental System Cost and Life Cycle Operating Cost for Heavy-Duty
Gasoline Vehicles x
Table ES-3. Per-EngineA Cost Effectiveness of the Standards for 2007 and Later MY Vehicles
xiv
Table ES-4. 30-year Net Present ValueA Cost Effectiveness of the Standards xiv
Table ES-5. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final HD
Engine/Diesel Fuel Rule xvi
Table ES-6. Key Alternative Benefits Calculations for the HD Engine/Diesel Fuel Rule in 2030
xvii
Table II.A-1. Comparison of eastern U.S. regional model performance statistics between the Tier
2/Sulfur modeling and the Heavy Duty Engine modeling n-10
Table II.A-2. Eastern Metropolitan Areas with Modeled Exceedances of the 1-Hour Ozone
Standard in 2007, 2020, or 2030 With and Without Emission Reductions
from this Rule H-14
Table HA-3. Nonattainment Areas For Which EPA Has Proposed Action On SIP Submissions
Containing 1-hour Ozone Attainment Demonstrations or Otherwise Has Considered
Results of Local Ozone Modeling n-18
Table II. A-4. Areas With Some Risk of Ozone Violations between 2007 and 2030 Based on
Information Other Than Predictive Ozone Modeling n-25
Table n.A-5. Metropolitan Areas With Established or Requested 2007 or 2010 Attainment
Deadlines H-30
Table II.A-6. Areas and 1999 Populations at Risk of Exceeding the Ozone Standard between
2007 and 2030 H-39
Table II.A-7. Controlled Exposure of Healthy Human Subjects to Ozone n-43
Table II.A-8. Ozone Exposure in Subjects with Preexisting Disease n-46
Table II. A-9. Pulmonary Function Effects After Prolonged Exposures to Ozone n-49
Table II.A-10. Increased Airway Responsiveness Following Ozone Exposures n-52
Table II.A-11. Bronchoalveolar Lavage Studies of Inflammatory Effects from Controlled Human
Exposure to Ozone n-53
Table H.A-12. Percent Contribution to PM2 5 by Component, 1998 H-59
Table II.A-13. PM10 Nonattainment Areas Violating the PM10 NAAQS in 1997- 1999 .... H-60
Table II.A-14. Areas with Significant Risk of Exceeding the PM10 NAAQS without Further
Emission Reductions between 2007 and 2030 n-63
Table II.A-15. Effect Estimates Per 50 //g/m3 Increase in 24-hour PM10 Concentrations From
U.S. And Canadian Studies H-68
Table II.A-16. Effect Estimates per Variable Increments in 24-hour Concentrations of Fine
Particle Indicators (PM25, SO, H+) From U.S. and Canadian Studies H-70
T-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.A-17. Effect Estimates per Increments* in Annual Mean Levels of Fine Particle
Indicators from U.S. and Canadian Studies n-71
Table II. A-21. Ambient Diesel Paniculate Matter Concentrations from Receptor Modeling,
Dispersion Modeling and Elemental Carbon Measurements n-82
Table II. A-22. Occupational and Population Exposure to Diesel Exhaust n-85
Table II.A-23. Occupational and Population Exposure to Diesel Exhaust, Environmental
Equivalent Exposures and Exposure Margins n-92
Table II.A-24. Metropolitan Areas and Regions Included in Toxic Emissions Modeling . . n-102
Table II.A-25. Modeled Average 50-State Ambient Exposure to Gaseous Toxics from All
Highway Motor Vehicles (|ig/m3) in 1990, 1996, 2007, and 2020 without 2007 HDV
Standards and for 2020 with 2007 HDV Standards n-105
Table II.B-1. HDDE Fuel Economy Estimates by Model Year (miles per gallon) II-l 13
Table II.B-2. Heavy-Duty Vehicle Exhaust Emissions Standards n-118
Table II.B-3. Heavy-Duty Vehicle Standards Phase-In (percent of production) n-118
Table II.B-4. Nationwide NOx Emissions from HDDEs H-l 19
Table II.B-5. Nationwide PM10 Exhaust and Break/Tire Wear Emissions from HDDEs Without
Existing Fleet Reductions n-120
Table II.B-6. Nationwide NMHC Exhaust Emissions from HDDEs n-121
Table II.B-7. HDDE Fuel Consumption Estimates by Calendar Year n-121
Table II.B-8. Consumption of Highway Diesel Fuel Including Spillover n-122
Table II.B-9. Existing Fleet PM Reductions From Low Sulfur Fuel n-123
Table II.B-10. Crankcase Emissions from Uncontrolled HDDEs n-124
Table H.B-11. Total Reductions from HDDEs for this Rule n-124
Table H.B-12. Reductions in CO from HDDEs n-128
Table H.B-13. Reductions in SOx from Low Sulfur Fuel n-128
Table II.B-14. Reductions in Air Toxics from HDDEs H-129
Table II.B-15.a. Estimated Nationwide NOx Emissions from HDGVs Based on the Air Quality
Analysis Case (thousand short tons per year) n-131
Table II.B-15.b. Estimated Nationwide NOx Emissions from HDGVs Based on the Updated
Control Case (thousand short tons per year) n-131
Table H.B-16.a. Estimated Nationwide Exhaust NMHC Emissions from HDGVs Based on the
Air Quality Analysis Case (thousand short tons per year) H-132
Table n.B-16.b. Estimated Nationwide Exhaust NMHC Emissions from HDGVs Based on the
Updated Control Case (thousand short tons per year) D-133
Table II.B-17.a. Estimated Nationwide Evaporative Emissions from HDGVs Based on the Air
Quality Analysis Case (thousand short tons per year) n-134
Table II.B-17.b. Estimated Nationwide Evaporative Emissions from HDGVs Based on the
Updated Control Case (thousand short tons per year) n-13 5
Table H.B-18. Estimated 49-State Reductions in Air Toxics from HDGVs (thousand short tons
per year) n-135
Table HB-19. Total NOx Emissions and Benefits for This Rule (thousand short tons per year)
T-14
-------
List of Tables
n-iss
Table HB-20. Total PM Emissions and Reductions for This Rule (thousand short tons per year)
H-138
Table HB-21. Total NMHC Emissions and Reductions for This Rule (thousand short tons per
year) H-139
Table II.B-22. Total Reductions in Air Toxics for This Rule (thousand short tons per year) n-139
Table II.B-23. Comparison of NPRM and FRM HDDE Baseline Inventories for NOx and
NMHC (thousand short tons per year) n-141
Table II.B-24. Comparison of NPRM and FRM HDDE Baseline Inventories for Exhaust PM and
SOx (thousand short tons per year) n-141
Table II.B-25. Comparison of NPRM and FRM HDGV Baseline Inventories for NOx and
NMHC (thousand short tons per year) n-142
Table II.B-26. Comparison of EMFAC2000 and EPA PM Deterioration Rates for HDDEs
(grams per mile per 10,000 miles) n-143
Table HB-27. 2007 Baseline Emissions Inventories for 48 Contiguous States n-145
Table HB-28. 2020 Baseline Emissions Inventories for 48 Contiguous States n-145
Table H.B-29. 2030 Baseline Emissions Inventories for 48 Contiguous States n-145
Table IH.A-1. PM Emissions from a Heavy-Duty Diesel Engine at the Indicated Fuel Sulfur
Levels IH-12
Table IH.A-2. SET Composite Test Results with the Dual Leg NOx Adsorber System . . . HI-45
Figure IHA-3. Timed Regeneration Schedule for Switching between NOx Adsorber Legs ni-46
Table IH.A-4. HDDE FTP Emissions from NVFEL Test Program IH-47
Table IHA-5. SO2 Oxidation Rates for a Platinum Oxidation Catalyst at the Indicated Catalyst
Inlet Temperatures 111-85
Table IHA-6. Estimated PM Emissions from a Heavy-Duty Diesel Engine at the Indicated
Average Fuel Sulfur Levels in-86
Table ILL A-7. Estimated Fuel Economy Impact from Desulfation of a 90 Percent Efficient NOx
Adsorber IH-89
Table IHB-1. Exhaust Emission Control Hardware and Technologies That May be Used to Meet
the 2008 Heavy-Duty Gasoline Standards IH-95
Table IHB-2. Emission Standards for Select Federal and California Gasoline Vehicles &
Engines IH-108
Table ffl.B-3. 2000 Model Year Vehicle Certification Data (gram/mile) ffl-111
Table IH.B-4. 2000 Model Year Engine Certification Data (g/bhp-hr) IH-112
Table ffi.B-5. New Heavy-Duty Evaporative Emission Standards ffi-119
Table IH.B-6. 1998-2000 Model Year Evaporative Emission Certification Data IH-122
Table IV.A-1. Volume Fraction of U.S. Highway Diesel Pool from each Blendstock
Component IV-4
Table IV.A-2. Sulfur Levels of Highway Diesel Blendstocks (CA Excluded) IV-5
Table IV.A-3. Average Highway Diesel Fuel Sulfur Levels by Geographic Area IV-7
Table IV.A-4. Distillation Characteristics of Diesel Blendstocks (CA Excluded) IV-9
T-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table IV.A-5. Properties of Diesel Blendstocks (CA Excluded) IV-10
Table IV.A-6. Average Highway Diesel Fuel Parameter Levels by Geographic Area IV-12
Table IV.A-7. Maximum Cost of Meeting the 15 ppm Cap IV-36
Table IV.A-8. Number of Refineries Producing 15 ppm Diesel by PADD IV-47
Table IV.B-1. Design and Construction Factors for Desulfurization Equipment IV-57
Table IV.B-2. Number of Gasoline Desulfurization Units Becoming Operational on January 1 of
the Indicated Year IV-58
Table IV.B-3. Distribution of Personnel Requirements Throughout the Project IV-59
Table IV.B-4. Maximum Monthly Demand for Personnel IV-61
Table IV.B-5. Capital Expenditures for Gasoline and Diesel Fuel Desulfurization IV-63
Table IV.D-1. Ratios Used in Comparing the Relative Difficulty in Limiting Contamination
During the Distribution of Various Fuels IV-84
Table V.A-1. Service Classes of Heavy-Duty Vehicles V-3
Table V. A-2. Summary of Near and Long Term Cost Estimates V-4
Table V.A-3. 2007 NOx Adsorber Cost Estimate V-l 1
Table V.A-4. 2007 Catalyzed Diesel Particulate Filter Cost Estimate V-l5
Table V.A-5. 2007 Diesel Oxidation Catalyst Cost Estimate V-16
Table V.A-6. 2007 Closed Crankcase Filtration System Cost Estimate V-17
Table V. A-7. Annualized and Non-Annualized R&D Costs for Light Heavy-Duty Diesel
Engines V-19
Table V. A-8. Annualized and Non-Annualized R&D Costs for Medium Heavy-Duty Diesel
Engines V-20
Table V. A-9. Annualized and Non-Annualized R&D Costs for Heavy Heavy-Duty Diesel
Engines and Urban Buses V-21
Table V.A-10. Annualized and Non-Annualized Tooling Costs for Light Heavy-Duty Diesel
Engines V-22
Table V.A-11. Annualized and Non-Annualized Tooling Costs for Medium Heavy-Duty Diesel
Engines V-23
Table V. A-12. Annualized and Non-Annualized Tooling Costs for Heavy Heavy-Duty Diesel
Engines and Urban Buses V-23
Table V.A-13. Annualized and Non-Annualized Certification Costs for Light Heavy-Duty Diesel
Engines V-24
Table V. A-14. Annualized and Non-Annualized Certification Costs for Medium Heavy-Duty
Diesel Engines V-25
Table V.A-15. Annualized and Non-Annualized Certification Costs for Heavy Heavy-Duty
Diesel Engines and Urban Buses V-25
Table V.A-16. Annualized Fixed Costs for Light Heavy-Duty Diesel Engines V-26
Table V.A-17. Annualized Fixed Costs for Medium Heavy-Duty Diesel Engines V-26
Table V.A-18. Annualized Fixed Costs for Heavy Heavy-Duty Diesel Engines and Urban Buses
V-27
T-16
-------
List of Tables
Table V.A-19. Projected Incremental Diesel Engine/Vehicle Costs V-36
Table V. A-20. Baseline Costs for Heavy-Duty Engines and Vehicles V-37
Table V. A-21. Estimated Annualized Nationwide Costs for Heavy-Duty Diesel Engines
Associated with the 2007 Emission Standard V-39
Table V. A-22. Estimated Non-Annualized Nationwide Costs for Heavy-Duty Diesel Engines
Associated with the 2007 Emission Standard V-40
Table V.B-1. 2005 (Phase 1) and Expected 2008 (Phase 2) Technology Packages for Heavy-
Duty Gasoline Vehicles excluding Medium-Duty Passenger Vehicles V-43
Table V.B-2. Costs Associated with the Increased Use of Precious Metals V-46
Table V.B-3. Costs Associated with Various Catalyst Configurations V-49
Table V.B-4. Summary of Hardware Costs for the Proposed 2007 Heavy-Duty Gasoline
Standards V-52
Table V.B-5. Summary of Incremental Costs to Meet the 2008 Heavy-Duty Gasoline Emission
Standards V-56
Table V.B-6. Estimated Annualized Nationwide Vehicle Costs Associated with the 2008 Heavy-
Duty Gasoline Emission Standards V-57
Table V.B-7. Estimated Non-Annualized Nationwide Vehicle Costs Associated with the 2008
Heavy-Duty Gasoline Emission Standards V-59
Table V.C-1. Presence of Light Cycle Oil in the Distillate of U.S. Refineries Producing Highway
Diesel Fuel V-64
Table V.C-2. Presence of Other Cracked Blendstocks in the Distillate of U.S. Refineries
Producing Highway Diesel Fuel V-65
Table V.C-3. Technology Projected to be Used to Achieve Various Diesel Fuel Sulfur Levels
V-69
Table V.C-4. Process Projections to Desulfurize a Typical Diesel Fuel V-72
Table V.C-5. Process Projections to Desulfurize 100% Straight Run Diesel Fuel V-73
Table V.C-6. Process Projections for Revamping an Existing Highway Diesel Hydrotreater for
Further Desulfurizing a Typical Diesel Fuel V-75
Table V.C-7. Process Projections for Revamping an Existing Highway Diesel Hydrotreater for
Desulfurizing 100% Straight Run Diesel Fuel V-78
Table V.C-8. Process Projections for Revamping an Existing Highway Diesel Hydrotreater for
Further Desulfurizing Diesel Fuel Blendstocks to Meet a 15 ppm Cap Standard . . . V-81
Table V.C-9. Process Projections for Installing a New Grassroots Unit for Desulfurizing
Untreated Diesel Fuel Blendstocks to Meet a 15 ppm Cap Standard V-84
Table V.C-10. Estimated Hydrogen Consumption to Desulfurize Nontreated Distillate, Stocks to
Meet the 15 ppm Highway Diesel Fuel Sulfur Cap V-86
Table V.C-11. Comparison of Calculated Hydrogen Consumption with the Hydrogen
Consumption provided by Vendors A and B for Specific Distillate Feeds V-87
Table V.C-12. Process Operations Information for Additional Units used in the Desulfurization
Cost Analysis V-90
Table V.C-13. Offsite and Location Factors Used for Estimating Capital Costs V-92
T-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table V.C-14. Economic Cost Factors Used in Calculating the Capital Amortization Factor
V-93
Table V.C-15. Summary of Costs From EIA Information Tables for 1999, and Other Cost
Factors V-95
Table V.C-16. Estimated Per-Refmery Capital, Operating and Per-Gallon Cost for Full
Implementation of Desulfurizing Highway Diesel Fuel to Meet a 15 ppm Cap Standard
V-98
Table V.C-17. Costs for Treating LCO, Coker, and Straight Run Diesel Feedstocks (1999
Dollars and 7% before tax ROI) V-100
Table V.C-18. Overall Estimated Per-Refmery Capital, Operating and Per-Gallon Cost for Years
2006 and 2010 for Implementation of Desulfurizing Highway Diesel Fuel to Meet a 15
ppm Cap Standard (1999 Dollars, 7% ROI before taxes) V-101
Table V.C-19. Projected U.S. Aggregate Operating and Capital Cost of Desulfurizing Highway
Diesel Fuel to Meet a 15 ppm Cap Standard (1999 Dollars, 7% ROI before taxes) V-103
Table V.C-20. Per-Gallon Cost for Desulfurizing Highway Diesel Fuel to Meet a 15 ppm Cap
Standard Based on Different Capital Amortization Rates V-104
Table V.C-21. PADD-Average Price Difference Between 500 ppm Highway and Non-Highway
Diesel (1999 Dollars, 7% ROI before taxes) V-106
Table V.C-22. Costs Under Nonhighway Production Shift Scenario (1999 Dollars, 7% ROI
before taxes) V-109
Table V.C-23. Estimated Costs of Nonhighway Production Shift Scenario versus Current
Highway Producer Scenario to Meet 15 ppm Highway Diesel Fuel Cap Standard (1999
dollars, 7% ROI before taxes) V-l 11
Table V.C-24. Comparison of Mathpro's and EPA's Costs for Meeting a 15 ppm Highway
Diesel Fuel Sulfur Cap Standard (7% ROI before taxes) V-l 14
Table V.C-25. Comparison of DOE and EPA Refining Costs for Meeting a 15 ppm Highway
Diesel Fuel Sulfur Cap Standard (7% ROI before taxes) V-l 18
Table V.C-26. Distribution Costs During the Initial Years of Our Sulfur Program and After the
Program Becomes Fully Effective V-l22
Table V.C-21'. Data Used to Calculate the Cost of Distributing the Additional Volume of
Highway Diesel Fuel Needed to Compensate for a Reduction in Energy Density . V-l24
Table V.C-28. MathPro Lubricity Additive Cost Estimates V-134
Table V.C-29. Components Potentially Affected by Lower Sulfur Levels in Diesel Fuel . V-136
Table V.C-30. Cost Savings to the Existing Fleet from Extend Oil Change Intervals Made
Possible by Low Sulfur Diesel Fuel V-138
Table V.C-31. Cost Savings to the Existing Fleet for Reduced EGR System Replacement Made
Possible by Low Sulfur Diesel Fuel V-139
Table V.C-32. Cost Savings to the New Fleet (2007 and later) for Reduced EGR System
Replacement Made Possible by Low Sulfur Diesel Fuel V-l40
Table V.C-33. Cost Savings to the Existing Fleet from Extend Exhaust System Replacement
Intervals Made Possible by Low Sulfur Diesel Fuel V-141
T-18
-------
List of Tables
Table V.C-34. Aggregate Savings to the Existing Fleet (pre-2007 fleet) Made Possible by Low
Sulfur Diesel Fuel V-143
Table V.C-35. Aggregate Savings for the New Fleet (2007 and later) Made Possible by Low
Sulfur Diesel Fuel V-144
Table V.C-36. Fleet Average Per-Engine Lifecycle Costs V-147
Table V.D-1. Total annualized costs of heavy-duty diesel engines, heavy-duty gasoline vehicles,
and 15 ppm diesel fuel V-149
Table VI.B-1. Fleet-average, Per-engine Costs for HDDE VI-5
Table VI.B-2. Fleet Average Per-Engine Costs for HDDE Used in Cost-effectiveness .... VI-8
Table VI.C-1. Per-engine Discounted Lifetime Tons for HDDE VI-10
Table VI.C-2. Engine Class Sales Weighting Factors for HDDE VI-10
Table VI.C-3. Fleet average, Per-engine Discounted Lifetime Tons for HDDE VI-11
Table VI.D-1. Fleet-average, Per-vehicle Costs for HDGV Used in Cost-effectiveness ... VI-13
Table VI.D-2. Per-vehicle Discounted Lifetime Tons for HDGV VI-14
Table VI.D-3. Vehicle Class Sales Weighting Factors for HDGV VI-14
Table VI.D-4. Fleet average, Per-vehicle Discounted Lifetime Tons for HDGV VI-15
Table VI.F-1. Per-vehicle Cost-effectiveness of the Standards VI-17
Table VI.F-2. 3 0-year Net Present Value Cost-effectiveness of the Standards VI-17
Table VI.F-3. Per-vehicle Cost-effectiveness of the Standards Using 3 Percent ROI and Discount
Rate VI-18
Table VI.F-4. 30-year Net Present Value Cost-effectiveness of the Standards Using 3 Percent
ROI and Discount Rate VI-19
Table VI.F-5. Cost-effectiveness of Previous Mobile Source Programs for NOx + NMHC
VI-19
Table VI.F-6. Cost-effectiveness of Previous Mobile Source Programs for PM VI-20
Table VII-1. Human Health and Welfare Effects of Pollutants Affected by the HD Engine/Diesel
Fuel Rule VH-6
Table VII-2. Summary of UAM-V Derived Ozone Air Quality Metrics Due to HD
Engine/Diesel Fuel Standards for Health Benefits EndPoints: Eastern U.S VD-16
Table VII-3. Summary of UAM-V Derived Ozone Air Quality Metrics Due to HD Engine/Diesel
Fuel Standards for Welfare Benefits Endpoints: Eastern U.S VII-17
Table VII-4. Summary of 2030 Base Case PM Air Quality and Changes Due to HD
Engine/Diesel Fuel Standards VII-23
Table VII-5. Distribution of PM2 5 Air Quality Improvements Over 2030 Population Due to HD
Engine/Diesel Fuel Standards VH-24
Table VII-6. Summary of Absolute and Relative Changes in PM Air Quality Due to HD
Engine/Diesel Fuel Standards VII-25
Table VII-7. Distribution of Populations Experiencing Visibility Improvements in 2030 Due to
HD Engine/Diesel Fuel Standards VH-26
Table VII-8. Summary of 2030 Baseline Visibility and Changes by Region: Residential . VII-27
T-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table VII-9. Summary of 2030 Baseline Visibility and Changes by Region: Recreational
(Annual Average Deciviews) VII-28
Table VII-10. Summary of 2030 Nitrogen Deposition in Selected Estuaries and Changes Due to
HD Engine/Diesel Fuel Rule (million kg/year) VII-30
Table VII-11. Elasticity Values Used to Account for Projected Real Income Growth .... VII-34
Table VII-12. Adjustment Factors Used to Account for Projected Real Income Growth . VD-35
Table VII-13. Primary Sources of Uncertainty in the Benefit Analysis VD-38
Table VD-14. Endpoints and Studies Included in the Primary Analysis VD-42
Table VII-15. Unit Values Used for Economic Valuation of Health Endpoints VD-44
Table VII-16. Alternative Concentration-Response Models Relating Premature Mortality and
Chronic Exposure to Fine Particulates VD-49
Table VII-17. Expected Impact on Estimated Benefits of Premature Mortality Reductions of
Differences Between Factors Used in Developing Applied VSL and Theoretically
Appropriate VSL VH-51
Table VD-18. Recent Studies on the Effects of Air Pollution on Asthma Symptoms .... VD-59
Table VD-19. Primary Estimate of Annual Health Benefits Associated With Air Quality Changes
Resulting from the HD Engine/Diesel Fuel Rule in 2030 VH-64
Table VII-20. Reduction Goals and 1998 Nitrogen Loads to Selected Eastern Estuaries (tons per
year) VH-73
Table VD-21. Estimated Annual Reductions in Nitrogen Loadings in Selected Eastern Estuaries
for the Final HD Engine/Diesel Fuel Rule in 2030 VH-73
Table VII-22. Primary Estimate of Annual Monetary Values for Welfare Effects Associated
With Improved Air Quality Resulting from the HD Engine/Diesel Fuel Rule in 2030711-75
Table VII-23. Primary Estimate of Annual Monetized Benefits Associated With Improved Air
Quality Resulting from the HD Engine/Diesel Fuel Rule in 2030 VH-77
Table VII-24. Alternative Estimates of Premature Mortality Benefits for the HD Engine/Diesel
Fuel Rule in 2030 VH-79
Table VII-25. Additional Alternative Benefits Calculations for the HD Engine/Diesel Fuel Rule
in 2030 VH-81
Table VH-26. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final HD
Engine/Diesel Fuel Rule VH-87
Table VII-A-1. Supplemental Benefit Estimates for the Final HD Engine/Diesel Fuel Rule for
the 2030 Analysis Year VH-101
Table VII-A-2. Sensitivity Analysis of Alternative Lag Structures for PM-related Premature
Mortality VH-103
Table VII-A-3. Sensitivity Analysis of Alternative Income Elasticities VD-105
Table Vffi-1. Industries Containing Small Businesses Potentially Affected by the Low Sulfur
Diesel Fuel Rule VIH-4
Table Vffi-2. Types and Number of Small Entities to Which the Diesel Sulfur Program Will
Apply VIH-4
T-20
-------
List of Figures
List of Figures
Figure ES-1. Total Annualized Costs xi
Figure II.B-1. Projected DSPM from Pre-2007 Engines Using Highway Diesel Fuel n-123
Figure H.B-2. Projected HDDE NOx Emissions Due to 2004 and 2007 Standards n-125
Figure HB-3. Projected Nationwide PM Emissions from HDDEs n-126
Figure H.B-4. Projected Nationwide NMHC Emissions from HDDEs n-127
Figure HB-5. Projected Nation wide Exhaust NOx Emissions from HDGVs n-130
Figure n.B-6. Projected Nationwide Exhaust NMHC Emissions from HDGVs n-132
Figure n.B-7. Proj ected Nationwide Evaporative Emissions from HDGVs n-134
Figure II.B-8. Proj ected NOx Inventory for Heavy-Duty Highway Vehicles n-136
Figure II.B-9. Projected PM Inventory for Heavy-Duty Highway Vehicles n-137
Figure II.B-10. Proj ected NMHC Inventory for Heavy-Duty Highway Vehicles n-137
Figure HB-11. Projected HDDE Exhaust PM Inventory with and without Consideration of
Tampering and Malmaintenance n-144
Figure IH.A-1. HD CDPF PM Removal Efficiency Over the Federal Test Procedure IH-8
Figure IHA-2. HD PM Removal Efficiency for a CDPF Over the Supplemental Emission Test
(SET) m-10
Figure IE. A-3. Schematic Representation of the Operation of a Dual-Bed NOx Adsorption
Catalyst IH-23
Figure IHA-4. A functional schematic representation of the PM and NOx exhaust emission
control system tested at NVFEL shown together with one possible approach having the
same functionality, but with further integration of components ffi-25
Figure IHA-5. NOx Adsorber Efficiency Characteristics versus Exhaust Temperature . . . 111-30
Figure IHA-6. Modal Definitions 111-36
Figure IH.A-7. SET & AVL Composites, and Temperature vs. NOx Chart for Adsorber A IH-39
Figure IHA-8. SET & AVL Composites, and Temperature vs. NOx Chart for Adsorber B 111-40
Figure IH.A-9. SET & AVL Composites, and Temperature vs. NOx Chart for Adsorber D IH-41
Figure ILL A-10. SET & AVL Composites, and Temperature vs. NOx Chart for Adsorber E 111-42
Figure IH.A-11. NOx Adsorber FTP Preconditioning Cycle used in NVFEL Testing IH-44
Figure in.A-12. Comparison of NOx Conversion Efficiency before and after Desulfati on . 111-58
Figure IE. A-13. Flow Reactor Testing of a NOx Adsorber with Periodic Desulfati ons ... ID-61
Figure IHA-14. Influence of Maximum Catalyst Bed Temperature During Desulfation . . . 111-62
Figure IHA-15. Integrated NOx Conversion Efficiency following Aging and Desulfation . 111-64
Figure IE. A-16. Integrated NOx Conversion Efficiency after Repeated Desulfation 111-65
Figure ILL A-17. Repeated Sulfur Poisoning and Desulfation on a Bench Pulsator 111-67
Figure ILL A-18. Effect of Fuel Sulfur on Regeneration Temperature ID-SI
Figure IE. A-19. NO and SO2 Conversion Rates Over Platinum IH-84
Figure IHA-20. Diesel Fuel Sulfur Effect on NOx Adsorber Performance after 150 hours 111-87
Figure in.A-21. Sulfate PM Emissions versus Diesel Fuel Sulfur Level with 3,500 ppm Sulfur
Engine Oil IH-93
T-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Figure in.B-1. Impact of Coating Architecture on HC andNOx Emissions III-104
Figure in.B-2. Emissions after an equivalent of 50,000 miles for various tested configurations of
Ford Expedition LDT4 SUVs with 5.4L V8 engines ffl-115
Figure in.B-3. Emissions after an equivalent of 50,000 miles for various tested configurations of
1999 GM Chevrolet Silverado LDT3 pickups with 5.3L V8 engines ffl-116
Figure IV.A-1. Diagram of a Typical Complex Refinery IV-2
Figure IV.A-2. Map of U.S. Petroleum Administrative Districts for Defense IV-7
Figure V.A-1. Distribution of Progress Ratios V-34
Figure V.C-1. Refinery Specific Costs for Fully Implemented 15 ppm Sulfur Cap Standard V-99
Figure V.C-2. Refinery Specific Production Rates of Highway Diesel versus No. 2 Oil Distillate
Pool V-107
Figure V.C-3. Lowest Refinery Costs for Converting NonHighway to 15 PPM Highway Diesel
Fuel V-108
Figure V.C-4. Refinery Costs per PADD for Current Highway Units Scenario for Meeting the 15
ppm Sulfur Highway Diesel Fuel Cap Standard V-l 11
Figure V.C-5. Refinery Costs per PADD under Converted NonHighway Units Shift Scenario for
Meeting the 15 ppm Highway Diesel Fuel Cap Standard V-l 12
Figure V.D-1. Total annualized costs of heavy-duty diesel engines, heavy-duty gasoline vehicles,
and 15 ppm diesel fuel V-148
Figure VII-1. Steps in the Heavy Duty Engine/Diesel Fuel Benefits Analysis VII-5
Figure VII-2. UAM-V Modeling Domain for Eastern U.S VH-13
Figure VII-3. REMSAD Modeling Domain for Continental U.S VII-18
Figure VII-4. Example of REMSAD 36 x 36km Grid-cells for Maryland Area VH-20
Figure VII-5. Recreational Visibility Regions for Continental U.S VD-28
Figure VII-6. Alternative Approaches for Assessing the Value of Reduced Mortality Risk VII-52
Figure VII-A-1. Impact of PM Health Effects Threshold on Avoided Incidences of Premature
Mortality Estimated with the American Cancer Society/Krewski, et al. (2000) C-R
Function
vn-104
T-22
-------
List of Acronyms
List of Acronyms
|ig/m3
A/F
ANPRM
API
ASTM
bbl
BTU
CAA or the Act
CARS
CASAC
CMB
CNMI
CO
CSMAs
DECSE
DOC
EGR
EHC
EIA
EPA or the Agency
FCC
FTP
g/bhp-hr
GVWR
HC
HD-FTP
HDE
HDV
HLDT
I/M
micrograms per cubic meter
air to fuel ratio
Advance Notice of Proposed Rulemaking
American Petroleum Institute
American Society for Testing and Materials
barrel
British Thermal Unit
Clean Air Act
California Air Resources Board
Clean Air Scientific Advisory Committee
chemical mass balance
Commonwealth of Northern Mariana Islands
carbon monoxide
consolidated metropolitan statistical areas
Diesel Emission Control Sulfur Effects
diesel oxidation catalyst
exhaust gas recirculation
electrically heated catalyst
Energy Information Administration
U.S. Environmental Protection Agency
fluidized catalytic cracker
federal test procedure
grams per brake-horsepower-hour
gross vehicle weight rating
hydrocarbon
heavy-duty federal test procedure
heavy-duty engine
heavy-duty vehicle
heavy light-duty truck
inspection/maintenance
T-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
IARC
IRFA
LCO
LOT
LDV
LEV
LLDT
LPG
MDPV
MECA
MECA
MSAs
MSCF
MY
NAAQS
NAICS
NCPs
NMHC
NMOG
NO2
NOx
NPC
NPRA
NPRM
NPV
NRC
NSR
NSWS
NTE
OAQPS
OBD
OMB
International Agency for Research on Cancer
initial regulatory flexibility analysis
light cycle oil
light-duty truck
light-duty vehicle
low emission vehicle
light light-duty truck
liquid petroleum gas
medium-duty passenger vehicle
Manufacturers of Emission Controls Association
Manufacturers of Emission Controls Association
metropolitan statistical areas
thousand standard cubic feet
model year
National Ambient Air Quality Standards
North American Industry Classification System
non-conformance penalties
non-methane hydrocarbons
non-methane organic gases
nitrogen dioxide
oxides of nitrogen
National Petroleum Council
National Petrochemical & Refiners Association
Notice of Proposed Rulemaking
net present value
National Research Council
New Source Review
National Surface Water Survey
not-to-exceed
Office of Air Quality Planning and Standards
on-board diagnostics
Office of Management and Budget
T-24
-------
List of Acronyms
QMS
PADD
Pd
PM
POM
ppm
Pt
R&D
RFA
RfC
Rh
RIA
ROI
SBA
SBARP or the Panel
SBREFA
SCR
SER
SFTP
SIC
SIGMA
SIP
SO2
SOF
SOx
TAC
TOG
TW
HDDS
ULEV
VMT
VOC
Office of Mobile Sources
Petroleum Administrative Districts for Defense
palladium
particulate matter
polycyclic organic matter
part per million
platinum
research and development
Regulatory Flexibility Act
reference concentration
rhodium
Regulatory Impact Analysis
return on investment
U.S. Small Business Administration
Small Business Advocacy Review Panel
Small Business Regulatory Enforcement Fairness Act
selective catalytic reduction
Small Entity Representative
supplemental federal test procedure
Standard Industrial Classification
Society of Independent Gasoline Marketers of America
State Implementation Plan
sulfur dioxide
soluble organic fraction
oxides of sulfur
toxic air contaminant
total organic gases
test weight
urban dynamometer driving schedule
ultra-low emission vehicles
vehicle miles traveled
volatile organic compound
T-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
T-26
-------
Chapter I: Introduction
Chapter I: Introduction
We prepared this Regulatory Impact Analysis (RIA) for our final rulemaking (FRM) on
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements. The purpose of this document is to present our estimates of the likely costs,
benefits, and industry impacts associated with the implementation of the final heavy-duty engine
and vehicle standards and the diesel sulfur requirements. Throughout this RIA we have
referenced a number of technical documents published by the Society of Automotive Engineers
(SAE). Information on how to obtain copies of these technical documents is available in the
docket for this rulemaking.a
This chapter provides an overview of the final rule. Subsequent chapters present the
following information:
Chapter II presents the health and welfare concerns associated with heavy-duty vehicle
emissions, and the expected emissions reductions resulting from the new standards.
Chapter III examines the engine and vehicle changes needed to meet the heavy-duty
emission standards and the feasibility of these changes under the implementation
schedule. It also presents the basis for the need for diesel fuel sulfur levels of 15 parts per
million or less.
Chapter IV examines the refinery and fuel distribution system changes needed to meet
the low sulfur highway diesel fuel requirement and the feasibility of these changes under
the implementation schedule.
• Chapter V estimates the economic impact of the engine/vehicle and fuels standards in
per-vehicle and per-gallon terms, and in the aggregate.
• Chapter VI discusses the cost-effectiveness of the program in achieving emission
reductions.
• Chapter VII discusses the cost-benefit analysis of the program.
a EPA Memorandum "Obtaining Society of Automotive Engineers (SAE) Technical Papers", William
Charmley, copy available in EPA Air Docket A-99-06.
1-1
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
• Chapter VIII presents the Regulatory Flexibility Analysis (RFA) for this FRM. This
analysis evaluates the impacts of the heavy-duty engine and vehicle standards and the
diesel sulfur requirements on small businesses.
• Chapter IX analyzes the issues surrounding how the engine and fuel standards should be
applied in Alaska and U.S. territories.
• Appendix A describes current regulatory requirements that affect diesel sulfur content
and explains our bases for controlling diesel sulfur under Section 21 l(c) of the Clean Air
Act.
Appendix B contains a table of vehicle miles traveled by heavy-duty diesel engine class.
The final rule implements the second of two phases of a comprehensive nationwide
program for controlling emissions from heavy-duty engines (FIDEs) and vehicles. It builds upon
the phase 1 program we recently finalized (65 FR 59896, October 6, 2000) . That action
reviewed and confirmed the 2004 model year emission standards set in 1997 (62 FR 54693,
October 21, 1997), finalized stringent new emission standards for gasoline-fueled heavy-duty
vehicles (HDVs), and finalized other changes to the heavy-duty program, including provisions to
ensure in-use emissions control.
This second phase of the program looks beyond 2004, based on the use of high-efficiency
exhaust emission control devices and the consideration of the vehicle and its fuel as a single
system. In developing the final rule, we took into consideration comments received in response
to a notice of proposed rulemaking (NPRM) published in June of this year (65 FR 35430, June 2,
2000), and comments we received in response to our discussion of future standards in the
heavy-duty 2004 standards rulemaking.
There are two basic parts to the final rule: (1) new exhaust emission standards for heavy-
duty highway engines and vehicles, and (2) new quality standards for highway diesel fuel. The
systems approach of combining the engine and fuel standards into a single program is critical to
the success of our overall efforts to reduce emissions, because the emission standards would not
be feasible without the fuel change. This is because the emission standards are expected to result
in the use of high-efficiency exhaust emission control devices that would be damaged by sulfur
in the fuel. This final rule, by providing extremely low sulfur diesel fuel, will also enable cleaner
diesel passenger vehicles and light-duty trucks. This is because the same pool of highway diesel
fuel also services these light-duty diesel vehicles, and these vehicles can employ technologies
similar to the high-efficiency heavy-duty exhaust emission control technologies that will be
enabled by the fuel change. We believe these technologies are needed for diesel vehicles to
comply with our recently adopted Tier 2 emissions standards for light-duty highway vehicles (65
FR 6698, February 10, 2000).
1-2
-------
Chapter I: Introduction
We believe that this systems approach is a comprehensive way to enable promising new
technologies for clean diesel affecting all sizes of highway diesel engines and, eventually, diesel
engines used in nonroad applications too. The fuel change, in addition to enabling new
technologies, will also produce emissions and maintenance benefits in the existing fleet of
highway diesel vehicles. These benefits include reduced sulfate and sulfur oxides emissions,
reduced engine wear and less frequent oil changes, and longer-lasting exhaust gas recirculation
(EGR) components on engines equipped with EGR. Heavy-duty gasoline vehicles will also be
expected to reach cleaner levels due to the transfer of recent technology developments for light-
duty applications, and the recent action taken to reduce sulfur in gasoline as part of the Tier 2
rule.
The basic elements of the final rule are outlined below. We are finalizing a PM emissions
standard for new heavy-duty engines of 0.01 grams per brake-horsepower-hour (g/bhp-hr), to
take full effect for diesels in the 2007 model year. We are also finalizing standards for NOx and
NMHC of 0.20 g/bhp-hr and 0.14 g/bhp-hr, respectively. These NOx and NMHC standards will
be phased in together between 2007 and 2010, for diesel engines. The phase-in will be on a
percent-of-sales basis: 50 percent from 2007 to 2009 and 100 percent in 2010. Gasoline engines
will be subject to these standards based on a phase-in requiring 50 percent compliance in the
2008 model year and 100 percent compliance in the 2009 model year. In addition, we are
finalizing our proposal to include turbocharged diesels in the existing crankcase emissions
prohibition, effective in 2007, with some revisions that allow some level of crankcase emissions
to be discharged as long as the sum total of crankcase and exhaust emissions remains below the
applicable standard.
Standards for complete HDVs will be implemented on the same schedule as for gasoline
engine standards. For certification of complete vehicles between 8500 and 10,000 pounds gross
vehicle weight rating (GVWR), the standards are 0.2 grams per mile (g/mi) for NOx, 0.02 g/mi
for PM, 0.195 g/mi for NMHC, and 0.016 g/mi for formaldehyde.13 For vehicles between 10,000
and 14,000 pounds, the standards are 0.4 g/mi for NOx, 0.02 g/mi for PM, and 0.230 g/mi for
NMHC, and 0.021 g/mi for formaldehyde. These standards levels are roughly comparable to the
engine-based standards in these size ranges. Note that these standards will not apply to vehicles
above 8500 pounds that we classify as medium-duty passenger vehicles as part of our Tier 2
program.
We are adopting new evaporative emissions standards for heavy-duty engines and
vehicles, effective on the same schedule as the gasoline engine and vehicle exhaust emission
standards. The new standards for 8500 to 14,000 pound vehicles are 1.4 and 1.75 grams per test
for the 3-day diurnal and supplemental 2-day diurnal tests, respectively. Standards levels of 1.9
b Vehicle weight ratings in this rule refer to GVWR (the curb weight of the vehicle plus its maximum
recommended load of passengers and cargo) unless noted otherwise.
1-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
and 2.3 grams per test will apply for vehicles over 14,000 pounds. These standards represent
more than a 50 percent reduction in the numerical standards as they exist today.
Finally, this rule specifies that, beginning September 1, 2006, diesel fuel sold for use in
model year 2007 and later highway vehicles must be limited in sulfur content to 15 parts per
million (ppm). This sulfur standard is based on our assessment of the impact of sulfur on
advanced exhaust emission control technologies, and a corresponding assessment of the
feasibility of low sulfur fuel production and distribution.
The new program includes a combination of flexibilities available to refiners to ensure a
smooth transition to low sulfur highway diesel fuel. First, refiners can take advantage of a
temporary compliance option which includes a banking and trading component. Under this
voluntary option, a restricted amount of highway diesel fuel may continue to be produced at the
existing 500 ppm sulfur maximum standard. At the end of the transition period all highway
diesel fuel must meet the 15 ppm sulfur standard. Second, we are providing additional
flexibilities for small refiners to minimize their economic burden in complying with the 15 ppm
sulfur standard. Third, we are including a provisions for refiners located in the Geographic
Phase-in Area (GPA) as defined in the Tier 2 program which will allow them to stagger their
gasoline and diesel investments. Finally, we are adopting a general hardship provision for which
any refiner may apply on a case-by-case basis under certain conditions.
With minor exceptions, existing compliance provisions for ensuring diesel fuel quality
that have been in effect since 1993 remain unchanged (55 FR 34120, August 21, 1990).
Additional compliance provisions have been established primarily during the first four years of
the program to verify refiners' compliance with the temporary compliance option and various
hardship provisions, to ensure the two grades of highway diesel fuel remain segregated, and to
prevent misfueling of model year 2007 and later diesel vehicles.
1-4
-------
Chapter II: Health and Welfare
Chapter II: Health and Welfare Concerns and
Emissions Benefits
This chapter describes the public health and welfare concerns associated with the
pollutants emitted by heavy-duty vehicles, and the emission reductions that are expected to occur
as a result of today's action. Specifically, we present information on the ambient air pollution
situation that is likely to exist without this rule between 2007 and 2030 for ambient pollutants of
concern (e.g., ozone, particulate matter). In addition, this chapter presents information on the
expected emission reductions based on our projected national heavy-duty vehicle emissions with
and without the new standards for nitrogen oxides (NOx), non-methane hydrocarbons (NMHC),
particulate matter (PM), sulfur dioxide (SOx), carbon monoxide (CO), and air toxics.
A. Health and Welfare Concerns
When revising emissions standards for heavy-duty vehicles, the Agency considers the
effects of air pollutants emitted from heavy-duty vehicles on public health and welfare.1 As
discussed in more detail below, the outdoor, or ambient, air quality in many areas of the country
is expected to violate federal health-based ambient air quality standards for ground level ozone
and particulate matter during the time when this rule will take effect. In addition, some studies
have found public health and welfare effects from ozone and PM at concentrations that do not
constitute a violation of their respective NAAQS. Other studies have associated diesel exhaust
with cancer and noncancer health effects. Of particular concern is human epidemiological
evidence linking diesel exhaust to an increased risk of lung cancer. Emissions from heavy-duty
vehicles also contribute to a variety of environmental and public welfare effects such as
impairment of visibility/ regional haze, acid deposition, eutrophication/ nitrification, and POM
deposition. As described in more detail throughout this chapter, the standards finalized in this
rule will result in a significant improvement in ambient air quality and public health and welfare.
1. Health and Welfare Concerns Raised During Public Hearings
Throughout the five public hearings held around the country on the heavy-duty engine
and diesel fuel rule, the Agency received strong public support at each venue for increasing the
stringency of heavy-duty truck and bus emission standards, and for further controls on sulfur in
diesel fuel, in order to enable the necessary exhaust emission control. Public officials and
representatives of environmental, public health, or community-based organizations testified
regularly about the link between public health ailments, such as asthma and lung cancer, and air
pollution caused by diesel exhaust and particulate matter. A common theme revolved around the
n-i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
notion that since asthma is an incurable disease, it was of utmost importance to help reduce the
severity and frequency of attacks by reducing environmental triggers such as ozone, particulate
matter and diesel exhaust. Many testifiers expressed a strong sentiment that the public and the
auto industry have done their fair share to clean up cars and keep them clean through regular
inspections and maintenance, and it was time for the diesel truck industry to do the same.
In different ways, many noted that the impact of diesel soot is compounded by the fact
that it is discharged at street level where people live and breathe. A common complaint was the
close proximity of bus depots, transfer terminals, and heavily-trafficked roadways to homes and
apartment buildings, and in particular, to hospitals, playgrounds, and schools. Cyclists described
the stinging eyes and choking caused by breathing fumes from buses and trucks along city streets,
especially when trucks accelerated after stopping at an intersection. Two testifiers cited to health
studies that they said reported an association between those living in homes located near heavily-
trafficked streets and increased incidences of childhood asthma and leukemia.
By far the most poignant testimony was about how air pollution has impaired the health
and well-being of children. As our recent reviews of the NAAQS have documented, we heard
concerns expressed by citizens that childhood asthma accounts for 10.1 million missed school
days in the United States each year, and that asthma is the leading cause of hospitalizations in
New York City for children aged 0-14. At times, parents and their children testified together.
Kyle Damitz, accompanied by his mother, entered the following testimony:
I have come here today to tell you how our bad air affects kids like
me with asthma. . . . During these ozone days, I will almost always
have an asthma attack if I go outside. On good days, I take two pills
in the morning and three pills at bedtime. I do an IV treatment every
two weeks. On a bad asthma day, I take four pills in the morning,
more at lunch, and again, more at bedtime.. . . I came here today to
. . . ask you to help make breathing for kids with asthma easier. By
making the air cleaner, you are giving asthmatics a chance to breathe
easier. If our air was cleaner, I would be able to take less medicine,
be able to play outside more. If you make our air cleaner, I will be
able to live longer.
Many testifiers took the Agency to task for not acting sooner on heavy-duty vehicles.
Reacting to industry testimony requesting additional time to comply with the standards, testifiers
representing their constituents, their communities, environmental organizations, or themselves,
expressed the simple desire to be healthy as soon as possible. Some compared the annual human
cost of air pollution - quantified by thousands of hospitalizations, emergency room and doctors
visits, asthma attacks — with per vehicle cost of $1,600, and stated their belief that the
n-2
-------
Chapter II: Health and Welfare
regulations are cost effective. Others suggested that several billion dollars spent on improving
the environmental performance of trucks and buses was reasonable in light of the petroleum
industry's multi-billion dollar profits in the first quarter of 2000.
Major industries represented during these public hearings were the heavy-duty vehicle
engine manufacturers, the oil industry, and the commercial truckers. While each had a different
perspective, most supported the underlying intent of this rule to improve public health and
welfare, and some also supported the specific requirements as proposed. For those who objected
to the proposal, the main thrust of their concerns related to the stringency and public health
necessity of the new standards and the diesel fuel sulfur requirement. Largely in their written
comments, these industries raised questions about the need for additional reductions in order to
meet existing ozone and PM national ambient air quality standards and took exception with the
Agency's characterization of diesel exhaust as a human carcinogen at environmental levels of
exposure. Some industry commenters also challenged the Agency's reliance on public welfare
and environmental effects such as visibility impairment and eutrophication of water bodies
because the Agency had insufficiently quantified the benefits that would result from new
standards on heavy-duty vehicles and diesel fuel.
The following subsections present the available information on the air pollution situation
that is likely to exist without this rule for each ambient pollutant. We also present information
on the improvement that is expected to result from this rule. The Agency received a significant
number of comments on this section during the public hearings and in written comments from
interested parties. Where appropriate, comments are addressed in this section, but the majority
are addressed only in the Response to Comment document that accompanies this preamble.
Interested parties should refer to the Response to Comment document for the Agency's response
to their specific comments.
2. Ozone
This section reviews health and welfare effects of ozone and describes the air quality
information that forms the basis of our belief that ozone concentrations in many areas across the
country face a significant risk of exceeding the ozone standard between 2007 and 2030.
Information on air quality was gathered from a variety of sources, including monitored ozone
concentrations from 1997-1999, air quality modeling forecasts conducted for this rulemaking,
ozone modeling and information from states that have recently submitted attainment
demonstrations, and other state and local air quality information. Studies have found that ozone
concentrations at levels that do not exceed the 1-hour ozone standard are associated with impacts
on public health and welfare, and this section also summarizes those health effects and provides
some information about the potential for ozone at these moderate levels to exist during the time
period when this rule will take effect.
n-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
a. Health and Welfare Effects of Ozone and its Precursors
Ground-level ozone, the main ingredient in smog, is formed by complex chemical
reactions of VOC and NOx in the presence of heat and sunlight. Ozone forms readily in the
lower atmosphere, usually during hot summer weather. Volatile Organic Compounds are emitted
from a variety of sources, including motor vehicles, chemical plants, refineries, factories,
consumer and commercial products, and other industrial sources. Volatile organic compounds
also are emitted by natural sources such as vegetation. Oxides of Nitrogen are emitted largely
from motor vehicles, off-highway 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.a 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, relatively small amounts of NOx enable ozone to form rapidly when
VOC levels are relatively high, but ozone production is quickly limited by removal of the NOx.
Under these conditions, NOx reductions are highly effective in reducing ozone while VOC
reductions have little effect. Such conditions are called "NOx limited." Because the contribution
of VOC emissions from biogenic (natural) sources to local ambient ozone concentrations can be
significant, even some areas where man-made VOC emissions are relatively low can be NOx
limited.
When NOx levels are relatively high and VOC levels relatively low, NOx forms
inorganic nitrates but relatively little ozone. Such conditions are called "VOC limited." Under
these conditions, VOC reductions are effective in reducing ozone, but NOx reductions can
actually increase local ozone under certain circumstances. Even in VOC limited urban areas,
NOx reductions are not expected to increase ozone levels if the NOx reductions are sufficiently
large. The highest levels of ozone are produced when both VOC and NOx emissions are present
in significant quantities on clear summer days.
a Carbon monoxide also participates in the production of ozone, albeit at a much slower rate than most
VOC and NOx compounds.
H-4
-------
Chapter II: Health and Welfare
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.2'3
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 (6
to 8 hours), repeated exposure to ozone can cause inflammation of the lung, impairment of lung
defense mechanisms, and possibly irreversible changes in lung structure, which over time could
lead to premature aging of the lungs and/or chronic respiratory illnesses such as emphysema and
chronic bronchitis.
Children and outdoor workers 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 prolonged
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 the possibility of a
positive 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
n-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 foliage 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.
Volatile organic compound 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. The lexicologically significant VOCs emitted in
substantial quantities from HDVs are discussed in detail in Section II. A.4 below.
Besides their role as an ozone precursor, NOx emissions produce a wide variety of health
and welfare effects.4 5 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. These environmental impacts are discussed further in Sections
II. A.6 and II. A.7.
Elevated levels of nitrates in drinking water pose significant health risks, especially to
infants. Studies have shown that a substantial rise in nitrogen levels in surface waters are highly
correlated with human-generated inputs of nitrogen in those watersheds.6 These nitrogen inputs
are dominated by fertilizers and atmospheric deposition. Nitrogen dioxide and airborne nitrate
also contribute to pollutant haze, which impairs visibility and can reduce residential property
values and the value placed on scenic views.
H-6
-------
Chapter II: Health and Welfare
b. Photochemical Ozone Modeling
In conjunction with this rulemaking, the Agency performed a series of ozone air quality
modeling simulations for nearly the entire Eastern U.S covering metropolitan areas from Texas
to the Northeast.13 The model simulations were performed for five emissions scenarios: a 2007
baseline projection, a 2020 baseline projection and a 2020 projection with heavy-duty vehicle
controls, a 2030 baseline projection, and a 2030 projection with heavy-duty vehicle controls.
The model outputs from the 2007, 2020 and 2030 baselines, combined with current air
quality data, were used to identify areas expected to exceed the ozone NAAQS in 2007, 2020 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. The
impacts of the heavy-duty vehicle controls were determined by comparing the model results in
the future year control runs against the baseline simulations of the same year. This modeling
supports the conclusion that there is a broad set of areas with predicted ozone concentrations at
or above 0.125 ppm between 2007 and 2030 in the baseline scenarios without additional
emission reductions.
The air quality modeling performed for this rule was based upon the same modeling
system as was used in the Tier 2 air quality analysis, with the addition of updated inventory
estimates for 2007, 2020 and 2030. Consistent with a commitment expressed in the rule
proposal, the Agency released the emissions inventory inputs for, and a description of ozone
modeling, into the public record (docket number A-99-06), and also onto a website developed
expressly for this purpose, on a continuous basis as they were developed. Further discussion of
this modeling, including evaluations of model performance relative to predicted future air
quality, is provided in the air quality modeling technical support document (TSD).
/'. Modeling Methodology, Domains, and Episodes
A variable-grid version of the Urban Airshed Model (UAM-V) was utilized to estimate
base and future-year ozone concentrations over the eastern 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
b EPA also performed ozone air quality modeling for the western United States but, as described further in
the air quality technical support document, model predictions were well below corresponding ambient
concentrations. Given that model performance was degraded to the extent that the directional response of the model
to controls may be questionable, and considering the performance relative to that for the East and what is typically
expected out of such regulation modeling applications, it was determined that the results of western ozone modeling
were not relied on for this rule.
H-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 air quality modeling TSD, which is located in the docket for this rule.
The eastern modeling domain covered that portion of the U.S. east of west longitude 99
degrees. The model resolution was 36 km over the outer portions of the domain and 12 km in the
inner portion of the grids. A modeling study considered the sensitivity of regional modeling
strategies to grid resolution (LADCO, 1999). 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 Ozone Transport Assessment Group (OTAG)C modeling application also investigated the
effects of grid resolution on national/regional control strategies. 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 final rule.
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 many areas of the eastern U.S.d In general,
these episodes do not represent extreme ozone events but, instead, are generally representative of
ozone levels near local design values. Each of the six emissions scenarios (1996 base year, 2007
baseline, 2020 base, 2020 control, 2030 baseline, 2030 control) were simulated for the three
episodes.
/'/'. 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. This model provided needed data
at every grid cell on an hourly basis. These meteorological modeling results were evaluated
0 The OTAG modeling project is used as a benchmark for this heavy-duty vehicle and low sulfur diesel
fuel 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.
d 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.
H-8
-------
Chapter II: Health and Welfare
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 Air Quality Technical Support Document, which is available in the docket to
this final action on heavy-duty vehicles.
Hi. Model Performance Evaluation
The purpose of the Heavy Duty Engine 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 air quality modeling is that a model which
adequately replicates observed pollutant concentrations in the base year can be used to support
future-year policymaking (i.e., assessing the effects of altering the original emissions state).
As with previous regional photochemical modeling studies, the accuracy of the Heavy
Duty Engine 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 day-to-day variations in synoptic-scale ozone
fairly closely.
The values of the two primary measures of model performance, mean normalized bias
and mean normalized gross error, indicate that the Heavy Duty Engine modeling over the eastern
U.S. is generally as good as the grid modeling done to support the Tier 2/Sulfur rulemaking, as
shown in Table II. A-1. In turn, the performance of the Tier 2/Sulfur modeling was determined to
be as good or better than the detailed OTAG regional modeling, which has served as a relative
benchmark for acceptable performance from a regional photochemical grid model.
H-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table II.A-1. Comparison of eastern U.S. regional model performance statistics between
the Tier 2/Sulfur modeling and the Heavy Duty Engine modeling. The units are
percentages.
Mean Normalized
Bias
Domain
Midwest
Northeast
Southeast
Southwest
Tier 2
June 95
-10
-11
-17
-4
+2
Tier 2
July 95
-6
-13
-9
+4
+8
Tier 2
August 95
+2
+7
-9
+7
+6
HDE
June 95
-13
-15
-20
-7
+1
HDE
July 95
-11
-16
-11
-3
+3
HDE
August 95
+5
+10
-15
+12
+11
Mean Normalized
Gross Error
Domain
Midwest
Northeast
Southeast
Southwest
Tier 2
June 95
24
24
27
20
24
Tier 2
July 95
24
26
22
24
27
Tier 2
August 95
23
22
24
22
24
HDE
June 95
22
22
27
18
22
HDE
July 95
23
24
23
21
24
HDE
August 95
24
22
24
25
27
Mean normalized bias is defined as the average difference between model predictions and
observations (paired in space and time) normalized by the magnitude of the observations. Mean
normalized gross error is defined as the average absolute difference between model predictions
and observations, paired in space and time, normalized by the magnitude of the observations.
EPA guidance on local ozone attainment demonstration modeling (not the purpose of the Heavy
Duty Engine modeling) suggests biases be no greater than 15 percent and errors be no greater
than 35 percent.
Model performance statistics for the Heavy Duty Engine base case simulations were
calculated for the entire grid and numerous smaller sub-grids. The model performance
evaluation consisted solely of comparisons against ambient surface ozone data. There was
insufficient data available in terms of ozone precursors or ozone aloft to allow for a more
complete assessment of model performance. From a regional perspective, the model generally
11-10
-------
Chapter II: Health and Welfare
underestimated observed ozone values (greater than 60 ppb) for the June and July episodes, but
predicted higher than observed amounts for the August episode. Errors average about 22-24
percent. The general tendency of the model, as discussed above, is to underestimate observed
ozone concentrations. This tendency should lead to a conservative estimate of future-year air
quality need.
c. Results of Photochemical Ozone Modeling
The determination that an area is at risk of exceeding the ozone standard in the future was
made for all areas with current design values greater than or equal to 0.125 ppm (or within a 10
percent margin) and with modeling evidence that exceedances will persist into the future. The
following sections provide background on methods for analysis of attainment and maintenance.
Those interested in greater detail should review the Air Quality Modeling Technical Support
Document, which is available in the docket to this rule.
/'. 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 0.125 ppm.
For a county, the design value is the highest design value from among all the monitors
with valid design values within that county. Jf 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
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 and are not considered in this analysis. For the purposes of defining the current design
value of a given area, the 1997-1999 design values were chosen to provide the most recent set of
air quality data for identifying areas likely to have an ozone problem in the future. The 1997-
1999 design values are listed in the Air Quality Modeling Technical Support Document, which is
available in the docket to this rule.
11-11
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
/'/'. Method for Projecting Future Exceedances
The exceedance method was used for interpreting the future-year modeling results to
determine where there is an appreciable risk of future nonattainment in the 2007, 2020 and 2030
Base and Control Cases. 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 greater than or equal to 0.125
ppm. Areas with current measured violations of the one-hour ozone standard (or within a 10
percent margin), and one or more model-predicted exceedances, are projected to have the
potential for a nonattainment problem in the future.
/'/'/'. Areas at Risk of Future Exceedances Based on Ozone Predictive Modeling
The Agency conducted ozone modeling based on inventories developed with and without
reductions from this rulemaking for three future years: 2007, 2020, and 2030. The year 2007 was
chosen because it is the first year of implementation for the new standards adopted in today's
action. It is also the year that ten major urban areas with a history of persistent and elevated
ozone concentrations must demonstration attainment. The year 2020 was chosen because of its
relevance to the ability of many areas to maintain the ozone standard. The year 2030 was chosen
to provide the reader with a full sense of the reductions in ambient ozone concentrations likely to
be achieved once the existing fleet of heavy-duty vehicles is replaced with vehicles meeting the
standards finalized today.
The predictive ozone modeling is based on emissions inventories which have been
updated and improved subsequent to the Agency's recent rulemaking on light duty vehicles and
gasoline sulfur, also known as Tier 2. Areas presented in Table n.A-2 have 1997-1999 air
quality data indicating violations of the 1-hour ozone NAAQS, or are within 10 percent of the
standard, and are predicted to have exceedances in 2007, 2020 or 2030 without the reductions
from this rule. Table II.A-2 lists those metropolitan areas with predicted exceedances of the 1-
hour ozone standard in 2007, 2020, or 2030 without emission reductions from this rule (i.e., base
cases). These areas are listed in columns with a "b" after the year (e.g., 2020b). Table n.A-2
also lists those metropolitan areas with predicted exceedances of the 1-hour ozone standard in
2020 and 2030, with emission reductions from this rule (i.e., control case). These areas are listed
in columns with a "c" after the year (e.g., 2020c). An area was considered likely to have future
exceedances if exceedances were predicted by the model, and the area is currently violating the
1-hour ozone standard, or is within ten percent of violating the 1-hour ozone standard.
Photochemical ozone modeling conducted for this rulemaking was based in part on
updated national emissions inventories for all sources. National emission trends for NOx predict
a significant decline from 1996 to 2007, a leveling off of the downward trend between 2007 to
2020, and an increase in NOx inventories from 2020 to 2030. By 2030, national NOx levels are
11-12
-------
Chapter II: Health and Welfare
estimated to reach levels that are within ten percent of 2007 levels. Predictions of national VOC
emissions indicate a reduction from 1996 to 2007, followed by an increase between 2007 and
2030 resulting in 2030 levels that are estimated to be 10 percent greater than VOC emissions
levels in 2007. In metropolitan ozone nonattainment areas, such as Charleston, Chicago and
Houston, NOx or VOC emissions in 2030 are predicted to reach or exceed 2007 levels. These
estimated national and metropolitan area emissions inventories of ozone precursors are consistent
with the conclusions reached by analysis of ozone modeling conducted for this rule that
additional reductions are needed in order to enable areas to reach and maintain attainment of the
ozone standard between 2007 and 2030.
In addition, the substantial reductions from today's rule will greatly lower ozone
concentrations which will help federal and State efforts to bring about attainment with the current
1-hour ozone standard. As described in the Air Quality Modeling Technical Support Document
for this rule, EPA performed regional scale ozone modeling for the Eastern U.S. to assess the
impacts of the controls in this rule on predicted 1-hour ozone exceedances. The results of this
modeling were examined for those 37 areas in the East for which EPA's modeling predicted
exceedances in 2007, 2020 and/or 2030 and current 1-hour design values are above the standard
or within 10 percent of the standard. The results for these areas combined indicate that there will
be substantial reductions in the number of exceedances and the magnitude of high ozone
concentrations in both 2020 and 2030 due to this rule. The modeling also indicates that without
the rule exceedances would otherwise increase by 37 percent between 2020 and 2030 as growth
in emissions offsets the reductions from Tier 2 and other current control programs.
For all areas combined, the rule is forecast to provide a 33 percent reduction in
exceedances in 2020 and a 38 percent reduction in 2030. The total amount of ozone above the
standard is expected to decline by nearly 37 percent in 2020 and 44 percent in 2030. Also, daily
maximum ozone exceedances are lowered by 5 ppb on average in 2020 and nearly 7 ppb in 2030.
The modeling forecasts an overall net reduction of 39 percent in exceedances from 2007, which
is close to the start of this program, to 2030 when controls fully in place. In addition, the results
for each individual area indicates that all areas are expected to have less exceedances in 2030
with the HDV controls than without this rule.
11-13
-------
Table II.A-2. Eastern Metropolitan Areas with Modeled Exceedances of the 1-Hour Ozone
Standard in 2007, 2020, or 2030 With and Without Emission Reductions from this Rule
VISA or CMSA / State
\tlanta, GA MSA
Barnstable-Yarmouth, MA MSA *
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA *
Biloxi-Gulfport-Pascagoula, MS MSA *
Birmingham, AL MSA
Boston- Worcester-Lawrence, MA CMSA
Charleston, WV MSA *
Charlotte-Gastonia-Rock Hill, NC MSA
Chicago-Gary-Kenosha, IL CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA *
Cleveland-Akron, OH CMSA *
Detroit-Ann Arbor-Flint, MI CMSA
Grand Rapids-Muskegon-Holland, MI MSA*
lartford, CT MSA
rlourna, LA MSA *
louston-Galveston-Brazoria, TX CMSA
luntington- Ashland, WV-KY-OH MSA
.ake Charles, LA MSA *
.ouisville, KY-IN MSA
Vlacon, GA MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
sfew London-Norwich, CT-RI MSA
sfew Orleans, LA MSA *
^few York-Northern NJ-Long Island, NY-NJ-
CT-PA CMSA
Norfolk- Virginia Beach-Newport News, VA-
s[C MSA *
3rlando, FL MSA *
-•ensacola, FL MSA
Philadelphia- Wilmington-Atlantic City, PA-
s[J-DE-MD CMSA
Drovidence-Fall River- Warwick, RI-MAMSA*
Richmond-Petersburg, VA MSA
5t. Louis, MO-IL MSA
Tampa-St. Petersburg, FL MSA *
Washington-Baltimore
Total number of areas
Copulation
2007b 2020b 2020c 2030b 2030c pop (1999)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
37
91.2
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
35
90.6
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
32
88.5
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
36
90.8
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
32
87.8
3.9
0.2
0.6
0.4
0.2
0.3
0.9
5.7
0.3
1.4
8.9
1.9
2.9
5.4
1.1
1.1
0.2
4.5
0.3
0.2
1
0.3
1.1
1.7
1.2
0.3
1.3
20.2
1.6
1.5
0.4
6
1.1
1
2.6
2.3
7.4
91.4
* These areas have registered 1997-1999 ozone concentrations within 10 percent of standard.
11-14
-------
Chapter II: Health and Welfare
The inventories that underlie the photochemical ozone modeling conducted for this
rulemaking included reductions from all current or committed federal, State and local controls
and, for the control case, the Heavy-Duty Vehicle and Diesel Fuel Sulfur Program itself. It did
not did not attempt to examine the prospects of areas attaining or maintaining the ozone standard
with possible future controls (i.e., controls beyond current or committed federal, State and local
controls). Therefore, Table HA-2 should be interpreted as indicating what areas are at risk of
ozone violations in 2007, 2020 or 2030 without federal or State measures that may be adopted
and implemented after this rulemaking is finalized. We expect many of the areas listed in Table
II. A-2 to adopt additional emission reduction programs, but the Agency is unable to quantify or
rely upon future reductions from additional State programs since they have not yet been adopted.
The Agency recently redesignated Cincinnati-Hamilton, OH-KY-IN to attainment on
June 19, 2000. This determination is based on four years of clean air quality monitoring data
from 1996 to 1999 (1999 data was not considered in the Tier 2 air quality analysis or the proposal
for this rulemaking), and a downward emissions trend. In today's action, Cincinnati-Hamilton is
considered to have some risk of registering exceedances of the 1-hour ozone standard during the
time period when the HD vehicle standards would take effect. This determination is based on air
quality monitoring analysis and 1999 data with concentrations within 10 percent of the standard.
Given these circumstances, the risk of future exceedances occurring in the Cincinnati-Hamilton
area is most prevalent in the time period beyond the end date of Cincinnati's proposed 10-year
maintenance plan (i.e., after 2010). As discussed in more detail in the relevant portions of the
response to comment document for the Cincinnati-Hamilton attainment determination, any
emissions and ozone modeling system used to predict future ozone involves approximations and
uncertainties, and are best treated as indicators of risk rather than absolute forecasts. Thus a
determination made in this rule that there is some risk of future exceedances during the relevant
time period is not inconsistent with EPA approval of Cincinnati's redesignation to attainment,
and its approval of Cincinnati's 10-year maintenance plan.7
d. Ozone Modeling and Analysis in 1-Hour State Implementation Plan
Submittals and Other Local Ozone Modeling
/'. Overview
We have compared and supplemented our own ozone modeling with other modeling
studies, submitted to us as state implementation plan (SIP) revisions, or brought to our attention
through our consultations with states on SIP revisions that are in development. The ozone
modeling in the SIP revisions has the advantage of using emission inventories that are more
specific to the area being modeled, and of using meteorological conditions selected specifically
for each area. Also, the SIP revisions included other evidence and analysis, such as analysis of
air quality and emissions trends, observation based models that make use of data on
concentrations of ozone precursors, alternative rollback analyses, and information on the
11-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
responsiveness of the air quality model. For some areas, we decided that the predictions of future
ozone concentrations from our modeling were less reliable than conclusions that could be drawn
from this additional evidence and analysis. For example, in some areas our episodes did not
capture the meteorological conditions that have caused high ozone, while local modeling did so.
Thus, these local analyses are considered to be more extensive than our own modeling for
estimating whether there would be NAAQS nonattainment without further emission reductions,
when interpreted by a weight of evidence method which meets our guidance for such modeling.
We have reviewed and recently proposed action on SIP submissions from 13 states and
the District of Columbia covering 10 serious and severe 1-hour ozone nonattainment areas. We
received these submissions as part of the three-phase SIP process described 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
conducted for this rule. We have also considered ozone modeling submitted as part of an
attainment date extension requests for Beaumont-Port Arthur, TX, and Dallas/Fort Worth, TX.
Finally, we have considered information in the most recent SIP submittal from California for the
South Coast Air Basin. Table II. A-3 lists the areas involved, whether the modeling indicates
attainment without further reductions and the Federal Register citation for our proposed action if
applicable. This section discusses the background for the submissions and our findings base on
them.
It is important to note that the information contained in this section on current and future
ozone nonattainment is current as of December 1, and there may have been recent developments
in some areas that are not incorporated here.
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. 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, we are proposing to find that the selection
of episode days met our guidance. The local modeling also may make use of location-specific
emission data and control programs than is practicable to include in regional-scale modeling by
EPA as described above.
The SIP submissions for the 13 states and the District of Columbia covering 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
11-16
-------
Chapter II: Health and Welfare
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 must 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.
11-17
-------
Table II.A-3. Nonattainment Areas For Which EPA Has 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
MS As)*
New York City*
Philadelphia*
Baltimore*
Washington, D.C.*
Atlanta*
Houston*
Chicago*
Milwaukee*
Benton Harbor
Grand Rapids
Dallas
Beaumont-Port Arthur
South Coast Air Basin
Affected
States
MA
CT
NY, CT,
NJ
PA, NJ,
DE,MD
MD
MD, VA,
D.C.
GA
TX
IL, IN
WI
MI
MI
TX
TX
CA
Attainment
Date
2003
(Requested
Extension)
2007
(Requested
Extension)
2007
2005
2005
2005
(Requested
Extension)
2003
(Requested
Extension)
2007
2007
2007
N/A
N/A
2007
(Requested
Extension)
2007
(Requested
Extension)
2010
Indicates Attainment Without "Further
Reductions"
Yes
Yes, but CT's extension request is based on
Greater CT's inabiity to attain because it is
affected by transport from the NYC
Metropolitan Area.
No
No
No
Yes
No
No
Revised SIP attainment modeling accounts
for reductions from this rule.
Revised SIP attainment modeling accounts
for reductions from this rule.
Revised SIP attainment modeling accounts
for reductions from this rule.
Revised SIP attainment modeling accounts
for reductions from this rule.
No
EPA weight of evidence proposed approval
based in part of reductions from this rule.
No
* Proposed for Action in December 16, 1999 Federal Register (64 FR 70318).
11-18
-------
Chapter II: Health and Welfare
/'/'. 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.8 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
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 is not
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 conditions that describe atmospheric conditions and emissions
11-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 at or 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
ozone concentrations for each modeled day6 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.
The initial, "ramp-up" days for each episode are excluded from this determination.
11-20
-------
Chapter II: Health and Welfare
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
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. In most of the local
ozone modeling in these SIP revisions, upwind NOx reductions have been assumed to occur
through implementation of 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/ For the purpose of this rule,
EPA has incorporated the emission reductions from the NOx SIP Call into its evaluation of
whether further reductions are needed. Absent such reductions, the need for additional
reductions is even greater.
f 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.
11-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Hi. Conclusions from the Local Modeling in SIP Submittals
As discussed previously, we have recently been able to review ozone modeling and other
evidence on the likelihood of attainment for ten major metropolitan nonattainment areas. The
local modeling only addresses the current and requested attainment date in each area. For the
areas involved, these dates fall between 2003 and 2007. The State and local ozone modeling
therefore does not address attainment prospects beyond 2007. In December, 1999, the Agency
proposed to approve attainment demonstrations for these 10 areas, in some cases with, and in
others without, a requirement that states adopt additional measures. More recently, we proposed
to approve an attainment demonstration for St. Louis.
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. 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. 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.
These State-specific conclusions are not final and we are not making them final via this
rule on heavy-duty vehicles. 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 below in analyzing the need for additional emission reductions in these areas.
For the New York Metro area, Philadelphia, Baltimore, and Houston nonattainment
areas, the EPA has proposed to determine that additional emission reductions beyond those
provided by the SIP submission are necessary for attainment. A portion of that reduction will be
achieved by federal actions, such as the Tier 2/Sulfur program. In the case of Washington DC,
the Tier 2/Sulfur program will provide additional emission reductions needed to keep local
emissions in 2005 at or below the levels needed to attain. However, as discussed subsequently,
there is still a risk of future nonattainment in the Washington, DC area in 2007 and later due to
inherent uncertainties in air quality forecasting and future exceedances predicted by Tier 2 air
quality modeling.
As a 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, Greater CT (Hartford and New
London metropolitan areas), Southeast Desert, and Milwaukee, WI areas demonstrated
attainment through the control measures contained in the submitted attainment strategy. Illinois,
Indiana, and Wisconsin must submit further SIP revisions, including updated modeling for the
Chicago and Milwaukee nonattainment areas by December 2000. For these areas, the updated
11-22
-------
Chapter II: Health and Welfare
regional ozone modeling conducted by the Lake Michigan Air Directors Consortium on behalf of
the states relies in part on reductions from this rulemaking.8 Thus, the 2007 attainment
demonstrations for these areas will be based in part on reductions from this rule.
Greater Connecticut and the Southeast Desert are subject to transport from upwind areas
that need additional reductions in order to reach attainment in 2007 (New York City), or 2010
(South Coast Air Basin).h If attainment is not achieved by New York and South Coast, it is
unlikely that Greater Connecticut and the Southeast Desert will achieve attainment. Since New
York and the South Coast need further reductions that this rule will in part satisfy, reductions
from this rule will also assist downwind nonattainment areas such as Greater Connecticut and the
Southeast Desert to reach attainment in 2007 and maintain the ozone NAAQS in future years.
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. All of the measures in that strategy -
as well as the measures identified as "additional measures" that were not modeled but needed for
attainment in the weight of evidence analyses - have been adopted. 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 (i.e., reductions in
the modeled strategy) were considered. The modeling for Atlanta assumed implementation of
the NOx SIP Call outside the local modeling domain. The NOx reductions relied on in the
Atlanta SIP local modeling domain are slightly greater than the NOx reductions that are expected
to be achieved under the NOx SIP call.
The specific reasons for reaching these conclusions are explained in the individual
Federal Register notices.
g Lake Michigan Air Directors Consortium. Midwest Subregional Modeling: 1-Hour Attainment
Demonstration - Tier II/Low Sulfur Controls. November 8, 1999.
h EPA approved the South Coast's "additional measures" relying on new technologies under Clean Air Act
section 182(e) in 1995. 60 FR 43379 (August 21, 1995). Emissions reduction shortfall was quantified at 60 tons
per day (tpd) for NOx, and 79 tps of VOC. These measures are discussed in the 1994 California SIP (Volume I, p.
133,1-35 and Volume II, p. 1-29,1-31). In addition, EPA found shortfalls remaining in the mobile source emissions
reductions needed for attainmetn of the 1-hour ozone standard in the South Coast (64 FR 39923-27, July 23, 1999).
11-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
iv. Other Local Ozone Modeling and Ozone Nonattainment Prospects
The photochemical ozone modeling conducted for this rule did not predict exceedances
for a number of areas for which other available information, such as local ozone modeling,
inventory and air quality trends, demonstrates a risk of future exceedances between 2007 and
2030. Table HA-4 lists these eight areas. These eight areas will be discussed in subsequent
sections along with the 37 areas with predicted ozone exceedances in 2007, 2020, or 2030 (Table
II.A-2).
We have received ozone modeling for the Beaumont-Port Arthur nonattainment area.9
Beaumont-Port Arthur is a moderate ozone nonattainment area which continues to have
concentrations above levels of the 1-hour ozone NAAQS. 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). Our Proposed Action indicated that we would approve the attainment date
extension request if Texas met several necessary conditions, one of which was submission of an
approvable demonstration of attainment showing attainment by that date. Texas submitted
revisions to the Beaumont/Port Arthur SIP on November 15, 1999 and April 28, 2000. The
modeling analysis in these SIP revisions indicates that after implementation of the State's
adopted control strategy nonattainment continues. The State supplied additional evidence to
indicate that the area is likely to attain the Standard based on EPA's weight of evidence
guidance. We have considered the additional reductions from this rule as part of our proposed
action on the State's attainment demonstration.
Texas also submitted a modeling analysis and attainment date extension request for the
Dallas-Fort Worth metropolitan area on April 28, 2000.10 The State has requested to extend the
attainment date to 2007. The SIP revision includes the State's adopted control strategy and a
modeling analysis and weight of evidence demonstration. Our preliminary finding is that the
combined modeling and weight of evidence analysis has little if any margin of safety for
demonstrating attainment. The Agency's expects that any future proposed determination of the
SIP attainment demonstration submitted by Dallas, TX, using meteorology conditions and other
inputs selected to be locally applicable, would rely in part on reductions from this rule.
11-24
-------
Chapter II: Health and Welfare
Table II.A-4. Areas With Some Risk of Ozone Violations
between 2007 and 2030 Based on Information Other Than Predictive Ozone Modeling
Metropolitan Areas
Dallas, TX
South Coast Air Basin, CA
(Los Angeles-Riverside-San
Bernardino)
San Diego, CA
Southeast Desert, CA
Sacramento, CA
Ventura County, CA
San Joaquin Valley, CA
San Francisco, CA
8 areas
Basis for Need of Additional
Reductions
Agency expects to rely on reductions
from this rule in its proposed weight of
evidence determination for the Dallas
SIP.
Emission reduction shortfalls (NOx and
VOC) identified in current SIP.
Transport from South Coast Air Basin.
Transport from South Coast Air Basin,
significant ozone levels and number of
exceedance days (1997-1999).
Significant ozone levels and number of
exceedance days (1997-1999).
Transport from South Coast Air Basin
Area needs to revise SIP for bump-up to
severe (2005); significant ozone levels
and number of exceedance days (1997-
1999)
Area needs to revise its SIP, significant
ozone levels and number of exceedance
days (1997-1999).
1999 Population
(in millions)
4.9
16.0
2.8
0.5
1.7
0.7
3.2
6.9
36.8
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. 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. Nine areas in California currently designated as nonattainment (and two counties
currently designated as being in attainment) with a population of approximately 32 million have
11-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1997-1999 design values above the 1-hour ozone NAAQS. Seven of the nonattainment areas
have approved SIPs, including demonstrations of attainment for their required date. Emissions
reductions expected from federal programs, such as the Tier 2/Sulfur rule, represent only a small
fraction of the emission reductions needed in the South Coast to attain the NAAQS.
Ozone levels in the South Coast Air Basin have declined over the last two decades, but
this area continues to register some of the highest ozone concentrations and the greatest number
of exceedance days in the nation. In the three year period from 1997 to 1999, the South Coast
recorded a peak ozone level of 0.211 ppm and averaged 39 days above the 1-hour ozone
standard. The South Coast has an approved SIP, but it contains shortfalls that must be filled if
the area is to reach attainment in 2010. The South Coast generates ozone and ozone precursors
that affect the air quality and attainment prospects of downwind areas such as Ventura County,1
San Diego, and the Southeast Desert.
The transport of ozone and its precursors from the South Coast to downwind areas such
as the San Diego, Ventura County and the Southeast Desert has been established by the
California Air Resources Board.j In addition to receiving transport from the South Coast, the
Southeast Desert registered a significant peak ozone concentrations of 0.170 ppm and exceeded
the 1-hour ozone standard 24 days on average between 1997-1999. While these areas may have
earlier attainment dates, their ability to attain the ozone NAAQS depends in part on attainment
by the South Coast. Reductions from this action will provide NOx and VOC reductions needed
to help fill shortfalls identified in the South Coast's approved SIP. By extension, since
attainment in the South Coast would assist efforts to reach attainment in Ventura and the
Southeast Desert by their respective deadlines (2005 and 2007), these two areas, along with San
Diego, are also dependent on South Coast reductions associated with today's action to maintain
attainment in the future.
We expect that California will be submitting one or more revisions since it appears that
one unclassified nonattainment area with a 2000 attainment deadline and one serious
classification nonattainment area in California with an attainment deadline of 1999 have not met
that date. These areas are San Francisco and the San Joaquin Valley. San Francisco had a
violation in its attainment year (2000), which may require the area to submit a revised attainment
1 Assessment and Mitigation of the Impacts of Transported Pollutants on Ozone Concentrations within
California, California Air Resources Board, June, 1990. This photochemical grid model analysis found that on
some days emissions from the South Coast Air Basin contribute in a significant way to ozone concentrations in
Ventura County.
3 Regulations on "Rulemaking on the Assessment of the Impacts of Transported Pollutants on Ozone
Concentrations in California" were approved by the California Office of Administrative Law on August 27, 1997.
Note that for purposes of the CARD's transport assessment, the Southeast Desert is divided into two air basins: the
Mojave Desert Air Basin and the Salton Sea Air Basin (Title 17, CA Code of Regulations, 60104, 60109, 60114).
11-26
-------
Chapter II: Health and Welfare
plan. From 1997 to 1999, the area registered a peak ozone value of 0.139 ppm and averaged
about 3 exceedance days per year. San Francisco does not currently have an attainment
classification.
San Joaquin has had too many exceedances to be eligible for an extension and EPA has
proposed to bump-up the area to severe classification with a 2005 attainment date. From 1997 to
1999, the area recorded a peak ozone level of 0.161 ppm and exceeded the ozone standard 14
days on average. In fact, since 1991, the area has consistently registered peak ozone levels of
around 0.160 ppm. The magnitude and persistence of peak ozone levels in San Joaquin Valley is
an important factor to consider when attempting to assess future attainment prospects.
Sacramento has an approved SIP. However, between 1997-1999, the area registered a
peak ozone concentration of 0.148 ppm and five exceedances days per year on average. These
ozone levels and number of exceedance days suggest that this area has some risk of not attaining
the standard by its attainment date.
San Diego is subject to transport from the South Coast Air Basin, and has registered
significant ozone levels from 1997-1999. This area was granted a 1-year attainment date
extension under the provisions of CAA section 181(a)(5), and appears to be eligible for a second
1-year attainment date extension based on clean air in 2000.
v. Areas At Risk of Exceeding the 1-Hour Ozone Standard in the Future
This section collects the information previously presented on the attainment prospects of
areas across the nation based on both photochemical ozone modeling and other local factors such
as magnitude and persistence of ozone exceedances, emissions inventory trends, local modeling,
SIP status, and transport from areas with later attainment dates.k The Agency's conclusions
about the risk of future nonattainment is provided in Table U.A-6, which is separated into two
broad groups: (1) those areas with attainment dates in 2007 or 2010 that will benefit from
reductions from this rule to attain and maintain the standard; and (2) those areas with attainment
dates prior to 2007 that will benefit from reductions from this rule to maintain the standard after
their attainment dates. Because ozone concentrations causing violations of the 1-hour ozone
standard are well established to endanger public health and welfare, this indicates that it is
appropriate for the Agency to set new standards for heavy-duty vehicles.
k In the proposal, we relied on photochemical ozone modeling performed for recently promulgated
standards on light duty vehicles, or Tier 2. The results presented in this final rulemaking for heavy-duty vehicles
and diesel fuel are largely consistent with the findings presented in the proposal, with small differences due to
updated emissions inventories. As stated in the proposal, the ozone modeling methodologies used in the proposal
and presented here in the final rule are identical.
11-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
vi. Areas with 2007 or 2010 Attainment Deadlines
The Clean Air Act requires states to submit a SIP to provide for attainment of the 1-hour
ozone standard which includes a demonstration of attainment (including air quality modeling) for
their nonattainment areas, as well as emission control measures needed to attain by the
attainment date. Once the attainment date arrives, areas that have not attained the standard based
on monitoring data are subject to applicable provisions of the Clean Air Act, including the
possibility of being required to adopt additional emission control measures. Areas that have
attained the standard have the option of applying for redesignation to attainment status, which
can permit adjustments in the emission control program.
Table II.A-5 identifies ten ozone nonattainment areas with attainment dates of 2007 or
2010. These ten areas are also listed on the top section of Table II.A-6, which is located at the
end of this subsection. Each of these areas will need additional reductions to attain the ozone
standard, and will also be able to rely on additional reductions from today's action in order to
maintain the standard. There are specific emission reduction shortfalls in attainment SIPs
submitted for New York, Houston and the South Coast Air Basin based on the local ozone
modeling and other evidence. The Agency has not identified a shortfall in the attainment
demonstrations submitted by Greater Connecticut (Hartford and New London, CT), but we have
proposed to approve an extension date to 2007 based on Greater Connecticut being unable to
attain because it is affected by transport from the New York metropolitan area. Transport of
ozone and its precursors from the South Coast to the Southeast Desert, San Diego, and Ventura
County hinders the ability of these areas to attain the standard. There is some risk that New York
will fail to attain the standard by 2007, and thus a transferred risk that Connecticut will also fail.
A similar situation exists in Southern California, where attainment of the South Coast is a
precondition of the ability of three downwind areas — Southeast Desert, San Joaquin Valley, and
Ventura County to reach attainment by their respective attainment dates. Additional reductions
from this rule will assist New York and Greater Connecticut, and the South Coast and its
downwind nonattainment areas, in reaching the standard by each areas' respective attainment
date, and maintaining the standard from attainment to 2030.
Chicago and Milwaukee originally submitted modeling which did not indicate a need for
additional local reductions. However, required, updated modeling for these two areas relies in
part on reductions from this rule.1 Moreover, the ozone modeling for this rulemaking predicted
exceedances in Chicago and Milwaukee in 2007, 2020 and 2030.
1 Technical Support Document, Midwest Subregional Modeling: 1-Hour Attainment Demonstration for
Lake Michigan Area and Emissions Inventory, Illinois Environmental Protection Agency, Indiana Department of
Environmental Management, Michigan Department of Environmental Quality, Wisconsin Department of Natural
Resources, September 27, 2000, at 14 and at 8.
11-28
-------
Chapter II: Health and Welfare
Dallas and Beaumont Port-Arthur, TX have requested attainment date extensions to 2007
on the grounds that 2007 is the attainment date for Houston and that local air quality is affected
by transport from Houston. We have proposed to grant an extension to Beaumont-Port Arthur.
We have not yet proposed any action on Dallas. The State of Texas has developed an attainment
plan for Dallas, which is a precondition for granting extensions based on transport. In a recently
proposed action on attainment SIP submitted for Beaumont-Port Arthur, we have proposed
approval based in part on the Agency's weight of evidence determination that included in its
consideration expected reductions from today's action. We expect that we will also consider
reductions from today's action in our action on the Dallas/Fort Worth plan. Furthermore, EPA's
ozone modeling indicated exceedances in Beaumont Port-Arthur in 2007. Although there were
no exceedances predicted in the future-year scenarios for Dallas in the modeling by EPA, the
episodes used by the state in their local modeling did predict future-year exceedances. We do
currently believe these two areas are likely to violate the NAAQS between 2007 and 2030,
without more emission reductions in the local areas, and/or from the upwind Houston area,
and/or from today's action.
The Los Angeles (South Coast Air Basin) ozone attainment demonstration is fully
approved, but it is based in part on reductions from new technology measures that have yet to be
identified. The 2007 attainment demonstration for the Southeast Desert area is also approved.
However, a transport situation exists between the Southeast Desert areas and the South Coast Air
Basin, such that attainment in the Southeast Desert depends on progress in reducing ozone levels
in the South Coast Air Basin.
11-29
-------
Table II.A-5. Metropolitan Areas With Established or Requested 2007 or 2010
Attainment Deadlines
Metropolitan Area
New York City, NY-
NJ-CT
Houston, TX
Hartford, CT
New London, CT
Chicago, IL-IN
Milwaukee, WI
Dallas, TX
Beaumont-Port
Arthur, TX
Los Angeles, CA
Southeast Desert, CA
Attainment Dates
2007
2007
2007 (requested
extension)
2007 (requested
extension)
2007
2007
2007 (requested
extension)
2007 (requested
extension)
2010
2007
Future Attainment
Prospects
VOC and NOx
Shortfall
NOx Shortfall
Contingent on New
York Attainment
Contingent on New
York Attainment
Updated modeling
relies in part on
reductions from this
rule.
Updated modeling
relies in part on
reductions from this
rule.
Local modeling
shows nonattainment
in 2007
Local modeling
shows nonattainment
in 2007
Approved attainment
demonstration, but
needs additional
reductions to attain
Approved attainment
demonstration, but
contingent on South
Coast Attainment
10 Metropolitan Areas Total Population (in millions)
Metropolitan Area
1999 Population
(in millions)
20.0
4.5
1.1
0.3
8.9
1.6
4.9
0.4
16
0.5
58.4
11-30
-------
Chapter II: Health and Welfare
Therefore, these 10 nonattainment areas with about 58 million people will need to rely in
part on the reductions from today's action to attain the 1-hour ozone standard by 2007 or 2010,
and maintain the standard from 2007/2010 and 2030. We expect to rely in part on these
reductions in reaching our final conclusion as to whether each area for which we have reviewed
an attainment demonstration is more likely than not to attain on its respective date, whether or
not the State formally relies on these reductions as part of its strategy to fill the identified
shortfall in its attainment demonstration, if any. This is especially true for those areas that have
shortfalls in their attainment demonstrations, or that have air quality modeling that suggests
additional reductions are needed. The NOx and VOC reductions in the early years of this
program may prove to be a critical part of a range of actions necessary for these areas to
overcome their shortfalls and reach attainment.
The emission reductions from this rule will also help these areas reach attainment at
lower overall cost, with less impact on small businesses. Following implementation of new
controls for regional NOx reductions, states will have already adopted emission reduction
requirements for most large sources of VOC and NOx for which cost-effective control
technologies are known and for which they have authority to control. Those that must adopt
measures to complete their attainment demonstrations therefore will have to consider their
remaining alternatives. Many of the alternatives that states may consider could be very costly,
and the emissions impact from each additional emissions source subjected to new emissions
controls could be considerably smaller than the emissions impact of the standards being proposed
today. Therefore, the emission reductions from the standards we are finalizing today will ease
the need for states to find first-time reductions from the mostly smaller sources that have not yet
been controlled, including area sources that are closely connected with individual and small
business activities. The emission reductions from the standards being finalized today will also
reduce the need for states to seek even deeper reductions from large and small sources already
subject to emission controls.
The Southeast Desert has an approved attainment demonstration, and we have proposed
to approve attainment demonstrations in some of the other nine areas without additional emission
reductions from local measures. Even if all shortfalls were filled for each area, this would not
mean that there is no danger that ozone levels in these areas will exceed the NAAQS, in the
absence of today's action Agency approval of an attainment demonstration generally indicates
our belief that a nonattainment area is reasonably likely to attain by the applicable attainment
date with the emission controls in the SIP. However, such approval does not indicate that
attainment is certain. Moreover, no ozone forecasting is 100 percent certain, so attainment by
these deadlines is not certain, even though we believe it is more likely than not. There are
significant uncertainties inherent in predicting future air quality, such as unexpected economic
growth, unexpected VMT growth, the year-to-year variability of meteorological conditions
conducive to ozone formation, and modeling approximations. Ozone formation is highly
dependent on local weather conditions. In fact, the variability in observed ozone due to
11-31
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
meteorology can be larger than the ozone reductions yielded from a significant emission
reduction."1 There is at least some risk in each of these ten areas that even assuming all shortfalls
are filled, attainment will not be reached by the applicable dates without further emission
reductions. The Agency's mid-course review in the SIP process — as well as the Clean Air Act's
provisions for contingency measures - is part of our strategy for dealing with some of these
uncertainties, but does not ensure successful attainment.
Where we have proposed a specific amount of additional reductions needed for
attainment, there is a risk that violations would occur in 2007 even if the additional measures for
this amount of reduction are adopted. In addition to all of the factors mentioned above in
connection with the Southeast Desert and the areas for which we did not identify a shortfall, there
is uncertainty in the conclusion about the existence and size of the shortfall. The shortfalls were
identified through consideration of a variety of evidence, without actual ozone modeling on the
effect of the additional emissions reductions.
Given the political, human, and economic factors involved, until the affected states
actually submit their emission control measures to make up the shortfalls, there is some risk that
the eight areas presently without approved attainment demonstrations will not adopt fully
approvable SIPs. In addition, some of these SIPs assume reductions in NOx emissions in upwind
areas in other states. Until those controls are adopted and implemented, those reductions are
somewhat uncertain. Also, success in implementing all the in-state measures in the SIPs once
they are developed and approved is somewhat uncertain, and this contributes to the risk that 2007
attainment will not happen. This possibility contributes to the risk that each of these areas will
have violations in 2007 despite all efforts to reach attainment.
If an area with a 2007 attainment date does fail to demonstrate actual attainment of the 1-
hour ozone NAAQS based on 2005-2007 ozone data, the Clean Air Act allows EPA to grant it
up to two one-year extensions, provided there has not been more than one exceedance of the
standard in the year prior to the attainment year. The emission reductions from the rule in 2008
and 2009 will be even larger than the reductions in 2007, and can play an important role in
allowing an area that needs these extensions to attain in 2008 or 2009.
m An analysis of ambient 1-hour design values for three, 3-year time periods between 1994 and 1998 for
monitoring sites in the East indicates a 10 percent swing in the 90 percent percentile design values. Thus, if an area
just attains in 2007, there is a risk that it could fall back into nonattainment in subsequent time periods due to year-
to-year variations in meteorology, assuming emissions do not change or change very little. The net NOx emissions
reductions due to Tier 2 in 2007 is 4 percent considering all Eastern States collectively. The Tier 2 modeling
indicates that this level of NOx reductions results in ozone reduction on the order of generally 1-3 ppb ozone. The
1-3 ppb reductions associated with the 4 percent Tier 2 NOx reductions are small compared to the effects of
variations in ozone due to meteorology. It is important to note that the episodes modeled by the Agency, though not
"worst case," may be somewhat more severe for most areas than meteorological conditions associated with recent
design values. Thus, modeling with these episodes that indicates attainment for an area is likely to be conservative.
11-32
-------
Chapter II: Health and Welfare
The Agency regards the continuing reductions from the 2007 heavy-duty rule as part of
the federal/State effort not only to reach attainment in the 2007 to 2009 time frame, but to ensure
that attainment is maintained in the future. The ability of states to maintain the ozone NAAQS
once attainment is reached has proved challenging, and the recent recurrence of violations of the
NAAQS in some other areas increases the Agency's concern about continuing maintenance of
the standard in these ten areas (and other areas discussed later) once attainment is achieved.
Agency uncertainty about the prospects of continued maintenance of the standard is also due, in
part, to the fact that State attainment demonstrations generally do not model beyond their
particular attainment date, and EPA does not insist that states prepare maintenance plans prior to
their request for redesignation to attainment after they have attained. Local modeling and our
review of the SIPs did not address whether additional reductions from fleet turnover would offset
factors that might cause violations after their attainment dates.
Recurrent nonattainment is especially problematic for areas where high population
growth rates lead to significant annual increases in vehicle trips and vehicle miles traveled.
Another factor that plays a role in long-term maintenance is meteorology. Our guidance to states
on ozone modeling for attainment demonstrations is to select high ozone days that are
representative of their current ozone design values. Analysis of these conditions are then used to
predict future ozone and in evaluating control strategies. When assessing the risk of air pollution
that would endanger public health and welfare during the period when the heavy-duty rule will
reduce emissions, we think it is appropriate to consider the possibility that year-to-year variability
of meteorological conditions conducive to ozone formation may be worse than this sometime in
the future. In considering the period for many years beyond 2007, it is possible that some years
will have meteorological conditions conducive to ozone formation substantially worse than
assumed in the ozone modeling in the attainment demonstration. Moreover, ozone modeling
conducted for this rule predicted exceedances in 2020 and 2030, which adds to the Agency's
uncertainty about the prospect of continued attainment for these areas.
To conclude, a total often metropolitan areas need additional measures to meet the
shortfalls in the applicable attainment demonstrations, or are subject to ozone transport from an
upwind area that has an identified shortfall. EPA finds that the states responsible may need,
among other reductions, the level of reductions provided by this rule in order to fill the shortfalls.
We expect to rely in part on these reductions in reaching our final conclusion as to whether each
of the eight areas for which we have recently reviewed an attainment demonstration is more
likely than not to attain on its respective date, whether or not the State formally relies on these
reductions as part of its strategy to fill the identified shortfall in its attainment demonstration. As
to all ten areas, even if all shortfalls were filled by the states, there is some risk that at least some
of the areas will not attain the standards by their attainment dates of 2007, or 2010 for Los
Angeles. In that event, the reductions associated with this program, which increase substantially
after 2007, would help assure that any residual failures to attain are remedied. Finally, there is
also some risk that the areas will be unable to maintain attainment after 2007. Considered
11-33
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
collectively, there is a significant risk that some areas would not be in attainment throughout the
period when the rule will reduce heavy-duty vehicle emissions.
vii. Areas with Pre-2007 Attainment Dates or No Attainment Date
The next group of 20 areas have required attainment dates prior to 2007, or have no
attainment date but are subject to a general obligation to have a SIP that provides for attainment
and maintenance. These 20 areas are found in the middle of Table n. A-6, which compiles
information about the 45 areas of concern. Table II. A-6 is located at the end of this subsection.
EPA and the states are pursuing the established statutory processes for attaining, and maintaining
the ozone standard, where it presently applies, and EPA has re-instated the ozone standard to the
remaining areas. The Agency's finding that there is a significant risk that future air quality
would exceed the ozone standard at some time in the 2007 and later period is based on three
factors: (1) recent exceedances in 1997- 1999, (2) predicted exceedances in 2007 or 2030 after
accounting for reductions from Tier 2 and other federal, local, state or regional controls currently
in place or required, and (3) our assessment of the magnitude of recent violations, the year-to-
year variability of meteorological conditions conducive to ozone formation, transport from areas
with later attainment dates, and uncertainty inherent in SIP attainment planning.
In addition, only a subset have yet adopted specific control measures that have allowed
the Agency to approve an attainment plan, and until the SIPs are actually submitted, reviewed
and approved, there is some risk that these areas will not adopt fully approvable SIPs.
Furthermore, some of these areas are not under a current requirement to obtain EPA approval for
an attainment plan. The mechanisms to get to attainment in areas without a requirement to
submit an attainment demonstration are less automatic, and more uncertain. Even with suitable
plans, implementation success is uncertain, and therefore there is some risk that 2007 attainment,
or maintenance thereafter, would not happen. Maintenance plans are not required to contain
enforceable measures beyond those in the conforming SIP, and all current maintenance plans will
expire prior to the time when the bulk of reductions from this rule will be achieved.
Seven metropolitan areas listed in Table n.A-6 contain a 1-hour ozone nonattainment
area, or areas, for which we have approved, or proposed to approve, an attainment demonstration
for an attainment date of 2003 or 2005 (including granted or requested extensions). These areas
include Atlanta, Philadelphia, Washington DC, Baltimore, Sacramento, Ventura County, and the
San Joaquin Valley. For Atlanta, Baltimore, and Philadelphia, we have proposed that specific
further emission reductions are needed in order to attain by the applicable attainment date. We
have proposed to approve Washington, D.C.'s attainment demonstration without requiring
additional local emission reductions beyond what the State is required to implement or has
already said it will implement. However, air quality modeling conducted for this rule predicted
11-34
-------
Chapter II: Health and Welfare
exceedances for Washington DC." Baltimore has predicted exceedances under our ozone
modeling and has a recognized emissions shortfalls in its attainment demonstration. We have
given final approval to the attainment demonstrations for the listed areas in California. Ventura
County's air quality (like that of Southeast Desert and San Diego) is greatly affected by transport
from the South Coast Air Basin, and has a significant risk of registering ozone exceedances until
the South Coast achieves attainment in 2010 or thereafter. Sacramento has a shortfall identified
in its approved SIP. The San Joaquin Valley has an approved SIP, but has recently registered
some of the highest ozone levels in the nation.
Subject to consideration of comments on our proposed approvals or other new
information, we consider it more likely than not that these seven areas with proposed or final
attainment demonstrations will attain by their deadlines, provided the identified additional
reductions are achieved. However, as noted above for the areas with 2007 or 2010 attainment
dates, there are inherent uncertainties in ozone modeling, attainment planning, and control plan
implementation. All of the uncertainties and risk factors discussed above in connection with the
2007 and 2010 areas also apply to these areas. As with most of the 2007 and 2010 areas, ozone
modeling predicted ozone exceedances in 2007 for many of these areas. There is some risk in
each of these areas that attainment will not be reached by its deadline. Furthermore,
nonattainment might persist beyond the deadline into the period when additional reductions from
the this rule can assist with attainment. Recurrent nonattainment is especially problematic for
areas like Atlanta, GA and Sacramento, CA, where high population growth rates lead to
significant annual increases in vehicle trips and vehicle miles traveled.
There are eight metropolitan areas still subject to the 1-hour ozone NAAQS which have
attainment dates of 1999 or earlier, but have experienced concentrations above the level of the 1-
hour ozone NAAQS. These are Baton Rouge, Birmingham, Cincinnati, Louisville, San Diego,
San Francisco (moderate, but with a 2000 attainment date), and St. Louis.0 Ozone modeling
n It is important to note that modeling conducted for this rulemaking shows that areas are
at risk of exceeding the ozone standard in 2007, 2020 and 2030, and that this modeling is not
related to the modeling analysis performed for the Washington, D.C. nonattainment area, which
demonstrates attainment by 2007 when combined with weight of evidence arguments.
0 Ozone monitoring data showing 1997-1999 violations in Baton Rouge, Phoenix, San Diego, Sacramento,
San Francisco, Southeast Desert, Ventura County and the San Joaquin Valley may in some cases still be in need of
final confirmation. San Diego had a 1999 attainment date, which it did not meet. However, it experienced only one
exceedance in 1999 and so is eligible for an extension to 2000, and then to 2001 if there is only one exceedance in
2000. The occurrence of only a single exceedance in 1999 arguably was attributable to unusual meteorology, and
there is a good risk that attainment will not be reached even by 2001. San Francisco was originally classified as a
moderate area with a 1996 attainment deadline. In 1995, the area was redesignated to attainment, but subsequently
violated the NAAQS. The area was again designated nonattainment and given a 2000 attainment deadline. Data
from 1998 make it clear that this area will not attain based on 1998-2000 monitoring data. Based on air quality
11-35
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
predicted 2007, 2020 and/or 2030 exceedances for all of the areas outside of California. The
California areas have recent exceedances. San Diego is impacted by South Coast's air quality
and recent violations prevent San Francisco from attaining in 2000. In addition, San Francisco is
without an approved attainment plan. For some of these areas, we have not yet received, or have
not proposed approval of, a SIP revision with a plan to correct the recent violations. Many of
these areas may require an attainment date extension while retaining their current classification,
or reclassification to a higher classification with a later attainment date. The present absence of
an attainment plan increases the risk that nonattainment will persist into the 2007 and later
period.
There are another eight areas of concern because of recent concentrations above the level
of the 1-hour ozone NAAQS and modeled predictions of 2007 nonattainment, for which the 1-
hour ozone NAAQS was recently reapplied, and are re-classified as attainment and have
maintenance plans spanning 10 year periods ending between 2005 and 2008. These 8 areas are
Charlotte, Grand Rapids, Huntington, Indianapolis, Memphis, Nashville, Houma, and Richmond.
Houma (LaFourche Parish), LA does not have a specific attainment date.
EPA has recently reinstated the 1-hour ozone standard. There were seven areas
designated attainment with maintenance plans that had violations since revocation between 1996-
1998. Four of these areas — Charlotte-Gastonia, NC, Huntington-Ashland, WV-KY, Nashville,
TN, Richmond, VA - also have predicted exceedances. Recent exceedances in these four areas
will likely trigger any contingency measures in the maintenance plans that are tied to new ozone
violations. However, contingencies tied to air quality were not always a required element or
enforced while the standard was revoked in these maintenance plans, and the SIPs may not yet
contain adequate provisions to bring these areas into consistent attainment. Our ozone modeling
predicted that, even with federal and regional controls in place at the time, these areas are likely
to exceed the standard in 2007, 2020 and 2030. EPA will monitor the situation in these areas,
and has options for working with the affected states towards further emission reductions if
needed. At this time, the Agency has not identified the specific next steps that states might
appropriately take to address this situation.
A group of four areas have had the ozone standard revoked, are without maintenance
plans, have experienced recent exceedances, and are predicted by ozone modeling to be
nonattainment in 2007 if more emission reductions are not implemented. The ozone standard
was reinstated for two of these areas - Boston and Providence. Benton Harbor was officially an
unclassifiable/ attainment area prior to the revocation of the 1-hour standard. Massachusetts and
Rhode Island have been required to develop and submit new attainment demonstrations for their
areas. For all the reasons discussed above in connection with other areas facing the need to
monitoring data not considered in the Tier 2 analysis and on 10 year emissions projections, the Agency has
proposed to redesignate Cincinnati into attainment.
11-36
-------
Chapter II: Health and Welfare
develop and implement an attainment plan, we find that there is some risk that these areas will
not consistently attain the standard in 2007 and beyond without additional controls such as those
proposed in this rulemaking. For Benton Harbor, there is no automatic requirement for
preparation of a new attainment demonstration, adding to the uncertainty about 2007 attainment.
There is some risk that these four areas will not attain the standard by 2007 or thereafter without
additional control from today's action.
As with other areas discussed above, the absence of enforceable local controls that are
demonstrated to be adequate to restore attainment in these areas on a long term basis supports the
Agency's finding that there is some risk in these areas that air quality may violate the ozone
standard in the 2007 and later period. There will remain risks even if a new plan is developed,
adopted, and implemented. All maintenance plans must be revisited eight years after
redesignation, and extended another 10 years. When these areas do face the task of planning for
maintenance in the period beyond their current maintenance plan, the emission reductions from
this rule will help them in reducing the risk of violations in that period.
For all of these 20 areas, EPA and the states are pursuing the established statutory
processes for attaining and maintaining the ozone standard. However, only a subset have yet
adopted specific control measures that has allowed or, we expect, will allow the Agency to
approve an attainment plan. Despite the presence of statutory and regulatory requirements for
those six areas, there is thus some uncertainty in whether states will adopt and implement
measures to provide the additional reductions needed to attain by 2007. Given the political,
human, and economic factors involved, until the SIPs are actually submitted there is some risk
that the areas presently without approved attainment demonstrations will not adopt fully
approvable SIPs. In addition, some of these SIPs assume reductions in NOx emissions in upwind
areas in other states, under the Regional Ozone Transport Rule. Until those controls are adopted
and implemented, those reductions are uncertain. Also, success in implementing all the in-state
measures in the SIPs once they are developed and approved is uncertain, and this contributes to
the risk that 2007 attainment will not happen. This possibility contributes to the risk that each of
these areas will have violations in 2007 or thereafter despite all efforts to achieve attainment.
viii. Areas within 10 percent of Violating the Ozone Standard
There are 15 additional metropolitan areas for which the available ozone modeling and
other evidence is less clear regarding the need for additional reductions. Our own ozone
modeling predicted these 15 areas to need further reductions to avoid exceedances in 2007, 2020
and/ or 2030. The recent air quality monitoring data for these areas shows ozone levels with less
than a 10 percent margin below the NAAQS. We believe there is still a risk of that future ozone
levels will be above the NAAQS because of the year-to-year variability of meteorological
conditions conducive to ozone formation.
11-37
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
ix. Conclusion
In sum, without these reductions, there is a significant risk that an appreciable number of
the 45 areas, with a population of 128 million people in 1999, will violate the 1-hour ozone
standard during the time period when these standards will apply to heavy-duty vehicles. The
evidence summarized in this section, and presented in more detail in the air quality modeling
TSD, supports the Agency's finding that emissions of NOx and VOC from heavy-duty vehicles
between 2007 and 2030 will contribute to a national ozone air pollution problem that warrants
regulatory action under section 202(a)(3) of the Act.
11-38
-------
Table II.A-6. Areas and 1999 Populations at Risk of Exceeding the Ozone Standard
between 2007 and 2030
MSA/ CMSA / State
1999 Population
(in millions)
Areas with 2007 '/ 2010 Attainment Dates (Established or Requested)
Beaumont-Port Arthur, TX
Chicago-Gary-Kenosha, IL-IN-WI
Dallas-Fort Worth, TX
Hartford, CT
Houston-Galveston-Brazoria, TX
Los Angeles-Riverside-Orange County, CA
Milwaukee-Racine, WI
New London-Norwich, CT-RI
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA
Southeast Desert, CA
10 areas
0.4
8.9
4.9
1.1
4.5
16.0
1.6
0.3
20.2
0.5
58. 4
Areas with Pre-2007 Attainment Dates or No Specific Attainment Date, with a Recent History
ofNonattainment.
Atlanta, GA
Baton Rouge, LA
Birmingham, AL
Boston-Worcester-Lawrence, MA-HN-ME-CT
Chaiiotte-Gastonia-Rock Hill, NC-SC
Detroit-Ann Arbor-Fling, MI MSA
Huntington-Ashland, WV-KY-OH
Louisville, KY-IN
Macon, GA MSA
Memphis, TN-AR-MS
Nashville, TN
Philadelphia- Wilmington- Atlantic City, PA-NJ-DE-MD
Richmond-Petersburg, VA
Sacramento- Yolo, CA
San Diego, CA
San Francisco-Oakland-San Jose, CA
San Joaquin Valley, CA
St. Louis, MO-IL
Ventura County, CA
Washington, DC-Baltimore, DC, MD, VA MSA
20 Areas
3.9
0.6
0.9
5.7
1.4
5.5
0.3
1.0
0.3
1.1
1.2
6
1
1.7
2.8
6.9
3.2
2.6
0.7
7.4
54.2
11-39
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table II.A-6. Areas and 1999 Populations at Risk of Exceeding the Ozone Standard
between 2007 and 2030
Areas with Pre-2007 Attainment Dates and Recent Concentrations within 10 Percent of an
Exceedance, But With No Recent History of Nonattainment.
Barnstable- Yarmouth, MA
Benton Harbor, MI
Biloxi-Gulfport-Pascagoula, MS MSA
Charleston, WV MSA
Cincinnati-Hamilton, OH-KY-IN
Cleveland-Akron, OH CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Houma, LA
Lake Charles, LA
New Orleans, LA MSA
Norfolk- Virginia Beach-Newport News, VA-NC MSA
Orlando, FL MSA
Pensacola, FL MSA
Providence-Fall River-Warwick, RI-MA
Tampa-St. Petersburg-Clearwater, FL MSA
15 areas
0.2
0.2
0.4
0.3
2.0
2.9
1.1
0.2
0.2
1.3
1.6
1.5
0.4
1.1
2.3
15.7
Total Areas: 45
Population: 128
e. Public Health and Welfare Concerns from Prolonged and Repeated
Exposures to Ozone
There exists a large body of scientific literature regarding health and welfare effects of
ozone. Initially, research indicates that there were harmful effects resulting from peak ozone
levels (e.g., one-hour concentrations above 0.125 ppm). However, in recent years, research has
shown that harmful effects can occur from much lower, sustained levels of exposure. Studies of
prolonged exposures, those lasting about 7 hours, showed health effects from exposures to ozone
concentrations as low as 0.08 ppm. Prolonged and repeated exposures to ozone at these levels
are common in areas that do not attain the 1-hour NAAQS, and also occur in areas where
ambient concentrations of ozone are in compliance with the 1-hour NAAQS. Thus, adverse
health effects from this type of ozone exposure can reasonably be anticipated to occur in the
future in the absence of this rule. Adverse welfare effects can also be anticipated, primarily from
damage to vegetation at ozone levels below peak levels.
11-40
-------
Chapter II: Health and Welfare
/'. Health and Welfare Effects
Studies of acute health effects from ozone have reported ozone exposure to cause or be
statistically associated with transient pulmonary function responses, transient respiratory
symptoms, effects on exercise performance, increased airway responsiveness, increased
susceptibility to respiratory infection, increased hospital and emergency room visits, and transient
pulmonary respiratory inflamation. Such acute health effects have been observed following
prolonged exposures at moderate levels of exertion at concentrations of ozone as low as 0.08
ppm, the lowest concentration tested. The effects are more pronounced as concentrations
increase, affecting more subjects or having a greater effect on a given subject in terms of
functional changes or symptoms. A detailed summary and discussion of the large body of ozone
health effects research may be found in Chapters 6 through 9 (Volume 3) of the 1996 Criteria
Document for ozone.11
The following is a brief summary focusing on studies on the effects of exposures to
concentrations of ozone just at and below peak ozone concentrations. Tables II. A-7 through
II.A-11 of this section are excerpted from the 1996 Criteria Document, with only studies that
used peak ozone concentrations or below retained.
It has long been established by exposure chamber studies that single, short-term (1 to 3
hour) exposures to ozone concentrations at or above peak levels produce a variety of respiratory
function effects in exposed subjects. Tables n.A-7 and HA-8 summarize these studies, for
healthy and diseased subjects, and also indicate that equally short-term exposures to
concentrations below peak levels have not shown these effects. More recent studies have sought
to investigate whether similar effects occur following longer exposures to lower levels of ozone.
These studies are summarized here in Tables II. A-9 and II. A-10. Exposures of 6.6 hours to
ozone concentrations of 0.08 ,0.10, and 0.12 ppm were used in these chamber exposures
studies, and are reported to cause decrements in lung function (reduced ability to take a deep
breath), increased respiratory symptoms (cough, shortness of breath, pain upon deep inspiration),
increased airway responsiveness (an indication that airways are predisposed to broncho-
constriction, which is characteristic of asthma), and increased airway inflammation in adults.
The effects are more pronounced as concentrations increase, affecting more subjects or having a
greater effect on a given subject in terms of functional changes or symptoms. Earlier studies
found these effects in heavily exercising adults exposed to ozone on a short-term basis, but the
level of exertion involved was high enough to be unusual among people conducting their normal
activities. The more recent studies with 6.6 hour exposures at 0.08 and 0.10 ppm observed these
functional changes and symptoms when subjects were exerting themselves at only moderate
levels. This means that much of the population could experience these effects from ambient
concentrations while conducting their normal activities at moderate exertion levels.
11-41
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
With regard to chronic health effects, the collective data from these chamber studies have
many ambiguities, but provide suggestive evidence of chronic effects in humans. Table II. A-11
summarizes studies associating a single prolonged exposure to ozone at 0.08 and 0.10 ppm with
lung inflammation. There is a biologically plausible basis for considering the possibility that
repeated inflammation associated with exposure to ozone over a lifetime, as can occur with
exposure to 8-hour ozone levels as low as 0.08 ppm, may result in sufficient damage to
respiratory tissue such that individuals later in life may experience a reduced quality of life,
although such relationships remain highly uncertain.
A number of "summer camp" studies of children and adolescents, and other types of
epidemiological studies involving exposure to ambient concentrations of ozone, confirm that
ozone concentrations are correlated with lung function changes, as indicated by the chamber
studies. The studies are not summarized in table form here. Changes reported at low ozone
concentrations in these studies are comparable to those observed in the chamber studies, although
comparisons are difficult because of differences in experimental design and analytical approach.
Studies published since 1986 have supported a direct association between ambient ozone/oxidant
concentrations and acute respiratory morbidity in asthmatics, although it is difficult to clearly
differentiate the independent effects of ozone from those of copollutants. Conclusions from the
field studies on asthmatics are based on observations over a range of ozone exposures extending
below the 0.12 ppm level of the 1-hour NAAQS.
Over 20 epidemiology studies of aggregate populations have investigated the relationship
between ozone concentrations and hospital admissions/ hospital visits. The studies are not
summarized in table form here. Significant associations are seen between ozone and hospital
admissions/visits at exposures below 0.12 ppm 1-hour daily maximum ozone.
Ozone also has many welfare effects, with damage to plants being of most concern. Plant
damage affects crop yields, forestry production, and ornamentals. The adverse effect of ozone on
forests and other natural vegetation can in turn cause damage to associated ecosystems, with
additional resulting economic losses. Ozone concentrations of 0.10 ppm can be phytotoxic to a
large number of plant species, and can produce acute injury and reduced crop yield and biomass
production. Ozone concentrations within the range of 0.05 to 0.10 ppm have the potential over a
longer duration of creating chronic stress on vegetation that can result in reduced plant growth
and yield, shifts in competitive advantages in mixed populations, decreased vigor, and injury.
Ozone effects on vegetation are presented in more detail in Chapter 5, Volume II of the 1996
Criteria Document.
11-42
-------
Table II.A-7. Controlled Exposure of Healthy Human Subjects to Ozone
Ozone
Concentration
ppm
Exposure
Duration and
' Activity
Number Subject
Exposure and Character-
Conditions Gender of Subjects istics
Observed Effect (s)
Reference
Healthy Exercising Adult Subjects
0.08
0.10
0.12
0.14
0.16
0.12
0.18
0.24
0.12
0.18
0.24
0.30
0.40
157
196
235
274
314
235
353
470
235
353
470
588
784
2hIE
(4 x 15 min
at VE =
68 L/min)
Ih
competitive
simulation
exposures at
mean VE =
87 L/min
2.5 h IE
(4 x 15 min
treadmill
exercise
[VE =
65 L/minl)
Tdb = 32 °C
RH = 38%
Tdb = 23 to
26 °C
RH = 45 to
60%
Tdb = 22 °C
RH = 40%
24 M
10 M
20 M
22 M
20 M
21 M
20 M
29 M
Young,
healthy adults,
18 to
33 years old
10 highly
trained
competitive
cyclists, 19 to
29 years old
Young,
healthy adults,
18 to
30 years old
No significant changes in pulmonary
function measurements.
Decrease in FVC and FEVj for 0.18- and
0.24-ppm O3 exposure compared with FA
exposure; decrease in exercise time for
subjects unable to complete the competitive
simulation at 0. 18 and 0.24 ppm O3,
respectively.
Significant decrease in FVC, FEVl5 and
FEF25_75% at 0. 12 ppm O3; decrease in VT
and increase in f and SRaw at 0.24 ppm O3.
Linn etal. (1986)
Schelegle and Adams
(1986)
McDonnell et al.
(1983)
11-43
-------
Table II.A-7. Controlled Exposure of Healthy Human Subjects to Ozone
0.12
0.18
0.24
0.30
0.40
0.12
0.18
0.24
0.30
0.40
0.12
0.18
0.24
0.30
0.40
235
353
470
588
784
235
353
470
588
784
235
353
470
588
784
2 x 2.5 h IE Tdb = 22°C
(4 x 15 min RH = 40%
treadmill
exercise
[VE =
35 L/min/m2
BSA]).
Exposure
separated by
48 ± 30 days
and 301
± 77 days
2 x 2.5 h IE Tdb = 22°C
(4 x 15 min RH = 40%
treadmill
exercise
[VE =
35 L/min/m2
BSA1)
2.5 h IE Tdb = 22°C
(4 x 15 min RH = 40%
treadmill
exercise
[VE =
25 L/min/m2
BSA])
8M
8M
5M
5M
6M
290 M
17 WM/15 BM/15
WF/ 15BF
15 WM/15 BM/15
WF/ 16BF
15 WM/17 BM/17
WF/ 15BF
16 WM/15 BM/17
WF/ 16BF
15 WM/15 BM/15
WF/ 15BF
15 WM/15 BM/15
WF/ 15BF
Young,
healthy adults,
18 to
30 years old
Young,
healthy adults,
18 to
32 years old
Young,
healthy whites
and blacks, 18
to 35 years old
Pulmonary function variables SRaw and VE
were not significantly different in repeat
exposures, indicating that the response to
0.18 ppm O3 or higher is reproducible.
O3 concentration and age predicted FEV{
decrements; it was concluded that age is a
significant predictor of response (older
subjects being less responsive to O3).
Decreases in FEVj for all levels of O3 as
compared with FA; increase in SRaw with
0.18 ppm O3 and greater compared with
FA; black men and women had larger FEVj
decrements than white men, and black men
had larger FEVj decrements than white
women.
McDonnell et al.
(1985b)
McDonnell et al.
(1993)
Seal etal. (1993)
11-44
-------
Table II.A-7. Controlled Exposure of Healthy Human Subjects to Ozone
0.12
0.20
0.10
0.15
0.20
0.25
235
392
196
294
392
490
IhCE Tdb = 31°C
(mean VE =
89 L/min)
2hIE Tdb = 22°C
(4 x 14 min RH = 50%
treadmill at
mean VE =
70.2 L/min>
15 M Highly trained
2 F competitive
cyclists, 19 to
30 years old
20 M Young,
healthy NS,
25.3 ±4.1
(SD) years old
Decrease in VEmax, VO2max, VTmax, work
load, ride time, FVC, and FEVj with 0.20
ppm O3 exposure during maximal exercise
conditions, but not significant with
0. 12 ppm O3 exposure, as compared to FA
exposure.
FVC, FEVb FEF25.75%, SGaw, 1C, and TLC
all decreased with (1) increasing O3
concentration, and (2) increasing time of
exposure; threshold for response was above
0.10 nnm but below 0.15 nnm O,.
Gong etal. (1986)
Kulleetal. (1985)
* See Appendix A of the 1996 Ozone Criteria Document for abbreviations and acronyms.
11-45
-------
Table II.A-8. Ozone Exposure in Subjects with Preexisting Disease
Ozone
Concentration
ppm ,ug/m3
Exposure
Duration and
Activity
Exposure
Condition
Number
and
Gender of
Subjects
Subject
Character-istics
Observed Effect(s)
Reference
Subjects with Chronic Obstructive Pulmonary Disease
0.12
236
IhlE (2 x
15 min light
bicycle
ergometry)
Tdb = 25 °C
RH = 50%
18 M, 7F
8 smokers,
14 ex-smokers,
3 nonsmokers;
FEVj/FVC = 32
to 66%
No significant changes in pulmonary function
measurements;
small significant decrease in arterial O2
saturation.
Linn et al.
(1982a)
Adult Subjects with Asthma
0.10
0.25
0.40
0.12
0.12
196
490
784
236
236
1 h light IE (2 x
15 min on
treadmill, VE =
27 L/min)
1 h rest
0.75 h IE
VE = 30 L/min
(15 min rest,
15 min exercise,
15 min rest)
followed by
15 min exercise
inhaling 0.10 ppm
SO,
Tdb = 21°C
RH = 40%
NA
Tdb = 22°C
RH = 75%
12 M, 9 F,
19 to 40 years
old
7M, 8F
8 M, 5 F,
12 to 18 years
old
Stable mild
asthmatics with
FEVj > 70% and
methacholine
responsiveness
Never smoked,
mild stable
asthmatics with
exercise-induced
asthma
Asthmatics
classified on
basis of positive
clinical history
and
methacholine
challenge.
Asymptomatic at
time of study.
No significant differences in FEVj or FVC
were observed for 0.10 and 0.25 ppm O3-FA
exposures or postexposure exercise challenge;
12 subjects exposed to 0.40 ppm O3 showed
significant reduction in FEV, .
Exposure to 0. 12 ppm O3 did not affect
pulmonary function. Preexposure to 0.12 ppm
O3 at rest did not affect the magnitude or time
course of exercise-induced
bronchoconstriction.
Filtered air followed by SO2 and O3 alone did
not cause significant changes in pulmonary
function. Ozone followed by SO2 resulted in
significant decrease in FEV{ (8%) and Vmax50%
(15%) and a significant increase in RT (19%).
Weymer et al.
(1994)
Fernandes et al.
(1994)
Koenig et al.
(1990)
11-46
-------
Table II.A-8. Ozone Exposure in Subjects with Preexisting Disease
0.12 236
0.24 472
0.12 236
0.12 236
1.5hIE,
VE =
25 L/min
6.5 h/day IE (6 x
50 min) (2 days
of exposure), VE
= 28 L/min
(asthmatic),
VE = 31 L/min
(healthy)
1 h rest
Tdb = 22°C 4M, 4F
RH = 65% (nonasthmatics
);
18 to 35 years
old;
5M, 5F
(asthmatics);
18 to 41 years
old
NA 8 M, 7 F
(nonasthmatics
);
22 to 41 years
old;
13 M, 17 F
(asthmatics);
18 to 50 years
old
NA 4 M, 3 F,
21 to 64 years
old
Physician-
diagnosed
asthma
confirmed with
methacholine
challenge test.
All nonsmokers
and
asymptomatic at
time of study.
Nine were
atopic.
Asthmatics
classified on
basis of positive
clinical history,
previous
physician
diagnosis, and
lowPD20. Mild
to severe
asthmatics.
Mild, stable
asthma
No significant changes in pulmonary and nasal McBride et al.
function measurements in either asthmatics or ( 1 994)
nonasthmatics. Significant increase in nasal
lavage white cell count and epithelial cell
following O3 exposure in asthmatics only.
Significant increase in bronchial reactivity to Linn et al.
methacholine in both asthmatics and ( 1 994)
nonasthmatics. FEVj decreased 8.6% in
asthmatics and 1.7% in nonasthmatics, with
difference not being significant.
Increase in bronchial responsiveness to Molfino et al.
allergen; no change in baseline airway (1991)
function.
Adolescent Subjects with Asthma
0.12 235
1 h rest
Tdb = 22 °C 4 M, 6 F
RH > 75% (normals),
13 to 18 years
old;
4M, 6F
(asthmatics),
11 to 18 years
old
Asthmatics had
a history of
atopic extrinsic
asthma
and exercise-
induced
bronchospasm
Decrease in FRC with O3 exposure in Koenig et al.
asthmatics; no consistent significant changes in (1985)
pulmonary functional parameters in either
group or between groups.
11-47
-------
Table II.A-8. Ozone Exposure in Subjects with Preexisting Disease
0.12 235
IhlE
(2 x 15 min
treadmill walking
at mean VE =
32.5 L/min)
Tdb = 22 °C
RH > 75%
5M, 8F
(normals),
12 to 17 years
old;
9M, 3F
(asthmatics),
12 to 17 years
old
Asthmatics
selected from a
clinical practice
and had
exercise-
induced
bronchospasm
Decrease in maximal flow at 50% of FVC in
asthmatics with O3 exposure compared to FA;
no significant changes with combined O3-NO2
exposure.
Koenig et al.
(1988)
0.12 235 40 min IE
0.18 353 (1x10 min
treadmill walking
at mean VE =
32.5 L/min)
NA
4M, 9F
(normals),
14 to 19 years
old;
8M, 8F
(asthmatics),
12 to 19 years
old
Asthmatics had
allergic asthma,
positive
responses to
methacholine,
and exercise-
induced
bronchospasm
Decrease in FEVj and increase in RT in
normals and asthmatics with 0.12 and 0.18
ppm O3 exposure compared to FA; no
consistent differences between normals and
asthmatics.
Koenig et al.
(1987)
11-48
-------
Table II.A-9. Pulmonary Function Effects After Prolonged Exposures to Ozone
„ ,-, . Number and
Ozone Concentration „ „ ^ , s n , •
, , Exposure Exposure (Bender of Subject
ppm tug/in
Duration and Activity Conditions Subjects Character-istics Observed Effect(s) Reference
0.08 157 6.6 h
0.10 196 IE(6x50min)
0.12 235 VE>= 39 L/min
18 °C 22 M Healthy NS, 18 FVC and FEVj decreased Horstman et al. (1990)
40% RH to 33 years old throughout the exposure; FEVj
decrease at end exposure was 7.0,
7.0, and 12.3%, respectively.
FEVj change >15% occurred in 3,
5, and 9 subjects at 0.08, 0.10, and
0.12 ppm, respectively.
Methacholine responsiveness
increased by 56, 89, and 121%,
respectively.
See Horstman et al. (1990)
and Folinsbee et al. (1988)
A lognormal model was fitted to
FEVj data. Model parameters
indicate O3 concentration had
greater effect than VE or duration
(estimated exponent for [O3] ~
4/3).
Larsenetal. (1991)
0.08
0.10
0.08
157
196
157
6.6 h
IE (6 x 50 min)
VE = 40 L/min
6.6 h
IE (6 x 50 min)
VE = 35 to 38 L/min
(1 day of air, 2 days of O,)
18 °C
40% RH
25 °C
48% RH
38M Healthy NS,
mean age
25 years old
5 F, 6 M Healthy NS, 30
to 45 years old
FEVb decreased 8.4% at 0.08 ppm McDonnell et
and 11. 4% at 0.10 ppm. (1991)
Symptoms of cough, PDI, and SB
increased with O, exposure.
FVC decreased 2. 1%, FEVj Horvath et al.
decreased 2.2% on first day of O3
exposure; no change on second O3
day.
al.
(1991)
11-49
-------
Table II.A-9. Pulmonary Function Effects After Prolonged Exposures to Ozone
0.12 235 6.6 h
IE (6 x 50 min)
VE = 42.6 L/min
0.12 235 6.5h/day
IE (6 x 50 min)
(2 days of exposure)
VE = 28 L/min (asthmatic)
VE = 31 L/min (healthy)
0.12 235 6.6 h
IE (6 x 50 min)
VE = 38.8 L/min
18 °C 10 M
40% RH
(1 exposure to
clean air;
1 exposure to
0,)
21 °C 15
50% RH (8 M, 7 F)
30
(13 M, 17 F)
18 °C 17 M
40% RH
(5 consecutive
days of
exposure to
03, 1 day
exposure to
CA)
Healthy NS, 18
to 33 years old
Healthy NS,
22 to 41 years
old
Asthmatic NS,
18 to 50 years
old
Healthy NS,
mean age 25 ± 4
years old
FEVj decreased by 13% after 6.6
h. FVC dropped 8.3%. Cough
and PDI increased with O3
exposure. Airway responsiveness
to methacholine doubled after O3
exposure.
Bronchial reactivity to
methacholine increased with
O3 exposure in healthy subjects.
FEVj decreased 2% (pre- to
postexposure) in healthy subjects
and 7.8% in asthmatics.
Responses were generally less on
the second day. Two healthy
subjects and four asthmatics had
FEV, decreases >10%.
FEVj decreased by 12.8, 8.7, 2.5,
and 0.6 and increased by 0.2 on
Days 1 to 5 of O3 exposure,
respectively. Methacholine airway
responsiveness increased by
>100% on all exposure days.
Symptoms increased on the first
O3 day, but were absent on the last
3 exposure days.
Folinsbeeetal. (1988)
Linnetal. (1994)
Folinsbeeetal. (1994)
11-50
-------
Table II.A-9. Pulmonary Function Effects After Prolonged Exposures to Ozone
(a) 0.12 235 8h
(b) Varied IE (8 x 30 min)
from 0.0 to VE = 40 L/min
0.24
(increased by
0.06 ppm/h
then
decreased by
0.06 ppm/h)
22 °C
40% RH
<3 //g/m3 TSP
23 M Healthy NS, 20 (a) FEVj decreased 5% by 6 h and Hazucha et al.
to 35 years old remained at this level through 8 h.
(b) FEVj change mirrored O3
concentration change with a lag
time of ~ 2 h. Max decrease of
10.2% after 6 h. FEVj change was
reduced in last 2 h of exposure.
(1992)
See Appendix A of the 1996 Ozone Criteria Document for abbreviations and acronyms.
11-51
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table II.A-10. Increased Airway Responsiveness Following Ozone Exposures *
Ozone
Concentration **
ppm /
0.08
0.10
0.12
0.10
0.32
1.00 1.
0.12
0.20
0.12
jg/m3
157
196
235
196
627
960
235
392
235
0.12 ppm O3-100ppbSO2
0.12ppmO3-0.12ppmO3
Air-100 ppb SO2
Air-antigen
0.12 ppm O3-antigen
Exposure Exposure
Duration and Activity Conditions
6.6 h 18 °C
IEat=39L/min 40% RH
2h NA
lhat VE = 89L/min 31 °C
followed by 3 to 4 min 3 5% RH
ats!50L/min
6.6 h with IE at NA
-25 L/min/m2 BSA
45 min in first 75% RH
atmosphere and 15 min 22 °C
in second
IE
1 h at rest NA
Number and Subject
Gender of Character-
Subjects istics
22 M Healthy NS,
18 to 32 years
old
14 Health NS,
24 ± 2 years
old
15 M, 2 F Elite
cyclists, 19 to
30 years old
10 M Healthy NS,
18 to 33 years
old
8 M, 5 F Asthmatic,
12 to 18 years
old
4 M, 3 F Asthmatic,
21 to 64 years
old
Observed Effect(s)
33, 47, and 55% decreases in cumulative dose
of methacholine required to produce a 100%
increase in SRaw after exposure to O3 at 0.08,
0.10, and 0.12 ppm, respectively.
Increased airway responsiveness to
methacholine immediately after exposure at
the two highest concentrations of O3.
Greater than 20% increase in histamine
responsiveness in one subject at 0.12 ppm
O3 and in nine subjects at 0.20 ppm O3.
Approximate doubling of mean methacholine
responsiveness after
exposure. On an individual basis, no
relationship between O3-induced changes in
airway responsiveness and FEV, or FVC.
Greater declines in FEV{ and V^^gy.
and greater increase in respiratory resistance
after O3-SO2 than after O3-O3 or air-SO2.
Increased bronchoconstrictor response to
inhaled ragweed or grass after O3 exposure
compared to air.
Reference
Horstman
etal.
(1990)
Konig
etal.
(1980)
Gong et al.
(1986)
Folinsbee
etal.
(1988)
Koenig
etal.
(1990)
Molfino
etal.
(1991)
* See Appendix A of the 1996 Ozone Criteria Document for abbreviations and acronyms.
11-52
-------
Table II.A-11. Bronchoalveolar Lavage Studies of Inflammatory Effects from Controlled Human Exposure to Ozone
Ozone
Concentration11
ppm //g/m3
Number
and
Exposure Activity Level Qender of
Duration (VF)
Observed Effect(s)
Reference
0.08 157 6.6 h IE (40 L/min)
0.10 196 six50-min
exercise
periods + 10
min rest;
35 min lunch
18 M, BAL fluid 18 h after exposure to 0.1 ppm O3 had Devlin
18 to significant increases in PMNs, protein, PGE2, etal.
35 years fibronectin, IL-6, lactate dehydrogenase, and cc-1 (1990,
old antitrypsin compared with the same subjects exposed to 1991)
FA. Similar but smaller increases in all mediators after Koren et
exposure to 0.08 ppm O3 except for protein and al. (1991)
fibronectin. Decreased phagocytosis of yeast by
alveolar macrophages was noted at both concentrations.
11-53
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
/'/'. Ozone Concentrations
This section summarizes the results of analyses of model-adjusted ozone air quality
concentrations and the anticipated air quality impact of reductions in emissions expected to result
from implementation of the heavy duty engine and vehicle standards and highway diesel fuel
sulfur control requirements. Specifically, it provides information on the number of people
estimated to live in metropolitan counties in which ozone monitors are predicted to repeatedly
experience certain levels of ozone of potential concern over prolonged periods, i.e., 8-hours.
Heavy-duty vehicles contribute a substantial fraction of ozone precursors in any
metropolitan area. Available health studies (summarized above) have indicated health effects
(e.g., lung function decrements, respiratory symptoms, and pulmonary inflammation) at ozone
concentrations between 0.08 ppm and 0.12 ppm over prolonged exposures (6.6 hours in most
chamber studies). An 8-hour averaging period was chosen as a convenient and appropriate
metric for describing current and future ozone patterns relevant to this concentration range.
Another important metric is the number of days with ozone levels between 0.08 and 0.12 ppm
because repeated exposure to ozone in this concentration range may be associated with long term
health effects related to pulmonary inflammation.
To provide a quantitative estimate of the number of people anticipated to reside in areas
in which ozone concentrations are predicted to experience multiple days with 8-hour ozone in the
range of 0.08 to 0.12 ppm and higher, we performed regional modeling for 6 different scenarios
(1996 base, 2007 base, 2020 base and control, 2030 base and control) for the eastern United
States. This modeling is further described in section A.2.6 "Photochemical Ozone Modeling."
Our analysis relies on projected county-level population from the U.S. Department of Census for
the period representing each year analyzed.
For each of the counties analyzed, we determined the number of days for periods on
which the highest model-adjusted 8-hour concentration at any monitor in the county was
predicted, for example, to be between 0.08 and 0.12 ppm (after rounding from 3 decimal places).
We then grouped the counties which had days with ozone in this range according to the number
of days this was predicted to happen, and summed their projected populations. We repeated this
for ozone ranges of 0.09 to 0.12 ppm, 0.10 to 0.12 ppm, 0.11 to 0.12 ppm and greater than or
equal to 0.12 ppm.
In the 2007 base case (i.e., before the application of emission reductions resulting from
this rule), we estimated that 116 million, or 93 percent of the total population considered in this
analysis, are predicted to live in areas with at least 2 days with model-adjusted 8-hour average
concentrations of 0.08 ppm or higher. The number of people involved is predicted to diminish as
the lower end of the concentration range increases or as the number of days predicted to
experience such peak 8-hour average concentrations increases. The number of people predicted
11-54
-------
Chapter II: Health and Welfare
to live in areas with at least 2 days with model-adjusted 8-hour average concentrations of 0.08
ppm or higher is estimated to increase in the 2020 base case to 122 million people, although this
is estimated to represent a smaller percentage (87 percent) of the total projected population
considered in the analysis. However, both the number of people (139 million) and the relative
percentage (91 percent) of the total population considered in the analysis is projected to grow in
the 2030 base case.
3. Particulate Matter
a. Health and Welfare Effects of Ambient 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.
Coarse PM are those particles which have a diameter in the range of 2.5 to 10 microns, and fine
particles are those particles which have a diameter less than 2.5 microns. Typically, PM is also
classified as PM10 (all particles less than 10 microns) or PM25 (all particles less than 2.5
microns). Human-generated sources of particles include a variety of stationary sources
(including power generating plants, industrial operations, manufacturing plants, waste disposal)
and mobile sources (light- and heavy-duty on-road vehicles, and off-highway vehicles such as
construction, farming, industrial, locomotives, marine vessels and other sources). Particles may
be emitted directly to the atmosphere (primary particles) or may be formed by transformations of
gaseous emissions of sulfur dioxide, nitrogen oxides or volatile organic compounds (secondary
particles). Secondary PM is dominated by sulfate in the eastern U.S. and nitrate in the western
U.S.12 Essentially all (>90 percent) of the direct mobile source PM emissions and their
secondary formation products are in the fine PM size range. 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.
The chemical and physical properties of PM vary greatly with time, region, meteorology,
and source category, thus complicating the assessment of health and welfare effects. At elevated
concentrations, particulate matter can adversely affect human health, visibility, and materials.
Components of particulate matter (e.g., sulfuric or nitric acid) also contribute to acid deposition,
nitrification of surface soils and water and eutrophication of surface water as will be discussed
below.
Key EPA findings regarding the health risks posed by ambient particulate matter can be
found in the Air Quality Criteria for Particulate Matter and are summarized as follows:
a. Health risks posed by inhaled particles are affected both by the penetration and deposition
of particles in the various regions of the respiratory tract, and by the biological responses
to these deposited materials.
11-55
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
b. The risks of adverse effects associated with deposition of ambient particles in the thorax
(tracheobronchial and alveolar regions of the respiratory tract) are markedly greater than
for deposition in the extrathoracic (head) region. Maximum particle penetration to the
thoracic regions occurs during oronasal or mouth breathing.
c. The key health effects categories associated with PM include premature death;
aggravation of respiratory and cardiovascular disease, as indicated by increased hospital
admissions and emergency room visits, school absences, work loss days, and restricted
activity days; changes in lung function and increased respiratory symptoms; changes to
lung tissues and structure; and altered respiratory defense mechanisms. Most of these
effects have been consistently associated with ambient PM concentrations, which have
been used as a measure of population exposure, in a large number of community
epidemiological studies. Additional information and insights on these effects are
provided by studies of animal toxicology and controlled human exposures to various
constituents of PM conducted at higher than ambient concentrations. Although
mechanisms by which particles cause effects are not well known, there is general
agreement that the cardio-respiratory system is the major target of PM effects.
d. Based on a qualitative assessment of the epidemiological evidence of effects associated
with PM for populations that appear to be at greatest risk with respect to particular health
endpoints, we have concluded the following with respect to sensitive populations:
1. Individuals with respiratory disease (e.g., chronic obstructive pulmonary disease,
acute bronchitis) and cardiovascular disease (e.g., ischemic heart disease) are at
greater risk of premature mortality and hospitalization due to exposure to ambient
PM.
2. Individuals with infectious respiratory disease (e.g., pneumonia) are at greater risk
of premature mortality and morbidity (e.g., hospitalization, aggravation of
respiratory symptoms) due to exposure to ambient PM. Also, exposure to PM
may increase individuals' susceptibility to respiratory infections.
3. Elderly individuals are also at greater risk of premature mortality and
hospitalization for cardiopulmonary problems due to exposure to ambient PM.
4. Children are at greater risk of increased respiratory symptoms and decreased lung
function due to exposure to ambient PM.
5. Asthmatic individuals are at risk of exacerbation of symptoms associated with
asthma, and increased need for medical attention, due to exposure to PM.
11-56
-------
Chapter II: Health and Welfare
e. There are fundamental physical and chemical differences between fine and coarse fraction
particles. The fine fraction contains acid aerosols, sulfates, nitrates, transition metals,
diesel exhaust particles, and ultra fine particles and the coarse fraction typically contains
high mineral concentrations, silica and resuspended dust. It is reasonable to expect that
differences may exist in both the nature of potential effects elicited by coarse and fine PM
and the relative concentrations required to produce such effects. Both fine and coarse
particles can accumulate in the respiratory system. Exposure to coarse fraction particles
is primarily associated with the aggravation of respiratory conditions such as asthma.
Fine particles are most closely associated with health effects such as premature death or
hospital admissions, and for cardiopulmonary diseases.
With respect to welfare or secondary effects, fine particles have been clearly associated
with the impairment of visibility over urban areas and large multi-State regions. Fine particles,
or major constituents thereof, also are implicated in materials damage, soiling and acid
deposition. Coarse fraction particles contribute to soiling and materials damage.
Particulate pollution is a problem affecting urban and non-urban localities in all regions
of the United States. Manmade emissions that contribute to airborne particulate matter (listed
above) result principally from combustion sources (stationary and mobile sources) and fugitive
emissions from industrial process and non-industrial processes (such as roadway dust from paved
and unpaved roads, wind erosion from cropland, construction, etc.). Natural sources also
contribute to particulate matter in the atmosphere and include sources such as wind erosion of
geological material, sea spray, volcanic emissions, biogenic emanation (e.g., pollen from plants,
fungal spores), and wild fires. Emission inventories for the relative contribution of diesel PM to
total ambient PM will be discussed below.
Secondary diesel PM includes particles containing sulfuric acid, nitric acid and organic
compounds of diesel exhaust origin. Sulfur dioxide (SO2) and nitrogen oxides (primarily nitric
oxide, or NO), are emitted from diesel engines. Sulfur dioxide is converted to sulfuric acid in the
presence of oxidizing reactants and water vapor to form (H2SO4) droplets which are less than 1
|im in diameter. Because SO2 is soluble in water, it is scavenged by fog, cloud water, and
raindrops. Sulfur emitted from diesel engines is predominantly (-98 percent) in the form of SO2,
a portion of which will form sulfate aerosols by the reaction described above. Off-road
equipment, typically use fuel containing 3300 ppm sulfur, and therefore emit more SO2 than on-
road diesel engines which use fuels currently containing an average of 340 ppm sulfur. We
estimate that mobile sources are responsible for about seven percent of nationwide SO2 emissions
with diesel engines contributing 80 percent of the mobile source total (the majority of the diesel
SO2 emissions originate from off-highway engines).13 The portion of this SO2 which is
subsequently converted to sulfuric acid will vary regionally and, especially in the eastern U.S.,
the contribution of diesel emissions will be minimal.
11-57
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Nitric oxide (NO) is also oxidized in the atmosphere to form NO2 and particulate nitrate
(nitric acid and ammonium nitrate primarily). Organic aerosols are also formed from atmospheric
transformation of hydrocarbons emitted in the gaseous phase from diesel engines. Little research
has been conducted to characterize the contribution of diesel exhaust to secondary organic
particulates in the ambient air. Some studies suggest that up to 38 percent of the organic aerosol
in an urban environment can be secondary in origin, a portion of which would come from diesel
exhaust.14 In a recent modeling study by Kleeman and Cass, 8.96 |ig/m3 PM25 (67 percent of the
diesel PM2.5 mass) at Riverside, CA was attributed to secondary formation from direct diesel
emissions.15 A portion of the secondary PM25 was attributed to primary emissions of
hydrocarbons (1 percent). The majority (70 percent) of the secondary diesel PM2 5 at Riverside
was attributed to nitrate formation.
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 SO2, NOx, and VOCs 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 off-highway 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.16
Heavy-duty vehicles contribute to fine particle formation through a number of pollutants.
The chemical composition of PM fine varies by region of the country (see Table n.A-12).
Sulfate plays a major role in the composition of fine particulate across the country, but typically
makes up over half the fine particles found in the Eastern United States. Organic carbon
accounts for a large portion of fine particle mass, with a slightly higher fraction in the west.
Diesel engines are the principle source of elemental carbon, which makes up about 5-6 percent of
particle mass.
11-58
-------
Chapter II: Health and Welfare
Table II.A-12. Percent Contribution to PM2 5 by Component, 1998 p
Sulfate
Elemental Carbon
Organic Carbon
Nitrate
Crustal Material
East
56
5
27
5
7
West
33
6
36
8
17
Nationally, nitrate plays a relatively small roles in the make up of fine particles, but it
plays a far larger role in southern California. Ammonium nitrate - formed secondarily from NOx
and ammonia emissions — is one of the most significant components of particulate matter
pollution in California. During some of the worst episodes of elevated particle levels in the
South Coast, ammonium nitrate can account about 65-75 percent of the PM2.5 mass.q Reducing
ammonium nitrate through controls on NOx sources is a critical part of California's parti culate
matter strategy. Nationally, the standards finalized in this rule will significantly reduce HDV
emissions of SOx, NOX, VOCs and elemental carbon, and thus contribute to reductions in
ambient concentrations of PM10 and PM2.5.
/'. Current PM10 Nonattainment
The most recent PM10 monitoring data indicates that 14 designated PM10 nonattainment
areas with a projected population of 23 million violated the PM10 NAAQS in the period 1997-
1999. Table n.A-13 lists the 14 areas, and also indicates the PM10 nonattainment classification,
and 1999 projected population for each PM10 nonattainment area. The projected population in
p National Air Quality and Emissions Trends Report, 1998, EPA 454/R-00-003, March, 2000.
q Southern California 1997 PM10 Air Quality Management Plan.
11-59
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
1999 was based on 1990 population figures which were then increased by the amount of
population growth in the county from 1990 to 1999.
Table II.A-13. PM10 Nonattainment Areas Violating the PM10 NAAQS in 1997- 1999
Nonattainment Area or County
Anthony, NM (Moderate)8
Clark Co [Las Vegas], NV (Serious)
Coachella Valley, CA (Serious)
El Paso Co, TX (Moderate) A
Hay den/Mi ami, AZ (Moderate)
Imperial Valley, CA (Moderate)
Los Angeles South Coast Air Basin, CA
(Serious)
Nogales, AZ (Moderate)
Owens Valley, CA (Serious)
Phoenix, AZ (Serious)
San Joaquin Valley, CA (Serious)
Searles Valley, CA (Moderate)
Wallula, WA (Moderate)8
Washoe Co [Renol, NV (Moderate)
Total Areas: 14
1999 Population
(projected, in millions)
0.003
1.200
0.239
0.611
0.004
0.122
14.352
0.025
0.018
2.977
3.214
0.029
0.052
0.320
23.167
EPA has determined that continuing PM10 nonattainment in El Paso, TX is attributable to such transport
under section 179(B).
T_>
The violation in this area has been determined to be attributable to natural events under section 188(f) of
the Act.
In addition to the 14 PM10 nonattainment areas that are currently violating the PM10
NAAQS listed in Table n.A-13, there are 25 unclassifiable areas that have recently recorded
ambient concentrations of PM10 above the PM10 NAAQS. EPA adopted a policy in 1996 that
allows areas with PM10 exceedances that are attributable to natural events to retain their
designation as unclassifiable if the State is taking all reasonable measures to safeguard public
health regardless of the sources of PM10 emissions. Areas that remain unclassifiable areas are not
required under the Clean Air Act 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. With respect to the monitored violations reported in 1997-99 in the 25 areas
designated as unclassifiable, we have not yet excluded the possibility that factors such as a one-
time monitoring upset or natural events, which ordinarily would not result in an area being
designated as nonattainment for PM10, may be responsible for the problem. Emission reductions
11-60
-------
Chapter II: Health and Welfare
from today's action will assist these currently unclassifiable areas to achieve ambient PM10
concentrations below the current PM10 NAAQS.
/'/'. Risk of Future Exceedances of the PM10 Standard
The new standards for heavy-duty vehicles will benefit public health and welfare through
reductions in direct diesel particles and NOx, VOCs, and SOx which contribute to secondary
formation of particulate matter. Because ambient particle concentrations causing violations of
the PM10 standard are well established to endanger public health and welfare, this information
supports the new standards for heavy-duty vehicles. The reductions from today's rule will assist
states as they work with the Agency through implementation of local controls including
development and adoption of additional controls as needed to move their areas into attainment by
the applicable deadline, and maintain the standards thereafter.
The Agency's PM inventory analysis performed for this rulemaking predicts that without
additional reductions 10 areas face a significant risk of failing to meet or to maintain the PM10
NAAQS even with federal, State and local controls currently in place. EPA has evaluated
projected emissions for this analysis rather than future air quality because REMSAD, the model
EPA has used for analyses related to this rule, was designed principally to estimate long-term
average concentrations of fine parti culate matter and its ability to predict short-term PM10
concentrations has not been satisfactorily demonstrated. In contrast with ozone, which is the
product of complex photochemical reactions and therefore difficult to directly relate to precursor
emissions, ambient PM10 concentrations are more heavily influenced by direct emissions of
particulate matter and can therefore be correlated more meaningfully with emissions inventories.
In the west, where most of the PM10 nonattainment areas are located, coarse PM is comprised of
70 percent particles composed of minerals, with only small fractions attributable to gaseous
pollutants such as SOx, NOx and ammonia/
Table II. A-14 presents information about these ten areas and subdivides them into two
groups. The first group of six areas are designated PM10 nonattainment areas which had recent
monitored violations of the PM10 NAAQS in 1997-1999 and increasing inventories of PM10 from
1996 to 2030. These areas have a population of 19 million. Included in the group are the
nonattainment areas that are part of the Los Angeles, Phoenix and Las Vegas (Clark County)
metropolitan areas, where traffic from heavy-duty vehicles is substantial. These six areas will
benefit from the reductions in emissions that will occur from the new standards for heavy-duty
vehicles, as will other areas impacted by heavy-duty vehicle emissions.
1 Air Quality Criteria for Particulate Matter, External Review Draft, EPA 600/P-99/002a, Volume 1,
October 1999, at 4.43.
11-61
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The second group of four counties listed in Table HA-14 with a total of nine million
people in 1999 also had predicted exceedances of the PM10 standard. These four areas registered,
in either 1997 or 1998, single-year annual average monitored PM10 levels of at least 90 percent of
the PM10 NAAQS, these areas did not exceed the formal definition of the PM10 NAAQS over the
three-year period ending in 1999. For each of these four areas (ie., Cuyahoga, Harris, New York,
and San Diego), inventories of total PM10 are predicted to increase between 1996, when these
areas recorded values within 10 percent of the PM10 standard, and 2030 when this rule will take
full effect. For some of these areas, total PM10 inventories are predicted to decline or stay
relatively constant from 1996 to 2007, and then increase after 2007. Based on these inventory
projections, the small margin of attainment which the four areas currently enjoy will likely erode
between 1996 and 2030, and for some areas before 2007, if additional actions to reduce the
growth of future emissions are not taken. We therefore consider these four areas to each
individually have a significant risk of exceeding the PM10 standard between 2007 and 2030
without further emission reductions. The emission reductions from the new standards for heavy-
duty vehicles will help these areas attain and maintain the PM10 NAAQS in conjunction with
other processes that are currently moving these areas towards attainment.
11-62
-------
Chapter II: Health and Welfare
Table II.A-14. Areas with Significant Risk of Exceeding the PM10 NAAQS without
Further Emission Reductions between 2007 and 2030
Area
Percent Increases in
PM10 Emissions
(1996-2030)
Areas Currently Exceeding the PM,n standard
Clark Co., NV (Las Vegas)
El Paso, TX *
Hayden/Miami, AZ
Los Angeles South Coast Air Basin, CA
Nogales, AZ
Phoenix, AZ
Subtotal for 6 Areas
41%
14%
4%
14%
3%
24%
Areas within 10% of Exceeding the PM10 Standard
Cuyahoga Co., OH (Cleveland)
Harris, Co., TX (Houston)
New York Co., NY
San Diego Co., CA
Subtotal for 4 Areas
28%
37%
14%
13%
1999 Population
(projected)
(millions)
1.217
0.611
0.004
14.352
0.025
3.012
79.22
1.37
3.26
1.55
2.83
9.01
10 Areas
28.23 million
* EPA has determined that PM10 nonattainment in this area is attributable to international transport. While
reductions in heavy-duty vehicle emissions cannot be expected to result in attainment, they will help reduce the
degree of PM10 nonattainment.
EPA recognizes that the SIP process is ongoing and that many of the 14 current
nonattainment areas in Table II. A-13 are in the process of, or will be adopting and implementing
additional control measures to achieve the PM10 NAAQS in accordance with their attainment
dates under the Clean Air Act. EPA believes, however, that as in the case of ozone, there are
uncertainties inherent in any demonstration of attainment that is premised on forecasts of
emission levels in future years. Even if these areas adopt and submit SIPs that EPA is able to
approve as demonstrating attainment of the PM10 standard, and attain the standard by the
appropriate attainment dates, the inventory analysis conducted for this rule and the history of
PM10 levels in these areas indicates that there is still a significant risk that these areas will need
the reductions from the heavy-duty vehicle standards adopted today to maintain the PM10
standards in the long term (ie, between 2007 and 2030). In addition, this list does not fully
consider the possibility that there are other areas which are now meeting the PM10 NAAQS that
have at least a significant probability of requiring further reductions to continue to maintain it.
11-63
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
/'/'/'. Conclusion
In sum, the Agency believes that ten areas listed in Table II.A-14 have a significant risk
of experiencing paniculate matter levels that violate the PM10 standard from 2007 to 2030. In
addition, this list does not fully consider the possibility that there are other areas which are now
meeting the PM10 NAAQS that have at least a significant probability of requiring further
reductions to continue to maintain it.
b. Public Health and Welfare Concerns from Exposure to Fine PM
/'. Health Effects Studies
There are many studies supporting the Agency's belief that ambient PM causes health and
welfare effects even in areas where PM10 concentrations are below the level of PM10 NAAQS.
This science points to fine PM in particular as being more strongly associated with serious health
effects, such as premature mortality, than coarse fraction PM. The health and welfare studies
support a conclusion that fine PM patterns, that can reasonably be anticipated to occur in the
future, are a serious public health and welfare concern warranting a requirement to reduce
emissions from heavy-duty vehicles, even where they may not constitute violation of the PM10
NAAQS.
The strongest evidence for ambient PM exposure health risks is derived from
epidemiologic studies. The following brief summary focuses on studies completed in the last 10
years on the health and welfare effects of PM. A detailed summary and discussion of the large
body of PM health effects research may be found in Chapters 10 to 13 of the 1996 Air Quality
Criteria for Particulate Matter (known as the Criteria Document or CD).
Many epidemiologic studies have shown statistically significant associations of ambient
PM levels with a variety of human health endpoints in sensitive populations, including mortality,
hospital admissions and emergency room visits, respiratory illness and symptoms, and
physiologic changes in mechanical pulmonary function. The epidemiologic science points to fine
PM as being more strongly associated with some health effects, such as premature mortality, than
coarse fraction PM, which is associated with other health effects.
Associations of both short-term and long-term PM exposure with most of these endpoints
have been consistently observed. Peer-reviewed studies in a variety of locations implicate PM
exposure in increased mortality at levels well below the current 24-hour PM10 NAAQS of 150
//g/m3 and annual PM10 NAAQS of 50 //g/m3. This section will briefly highlight the short-term
exposure studies first and then some of the longer-term exposure studies.
11-64
-------
Chapter II: Health and Welfare
The general internal consistency of the epidemiologic data base and available findings
have led to increasing public health concern, due to the severity of several studied endpoints and
the frequent demonstration of associations of health and physiologic effects with ambient PM
levels at or below the current PM10 NAAQS. Time-series analyses strongly suggest a positive
effect on daily mortality across the entire range of ambient PM levels. Relative risk (RR)
estimates for daily mortality in relation to daily ambient PM concentration are consistently
positive, and statistically significant (atP < 0.05), across a variety of statistical modeling
approaches and methods of adjustment for effects of relevant covariates such as season, weather,
and co-pollutants. Questions remain about the influence of other factors and other issues, and are
described in detail in the Criteria Document. However, even considering the uncertainties, the
Agency believes that the weight of epidemiologic evidence suggests that ambient PM exposure
has affected the public health of U.S. populations.
Within the body of evidence, there is considerable agreement among different studies that
the elderly are particularly susceptible to effects from both short-term and long-term exposures to
PM, especially if they have underlying respiratory or cardiac disease. These effects include
increases in mortality and increases in hospital admissions. Children, especially those with
respiratory diseases, may also be susceptible to pulmonary function decrements associated with
exposure to PM or acid aerosols. Respiratory symptoms and reduced activity days have also
been associated with PM exposures in children.
Numerous time-series analyses published in the late 1980s and early 1990s demonstrate
significant positive associations between daily mortality or morbidity and 24-hour concentrations
of ambient particles indexed by various measures (black smoke, TSP, PM10, PM25, etc.) in
numerous U.S. metropolitan areas and in other countries (e.g., Athens, Sao Paulo, Santiago).8
These studies collectively suggest that PM alone or in combination with other commonly
occurring air pollutants (e.g., SO2) is associated with daily mortality and morbidity, the effect of
PM appearing to be most consistent. In both the historic and recent studies, the association of
PM exposure with mortality has been strongest in the elderly and for respiratory and
cardiovascular causes of death.
Table II.A-15 summarizes effect estimates (relative risk information) derived from
epidemiologic studies demonstrating health effects associations with ambient 24-hour PM10
concentrations in U.S. and Canadian cities. The evidence summarized in Table HA-17 leaves
little doubt that PM concentrations typical of contemporary U.S. urban air sheds are correlated
with detectable increases in risk of human mortality and morbidity. Evidence from studies that
looked at PM indicators other than PM10, summarized in Table II. A-16, also suggests that fine
s In the tables summarizing the studies, relative risks with lower confidence intervals greater than 1.0 are
statistically significant at the 95 percent confidence level. In Table IIA-17, for example, the first entry showing
Portage, WI, with a confidence interval of 0.98 -1.09 is not statistically significant.
11-65
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
particles may be important contributors to the observed PM-health effects associations given the
increased risks (of mortality, hospitalization, respiratory symptoms, etc.) associated with several
different fine particle indicators (e.g., PM2 5, SOJ, H+). In particular, more recent reanalyses of
the Harvard Six-City Study by Schwartz et al. (1996a) examined the effects on daily mortality of
24-hour concentrations of fine particles (PM25), inhalable particles (PM15/10), or coarse fraction
particles (PM15/10 minus PM2 5) as exposure indices. Overall, these analyses suggest that, in
general, the association between excess mortality and thoracic particles appears to be stronger for
the fine than the coarse fraction.
In addition to short-term exposure effects, mortality and morbidity effects associated with
long-term exposure to PM air pollution have been assessed in cross-sectional studies and more
recently, in prospective cohort studies. A number of older cross-sectional studies provided
indications of increased mortality associated with chronic exposures to ambient PM (indexed
mainly by TSP or sulfate measurements); however, unresolved questions regarding adequacy of
statistical adjustments for other potentially important covariates tended to limit the degree of
confidence that could be placed on such studies.
Table HA-17 summarizes some more recent studies using improved methods to examine
relationships between chronic PM exposures indexed by different particle size indicators (PM15,
PM2 5, PM15 to PM2 5). These studies observed associations between increased risk of
mortality/morbidity and chronic (annual average) exposures to PM10 or fine particle indicators in
contemporary North American urban air sheds.
Since the completion of the 1996 PM Criteria Document (CD), many new
epidemiological studies have been published. The PM Criteria Document for the current PM
NAAQS review is now being prepared, and in the CD these many new studies will be reviewed,
summarized, and integrated with what was learned in previous reviews. EPA will await the
completion of the current PM CD before drawing conclusions regarding the findings of this new
body of literature regarding the PM NAAQS.
Separate from the NAAQS review, however, new peer-reviewed studies may be
considered for use in Regulatory Impact Analyses or other such analyses. EPA believes it
appropriate to use the more recent scientific findings for these purposes, especially where the
new information adds value to the analyses. Some of these new studies are described below, and
the findings of these studies will be incorporated in the larger review of the literature contained
in the next PM CD.
Two new Health Effects Institute (HEI) funded studies have received substantial attention
from both scientists and the public: a multi-city analysis of mortality and morbidity associations
with PM10 and other air pollutants (Samet et al., 2000) and the reanalysis of two previous studies
of mortality associations with long-term exposure to PM (Krewski et al., 2000).
11-66
-------
Chapter II: Health and Welfare
The multi-city study, National Morbidity, Mortality and Air Pollution Study (NMMAPS),
evaluated associations between air pollutants and mortality in 90 U.S. cities, and also evaluated
associations between air pollutants and hospital admissions among the elderly in 14 U.S. cities.
The authors report: "Together, the 2 sets of analyses - that of mortality in 90 cities and
hospitalization in 14 cities - provide new and strong evidence linking particulate air pollution to
adverse health effects." (Samet et al., 2000, p. 42)
In the Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study
of Parti culate Air Pollution and Mortality, data were obtained from the original investigators for
two previous studies (Dockery et al., 1993; Pope et al., 1995). The extensive analyses included
replication and validation of the previous findings, as well as sensitivity analyses using
alternative analytic techniques, including different methods of covariate adjustment, exposure
characterization, and exposure-response modeling. The authors concluded: "The risk estimates
reported by the Original Investigators were remarkably robust to alternative specifications of the
underlying risk models, thereby strengthening confidence in the original findings." (Krewski et
al., 2000, p. 234)
Some of these new epidemiology studies have presented interesting new findings related
to mobile source emissions. For example, Laden et al. (2000) used factor analysis with
indicators of parti culate matter from several sources, and reported that among these sources,
parti culate matter from mobile sources had the largest association with mortality in six U.S.
cities. Mar et al. (2000) conducted a similar analysis using data from Phoenix, Arizona, and
report that mortality from cardiovascular diseases was associated with motor vehicle exhaust-
related pollutants. An additional new analysis uses the results of a number of new
epidemiological studies to assess the public health impact of outdoor and traffic-related pollution
for three European countries. The authors report findings of "considerable" public health
impacts for both mortality and morbidity (e.g., bronchitis, exacerbation of existing asthma)
effects (Kunzli et al., 2000). These new studies suggest that particles from mobile source
emissions play a role in ambient PM-related health effects.
In conclusion, the weight of epidemiologic evidence suggests that PM exposures are
correlated with a variety of serious health effects at levels well below the current 24-hour PM10
NAAQS of 150 //g/m3 and annual PM10 NAAQS of 50 //g/m3. Similarly, although relatively few
cohort studies of long-term PM exposure and mortality are available, they are consistent in
direction and magnitude of excess risk with a larger body of cross-sectional annual mortality
studies, and most show positive associations of PM exposure with mortality.
11-67
-------
Table II.A-15. Effect Estimates Per 50 Aig/m3 Increase in 24-hour PM10 Concentrations
From U.S. And Canadian Studies
Study Location
RR (± CI*)
OnlyPM
in Model
Reported
PM10 Levels
Mean (Min/Maxf
increased Total Short-term Exposure Mortality
Six CitiesA
Portage, WI
Boston, MA
Topeka, KS
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
St. Louis, MOC
•Cingston, TNC
Chicago, ILH
Chicago, ILG
Utah Valley, UTB
Birmingham, ALD
^os Angeles, CAF
increased Hospital Admissions
Hespiratorv Disease
Toronto, CAN1
lacoma, WAJ
s[ew Haven, CTJ
Cleveland, OHK
Spokane, WAL
Chronic Obstructive
3ulmonarv Disease
Minneapolis, MNN
Birmingham, ALM
Spokane, WAL
Detroit. MI°
1.04 (0.98, 1.09)
1.06(1.04, 1.09)
0.98 (0.90, 1.05)
1.03 (1.00, 1.05)
1.05 (1.00, 1.09)
1.05 (1.00, 1.08)
1.08(1.01, 1.12)
1.09 (0.94, 1.25)
1.04(1.00, 1.08)
1.03 (1.02, 1.04)
1.08(1.05, 1.11)
1.05(1.01, 1.10)
1.03 (1.00, 1.055)
(for Elderly > 65 yrs.)
1.23(1.02, 1.43)*
1.10(1.03, 1.17)
1.06(1.00, 1.13)
1.06(1.00, 1.11)
1.08(1.04, 1.14)
1.25(1.10, 1.44)
1.13(1.04, 1.22)
1.17(1.08, 1.27)
1.10(1.02. 1.171
18 (±11.7)
24 (±12.8)
27 (±16.1)
31 (±16.2)
32 (±14.5)
46 (±32.3)
28 (1/97)
30 (4/67)
37 (4/365)
38 (NR/128)
47(11/297)
48 (21, 80)
58( 15/177)
30-39 §
37 (14, 67)
41 (19,67)
43 (19, 72)
46 (16, 83)
36 (18, 58)
45 (19, 77)
46 (16, 83)
48 T22. 821
Table II.A-15 continues on next page.
11-68
-------
Table II.A-15 (cont'd). Effect Estimates per 50 Atg/m3 Increase in 24-hour PM10
Concentrations from U.S. and Canadian Studies
Study Location
3neumonia
Minneapolis, MNN
Birmingham, ALM
Spokane, WAL
Detroit, MI°
^schemic HD
Detroit, MIP
RR (± CI*)
OnlyPM
in Model
1.08(1.01, 1.15)
1.09(1.03, 1.15)
1.06(0.98, 1.13)
—
1.02(1.01, 1.03)
RR (± CI*) Reported
Other Pollutants PM10 Levels
in Model Mean (Min/Maxf
— 36(18,58)
— 45 (19, 77)
— 46 (16, 83)
1.06(1.02,1.10) 48(22,82)
1.02 (1.00, 1.03) 48 (22, 82)
increased Respiratory Symptoms
^ower Respiratory
Six CitiesQ
Utah Valley, UTR
Utah Valley, UTS
Cough
Denver, COX
Six CitiesQ
Utah Valley, UTS
Decrease in Lung Function
Utah Valley, UTR
Utah Valley, UTS
Utah Valley, UTW
2.03 (1.36, 3.04)
1.28(1.06, 1.56)T
1.01 (0.81, 1.27)"
1.27(1.08, 1.49)
1.09(0.57,2.10)
1.51(1.12,2.05)
1.29(1.12, 1.48)
55 (24, 86)**
30 (10, 50)**
29(7,51)***
Similar RR 30(13,53)
— 46(11/195)
— 76(7/251)
— 22 (0.5/73)
Similar RR 30(13,53)
— 76(7/251)
— 46(11/195)
— 76(7/251)
— 55(1,181)
* CI = Confidence Interval.
t Min/Max 24-h PM10 in parentheses unless noted
otherwise as standard deviation (± S.D), 10 and 90
percentile (10, 90). NR = not reported.
T Children.
Table II.A-15 References
A Schwartz etal. (1996a).
B Pope etal. (1992, 1994)/O3.
c Dockeryetal. (1992)/O3.
D Schwartz (1993).
F Kinney et al. (1995)/O3, CO.
G Ito and Thurston (1996)/O3.
H Styeretal. (1995).
1 Thurston etal. (1994)/O3.
J Schwartz (1995)/SO2.
r" Asthmatic children and adults.
§ Means of several cities.
* RR refers to total population, not just>65 years.
PEFR decrease in ml/sec.
FEVj decrease.
K Schwartz etal. (1996b).
L Schwartz (1996).
M Schwartz (1994e).
N Schwartz (19941).
0 Schwartz (1994d).
p Schwartz and Morris
(1995)/O3, CO, SO2.
Q Schwartz et al. (1994).
R Pope etal. (1991).
s Pope and Dockery (1992).
T Schwartz (1994g).
w Pope and Kanner (1993).
x Ostro etal. (1991).
11-69
-------
Table II.A-16. Effect Estimates per Variable
Particle Indicators (PM, „ SOI, w
Increments in 24-hour Concentrations of Fine
) From U.S. and Canadian Studies
Short-term Exposure
Mortality
Six CityA
Portage, WI
Topeka, KS
Boston, MA
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
Indicator
PM25
PM25
PM25
PM25
PM25
PM,,
RR (± CI*) per 25 /jg/m3
PM Increase
1.030(0.993, 1.071)
1.020(0.951, 1.092)
1.056(1.038, 1.0711)
1.028(1.010, 1.043)
1.035 (1.005, 1.066)
1.025 (0.998, 1.053)
Reported PM
Levels Mean
(Min/Maxf
11. 2 (±7.8)
12.2 (±7.4)
15.7 (±9.2)
18.7 (±10.5)
20.8 (±9.6)
29.6 (±21.9)
Increased Hospitalization
Ontario, CANB
Ontario, CAN0
NYC/Buffalo, NYD
so;
so;
03
so;
Toronto0 H+ (Nmol/m3)
SO;
PM,,
1.03 (1.02, 1.04)
1.03 (1.02, 1.04)
1.03 (1.02, 1.05)
1.05(1.01, 1.10)
1.16(1.03, 1.30)*
1.12(1.00, 1.24)
1.15 (1.02, 1.78)
R= 3. 1-8.2
R = 2.0-7.7
NR
28.8 (NR/391)
7.6 (NR, 48.7)
18.6 (NR, 66.0)
Increased Respiratory Symptoms
Southern CaliforniaF
Six Cities0
(Cough)
Six Cities0
(Lower Resp. Symp.)
so;
PM25
PM2 5 Sulfur
H+
PM25
PM2 5 Sulfur
H+
1.48(1.14, 1.91)
1.19(1.01, 1.42)"
1.23 (0.95, 1.59)"
1.06 (0.87, 1.29)"
1.44(1.15-1.82)**
1.82 (1.28-2.59)**
1.05 (0.25-1.30)**
R = 2-37
18.0 (7.2, 37)"*
2.5(3.1,61)*"
18.1 (0.8,5.9)*"
18.0 (7.2, 37)"*
2.5 (0.8, 5.9)"*
18.1(3.1,61)"*
Decreased Lung Function
Uniontown, PAE
PM25
PEFR23.1 (-0.3, 36.9) (per 25 ,ug/m3)
25/88 (NR/88)
* CI = Confidence Interval.
* Min/Max 24-h PM indicator level shown in parentheses unless otherwise noted as (± S.D.), 10 and 90
percentile (10,90) or R = range of values from min-max, no mean value reported. NR = not reported.
* Change per 100 nmoles/m3.
Change per 20 //g/m3 for PM2 5; per 5 //g/m3 for PM2 5 sulfur; per 25 nmoles/m3 for H+.
50th percentile value (10,90 percentile).
Table II.A-16 References
A Schwartz etal. (1996a).
B Burnett etal. (1994).
c Burnett etal. (1995) O3.
D Thurston et al. (1992, 1994)
E Neasetal. (1995).
F Ostro et al. (1993).
Schwartz et al. (1994).
11-70
-------
Table II.A-17. Effect Estimates per Increments* in Annual Mean Levels of Fine Particle
Indicators from U.S. and Canadian Studies
Type of Health
Effect & Location
Increased total chronic
Six CityB
ACS Study0
(151 U.S. SMSA)
Increased bronchitis in
Six CityD
Six CityE
24 CityF
24 CityF
24 CityF
24 CityF
Southern California0
Indicator
mortality in adults
PM15/10
PM25
so;
PM25
so:
children
PM15/10
TSP
H+
so;
PM21
PM10
so:
Change in Health Indicator per
Increment in PM*
Relative Risk (95% CI)
1.42(1.16-2.01)
1.31(1.11-1.68)
1.46(1.16-2.16)
1.17(1.09-1.26)
1.10(1.06-1.16)
Odds Ratio (95% CI)
3.26(1.13, 10.28)
2.80(1.17,7.03)
2.65 (1.22, 5.74)
3.02(1.28,7.03)
1.97(0.85,4.51)
3.29(0.81, 13.62)
1.39 (0.99, 1.92)
Range of City
PMLevels
Means (,ug/m3)
18-47
11-30
5-13
9-34
4-24
20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
—
Decreased lung function in children
Six CityD
Six CityE
24 Cityu
24 City1
24 City1
24 City1
PM15/10
TSP
H+ (52 nmoles/m3)
PM,, (15 //g/m3)
SOI (7 /^g/m3)
PM10(17,ug/m3)
No significant changes
No significant changes
-3.45% (-4.87, -2.01) FVC
-3.21% (-4.98, -1.41) FVC
-3.06% (-4.50, -1.60) FVC
-2.42% (-4.30, -.0.51) FVC
20-59
39-114
—
—
—
—
* Estimates calculated annual-average PM increments
for PM10 and PM15; a 25 //g/m3 increase for PM2 5; and
100 nmole/m3 increase for H+.
assume: a 100 //g/m3 increase for TSP; a 50 //g/m3 increase
a 15 //g/m3 increase for SO:, except where noted otherwise; a
Table II.A-17 References
B Dockeryetal. (1993)
c Pope etal. (1995)
D Dockeryetal. (1989)
E Ware etal. (1986)
F Dockeryetal. (1996)
G Abbey etal. (1995a,b,c)
1 Raizenne et al. (1996)
J Pollutant data same as for
Dockeryetal. (1996)
11-71
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Statistically significant increased mortality from daily exposures to fine PM was observed
in cities with longer-term average fine PM concentrations in the range of 16 to 21 ug/m3. It is
reasonable to anticipate that populations exposed to similar or higher levels, now and in the 2007
and later time frame, will also experience cases of premature mortality attributable to short term
exposures to fine PM. In addition to mortality, statistically significant relationships between
daily fine PM levels (or close indicators of fine PM) and increased respiratory symptoms,
decreased lung functions, and increased hospitalizations, have also been observed in U.S. cities.
/'/'. Current and Future Exposures
At the beginning of 1999, State environmental agencies began operating a broad network
of monitoring stations for the measurement of fine particulate matter (measured as particulate
matter having an aerometric diameter less than or equal to 2.5 micrometers, or PM2.5), using the
Federal Reference Method for PM2.5 mass established when the PM2.5 national ambient air
quality standard was promulgated (62 FR 38763, July 18, 1997). The data that have been
submitted to EPA from this network are available in summary form via the internet on EPA's
website (http://www.epa.gov/aqspubl l/annual_summary.html). Copies of raw data may be
obtained by contacting the Information Management Group, Information Transfer and Program
Integration Division within the Office of Air Quality Planning and Standards. Monitors are
generally located within metropolitan statistical areas, although some monitors intended to
measure upwind PM2.5 concentrations are located outside of metropolitan areas. Monitors in
this network report a 24-hour average PM2.5 concentration for each day of successful
monitoring.
At present, virtually all States have completed the quality assurance review and
certification process. Data which have been certified as valid are considered to be reliable,
although for the purposes of characterizing air quality in areas to which people may be exposed,
there must also be a sufficient number of valid samples during the period in question. For the
purposes of this analysis, we have only included data certified by the States as valid, and have
included only data from sites recording eleven or more valid samples in each calendar quarter.
These data are not sufficient for determining whether given areas should be designated under the
Clean Air Act as attainment or nonattainment with the PM2.5 NAAQS. Under EPA regulations,
this would require consideration of 3 years of valid data. However, these data provide a
sufficient basis to estimate the number of people who lived in monitored counties in 1999 in
which annual average concentrations of PM2.5 equaled or exceeded certain specified values.
In this analysis, we focus on the long-term average concentrations of PM25. Accordingly,
we analyze the 1999 PM25 monitoring data, as available, quality assured, and certified by the
states, to estimate the long-term average concentration at each monitor for the final rule. These
data will not be sufficient for predicting attainment or nonattainment with the PM2 5 NAAQS,
which requires three years of data. However, for the purpose of this analysis, the currently
available monitor data will suffice.
11-72
-------
Chapter II: Health and Welfare
Current 1999 PM2.5 monitored values, which cover about a third of the nation's counties,
indicate that at least 40 million people live in areas where long term ambient fine particulate
matter levels are at or above 16 |ig/m3 (37 percent of the population in the areas with monitors),
which is the low end of the range of long term average PM2.5 concentrations in cities where
statistically significant associations were found with serious health effects, including premature
mortality (EPA, 1996).'
Our REMSAD modeled predictions allow us to also estimate the affected population for
the counties which do not currently have PM2.5 monitors. According to our national modeled
predictions, there were a total of 76 million people (1996 populations) living in areas with
modeled annual average PM2.5 concentrations at or above 16 ug/m3 (29 percent of the
population)."
The REMSAD model also allows us to estimate future PM2.5 levels. However, the most
appropriate method of making these projections relies on the model to predict changes between
current and future states. Thus, we have estimated future conditions only for the areas with
current PM2.5 monitored data (which, as just noted, covers about a third of the nation's
counties). For these counties, REMSAD predicts the current level of 37 percent of the
population living in areas where fine PM levels are at or above 16 |ig/m3 to increase to 59
percent in 2030.
It is reasonable to anticipate that sensitive populations exposed to similar or higher levels,
now and in the 2007 and later time frame, will also be at increased risk of premature mortality
associated with exposures to fine PM. In addition, statistically significant relationships have also
been observed in U.S. cities between PM levels and increased respiratory symptoms and
decreased lung functions in children.
Since EPA's examination in the mid-1990s of the epidemiological and toxicological
evidence of the health effects of PM, many new studies have been published that reevaluate or
extend the initial research. The Agency is currently reviewing these new studies to stay abreast
of the literature and adjust as necessary its assessment of PM's health effects. It is worth noting
that within this new body of scientific literature, there are two new studies funded by the Health
Effects Institute, a EPA-industry jointly funded group, that have generally confirmed the mid-
1990s findings of the Agency about the association of fine particles and premature mortality and
various other respiratory and cardiovascular effects. HEI's National Morbidity, Mortality and
Air Pollution Study (NMMAPS), evaluated associations between air pollutants and mortality in
90 U.S. cities, and also evaluated associations between air pollutants and hospital admissions
4 EPA (1996) Review of the National Ambient Air Quality Standards for Particulate Matter: Policy
Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA-452VR-96-013.
11 REMSAD modeling for PM2.5 annual average concentrations. Total 1996 population in all REMSAD
grid cells is 263 million
11-73
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
among the elderly in 14 U.S. cities.v In HEI's Reanalysis of the Harvard Six Cities Study and
the American Cancer Society Study of Paniculate Air Pollution and Mortality, data were
obtained from the original investigators for two previous studies.™x The extensive analyses
included replication and validation of the previous findings, as well as sensitivity analyses using
alternative analytic techniques, including different methods of covariate adjustment, exposure
characterization, and exposure-response modeling/
In conclusion, we believe that in the period 2007 to 2030, when the standards adopted in
today's action will help reduce ambient PM2.5 concentrations, a significant portion of the US
population may be exposed to ambient PM2.5 concentrations that studies have found may cause
adverse health effects.
4. Diesel Exhaust
The following section presents information about the health hazard and potential risk to
public health and welfare posed by exposure to diesel exhaust. The finding of a health hazard
addresses the question of whether exposure to an agent is likely to cause an adverse human
effect, whereas a discussion of risk is an attempt to provide information on the possible
exposure-related impact of the hazard for an exposed population. In this section, we describe in
some detail the cancer, chronic noncancer, and acute health effects associated with exposure to
diesel exhaust and provide the Agency's current position on the potential for environmental
concern. Ambient concentrations and exposure to diesel particulate matter are also described to
put the hazard conclusions in perspective.
a. Cancer and Noncancer Effects of Diesel Exhaust
The EPA has concluded that diesel exhaust is likely to be carcinogenic to humans by
inhalation at occupational and environmental levels of exposure.17 Available evidence shows
that exposure to diesel exhaust may also cause adverse noncancer health effects with episodic,
v Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A. 2000.
The National Morbidity, Mortality and Air Pollution Study: Part II: Morbidity, Mortality and Air Pollution in the
United States. Research Report No. 94, Part II. Health Effects Institute, Cambridge MA, June 2000.
w Dockery, D.W., Pope, C.A., III, Xu, X., Spengler, J.D., Ware, J.H., Fay, M.E., Ferris, E.G., Speizer,
F.E. (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329:1753-1759.
x Pope, C. A., Ill, Thun, M. J., Namboodiri, M. M., Dockery, D. W., Evans, J. S., Speizer, F. E., Heath, C.
W., 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.
y Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M, White
WH.(2000) Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air
Pollution and Mortality. Special Report to the Health Effects Institute, Cambridge MA, July 2000
11-74
-------
Chapter II: Health and Welfare
acute exposures, as well as noncancer and cancer effects to the respiratory system at longer term,
chronic exposures. The draft Health Assessment Document for Diesel Exhaust (draft
Assessment), was reviewed in public session by the Clean Air Scientific Advisory Committee
(CASAC) on October 12-13, 2000.18 CASAC found that the Agency's conclusion that diesel
exhaust is likely to be carcinogenic to humans is scientifically sound. CASAC concurred with
the draft Assessment's findings with the proviso that EPA provide modifications and
clarifications on certain topics. The Agency expects to produce the finalized Assessment in early
2001. Information presented here is consistent with that to be provided in the final Assessment.
In the draft Assessment, the Agency presents evidence to support its determination that
exposure to diesel exhaust is likely to pose a carcinogenic hazard to humans. The most
compelling information to suggest a carcinogenic hazard is the consistent association that has
been observed between increased lung cancer and diesel exhaust exposure in certain
occupationally exposed workers working in the presence of diesel engines. In its review of the
published literature, EPA found that about 30 individual epidemiologic studies show increased
lung cancer risk associated with diesel emissions. In the draft Assessment EPA evaluated 22
studies that were most relevant for risk assessment, 16 of which reported significant increased
lung cancer risks, ranging from 20 to 167 percent, associated with diesel exhaust exposure.
These studies are of varying quality in terms of design and controlling for factors that might
confound a lung cancer response.
Published analytical results of pooling the positive study results show that on average the
lung cancer risks were increased by 33 to 47 percent within a range of 20-89 percent across the
studies. Individual epidemiological studies numbering about 30 show increased lung cancer risks
of 20 to 89 percent within the study populations depending on the study. The magnitude of the
pooled risk increases is not precise owing to uncertainties in the individual studies, the most
important of which is a continuing concern about whether smoking effects have been accounted
for adequately and in some cases whether other PM exposures were also present. While not all
studies have demonstrated an increased risk (six of 34 epidemiological studies summarized by
the Health Effects Institute19 reported relative risks less than 1.0), the fact that an increased risk
has been consistently noted in the majority of epidemiological studies strongly supports the
determination that exposure to diesel exhaust is likely to pose a carcinogenic hazard to humans.
Additional evidence supporting the identification of a cancer hazard for diesel exhaust
includes the observation tumors in animals following applications of various fractions of the
diesel exhaust mixture to skin, and implantation of diesel particles in respiratory tissue.
Recognizing that diesel exhaust is a complex mixture of carbon particles and associated organics
and other inorganics, it is unclear what fraction or combination of fractions is responsible for the
carcinogenicity and other respiratory effects. It has been shown, however, that the carbon
particles as well as the organics have the potential to be active toxicological agents, either
because of the potential to be irritants which cause inflamation, or because of a capacity to
produce mutagenic and/or carcinogenic activity. In the case of the organics (which exist both in
particle and gaseous states in diesel exhaust) some have potent mutagenic and carcinogenic
11-75
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
properties. In addition, some evidence for the bioavailability of these particle adsorbed
compounds has been demonstrated which supports a hypothesis that the adsorbed organics are
bioavailable to the lung as well as being transported to sites distant from the lung.
While much of the available evidence for a cancer hazard in humans comes from
occupational exposures which generally have higher exposures than in the ambient environment,
there is a basis to infer that the lung cancer hazard extends to ambient environmental exposures.
The basis for the ambient environmental cancer hazard recommendation is due, in part, to the
observation that some ambient environmental concentrations and thus exposures are close to or
overlap low-end occupational exposure estimates as discussed below. This potential overlap in
exposures suggests that little extrapolation is necessary or, conversely, that there is no margin or
only a small margin of safety for some in the general population when compared to occupational
exposures where increased cancer risk is observed. Key to the extrapolation is the assumption
that across any population showing a risk, that risk would be proportional to total lifetime
exposure. The proportional assumption is always made by EPA unless there is evidence to the
contrary, and in the case of diesel exhaust, the extrapolation of occupational risk to
environmental exposure levels is more confidently judged to be appropriate due to the potential
for small exposure differences.
Additional evidence for treating diesel exhaust as a carcinogen at ambient levels of
exposure is provided by the observation of the presence of small quantities of many mutagenic
and some carcinogenic compounds in the diesel exhaust. A carcinogenic response believed to be
caused by such agents is assumed not to have a threshold unless there is direct evidence to the
contrary. This is an EPA risk assessment policy choice in the absence of clear contrary evidence.
In addition, there is evidence that at least some of the organic compounds associated with diesel
particulate matter are extracted by lung fluids (i.e., are bioavailable) and, therefore, are available
in some quantity to the lungs as well as entering the bloodstream and being transported to other
sites in the body.
In the late 1980s, the International Agency for Research on Cancer (IARC) determined
that diesel exhaust is "probably carcinogenic to humans" and the National Institute for
Occupational Safety and Health classified diesel exhaust a "potential occupational carcinogen."20
21 Based on IARC findings, the State of California identified diesel exhaust in 1990 as a
chemical known to the State to cause cancer. In 1996, the International Programme on Chemical
Safety of the World Health Organization listed diesel exhaust as a "probable" human
carcinogen.22 In 1998, the California Office of Environmental Health Hazard Assessment
(OEHHA, California EPA) identified diesel PM as a toxic air contaminant due to the noncancer
and cancer hazard and because of the potential magnitude of the cancer risk.23 Most recently, the
U.S. Department of Health and Human Services National Toxicology Program designated diesel
exhaust particles as "reasonably anticipated to be a human carcinogen" in its Ninth Report on
Carcinogens.24 The concern for a carcinogenicity hazard resulting from diesel exhaust exposures
is longstanding and widespread.
11-76
-------
Chapter II: Health and Welfare
The adverse noncancer effects of diesel exhaust are also of concern to the Agency. Acute
(usually episodic, short duration, high concentration) exposures to diesel exhaust have been
associated with a variety of inflammation-related symptoms such as headache, eye discomfort,
asthma-like reactions, nausea and exacerbation or initiation of allergenic hypersensitivity. No
specific recommendations are made by EPA at this juncture about safe or unsafe exposures to
protect from acute effects, since the onset of acute effects is so variable in the population and the
available acute health effects data lacks adequate detail regarding dose-response relationships.
The potential allergenic effects area of diesel exhaust are of growing interest in the health
research community and as additional information emerges, additional review may be warranted.
Chronic (frequent or continuous, long duration, lower concentrations) diesel exhaust
exposure, at sufficient inhalation levels, is judged to constitute a chronic noncancer respiratory
hazard for humans. For chronic diesel exhaust exposure, EPA is completing the development of
an inhalation reference concentration (RfC) for diesel exhaust exposure. The RfC is an estimate
of the continuous human inhalation exposure (including sensitive subgroups) that is likely to be
without an appreciable risk of deleterious noncancer effects during a lifetime. While the limited
amount of human data are suggestive of respiratory distress, animal test data are quite definitive
in providing a basis to anticipate a hazard to the human lung based on the irritant and
inflammatory reactions in the lung of test animals. Thus, EPA believes that chronic diesel
exhaust exposure, at sufficient exposure levels, increases the hazard and risk of an adverse health
effect. Based on CAS AC advice regarding the use of the animal data to derive the RfC, the
Agency will provide an RfC based on diesel exhaust effects in test animals of approximately 5
//g/m3.
In addition, it is also instructive to recognize that diesel exhaust paniculate matter is part
of ambient fine PM. A qualitative comparison of adverse effects of exposure to ambient fine PM
and diesel exhaust paniculate matter shows that the respiratory system is adversely affected in
both cases, though a wider spectrum of adverse effects has been identified for ambient fine PM.
Relative to the diesel PM database, there is a wealth of human data for fine PM noncancer
effects. Since diesel exhaust PM is a component of ambient fine PM, the fine PM health effects
data base can be informative. The final Assessment will discuss the fine PM health effects data
and its relation to evaluating health effects associated with diesel exhaust.
b. The Link Between Diesel Exhaust and Diesel Particulate Matter
Diesel exhaust includes components in the gas and particle phases. Gaseous components
of diesel exhaust include nitrogen compounds, sulfur compounds, organic compounds, carbon
monoxide, carbon dioxide, water vapor, and excess air (nitrogen and oxygen). Among these gas-
phase constituents, at least one of the organic compounds is a known human carcinogen (e.g.,
benzene) while possible or probable human carcinogens are present (e.g., formaldehyde,
acetaldehyde, 1,3-butadiene), along with compounds for which the Agency has set inhalation
reference concentrations as a guidance to protect the public from noncancer health effects (e.g.,
acetaldehyde, acrolein, naphthalene).
11-77
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Diesel particulate matter is either directly emitted from diesel-powered engines (primary
particulate matter) or is formed from the gaseous compounds emitted by a diesel engine
(secondary particulate matter). After emission from the tail-pipe, diesel exhaust undergoes
dilution, reaction and transport in the atmosphere. The primary emission is considered 'fresh',
while 'aged' diesel exhaust is considered to have undergone chemical and physical
transformation. In an urban or industrial environment, or downwind of an area with large
emission sources, diesel exhaust may enter an atmosphere with high concentrations of
compounds capable of transforming some diesel particulate matter organic constituents into
compounds which exhibit greater toxicity than the primary emitted particle. The formation of
nitroarenes is one example of atmospheric transformation of a diesel exhaust organic compound
to a more lexicologically significant compound.25 Some assessments report up to 16 organic
compounds in primary and secondary diesel exhaust with known or suspected carcinogenic
activity or other lexicologically significant effects.26
Primary diesel particles mainly consist of carbonaceous material, with a small
contribution from sulfuric acid and ash (trace metals). Many of these particles exist in the
atmosphere as a carbon core with a coating of organic carbon compounds, or as sulfuric acid and
ash, sulfuric acid aerosols, or sulfate particles associated with organic carbon.27 While
representing a very small portion (less than one percent) of the national emissions of metals, and
representing a small portion of diesel particulate matter (one to five percent), we note that several
trace metals that may have general toxicological significance depending on the specific species
are also emitted by diesel engines in small amounts including chromium, manganese, mercury
and nickel. In addition, small amounts of dioxins have been measured in diesel exhaust, some of
which may partition into the particle phase.
Approximately 80-95 percent of diesel particle mass is in the size range from 0.05-1.0
micrometers with a mean particle diameter of about 0.2 micrometers. These fine particles have a
very large surface area per gram of mass, which make them excellent carriers for adsorbed
inorganic and organic compounds that can effectively reach the lowest airways of the lung.
Approximately 50-90 percent of the number of particles in diesel exhaust are in the ultrafine size
range from 0.005-0.05 micrometers, averaging about 0.02 micrometers. While accounting for
the majority of the number of particles, ultrafine diesel particulate matter accounts for 1-20
percent of the mass of diesel particulate matter.
Diesel particulate matter is mainly attributable to the incomplete combustion of fuel
hydrocarbons as well as engine oil and other fuel components such as sulfur. Diesel exhaust
particles are part of ambient PM2 5, since diesel engines are used to power numerous types of
equipment in many places. Some geographic areas may have higher diesel particulate loading
because of the number of engines that exhaust into the ambient air. While diesel particulate
matter contributes to ambient levels of PM25, the high content of elemental carbon with the
adsorbed organic compounds and the high number of ultrafine particles (organic carbon and
sulfate) in diesel exhaust distinguish it from other noncombustion sources of PM25. In addition,
diesel particulate matter from mobile source diesel engines is emitted into the breathing zone of
11-78
-------
Chapter II: Health and Welfare
humans and thus has a greater potential for human exposure (per kg of emissions) compared to
other combustion particles emitted out of stacks.
While some of the cancer risk may be associated with exposure to the gaseous
components of diesel exhaust, studies suggest that the paniculate component plays a substantial
role in carcinogenicity and noncancer effects. Investigations show that diesel particles (the
elemental carbon core plus the adsorbed organics) induce lung cancer at high doses, and that the
particles, independent of the gaseous compounds, elicit an animal lung cancer response. The
presence of non-diesel elemental carbon particles, as well as the organic-laden diesel particles,
correlate with an adverse inflammatory effect in the respiratory system of animals. Additional
evidence suggesting the importance of the role of paniculate matter in diesel exhaust includes the
observation that the extractible particle organics collectively produce cancer and mutagenic
toxicity in experimental test systems. Many of the individual organic compounds are mutagenic
or carcinogenic in their own right. EPA believes that exposure to whole diesel exhaust is best
described, as many researchers have done over the years, by diesel exhaust concentrations
expressed in units of mass concentration, i.e., micrograms/m3. This dosimeter does not directly
quantify the gaseous component of diesel exhaust exposure.
Overall, information suggests that the diesel particle may be playing a key role(s) in
contributing to the chronic noncancer and carcinogenicity hazards associated with exposure to
diesel exhaust: both as a mechanism of delivery for many of the organics and trace metals into
the respiratory system, and as a physical irritant in and of itself. Given the available information,
it is a reasonable and prudent step to protect public health by proposing regulations on diesel
exhaust. Today's action will reduce exposure to both the particulate phase and the gaseous
component of diesel exhaust as a result of the parti culate matter and NMHC standards adopted.
The emission standards and fuel sulfur limit would not directly limit emissions of trace metals,
but may indirectly do so by encouraging engine designs with better control of engine oil
consumption.2
c. Ambient Concentrations and Exposure to Diesel Exhaust
As stated previously, the current Agency position is that diesel exhaust is likely to be
carcinogenic to humans and that this cancer hazard exists for occupational as well as ambient
levels of exposure. To provide a context in which to assess the potential hazard from ambient
levels of diesel exhaust, EPA uses the mass concentration of diesel particulate matter (as do
many researchers) as the exposure metric for whole diesel exhaust. A summary of diesel
particulate matter concentrations is found in Table HA-21 and levels of ambient exposure and
occupational exposure for some job categories are presented in Table II.A-22.
z We are also proposing in today's action to prohibit the introduction of used motor oil into the fuel
delivery system which would reduce the trace metal content of the fuel (See Section VIII).
11-79
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
/'. Ambient Concentrations
Information about ambient concentrations of diesel particulate matter and the relative
contribution of diesel engines to ambient particulate matter levels is available from source-
receptor models, dispersion models, and elemental carbon measurements. The most commonly
used receptor model for quantifying concentrations of diesel particulate matter at a receptor site
is the chemical mass balance model (CMB). Input to the CMB model includes particulate matter
measurements made at the receptor site as well as measurements made of each of the source
types suspected to impact the site. Because of problems involving the elemental similarity
between diesel and gasoline emission profiles and their co-emission in time and space, it is useful
to carefully quantify chemical molecular species that provide markers for separation of these
sources. Recent advances in chemical analytical techniques have facilitated the development of
sophisticated molecular source profiles, including detailed speciation of organic compounds
which allow the apportionment of particulate matter to gasoline and diesel sources with increased
certainty. Older studies that made use of only elemental source profiles have been published and
are summarized here, but are subject to more uncertainty. It should be noted that since receptor
modeling is based on the application of source profiles to ambient measurements, the CMB
estimates of diesel particulate matter concentrations do not distinguish between on-road and off-
highway sources. In addition, this model accounts for primary emissions of diesel particulate
matter only; the contribution of secondary aerosols is not included.
Dispersion models estimate ambient levels of particulate matter at a receptor site on the
basis of emission factors for the relevant sources and the investigator's ability to model the
advection, mixing, deposition, and chemical transformation of compounds from the source to the
receptor site. Dispersion models can provide the ability to distinguish on-road from off-highway
diesel sources and can be used to estimate the concentrations of secondary aerosols from diesel
exhaust. Dispersion modeling is being conducted by EPA to estimate concentrations of, and
exposures to several toxic species, including diesel particulate matter.
Elemental carbon (EC) is a major component of diesel particulate matter, contributing
approximately 60 to 80 percent of diesel particulate mass, depending on engine technology, fuel
type, duty cycle, lube oil consumption, and state of engine maintenance.28 29 30 31 In most ambient
environments, diesel particulate matter is one of the major contributors to EC, with other
potential sources including gasoline exhaust; combustion of coal, oil, or wood; charbroiling;
cigarette smoke; and road dust. Because of the large portion of EC in diesel particulate matter,
and the fact that diesel exhaust is one of the major contributors to EC in most ambient
environments, diesel particulate matter concentrations can be bounded using EC measurements.
One approach for calculating diesel particulate matter concentrations from EC measurements is
presented in the draft Health Assessment for Diesel Exhaust^2 The surrogate diesel particulate
matter calculation is a useful approach for estimating diesel particulate matter in the absence of a
more sophisticated modeling analysis for locations where EC concentrations are available.
11-80
-------
Chapter II: Health and Welfare
Annual average diesel paniculate matter concentrations measured during or after 1988 in
urban areas are generally greater than 0.6 micrograms/m3 and range up to 3.6 micrograms/m3 in
the South Coast Air Basin and 2.4 micrograms/m3 in Phoenix, AZ (Table II.A-21). Diesel
particulate matter concentrations measured on individual days in urban areas are as high as 46.7
micrograms/m3 in Manhattan, NY, 22 micrograms/m3 in Phoenix, AZ and 13.3 micrograms/m3 in
Riverside, CA, the latter of which includes both primary and secondary diesel particulate matter.
In two dispersion model studies in Southern California, secondary formation of diesel particulate
matter accounted for 27 to 67 percent of the total diesel particulate matter concentrations on
individual days of 2.6 micrograms/m3 and 13.3 micrograms/m3, respectively.33 34 Off-highway
diesel engines also operate in urban areas, and may have contributed to the ambient diesel
particulate matter concentrations reported for CMB studies, depending on the sampling location.
Dispersion modeling conducted in Southern California reported that the on-road contribution to
the reported diesel particulate matter levels ranged from 63-89 percent of the total diesel
particulate matter.35
11-81
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.A-21. Ambient Diesel Particulate Matter Concentrations from Receptor
Modeling, Dispersion Modeling and Elemental Carbon Measurements
Location
West LA, CA
Pasadena, CA
Rubidoux, CA
Downtown LA, CA36
Phoenix area, AZ37
Phoenix, AZ38
California, 15 Air Basins39
Manhattan, NY40
Welby, CO
Brighton, CO41
Azusa, CA
Pasadena, CA
Anaheim, CA
Long Beach, CA
Downtown LA, CA
Lennox, CA
West LA, CA42
Claremont, CA43
Long Beach, CA
Fullerton, CA
Riverside, CA44
Boston, MA
Rochester, NY45
Washington, DC46
South Coast Air Basin47
Year of Sampling
1982, annual
1982, annual
1982, annual
1982, annual
1989-90, Winter
1 994-95, Nov-
1988-92, annual
1993, Springs
1996-97, Winter 60
days
1996-97, Winter 60
days
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
18-19 Aug 1987
24 Sept 1996
24 Sept 1996
25 Sept 1996
1995, annual
1995, annual
1992-1995,
1995-1996.
Diesel PM10
&PM25
/;g/m3
(mean)
4.4
5.3
5.4
11.6
4-22A
0-5.3 (2.4)
0.2-3. 6A
13.2-46.7A
0-7.3(1.7)
0-3.4(1.2)
1.4D
2.0D
2.7D
3.5D
3.5D
3.8D
3.8D
2.4 (4.0)c D
1.9(2.6)°
2.4(3.9)°
4.4(13.3)°
0.8-1.7(1.1)
0.4-0.8 (0.5)
1.0-2.2(1.5)
2.4-4.5E
DieselPM
% of Total
PM
18
19
13
36
9-20
0-27
31-68
0-26
0-38
5
7
12
13
11
13
16
8
8
9
12
6-12
3-6
5-12
Type of Data
Source-
Receptor
Model: Based
on ambient
measurements
at receptor
sites.
Dispersion
Model: Based
on emission
rates from the
majority of
PM sources
contributing to
the area
studied.
Diesel PM
based on EC
measurements.
PM10 The reader should note that 80-95 percent of diesel PM is PM2 5
B Not Available.
° Value in parenthesis includes secondary diesel PM (nitrate, ammonium, sulfate and hydrocarbons) due
to atmospheric reactions of primary diesel emissions of NOx, SO2 and hydrocarbons.
D On-road diesel vehicles only; All other values are for on-road plus off-highway diesel emissions.
E The Multiple Air Toxics Exposure Study in the South Coast Air Basin reported average annual values
for 8 sites in the South Coast Basin.
11-82
-------
Chapter II: Health and Welfare
In addition to these studies, investigations of the concentrations of diesel particulate
matter in some microenvironments and "hotspot" areas have been conducted. One such study in
Manhattan, NY collected ambient particulate matter near a bus stop on Madison Avenue during a
three day period in 1993.48 Source apportionment applied to these samples indicated that diesel
particulate matter concentrations ranged from 13.2 to 46.7 micrograms/m3 and this study
attributed, on average, 53 percent of the total PM10 to diesel exhaust. Interpretation of the results
of this study require some caution due to the methods used to apportion sources. Concentrations
of diesel particulate matter in the vicinity of bus stops may be indicative of concentrations also
experienced by urban dwellers who live and/or work in the vicinity of large on-road diesel
emission sources and these concentrations may contribute significantly to exposures among some
urban dwellers. Kinney et al. (2000) reported elemental carbon concentrations from personal
monitors worn by study participants who were located on sidewalks at four intersections in
Harlem, NY. The elemental carbon concentrations ranged from 1.5 micrograms/m3 to 6
micrograms/m3 and were reported to be associated with diesel bus and truck counts.
In an additional study to assess diesel particulate matter concentrations near heavily
traveled roadways, the California Air Resources Board (ARE) collected data on ambient
elemental carbon concentrations near the Long Beach Freeway for 3 days in December 1993.49
Using emission estimates from their mobile source emissions model, and elemental/organic
carbon composition profiles for diesel and gasoline exhaust, tire wear, and road dust, ARB
estimated that the contribution of freeway diesel traffic resulted in diesel particulate matter
concentrations ranging from 0.7 micrograms/m3 to 4.0 micrograms/m3 above background
concentrations.
A study designed to investigate relationships between diesel exhaust exposure and
respiratory health of children in the Netherlands found that schools within 400 meters of a
freeway had average elemental carbon concentrations of 3.4 micrograms/m3, while schools more
than 400 meters from freeways had average elemental carbon concentrations of 1.4
mi crogram s/m3.50
Recently the South Coast Air Quality Management District completed their Multiple Air
Toxics Exposure Study in the South Coast Air Basin (MATES-II) to investigate spatial
differences in risk from air toxics exposures in the Basin.51 For this study, elemental carbon
concentrations were measured as a surrogate for diesel particulate matter every sixth day for a
one year period from April 1998 through March 1999 at eight locations throughout the South
Coast Basin. Annual average elemental carbon concentrations ranged from 2.4 micrograms/m3
to 4.5 micrograms/m3 across the eight-site network. Monthly mean elemental carbon values
peaked during winter months with maximum monthly elemental carbon reaching 13.4
micrograms/m3.
In a separate study, the California ARB measured elemental carbon concentrations in
vehicles on Los Angeles roadways as a surrogate for diesel particulate matter. In-vehicle
concentrations of diesel particulate matter are an important microenvironmental exposure for
11-83
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
many people.52 Diesel particulate matter concentrations in the vehicle were estimated to range
from approximately 2.8 micrograms/m3 to 36.6 micrograms/m3 with the higher concentrations
measured when the vehicle followed a HDDV.
/'/'. Occupational and Population Exposures
A distinction must be made between ambient concentrations and the concentration of
diesel particulate matter to which people are exposed. Ambient concentrations reflect outdoor
levels of diesel particulate while exposure depends on both the concentrations of diesel
particulate matter and the time spent in various microenvironments where people are exposed.
Since people typically spend a large portion of their day indoors and indoor diesel particulate
matter concentrations are lower than outdoor concentrations (in the absence of an indoor diesel
PM source), then the concentrations to which most people are exposed are expected to be lower
than ambient diesel particulate matter concentrations. Exposure to diesel exhaust is most
commonly measured in terms of diesel particulate matter and is reported as such in the following
section. This information is summarized in the draft Health Assessment for Diesel Exhaust and
briefly summarized here.
Exposure to diesel exhaust has been measured for several occupationally exposed groups
including miners, railroad workers, diesel forklift operators, firefighters, truck drivers,
dockworkers and mechanics. Diesel exhaust occupational exposures (typically measured as
respirable dust) reported for workers in non-coal mines using diesel-powered shuttle cars range
from approximately 38 to 1,280 micrograms/m3.53 54 Diesel exhaust exposures measured among
railroad workers (as smoking-adjusted respirable particulate) ranged from 39 micrograms/m3 for
engineers/firers, to 134 micrograms/m3 for locomotive shop workers and 191 micrograms/m3 for
hostlers.55 Diesel exhaust exposure among firefighters operating diesel engine vehicles ranges
from 4-748 micrograms/m3 which also encompasses the range of diesel exhaust exposures
reported for diesel forklift dockworkers (18.6-64.7 micrograms/m3).56 57 58 59 Diesel exhaust
exposures measured for truck drivers, mechanics and dockworkers using elemental carbon as a
surrogate for diesel particulate matter ranged from 2.0-7.0 micrograms/m3 for road and local
truckers and from 4.8 to 28.0 micrograms/m3 for dockworkers and mechanics.60
For several occupational categories, the occupational exposure and/or environmental
equivalent of the occupational exposure overlap with some current ambient concentrations and
also overlap with exposure estimates provided by the Hazardous Air Pollutant Exposure Model
described below (Table n.A-22). The relevance of the comparison between estimated
occupational exposures and ambient exposures to diesel exhaust is discussed in section d.
Potential for Cancer Risk, below.
11-84
-------
Chapter II: Health and Welfare
Table II.A-22. Occupational and Population Exposure to Diesel Exhaust
Year of
Sampling
Locations
Occupational Exposure for a Minimum 8-Hour Workday
1980's
1980's
1980's
1980 and 1990's
1990's
1990
Non-coal Miners A
Railroad Workers B
Diesel Forklift Dockworkers c
Firefighters/Fire Station Employees D
Public Transit Workers, Airport Ground Crew E
Long- and Short-Haul Truckers, Dockworkers, Mechanics11
Diesel PM,
ug/m3
38 - 1,280
39- 191
9-61
4-748
7-98
2-28
Ambient Exposure Estimates (On-Road) G
1990
1990
1990
1990
National Annual Average
Urban Annual Average
Urban Annual Average Outdoor Workers
Range of Annual Average for Most Highly Exposed by City
0.84
0.92
1.1
0.83 -4.0
California Exposure Estimates (On-Road & Nonroad) H
1990
1995
2000
2010
California Annual Average
Projected California Annual Average
Projected California Annual Average
Projected California Annual Average
2.1
1.5
1.3
1.2
A Watts (1995) and Saverin et al, (1999)
B Woskie et al. (1988)
c NIOSH (1990); Zaebst et al. (1991)
D Friones et al. (1991); NIOSH (1992); Birch and Carey (1996)
E Birch and Carey (1996)
F Zaebst etal. (1991)
G HAPEM-MS3 exposure results for 1990 for on-road sources only. Methodology is described below. These
estimates are for the average population and the uncertainty associated with them is large. In particular, in areas
where diesel vehicles comprise a higher-than-average portion of the vehicle fleet, exposures will be substantially
higher than predicted average exposure estimates.
H California EPA (1998).
To estimate population exposures to diesel paniculate matter the EPA currently uses the
Hazardous Air Pollutant Exposure Model - Mobile Source 3 (HAPEM-MS3).61 This model
provides national and urban-area specific exposures to diesel particulate matter from on-road
sources only. Results for 1990 are presented in Table n.A-22. Modeled atmospheric
concentrations and exposure estimates of diesel PM from on-highway and nonroad sources have
11-85
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
recently been developed as part of the National Air Toxics Assessment (NAT A) National-Scale
Analysis. Results from the National-Scale Analysis are currently in draft form and are
undergoing technical review by States and EPA's Scientific Advisory Board after which time the
data may change. Information on the National-Scale Analysis can be found on the Agency's
Urban Air Toxics Website.62 Table HA-22 also includes exposure estimates for on-road and
nonroad sources modeled by the California EPA's California Population Indoor Exposure Model
(CPIEM). Results from this model are presented below and described in more detail in
California ARB's "Proposed Identification of Diesel Exhaust as a Toxic Air Contaminant
Appendix in Part A: Exposure Assessment".63
The HAPEM-MS3 model estimates personal exposures to diesel particulate matter using
a ratio to ambient CO measurements. Since most ambient CO comes from motor vehicles, we
believe CO exposure is a reasonable surrogate for exposure to other motor vehicle emissions,
including emissions of toxic compounds. The HAPEM-MS3 model is based on the carbon
monoxide (CO) probabilistic NAAQS exposure model (pNEM/CO), which is used to estimate
the frequency distribution of population exposures to CO and the resulting carboxyhemaglobin
levels. The pNEM/CO model has undergone evaluation and the results of this evaluation are
considered applicable to HAPEM-MS3.64 The HAPEM-MS3 model simulates the movement of
individuals between home and work and through 37 microenvironments. CO concentrations are
based on ambient measurements made in 1990 and are related to exposures of individuals in a ten
km radius around the sampling site.
Exposure modeling was conducted for 1990. CO concentration data from ten urban areas
were used to model 1990 exposures. These areas were Atlanta, GA, Chicago, IL, Denver, CO,
Houston, TX, Minneapolis, MN, New York, NY, Philadelphia, PA, Phoenix, AZ, Spokane, WA,
and St. Louis, MO. These areas were selected because a large percentage of the population lived
within reasonable proximity to CO monitors, and also to represent good geographic coverage of
the U.S. 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+) 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 derived from regression
analysis of ambient and personal monitor data. Each microenvironmental factor has a
multiplicative term, which represents ambient exposure, and an additive term, which represents
exposure to emissions originating within microenvironments. These factors were derived by IT
Corporation using paired ambient and personal exposure monitor measurements from CO studies
in Denver and Washington.65 66 In our modeling, we set the additive term to zero, to eliminate
non-ambient sources of CO, such as gas stoves. The multiplicative term has a component that
represents penetration from the ambient air into the microenvironment, and a factor that
represents the proximity of the microenvironment to monitors. Thus, even though a compound
may have a penetration of close to one, the microenvironmental factor could be significantly less
11-86
-------
Chapter II: Health and Welfare
than one if the microenvironment is typically found a significant distance from where CO
monitors are located.
With the 1990 CO exposure estimates generated by the HAPEM-MS3 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.67 68 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 estimates of urban versus rural exposure from the 1993 Motor
Vehicle-Related Air Toxics Study69
Motor vehicle diesel particulate matter and CO emission rates reported by EPA70 are used
to calculate mobile source diesel particulate matter exposures. Methods for the development of
particulate matter emissions used to calculate population exposures can be found in "Analysis of
the Impacts of Control Programs on Motor Vehicle Toxic Emissions and Exposure in Urban
Areas and Nationwide: Volumes I and IT'.71 Diesel particulate matter (DPM) exposures are
calculated as in Equation 1, using a ratiometric approach to CO.
»™uglm>=(COuglm3ICOglml)xDPMglml
Equation 1. Ratiometric Calculation of Diesel Particulate Matter Based on CO Exposures.
To estimate diesel particulate matter emissions, we used EPA's PARTS model. PARTS
is similar in structure and function to the MOBILE series of models and calculates exhaust and
non-exhaust (e.g., road dust) particulate emissions for each vehicle class included in the
MOBILE models. PARTS is currently being modified to account for deterioration, in-use
emissions, poor maintenance and tampering effects, all of which would increase emission factors.
As a result, we believe that HAPEM-MS3 exposure estimates, based on PARTS emission
factors, may underestimate true exposures. A comparison of PARTS HDDV emission factors
with a comprehensive review of HDDV emission factors reported from in-use chassis
dynamometer testing72 and modeling performed by CARB suggests that PARTS may
underestimate HDDV emissions by up to 50 percent. Diesel PM exposures reported here were
adjusted to account for new data demonstrating higher HDDV VMT compared with the HDDV
VMT presented in the "Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volumes I and IF'. A complete
description of the HAPEM-MS3 model can be found in "Final Technical Report on the Analysis
of Carbon Monoxide Exposure for Fourteen Cities Using HAPEM-MS3".73
11-87
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Our methodology for modeling exposure to diesel paniculate matter using HAPEM-MS3
has certain limitations and uncertainties. Our use of HAPEM-MS3 to estimate population
exposures to air toxics was peer reviewed for the 1993 Motor Vehicle Related Air Toxics Study74
and more recently for the EPA (1999) report summarized here.75 76 77 Important aspects of our
modeling approach are addressed in these comments and are summarized briefly here.
A validation study conducted for the pNEM/CO model on which HAPEM-MS3 is based,
indicates that CO exposures for the population in the 5"1 percentile were overestimated by
approximately 33 percent, while those with exposures in the 98tn percentile were underestimated
by about 30 percent. Based on this finding, we expect that HAPEM-MS3 also underestimates
exposures in the highly exposed populations. To assess exposures for highly exposed
populations, we have used 1990 CO concentrations relevant to the most highly exposed
populations to estimate 1990 diesel paniculate matter exposures for different demographic
groups in this population.78
Two aspects of the HAPEM-MS3 model which result in some uncertainty in diesel
particulate matter exposure estimates are: 1) HAPEM-MS3 assumes that the highway fleet
(gasoline plus diesel) emissions ratio of CO to diesel particulate matter can be used as an
adjustment factor to convert estimated CO personal exposure to diesel particulate matter
exposure estimates; and 2) the model does not account for physical and chemical differences
between diesel particulate matter and CO. Even though gasoline vehicles emit the large majority
of CO, gasoline and diesel highway vehicles travel on the same roadways and we are making the
assumption that diesel vehicles will comprise a constant fraction of on-road traffic. Diesel
particulate matter and CO are both relatively long-lived atmospheric species (1-3 days) except
under certain conditions such as precipitation which will more readily remove particulate matter.
Our exposure modeling assumes that for the average person in a modeled air district, CO and
diesel particulate matter are well mixed. We are not attempting to assess exposure in microscale
environments in which these assumptions may not be valid. While our assumptions have
inherent uncertainties, we find that exposure estimates provided by the HAPEM-MS3 model are
lower than the majority of ambient diesel particulate matter concentrations. This comparison
provides some indication that HAPEM-MS3 exposure estimates are in the range of reasonable
exposure estimates for the average population. It is noteworthy that these exposure estimates
underestimate exposures for the more highly exposed populations in part due to the
underestimate of CO exposures in the 98th percentile (discussed above), underestimates of
emission factors by PARTS, and the inability to assess small spatial and temporal scale
environments.
While EPA continues efforts toward improving exposure estimates, the results of current
HAPEM-MS3 exposure modeling are used here to compare exposure ranges to ambient
concentration data for the purposes of characterizing potential environmental risk.
Diesel particulate matter exposure was assessed by on-road vehicle class and found to be
due almost entirely to emissions from HDD Vs. Nationally in 1996, 99 percent of diesel
11-88
-------
Chapter II: Health and Welfare
particulate matter exposure from on-road vehicles is attributable to HDDVs and the rest is
generated mainly by LDDTs. We estimate that in 1990, exposure to diesel particulate matter
ranged from 0.84 micrograms/m3 for the general population to 1.1 micrograms/m3 for outdoor
workers (Table II. A-22). Since HDDV traffic, and therefore exposure to diesel particulate
matter, varies for different urban areas, we used HAPEM-MS3 to estimate annual average
population exposures for ten urban areas.79 Modeled 1990 diesel particulate matter exposures in
Minneapolis, MN (1.0 micrograms/m3), New York, NY (1.6 micrograms/m3), Phoenix, AZ (1.3
micrograms/m3), and Spokane, WA (1.2 micrograms/m3) were all higher than the 1990 urban
exposure average of 0.92 micrograms/m3 for 1990.aa
Since HAPEM-MS3 is suspected to underestimate exposures in the highly exposed
populations, we have used 1990 CO concentrations relevant to the most highly exposed
populations to estimate 1990 diesel particulate matter exposures for different demographic
groups in this population.80 The highest estimated diesel particulate matter exposures ranged up
to 4.0 micrograms/m3 for outdoor children in New York. The highest exposed demographic
groups were those who spend a large portion of their time outdoors. It is important to note that
these exposure estimates are lower than the total exposure to diesel particulate matter since they
reflect only diesel particulate matter from on-road sources.
Annual average exposure to on-road HDDV particulate matter was modeled for 1990 and
1996. We expect annual average nationwide exposures to change proportionally with the change
in the PM emissions inventory. These estimates are for the average population and the
uncertainty associated with them is significant. In particular, in areas where diesel vehicles
comprise a higher-than-average portion of the vehicle fleet, exposures may be substantially
higher than predicted average exposure estimates.
The exposure estimates using HAPEM-MS3 are substantially lower than those reported
by California EPA which range from 1.5 micrograms/m3 in 1995, to 1.3 micrograms/m3 in
2000.81 One significant reason for the difference is that the California estimate is for diesel PM10
from all sources, including off-highway while HAPEM estimates exposures for highway
vehicles only. 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.
HAPEM-MS3 exposure estimates for the general population are also lower than annual
average diesel particulate matter concentrations reported from most receptor and dispersion
models. We have modeled exposure for two urban areas for which there is an estimate of
ambient diesel particulate concentrations (Phoenix, AZ and Denver, CO). In these locations, the
annual average exposure estimates are up to a factor of two lower than ambient concentrations.
aa Memorandum to air docket, May 1, 2000, Determination of demographic groups with the highest annual
averaged modeled diesel PM exposure. Pamela Brodowicz, Office of Transportation and Air Quality.
11-89
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
For example, the modeled annual average exposure for the general population in Phoenix in 1996
is 1.3 |ig/m3 and recent sampling conducted in 1994-1995 in Phoenix indicates that
concentrations of diesel paniculate matter are 2.4 micrograms/m3. In Denver, CO the 1996
exposure estimate for the general population is 0.8 micrograms/m3 and the winter sampling
conducted during the Northern Front Range Air Quality Study indicates that in Welby and
Brighton, CO, average ambient concentrations of diesel particulate matter are 1.7 micrograms/m3
and 1.2 micrograms/m3, respectively. This difference in exposure estimates and ambient
concentrations is expected since a large portion of time is spent indoors by most people (where
diesel PM concentrations are lower than outdoors) and the HAPEM-MS3 exposure estimates do
not include the influence of off-highway sources of diesel particulate matter. Our emissions
inventory suggests that mobile sources account for approximately 98 percent of all diesel
particulate matter emissions and that on-road FtDDVs emit approximately one-third of the diesel
particulate matter with the rest attributable to off-highway equipment.82 Reductions in on-road
diesel particulate matter emissions resulting from today's action will have a substantial impact on
population exposure to diesel particulate matter.
The discrepancy between exposure and ambient concentrations is small for those who
spend a large portion of their day out-of-doors or for those whose microenvironmental exposures
permit greater intrusion of outdoor air (such as those whose occupations require that they spend
substantial time in motor vehicles). For these more highly exposed demographic groups
HAPEM-MS3 still underestimates exposure. Given the ambient concentration data available
from some hotspot studies, exposure to diesel particulate matter for the highly exposed subset
could be quite large and is likely to overlap some occupational exposures to a large degree.
d. Potential for Cancer Risk
The EPA has concluded that diesel exhaust is likely to be carcinogenic to humans by
inhalation at occupational and ambient levels of exposure. While the available evidence leads to
EPA's conclusion that diesel exhaust is a likely human lung carcinogen, the evidence is
insufficient to develop a confident estimate of cancer unit risk. The absence of quantitative
estimates of the lung cancer unit risk for diesel exhaust limits our ability to characterize the
precise magnitude of the cancer impact. Given the absence of a unit risk estimate, we provide a
perspective on the possible risks to gain a better understanding of the potential significance of the
cancer hazard for the general population.
With respect to the estimation of a unit risk for diesel exhaust, risk assessments using
epidemiological studies in the peer-reviewed literature which have attempted to assess the
lifetime risk of lung cancer in workers occupationally exposed to diesel exhaust suggest that lung
cancer risk may range from 10"4 to 10"2. 83 84 85 The Agency recognizes the significant
uncertainties in these studies, and has not used these estimates to assess the possible cancer unit
risk associated with ambient exposure to diesel exhaust.
11-90
-------
Chapter II: Health and Welfare
In the draft Assessment, EPA acknowledged the limitations in confidently characterizing
a unit risk and provided a discussion of the possible cancer risk consistent with occupational
epidemiological findings of increased risk and relative exposure ranges in the occupational and
environmental settings. Such an approach does not produce estimates of cancer unit risk.
Rather, this approach provides a perspective on the possible magnitude of environmental cancer
risk and thus insight about the possible significance of the hazard. We describe here two
approaches to gauge the magnitude of potential cancer risk from ambient exposure to diesel
exhaust. A more complete description of the approaches and the methods used can be found in
the draft Assessment.
One approach to provide a perspective on the possible magnitude of the environmental
cancer risk involves examining the differences between the levels of occupational and ambient
exposures, and assuming that cancer risk posed by exposure to diesel exhaust is linearly
proportional with cumulative lifetime exposure. Risks to the general public are considered to be
of concern if the differences between occupational and ambient exposure are small (i.e., within
one to two orders of magnitude), as they would approach workers' risk as observed in
epidemiologic studies of past occupational exposures.
To compare differences between occupational and ambient exposures, it is necessary to
convert occupational exposure estimates to continuous exposure (e.g., an environmental
equivalent exposure). The relationship between occupational exposure and environmental
equivalent exposure is calculated based on a typical set of assumptions to account for the
difference between the amount of air breathed by a worker during their working lifetime
compared to an individual in the general population during their 70-year lifetime (environmental
equivalent exposure = 0.21 x occupational exposure).bb The environmental equivalent exposures
for the occupational exposures presented in Table II. A-23 range from 0.4 to 269 micrograms/m3.
The environmental equivalent exposure is then compared to ambient diesel exhaust
exposure by calculating an exposure margin (EM) which is the ratio of the environmental
equivalent exposure to ambient exposure. Table II. A-23 presents the ratios of environmental
equivalent exposure to ambient exposures. An EM of one or less indicates that ambient exposure
is comparable to occupational exposure (expressed as the environmental equivalent exposure).
An EM greater than one means that the occupational exposure is greater than the ambient
exposure. Table II. A-23 shows that the EMs based on the average nationwide ambient exposure
(0.84 |ig/m3) may be less than one for low-end occupational exposure and start to approach three
orders of magnitude for high end occupational exposure. The EMs based on a high-end ambient
exposure (i.e., 4.0 |ig/m3) range from less than one to less than two orders of magnitude. This
exposure analysis only addresses on-road sources for DE exposure. With additional diesel
bb The fraction of a worker exposure relevant to a 70-year lifetime exposure is typically calculated by
multiplying the fraction of air inhaled during a typical work shift by the fraction of a week, year and life during
which a worker is exposed: (lOnf/shift / 20nf/day) * (5 days / 7days) * (48 weeks / 52 weeks) * (45 years / 70
years) = 0.21.
11-91
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
exhaust exposures from nonroad sources, there is a potential small margin of exposure and hence
a greater concern for diesel exhaust-induced cancer risk.
Table II.A-23. Occupational and Population Exposure to Diesel Exhaust, Environmental
Equivalent Exposures and Exposure Margins
Occupational Group
Non-coal MinersA
U.S. Railroad Workers8
Firefighters0
Public Transit Workers,
Dockworkers0
Estimated
Occupational
Exposure,
Mg/ni3
38-1,280
39-191
4-748
2-98
Environmental
Equivalent
Exposure,
Mg/ni3
8-269
8-40
0.8-157
0.4-21
Exposure margin
ratio using 0.84
Mg/m3 ambient
exposure
10-320
10-48
1-187
0.5-25
Exposure margin
ratio using 4.0
Mg/m3 ambient
exposure
2-67
2-10
0.2-39
0.1-5
Watts (1995) and Saverin et al, (1999).
B Woskie et al. (1988).
c Friones et al. (1991); NIOSH (1992); Birch and Carey (1996).
D Birch and Carey (1996); Zaebst et al. (1991); NIOSH (1990).
The potential overlap and small margins between occupational and ambient diesel
exhaust exposures demonstrated in this analysis, is a significant public health concern for an
environmental pollutant that is viewed as a likely human carcinogen. Several factors including
the carcinogenicity of diesel, differences in human susceptibility, and our current lack of
information regarding exposure to diesel exhaust from non-road sources all affirm the Agency's
concern regarding the small difference between ambient concentrations and exposures and
occupational exposure levels where the presence of diesel exhaust correlates with an increased
risk of lung cancer.
To further characterize the significance of the potential environmental cancer hazard, the
Agency is using a three step process based on general epidemiological principles to evaluate the
available information. First, the risk of excess lung cancer attributed to occupational exposure to
diesel exhaust is estimated. Second, the exposure margin between occupational and ambient
exposures is considered. Finally, a perspective on the diesel exhaust hazard significance is
derived by proportioning the excess risk from step one by the diesel exhaust exposure margins
provided from step two. This approach is expanded upon below and is explained in more detail
in the draft Assessment.86
11-92
-------
Chapter II: Health and Welfare
In its review of the published literature, EPA found that about 30 individual
epidemiological studies show increased lung cancer risk associated with diesel emissions. In the
draft Assessment EPA evaluated 22 studies that were the most relevant for risk assessment, 16 of
which reported significant increased lung cancer risks, ranging from 20 to 167 percent,
associated with diesel exhaust exposure. Questions remain about the influence of other factors
(e.g., effect of smoking, other particulate sources), the quality of the individual epidemiologic
studies, exposure levels, and consequently the precise magnitude of the increased risk of lung
cancer. Two published analytic studies pooled many of the 30 individual epidemiological studies
and after adjusting for smoking reported a relative risk increase of 1.35 and 1.47. 87 88 For the
purpose of this analysis, we have used these pooled studies to select a relative risk of 1.4 as a
reasonable estimate of the increased lung cancer attributed to exposure to diesel exhaust in
occupational settings.
The relative risk of 1.4 means that the occupationally exposed workers experienced an
extra risk that is 40 percent higher than the 5 percent background lifetime lung cancer risk in the
U.S. population.cc Thus, using the relationship [excess risk = (relative risk-1) x background
risk], the diesel exhaust-exposed workers would have an excess risk of developing lung cancer of
2 percent (10"2) due to occupational exposure to diesel exhaust [(1.4 -1) x 0.05)= 0.02]. In this
analysis, we refer to this value as the occupational population risk.dd This is not a unit risk value.
Since the risk is assumed to be proportional to cumulative lifetime exposure, lower
exposures among the general population compared to the occupational population, decrease the
occupational population risk proportionally. As discussed above, occupational and ambient
exposure estimates indicate that the exposure margins (i.e., the EM ratio) between occupational
and ambient exposures may range from 0.5-320 when comparing occupational environmental
equivalent exposure to the nationwide average ambient exposure of 0.84 //g/m3. If lifetime risks
decrease proportionately with reduced exposure, and if one assumes that past occupational
exposures were at the high end, then the risk from average ambient exposure could be between
10"5 and 10"4 (0.02 H- 320 = 6 x io~5). If occupational exposures were closer to 50 |ig/m3, a value
00 The background rate of 0.05 is an approximated lifetime risk calculated by the method of lifetable
analysis using age-specific lung cancer mortality data and probability of death in the age group taken from the
National Health Statistics (HRS) monographs of Vital Statistics of the U.S. (Vol. 2, Part A, 1992). Similar values
based on two rather crude approaches can also be obtained: (1) 59.8 x 10'5 / 8.8 x 10'3 = 6.8 x 10'2 where 59.8 x 10'5
and 8.8 x 10"3 are respectively the crude estimates of lung cancer deaths (including intrathoracic organs, estimated to
be less than 105 of the total cases) and total deaths for 1996 reported in Statistical Abstract of the U.S. (Bureau of
the Census, 1998, 118th Edition), and (2) 156,900/270,000,000 x 76 = 0.045, where 156,900 is the projected lung
cancer deaths for the year 2000 as reported in Cancer Statistics 9J of American Cancer Society, Jan/Feb 2000),
270,000,000 is the current U.S. population, and 76 is the expected lifespan.
dd As used in this document, population risk is defined as the risk (i.e. a mathematical probability) that lung
cancer might be observed in the population after a lifetime exposure to diesel exhaust. Exposure levels may be
occupational lifetime or environmental lifetime exposures. A population risk in the magnitude of 10"2 translates as
the risk of lung cancer being evidenced in one person in one hundred over a lifetime exposure.
11-93
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
that is represented in several data sets shown in Table HA-23 (with an equivalent environmental
exposure of 11 |ig/m3 and a corresponding EM of 13), then risks from ambient exposure would
approach 10'3 (0.02 + 13 = 2 x 1Q-3).
This analysis establishes a reasonable basis for concern that the general population faces
possible lifetime environmental cancer risk ranging from 10~5 to 10~3. Adding to this concern is
recognition that segments of the population may be additionally exposed to nonroad sources of
diesel exhaust which would increase the risk.
The environmental risk estimates included in the Agency's risk perspective are meant
only to gauge the possible magnitude of risk to provide a means to understand the potential
significance of the lung cancer hazard. The conversion of relative risk to population risk is not
specific to the diesel exhaust data as it would apply to any pollutant exposure for which cancer
risk increases are observed and there is a known background rate for the cancer in question. The
environmental risk estimates are not to be construed as cancer unit risk estimates and are not
suitable for use in analyses which would estimate possible lung cancer cases in exposed
populations.
EPA recognizes that, as in all such risk assessments, there are uncertainties in the
assessment of an environmental risk range. For diesel exhaust, these uncertainties include
limitations in exposure data, uncertainty with respect to the most accurate characterization of the
risk increases observed in the occupational epidemiological studies, chemical changes in diesel
exhaust over time, and extrapolation of the risk from occupational to ambient exposures. As
with any such risk assessment for a carcinogen, despite EPA's thorough examination of the
available epidemiologic evidence and exposure information, at this time EPA can not rule out the
possibility that the lower end of the risk range includes zero.ee However, it is the Agency's best
scientific judgement that the assumptions and other elements of this analysis are reasonable and
appropriate for identifying the risk potential based on the scientific information currently
available.
The Agency believes that the risk estimation techniques that were used in the draft
Assessment to gauge the potential for and possible magnitude of risk are reasonable and the
CASAC panel has concurred with the Assessment's discussion of the possible environmental
ee EPA's scientific judgment (which CASAC has supported) is that diesel exhaust is likely to be
carcinogenic to humans. Notably, similar scientific judgements about the carcinogenicity of diesel exhaust have
been recently made by the National Toxicology Program of the Department of Health and Human Services, NIOSH,
WHO, and OEHA of the State of California. In the risk perspective discussed above, EPA recognizes the possibility
that the lower end of the environmental risk range includes zero. The risks could be zero because (1) some
individuals within the population may have a high tolerance level to exposure from diesel exhaust and therefore are
not susceptible to the cancer risks from environmental exposure and (2) although EPA has not seen evidence of this,
there could be a threshold of exposure below which there is no cancer risk.
11-94
-------
Chapter II: Health and Welfare
risk range with an understanding that some clarifications and caveats would be added to the final
version of the Assessment.
In the absence of having a unit cancer risk to assess environmental risk, EPA has
considered the relevant epidemiological studies and principles for their assessment, the risk from
occupational exposure as assessed by others, and relative exposure differences between
occupational and ambient levels of diesel exhaust exposure.
While uncertainty exists in estimating the possible magnitude of the environmental risk
range, the likely hazard to humans together with the potential for significant environmental risks
leads the Agency to believe that diesel exhaust emissions should be reduced in order to protect
the public's health. We believe that this is a prudent measure in light of:
• the designation that diesel exhaust is likely to be carcinogenic to humans,* the exposure
of the entire population to various levels of diesel exhaust,
• the consistent observation of significantly increased lung cancer risk in workers exposed
to diesel exhaust, and
• the potential overlap and/or relatively small difference between some occupational
settings where increased lung cancer risk is reported and ambient exposures.
Today's action will reduce exposure to the toxic gaseous component of diesel exhaust as a
result of the NMHC standard and we expect that the particulate matter standard in today's action
will result in the implementation of particulate matter control technology (catalyzed particulate
traps) that will significantly reduce particulate matter and additionally remove gaseous
hydrocarbons.
5. Gaseous Air Toxics
This section summarizes our analysis of the impact of the proposed HDV standards on
exposure to gaseous air toxics. Heavy-duty vehicle emissions contain several substances that are
known, likely, or possible human or animal carcinogens, or that have serious noncancer health
effects. These substances include, but are not limited to, benzene, formaldehyde, acetaldehyde,
1,3-butadiene, acrolein, and dioxin. For the purposes of the exposure estimates presented in this
section, we have chosen to focus on those compounds in heavy duty vehicle exhaust that are
known, likely, or possible carcinogens and that have significant emissions from heavy-duty
vehicles.
11-95
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
a. Health Effects
/'. Benzene
Highway mobile sources account for 48 percent of nationwide emissions of benzene and
HDVs account for 7 percent of all highway vehicle benzene emissions.89 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 other
aromatics in the fuel, but is generally about three to five percent. The benzene fraction of
evaporative emissions depends on control technology and fuel composition and characteristics
(e.g., benzene level and the evaporation rate) and is generally about one percent.90
The EPA has recently reconfirmed that benzene is a known human carcinogen by all
routes of exposure.91 Respiration is the major source of human exposure. Long-term respiratory
exposure to high levels of ambient benzene concentrations has been shown to cause cancer of the
tissues that form white blood cells. Among these are acute nonlymphocytic leukemia,ff 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.92'93
Leukemias, lymphomas, and other tumor types have been observed in experimental animals
exposed to benzene by inhalation or oral administration. Exposure to benzene and/or its
metabolites has also been linked with genetic changes in humans and animals94 and increased
proliferation of mouse bone marrow cells.95 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.96
The latest assessment by EPA places the excess risk of developing acute nonlymphocytic
leukemia at 2.2 x 10"6 to 7.7 x 10"6/|ig/m3. There is a risk of about two to eight excess acute
nonlymphocytic leukemia cases in one million people exposed to 1 |ig/m3 over a lifetime (70
years).97 This range of unit risk represents the maximum likelihood (MLE) estimate of risk, not
an upper confidence limit (UCL).
ff 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.
11-96
-------
Chapter II: Health and Welfare
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.98
People with long-term exposure to benzene may experience harmful effects on the blood-forming
tissues, especially the bone marrow. These effects can disrupt normal blood production and
cause a decrease in important blood components, such as red blood cells and blood platelets,
leading to anemia (a reduction in the number of red blood cells), leukopenia (a reduction in the
number of white blood cells), or thrombocytopenia (a reduction in the number of blood platelets,
thus reducing the ability for blood to clot). Chronic inhalation exposure to benzene in humans
and animals results in pancytopenia,gg a condition characterized by decreased numbers of
circulating erythrocytes (red blood cells), leukocytes (white blood cells), and thrombocytes
(blood platelets)."400 Individuals that develop pancytopenia and have continued exposure to
benzene may develop aplastic anemia,1* whereas others exhibit both pancytopenia and bone
marrow hyperplasia (excessive cell formation), a condition that may indicate a preleukemic
state.101102 The most sensitive noncancer effect observed in humans is the depression of absolute
lymphocyte counts in the circulating blood.103
/'/'. 1,3-Butadiene
Highway mobile sources account for approximately 42 percent of the annual emissions of
1,3-butadiene and HDVs account for approximately 15 percent of the highway vehicle portion.104
1,3-Butadiene is formed in vehicle exhaust by the incomplete combustion of fuel. It is not
present in vehicle evaporative emissions, because it is not present in any appreciable amount in
fuel. 1,3-Butadiene accounts for 0.4 to 1.0 percent of total organic gas exhaust, depending on
control technology and fuel composition.105
1,3-Butadiene was classified by EPA as a Group B2 (probable human) carcinogen in
1985.106 This classification was based on evidence from two species of rodents and
epidemiologic data. In the EPA1998 draft Health Risk Assessment of 1,3-Butadiene, that was
reviewed by the Science Advisory Board (SAB), the EPA proposed that 1,3-butadiene is a known
human carcinogen based on human epidemiologic, laboratory animal data, and supporting data
gg 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.
tt Aplastic anemia is a more severe blood disease and occurs when the bone marrow ceases to function,
i.e.,these stem cells never reach maturity. The depression in bone marrow function occurs in two stages -
hyperplasia, or increased synthesis of blood cell elements, followed by hypoplasia, or decreased synthesis. As the
disease progresses, the bone marrow decreases functioning. This myeloplastic dysplasia (formation of abnormal
tissue) without acute leukemiais known as preleukemia. The aplastic anemia can progress to AML (acute
mylogenous leukemia).
11-97
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
such as the genotoxicity of 1,3-butadiene metabolites.107 The Environmental Health Committee
of EPA's Scientific Advisory Board (SAB), reviewed the draft document in August 1998 and
recommended that 1,3-butadiene be classified as a probable human carcinogen, stating that
designation of 1,3-butadiene as a known human carcinogen should be based on observational
studies in humans, without regard to mechanistic or other information.108 In applying the 1996
proposed Guidelines for Carcinogen Risk Assessment, the Agency relies on both observational
studies in humans as well as experimental evidence demonstrating causality and therefore the
designation of 1,3-butadiene as a known human carcinogen remains applicable.109 The Agency
has revised the draft Health Risk Assessment of 1,3-Butadiene based on the SAB and public
comments. The draft Health Risk Assessment of 1,3-Butadiene will undergo the Agency
consensus review, during which time additional changes may be made prior to its public release
and placement on the Integrated Risk Information System (IRIS).
The SAB panel recommended that EPA calculate the lifetime cancer risk estimates based
on the human data from Delzell et al. 1995110 and account for the highest exposure of "360 ppm-
year" for 70 years. Based on this calculation111 the maximum likelihood estimate of lifetime
cancer risk from continuous 1,3-butadiene exposure is 2.21 x 10"6/microgram/m3. This estimate
implies that approximately 2 people in one million exposed to 1 microgram/m3 1,3-butadiene
continuously for their lifetime (70 years) would develop cancer as a result of their exposure.
An adjustment factor of 3 can be applied to this potency estimate to reflect evidence from
rodent studies suggesting that extrapolating the excess risk of leukemia in a male-only
occupational cohort may underestimate the total cancer risk from 1,3-butadiene exposure in the
general population.112 First, studies in both rats and mice indicate that 1,3-butadiene is a multi-
site carcinogen. It is possible that humans exposed to 1,3-butadiene may also be at risk of
cancers other than leukemia and that the epidemiologic study had insufficient power to detect
excess cancer risks for other tissues or sites in the body. Second, both the rat and mouse studies
suggest that females are more sensitive to 1,3-butadiene-induced carcinogenicity than males, and
the female mammary gland was the only 1,3-butadiene-related tumor site common to both
species. Use of a 3-fold adjustment to the potency estimate of 2.21 x 10"6/microgram/m3 derived
from the occupational epidemiologic study yields a upper bound cancer potency estimate of 1.4 x
10"5/microgram/m3, which roughly corresponds to a combination of the human leukemia and
mouse mammary gland tumor risk estimates, at least partially addressing the concerns that the
leukemia risk estimated from the occupational data may underestimate total cancer risk to the
general population, in particular females.
1,3-Butadiene also causes a variety of noncancer reproductive and developmental effects
in mice and rats (no human data) when exposed to long-term, low doses of butadiene.113 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 4.0 microgram/m3 to avoid appreciable risks of these
11-98
-------
Chapter II: Health and Welfare
reproductive and developmental effects.114 EPA has developed a draft chronic, subchronic, and
acute RfC values for 1,3-butadiene exposure as part of the draft risk characterization mentioned
above. The RfC values will be reported on IRIS.
Hi. Formaldehyde
Highway mobile sources contribute approximately 24 percent of the national emissions of
formaldehyde, and HDVs account for approximately 36 percent of the highway portion.115
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 organic
gaseous emissions, depending on control technology and fuel composition. It is not found in
evaporative emissions.
Formaldehyde exhibits extremely complex atmospheric behavior.116 It is 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).
EPA has classified formaldehyde as a probable human carcinogen based on limited
evidence for carcinogenicity in humans and sufficient evidence of carcinogenicity in animal
studies, rats, mice, hamsters, and monkeys.117 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. 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. 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.118 Research has demonstrated that formaldehyde produces
mutagenic activity in cell cultures.119
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. The agency is currently conducting a reassessment of
risk from inhalation exposure to formaldehyde.
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
11-99
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
reported nasal symptoms such as rhinitis (inflammation of the nasal membrane), nasal
obstruction, and nasal discharge following chronic exposure.120 In persons with bronchial
asthma, the upper respiratory irritation caused by formaldehyde can precipitate an acute
asthmatic attack, sometimes at concentrations below 5 ppm.121 Formaldehyde exposure may also
cause bronchial asthma-like symptoms in non-asthmatics.122123
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 in these studies
was decreased maternal body weight gain at the high-exposure level. 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 noncancer health risks, is not available for formaldehyde at this time.
iv. Acetaldehyde
Highway mobile sources contribute 29 percent of the national acetaldehyde emissions and
HDVs are responsible for approximately 33 percent of the highway emissions.124 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 total organic gas exhaust, depending on control
technology and fuel composition.125
The atmospheric chemistry of acetaldehyde is similar in many respects to that of
formaldehyde.126 Like formaldehyde, it is produced and destroyed by atmospheric chemical
transformation. Mobile sources contribute to ambient acetaldehyde levels both by their primary
emissions and by secondary formation resulting from their VOC emissions. Acetaldehyde
emissions are classified as a probable human carcinogen. Studies in experimental animals
provide sufficient evidence that long-term inhalation exposure to acetaldehyde causes an increase
in the incidence of nasal squamous cell carcinomas (epithelial tissue) and adenocarcinomas
(glandular tissue).11 •" 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. The agency is currently conducting a
reassessment of risk from inhalation exposure to acetaldehyde.
n Environmental Protection Agency, Health assessment document for acetaldehyde, Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Research Triangle Park, NC, EPA-
600/8-86/015A (External Review Draft), 1987.
s Environmental Protection Agency, Integrated Risk Information System (IRIS), Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH, 1992.
n-ioo
-------
Chapter II: Health and Welfare
Noncancer effects in studies with rats and mice showed acetaldehyde to be moderately
toxic by the inhalation, oral, and intravenous routes.127128129 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 noncancer health effects.130
v. Acrolein
Highway mobile sources contribute 16 percent of the national acrolein emissions and
HDVs are responsible for approximately 39 percent of these highway mobile source emissions.
Acrolein is extremely toxic to humans from the inhalation route of exposure, with acute exposure
resulting in upper respiratory tract irritation and congestion. The Agency developed a reference
concentration for inhalation (RfC) of acrolein of 0.02 micrograms/m3 1993. Although no
information is available on its carcinogenic effects in humans, based on laboratory animal data,
EPA considers acrolein a possible human carcinogen.131
vi. Dioxins
Recent studies have confirmed that dioxins are formed by and emitted from heavy-duty
diesel trucks and are estimated to account for 1.2 percent of total dioxin emissions in 1995. In
the environment, the pathway of immediate concern is the food pathway (e.g., human ingestion
of certain foods, e.g. meat and dairy products contaminated by dioxin) which may be affected by
deposition of dioxin from the atmosphere. EPA classified dioxins as probable human
carcinogens in 1985. Recently EPA has proposed, and the Scientific Advisory Board has
concurred, to classify one dioxin compound, 2,3,7,8-tetrachlorodibenzo-p-dioxin as a human
carcinogen and the complex mixtures of dioxin-like compounds as likely to be carcinogenic to
humans using the draft 1996 carcinogen risk assessment guidelines.132 Using the 1986 cancer
risk assessment guidelines, the hazard characterization for 2,3,7,8-tetrachlorodibenzo-p-dioxin is
'known' human carcinogen and the hazard characterization for complex mixtures of dioxin-like
compounds is 'probable' human carcinogens. Acute and chronic noncancer effects have also
been reported for dioxin.
b. Assessment of Exposure
This subsection describes the analysis conducted by the Agency to evaluate the impact of
HDV standards on exposure to gaseous toxics present in significant quantities in heavy duty
vehicle exhaust: benzene, formaldehyde, acetaldehyde, and 1,3-butadiene. The information in
n-ioi
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
this section is based on the 1999 'Analysis of the Impacts of Control Programs on Motor Vehicle
Toxics Emissions and Exposure in Urban Areas and Nationwide' ('1999 Study').133
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 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.
We modeled toxic emissions for 10 urban areas and 16 geographic regions selected to
encompass a broad range of I/M programs, fuel parameters, and temperature regimes. These
urban areas and geographic regions are listed in Table II.A-24. The intent of the selection was to
best characterize the different combinations of I/M programs, fuel parameters, and temperature
regimes needed to perform accurate nationwide toxic emissions estimates. Every U.S. county in
the country was then "mapped" to one of these modeled areas or regions (i.e., the emission factor
for the modeled area was also used for the area "mapped" to it). Mapping was done based on a
combination of geographic proximity, I/M program, and fuel control programs. Details of this
process are provided in the 1999 Study. We then multiplied the resulting county level emission
factors by county-level VMT estimates from EPA's Emission Trends Database and summed the
results across all counties to come up with nationwide emissions in tons.
Table II.A-24. Metropolitan Areas and Regions 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/WY
UT/NM/NV
West TX
ND/ SD/ NB/ IA/ KS/ Western MO
APJ 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
n-102
-------
Chapter II: Health and Welfare
Modeling for these areas was accomplished 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.134 Data from the EPA Emission Trends Database and other agency sources
were used to develop appropriate local modeling parameters for I/M programs, Stage II refueling
controls, fuel RVP, average ambient temperature, and other inputs.
These emissions data were used as input to the HAPEM-MS3 exposure model to assess
ambient exposures to the four gaseous toxics discussed in this section. With the 1990 CO
exposure estimates generated by the HAPEM-MS3 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.135136 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 estimates of urban versus rural exposure from the 1993 Motor Vehicle-Related
Air Toxics Study131
Modeled on-road 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 on-road toxic
emissions, as shown in Equation 2:
TOXExposure(tlg/m3) — L^OExposure(tlg/m3)/COEF(g/mi)J1990 x TOXEF(g/mi) (2)
where TOX reflects one of the four toxic pollutants considered in this study.
The ambient exposure estimates for calendar years 1996, 2007, and 2020 were adjusted
for VMT growth relative to 1990. Exposure estimates were adjusted to account for the VOC
emissions modeling conducted for this rulemaking.
To account for atmospheric loss of 1,3-butadiene that varies seasonally1*, exposure
estimates were adjusted using the following multiplicative factors: 0.44 for summer, 0.70 for
spring and fall, and 0.96 for winter.138 These factors account for the difference in reactivity
kk Seasons were defined as Spring (March, April, May); Summer (June, July, August); Fall (September,
October, November); Winter (December, January, February).
n-ios
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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. We evaluated the validity of this assumption by comparing
our results to draft average ambient concentration estimates from the 1996 National Air Toxics
Assessment (NATA). The NATA for 1996 used the same inventory applied to the analysis
presented here for motor vehicle toxics. The Assessment System for Population Exposure
Nationwide (ASPEN) dispersion model was used in the NATA to estimate ambient
concentrations of several mobile source toxics, including aldehydes. Assumptions applied in the
ASPEN model include an estimate that 68 percent of formaldehyde is primary emissions (i.e.
direct emission as opposed to secondary formation in the atmosphere), while only about 20
percent of acetaldehyde is assumed to be primary emissions. The comparison between ASPEN
concentrations and HAPEM-MS3 exposures indicated fairly good agreement for formaldehyde,
but suggested the HAPEM-MS3 exposure estimates for acetaldehyde may be low by a factor of
three. Thus, our acetaldehyde exposure estimates were adjusted upward by a factor of three to
match draft ambient concentration estimates from the National Air Toxics Assessment.
HAPEM-MS3 does not account for exposures originating within microenvironments. For
instance, the model would not account for exposure to evaporative benzene emissions indoors
from vehicles parked in attached garages, or to vehicles during refueling.
Table II. A-25 presents annual average nationwide exposure estimates from all highway
motor vehicles for benzene, acetaldehyde, formaldehyde and 1,3-butadiene. The projected
contribution of HDVs to the highway motor vehicle exposures estimates in 2007 is 13 percent for
benzene, 51 percent for acetaldehyde, 59 percent for formaldehyde, and 9 percent for 1,3-
butadiene. With today's standards in place, exposure to toxics from all HDVs in 2020 would be
reduced by 7 percent for benzene, 20 percent for acetaldehyde, 23 percent for formaldehyde, and
7 percent for 1,3-butadiene. And exposure to toxics from all highway sources in 2020 (Table
II. A-25) would be reduced by 2 percent for benzene, 15 percent for acetaldehyde, 18 percent for
formaldehyde, and 5 percent for 1,3-butadiene.
H-104
-------
Chapter II: Health and Welfare
Table II.A-25. Modeled Average 50-State Ambient Exposure to Gaseous Toxics from All
Highway Motor Vehicles (jig/m3) in 1990,1996, 2007, and 2020 without 2007 HDV
Standards and for 2020 with 2007 HDV Standards
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
1990
1.07
0.51
0.57
0.11
1996
0.71
0.38
0.37
0.08
2007
0.38
0.21
0.18
0.03
2020
0.28
0.22
0.17
0.03
20204
0.28
0.18
0.14
0.03
Percent
Reduction in
2020 with
2007 HDV
Standards8
2%
15%
18%
5%
A Exposure estimates with the 2007 Heavy-Duty Vehicle Standards.
B Percent reductions use exposures calculated to four decimal places.
Separately, exposure estimates were also generated for the 10 urban areas listed in Table
II. A-24. In Denver, CO, Minneapolis, MN, Spokane, WA, Atlanta, GA and Phoenix, AZ,
exposure to these four gaseous toxic compounds resulting from HDV emissions is projected to
be higher than the national average in 2007. Of the cities modeled, Denver, and Phoenix are
projected to have two-fold higher exposure estimates for acetaldehyde, formaldehyde and 1,3-
butadiene from HDVs compared with the national average in 2007.
6. Visibility/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 natural conditions and also 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.
n-105
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 estimated total light extinction in most regions. In some areas, such as the
Cascades region of Oregon, sulfates are estimated to account for over 50 percent of annual
average light extinction. Organic carbon typically is estimated to be 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.139
The CAA requires EPA to address visibility impairment, 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
PM25. The national visibility program established in 1980 requires the protection of visibility in
156 mandatory Federal Class I areas across the country (primarily national parks and wilderness
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 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 would 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
n-106
-------
Chapter II: Health and Welfare
recommendations related to the ... federal national mobile source emissions control strategies",
including Tier 2 vehicle emissions standards.140 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.141 The Commission identifies mobile source
pollutants of concern as VOC, NOX, and elemental and organic carbon.
Visibility is greatly affected by ambient PM2 5 concentration, with PM2 5 concentrations
below the NAAQS being sufficient to impair visibility. Black elemental carbon particles are a
dominant light adsorbing species in the atmosphere 142, and a major component of diesel exhaust.
The reductions in ambient PM2 5 from the standards in this rulemaking are expected to 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.
7. Acid Deposition
Acid deposition, or acid rain as it is commonly known, occurs when SO2 and NOx react
in the atmosphere with water, oxygen, and oxidants to form various acidic compounds that later
fall to earth in the form of precipitation or dry deposition of acidic particles." It contributes to
damage of trees at high elevations and in extreme cases may cause lakes and streams to become
so acidic that they cannot support aquatic life. In addition, acid deposition accelerates the decay
of building materials and paints, including irreplaceable buildings, statues, and sculptures that are
part of our nation's cultural heritage. To reduce damage to automotive paint caused by acid rain
and acidic dry deposition, some manufacturers use acid-resistant paints, at an average cost of $5
per vehicle—a total of $61 million per year if applied to all new cars and trucks sold in the U.S.
Acid deposition primarily affects bodies of water that rest atop soil with a limited ability
to neutralize acidic compounds. The National Surface Water Survey (NSWS) investigated the
effects of acidic deposition in over 1,000 lakes larger than 10 acres and in thousands of miles of
streams. It found that acid deposition was the primary cause of acidity in 75 percent of the acidic
lakes and about 50 percent of the acidic streams, and that the areas most sensitive to acid rain
were the Adirondacks, the mid-Appalachian highlands, the upper Midwest and the high elevation
West. The NSWS found that approximately 580 streams in the Mid-Atlantic Coastal Plain are
acidic primarily due to acidic deposition. Hundreds of the lakes in the Adirondacks surveyed in
the NSWS have acidity levels incompatible with the survival of sensitive fish species. Many of
the over 1,350 acidic streams in the Mid-Atlantic Highlands (mid-Appalachia) region have
already experienced trout losses due to increased stream acidity. Emissions from U.S. sources
n Much of the information in this subsection was excerpted from the EPA document, Human Health
Benefits from Sulfate Reduction, written under Title IV of the 1990 Clean Air Act Amendments, U.S. EPA, Office
of Air and Radiation, Acid Rain Division, Washington, DC 20460, November 1995.
H-107
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
contribute to acidic deposition in eastern Canada, where the Canadian government has estimated
that 14,000 lakes are acidic. Acid deposition also has been implicated in contributing to
degradation of high-elevation spruce forests that populate the ridges of the Appalachian
Mountains from Maine to Georgia. This area includes national parks such as the Shenandoah
and Great Smoky Mountain National Parks.
The SOx and NOx reductions from today's rule would help reduce acid rain and acid
deposition, thereby helping to reduce acidity levels in lakes and streams throughout the country
and help accelerate the recovery of acidified lakes and streams and the revival of ecosystems
adversely affected by acid deposition. Reduced acid deposition levels would also help reduce
stress on forests, thereby accelerating reforestation efforts and improving timber production.
Further deterioration of our historic buildings and monuments, and of buildings, vehicles, and
other structures exposed to acid rain and dry acid deposition also would be slowed, and the costs
borne to prevent acid-related damage may also decline. While the reduction in sulfur and
nitrogen acid deposition would be roughly proportional to the reduction in SOx and NOx
emissions, respectively, the precise impact of today's rule would differ across different areas.
8. Eutrophication and Nitrification
Nitrogen deposition into bodies of water can cause problems beyond those associated
with acid rain. The Ecological Society of America has included discussion of the contribution of
air emissions to increasing nitrogen levels in surface waters in a recent major review of causes
and consequences of human alteration of the global nitrogen cycle in its Issues in Ecology
series."™ Long-term monitoring in the United States, Europe, and other developed regions of the
world shows a substantial rise of nitrogen levels in surface waters, which are highly correlated
with human-generated inputs of nitrogen to their watersheds. These nitrogen inputs are
dominated by fertilizers and atmospheric deposition.
Human activity can increase the flow of nutrients into those waters and result in excess
algae and plant growth. This increased growth can cause numerous adverse ecological effects
and economic impacts, including nuisance algal blooms, dieback of underwater plants due to
reduced light penetration, and toxic plankton blooms. Algal and plankton blooms can also
reduce the level of dissolved oxygen, which can also adversely affect fish and shellfish
populations. This problem is of particular concern in coastal areas with poor or stratified
circulation patterns, such as the Chesapeake Bay, Long Island Sound, or the Gulf of Mexico. In
such areas, the "overproduced" algae tends to sink to the bottom and decay, using all or most of
the available oxygen and thereby reducing or eliminating populations of bottom-feeder fish and
Tam Vitousek, Peter 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.
H-108
-------
Chapter II: Health and Welfare
shellfish, distorting the normal population balance between different aquatic organisms, and in
extreme cases causing dramatic fish kills.
Collectively, these effects are referred to as eutrophication, which the National Research
Council recently identified as the most serious pollution problem facing the estuarine waters of
the United States (NRC, 1993). Nitrogen is the primary cause of eutrophi cation in most coastal
waters and estuaries.™1 On the New England coast, for example, the number of red and
browntides and shellfish problems from nuisance and toxic plankton blooms have increased over
the past two decades, a development thought to be linked to increased nitrogen loadings in
coastal waters. We believe that airborne NOx contributes from 12 to 44 percent of the total
nitrogen loadings to United States coastal water bodies. For example, some estimates assert that
approximately one-quarter of the nitrogen in the Chesapeake Bay comes from atmospheric
deposition.
Excessive fertilization with nitrogen-containing compounds can also affect terrestrial
ecosystems.00 Research suggests that nitrogen fertilization can alter growth patterns and change
the balance of species in an ecosystem, providing beneficial nutrients to plant growth in areas
that do not suffer from nitrogen over-saturation. In extreme cases, this process can result in
nitrogen saturation when additions of nitrogen to soil over time exceed the capacity of the plants
and microorganisms to utilize and retain the nitrogen. This phenomenon has already occurred in
some areas of the U.S.
Deposition of nitrogen from heavy-duty vehicles contributes to these effects. In the
Chesapeake Bay region, modeling shows that mobile source deposition occurs in relatively close
proximity to highways, such as the 1-95 corridor which covers part of the Bay surface. The NOx
reductions from the standards for heavy-duty vehicles should reduce the eutrophi cation problems
associated with atmospheric deposition of nitrogen into watersheds and onto bodies of water,
particularly in aquatic systems where atmospheric deposition of nitrogen represents a significant
portion of total nitrogen loadings.
m Much of this information was taken from the following EPA document: Deposition of Air Pollutants to
the Great Waters-Second Report to Congress, Office of Air Quality Planning and Standards, June 1997, EPA-
453/R-97-011.
00 Terrestrial nitrogen deposition can act as a fertilizer. In some agricultural areas, this effect can be
beneficial.
n-109
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
9. POM Deposition
EPA's Great Waters Program has identified 15 pollutants whose deposition to water
bodies has contributed to the overall contamination loadings to the these Great Waters.pp One of
these 15 compounds, a group known as polycyclic organic matter (POM), are compounds that are
mainly adhered to the particles emitted by mobile sources and later fall to earth in the form of
precipitation or dry deposition of particles. The mobile source contribution of the 7 most toxic
POM is at least 62 tons/yearqq and represents only those POM that are adhered to mobile source
particulate emissions. The majority of these emissions are produced by diesel engines.
POM is generally defined as a large class of chemicals consisting of organic compounds
having multiple benzene rings and a boiling point greater than 100°C. Polycyclic aromatic
hydrocarbons are a chemical class that is a subset of POM. POM are naturally occurring
substances that are byproducts of the incomplete combustion of fossil fuels and plant and animal
biomass (e.g., forest fires). Also, they occur as byproducts from steel and coke productions and
waste incineration.
Evidence for potential human health effects associated with POM comes from studies in
animals (fish, amphibians, rats) and in human cells culture assays. Reproductive, developmental,
immunological, and endocrine (hormone) effects have been documented in these systems. Many
of the compounds included in the class of compounds known as POM are classified by EPA as
probable human carcinogens based on animal data.
The particulate reductions from today's rule would help reduce not only the particulate
emissions from highway diesel engines but also the deposition of the POM adhered to the
particles, thereby helping to reduce health effects of POM in lakes and streams, accelerate the
recovery of affected lakes and streams, and revive the ecosystems adversely affected.
10. Carbon Monoxide
We believe that the aftertreatment technology that would be used to meet the standards
for NOx, and diesel particles would result in a per-vehicle reduction in excess of 90 percent in
CO from baseline levels. As of December 1999, there were 17 CO nonattainment areas with a
population of about 30 million people.143 An additional 24 areas with a combined population of
22 million are designated as CO maintenance areas. The broad trends indicate that ambient
levels of CO are declining. The standards being promulgated today would help reduce levels of
carbon monoxide (CO).
pp Much of this information was taken from the following EPA document: Deposition of Air Pollutants to
the Great Waters-Second Report to Congress, Office of Air Quality Planning and Standards, June 1997,
EPA-453/R-97-011. You are referred to that document for a more detailed discussion.
qq The 1996 National Toxics Inventory, Office of Air Quality Planning and Standards, October 1999.
n-no
-------
Chapter II: Health and Welfare
B. Heavy-Duty Diesel Inventory Impacts
This part of the environmental impact chapter presents the emission inventory reductions
we anticipate from heavy-duty vehicles as a result of our NMHC, NOx, and PM emission
standards and as a result of our fuel sulfur standards. This section provides detail on our
emissions inventory calculations and catalogs changes from the NPRM analysis. In addition, this
section contains a sensitivity analysis of deterioration, tampering, and malmaintenance on PM
emissions.
1. Description of Calculation Method
We calculated our emissions reductions by first determining baseline emissions from
HDVs then determining the percent reduction by calendar year. The determination of the
baseline and controlled inventories is described below.
For the controlled emission inventory, we actually present two cases. These two control
cases are labeled as Air Quality Analysis Case and Updated Control Case. The Air Quality
Analysis Case is used in the county-by-county, hour-by-hour air quality analyses associated with
this rule. This inventory was developed using the assumptions and proposed standards presented
in the NPRM for this rule. Because the detailed air quality analyses take several months to
perform, we had to begin as soon as the NPRM was finalized and were not able to incorporate
any changes in the final standards.
The Updated Control Case incorporates changes in the standards and assumptions from
the NPRM to the FRM. Although the differences are fairly small, the Updated Control Case
more precisely represents the reductions associated with the final standards and is used in our
cost-effectiveness analysis. These updates only affect the control inventory and do not affect the
baseline inventory.
a. Baseline Emissions Inventory
i. HC, NOx, CO, PM, andSOx
In modeling emissions from heavy-duty diesel engines, our intent is to be consistent with
the upcoming MOBILE6 model. MOBILE6 is the upcoming version of the MOBILE model that
we historically use to develop calendar year specific emission factors for highway vehicles. This
model will be publically available in early 2001. However, the new data used develop this model
has been made publicly available for stakeholder review. Therefore, we use new published data
that was developed for use in the upcoming MOBILE6.
H-lll
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Emissions inventories from HDVs were calculated at the county level for 1996, 2007,
2020, and 2030. MOBILES was used to calculate VOC,rr NOx, and CO emissions inventories;
PARTS was used to calculate PM and SOx. Adjustment factors were then applied to account for
the new data collected as part of the development of the upcoming MOBILE6 emission factor
model. This methodology is described and detailed inventories are presented in the docket.144145
The adjustment factors used to incorporate the new data and the development of these adjustment
factors are also described in the docket.146147 No adjustments were made to the brake and tire
wear calculations.
/'/'. Fuel Consumption
To determine the impact of the low sulfur diesel fuel requirement on vehicle operation
costs and on emissions, we first need to calculate the diesel fuel consumption. We calculated
HDDE fuel consumption using Equation 1:
Gallons^ = ]class { VMT^]m/age [FC x TFj } (1)
where:
GallonsCY - fuel consumption in gallons/year
class - LHDDE, MHDDE, HHDDE, and urban bus
VMT- total vehicle miles traveled in a given calendar year by class
MY/age - distribution of vehicles in a calendar year by vehicle age
FC - fuel consumption in gallons per mile
TFage - travel fraction of vehicles from each model year in a given calendar year
VMT projections are described in the same report as the calculations of VOC, NOx, CO,
PM, and SOx.148 The travel fraction is described in the memo which details the adjustment
factors.149
Historical fuel consumption estimates (1987-1996) come from a report performed to
support the upcoming MOBILE6 model.150 These historical fuel consumption estimates suggest
that fuel economy is improving. For future fuel consumption estimates, we extrapolate the
historical estimates into the future using a constant, linear improvement in terms of miles per
gallon. We use a single, weighted average, growth rate for MHDDEs and HHDDEs. This is
because a straight projection of the MHDDE and HHDDE fuel economies would suggest that
HHDDEs would have better fuel economy than MHDDEs beginning in 2020. We don't believe
this is likely because of the lower weight of MHDDEs. Table II.B-3 presents per-vehicle the
HDDE fuel economy estimates for selected years.
Volatile Organic Compounds-This includes exhaust and evaporative hydrocarbon emissions.
H-112
-------
Chapter II: Health and Welfare
Table II.B-1. HDDE Fuel Economy Estimates by Model Year (miles per gallon)
Model Year
1990
2000
2010
2020
LHDDE
10.7
11.8
12.9
14.0
MHDDE
7.7
8.1
8.7
9.4
HHDDE
5.9
6.6
7.3
7.9
Urban Bus
3.6
4.0
4.4
4.8
To fully evaluate the effects of the fuel sulfur level standards, we also need to consider
other sources that will likely consume low sulfur fuel produced for HDDEs. These sources
include light-duty vehicles, off-highway engines, and stationary sources. We refer to the low
sulfur fuel used in sources other than highway engines as spillover.
To include the gallons consumed by light-duty diesel vehicles, we use estimates
developed for our Tier 2 final rule151 and fuel economy estimates of 25 mpg and 16.7 mpg for
light-duty diesel vehicles (LDDV) and light-duty diesel trucks (LDDT), respectively.152 We
divided the VMT values within each of these light-duty diesel fuel categories by the
corresponding MOBILE6 projected fuel economy estimates to derive the diesel fuel consumption
for each category per year.
Highway engines are not the only sources that burn highway diesel fuel. Due to
limitations of the fuel production and distribution system, a considerable amount of low sulfur
diesel fuel is currently consumed in off-highway and other applications. To estimate the amount
of highway diesel fuel consumed by other sources, we used data compiled by the Energy
Information Administration (EIA) which showed that combined 1996 production plus
importation minus exportation of highway diesel fuel was 32.8 billion gallons.153 We then
subtracted our estimates of HDDE and LDV diesel fuel consumption to determine the spillover
to sources other than highway engines.
For future years we estimate that spillover will increase as fuel production increases. We
recognize that spillover could decrease in future years if the highway fuel cost were to increase
significantly with respect to the off-highway fuel cost and if the fuel were redistributed
economically. However, we believe the proportion of spillover is largely driven by the
limitations of the fuel distribution system and that it is not likely to change substantially in
response to this rule.
Hi. Crankcase Emissions
H-113
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
We anticipate some benefits in NMHC, NOx, and PM from the closed crankcase
requirements for turbocharged HDDEs. Based on limited engine testing, we estimate that
crankcase emissions of NMHC and PM from HDDEs are each about 0.01 g/bhp-hr.154 NOx data
varies, but crankcase NOx emissions may be as high as NMHC and PM. Therefore, we use the
same crankcase emission factor of 0.01 g/bhp-hr for each of the three constituents.
iv. Air Toxics
We use baseline gaseous toxic emission estimates for heavy duty gasoline vehicles
prepared by Sierra Research. Sierra developed inventory estimates for several gaseous mobile
source air toxics (MS AT), including acetaldehyde, benzene, 1-3 butadiene, and formaldehyde.155
The Sierra study provided estimates of toxic emissions under various control scenarios for
several years. These specific MSATs were addressed because detailed information on the
emission impacts of emission control technologies, fuel properties, and other parameters were
available for these compounds.
The 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 available elements from MOBILE6 used to develop the VOC inventory
for the Tier 2 final rule. The model accounted for differences in toxic fractions between
technology groups, driving cycles, and normal versus high emitting vehicles and engines ("high
emitters"). Impacts of fuel formulations were also addressed in the modeling.
Sierra modeled toxic emissions for 10 urban areas and 16 geographic regions. The areas
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. Every U. S. county in the
country was then "mapped" to one of these modeled areas or regions (i.e., the emission factor for
the modeled area was also used for the area "mapped" to it). Mapping was done based on a
combination of geographic proximity, I/M program, and fuel control programs.
Modeling for these areas was done on a seasonal basis. Information on fuel properties for
was obtained from surveys conducted by the National Institute for Petroleum and Energy
Research (NIPER) and the American Automobile Manufacturers Association (AAMA) and
additional information from the American Petroleum Institute. Data from the EPA Emission
Trends Database and other agency sources were used to develop appropriate local modeling
parameters for I/M programs, Stage II refueling controls, fuel RVP, average ambient temperature,
and other inputs.
To estimate the effect of the 2007 and later model year heavy-duty engine standards on
toxics inventories, we started with the toxics inventories estimated in the Sierra study assuming
H-114
-------
Chapter II: Health and Welfare
all heavy-duty engine programs up until the 2004 model year standards are in effect. Using these
"baseline" inventory estimates for 2007 and 2020 and the nationwide vehicle miles traveled
estimates from the same study, we then estimated the "baseline" gram per mile emissions for the
five toxics (on a nationwide, average basis) for 1996, 2007 and 2020. The emission factors for
other years, were interpolated from these estimates.
Finally, we then multiplied the gram per mile estimates by the nationwide vehicle mile
traveled estimates developed for this rule, to obtain the heavy-duty gasoline and diesel vehicle
toxic inventories used in this analysis. Because benzene has an exhaust and an evaporative
component, we applied the percent reduction based on total (exhaust and evaporative) NMHC
benefits. For formaldehyde, acetaldehyde, and 1,3-butadiene, which do not have an evaporative
component, we applied the percent reduction based on exhaust NMHC only.
b. Controlled Emissions Inventory (Air Quality Analysis Case)
i. HC, NOx, CO, PM, andSOx
To determine the emissions reductions in NMHC, NOx, CO, and PM we look at the
percent emission reductions expected from new engines then calculate percent reductions by
calendar year using the travel fractions discussed above. For the Air Quality Analysis Case we
base the calculations on the proposed HDV standards. This methodology is described and
detailed inventories are presented in the docket.156157 We assume that manufacturers will design
their engine with a compliance margin below the standards. Based on historical certification
data, we use an eight percent compliance margin for HDDEs and a 25 percent compliance margin
for HDGVs.
Based on our analysis of the aftertreatment technology described in Chapter IE, HDDEs
meeting the standards should have very low levels of CO. Although the standards give
manufacturers the same phase-in for NMHC as for NOx, we model the NMHC reductions to be
fully in place for diesel engines in 2007. We believe the use of aftertreatment for PM control
will result in HDDEs meeting the NMHC standards in 2007 and will result in 90 percent
reductions in CO levels soon as the PM standard goes into effect in 2007. In the Air Quality
Analysis Case, we assume that particulate traps will result in a 90 percent reduction in NMHC;
however, as discussed later, we changed this assumption in the Updated Control Case.
We assume that hot soak, diurnal and resting loss emissions from HDGVs would be
reduced proportionally to the reduction in the evaporative emission standard. However, we only
apply these reductions to the emissions of HDGVs which pass the EPA pressure and EPA purge
functional test procedures. We do not claim any benefits from HDGVs which fail these tests.
The majority of the projected PM reductions from HDDEs are directly a result of the PM
standard. However, some PM reductions will come from reducing sulfur in the fuel. Reducing
sulfur in the fuel decreases the amount of direct sulfate PM (DSPM) emitted from heavy-duty
H-115
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
diesel engines and other engines using highway fuel. This section describes the calculations
exhaust emission PM benefits that are directly the result of the 2007 standards. DSPM benefits
from the existing fleet are calculated separately and are discussed later. For SOx reductions, we
look at the reduction of sulfur in the fuel and the amount of sulfur in the fuel that can be assumed
to be converted to SO2.
The control emission factors and percent reductions by calendar year are described in
more detail in the docket.158159
ii. Direct Sulfate PM
Once the low sulfur diesel fuel requirements go into effect, pre-2007 model year HDDEs
will be using low sulfur fuel, as will engines using new PM control technology. Because these
pre-2007 engines will be certified with high sulfur fuel, they will achieve reductions in PM
beyond their certification levels.
For engines built prior to 2007 that use low sulfur fuel in 2007 and later, we need to
calculate the PM benefit associated with the reduction of direct sulfate PM. Equation 2 shows
how we calculate this benefit and express it in terms of an emission factor. We did not consider
deterioration for DSPM which is consistent with our analysis of total PM. We must calculate the
per-vehicle average g/mi reduction independently for each class and calendar year.
DSPMTONS = 10-6 x ppmS x MWR x Sconv xFFxFCx density/2000 (2)
where:
DSPMTONS - direct sulfate PM for a given calendar year [short tons]
ppmS = average fuel sulfur level expressed in parts per million
MWR - molecular weight ratio of DSPM measured on a filter to sulfur in the fuel
= 224/32 (224 is the molecular weight of H2SO4hydrated seven times)
SConv - % of sulfur in fuel converted to direct sulfate PM
FF - fraction of VMT from pre-2007 MY fleet
FC - total consumption of fuel intended for UDDEs in gallons
density - fuel density = 7.1 Ibs/gallon
For the reduction in average fuel sulfur level, we use 334 ppm. We base this reduction on
an average baseline fuel level of 340 ppm S and an average low sulfur fuel level of 7 ppm S with
adjustments for sulfur in the oil. We estimate that oil adds the equivalent of about 1 ppm S to the
fuel. In the baseline case most of the crankcase vapor is vented to the atmosphere which
minimizes the oil burned in the cylinder. In the control case where there are closed crankcase
requirements, we consider the oil recovery system discussed in Chapter HI.
H-116
-------
Chapter II: Health and Welfare
We use the fuel consumption estimates described above in Section B.l.a.iv. This fuel
consumption includes highway fuel burned in heavy-duty engines, light-duty vehicles, and other
sources which use distillate fuel.
For engines not using aftertreatment, we assume that 2 percent of the sulfur in the fuel is
converted to direct sulfate PM. This conversion rate is consistent with the PARTS emission
model. We model the use of low sulfur fuel to begin in mid-2006.
Hi. Crankcase Emissions
By routing crankcase vapors to the exhaust upstream of the aftertreatment systems,
HDDE manufacturers should be able to reduce crankcase emissions by about the same
percentage as for engine-out exhaust. For this analysis, we recognize that the crankcase
emissions will be included in the total exhaust emissions when the engine is designed to the
standards. Because exhaust emissions would have to be reduced slightly to offset any crankcase
emissions, the crankcase emission control is functionally equivalent to a 100 percent reduction in
crankcase emissions.
The engine data we use to determine crankcase emission levels is based on new FtDDEs.
We do not have data on the effect of in-use deterioration of crankcase emissions. However, we
expect that these emissions would increase as the engine wears. Therefore, this analysis may
underestimate the benefits that would result from our crankcase emission requirements.
iv. Air Toxics
We use the same methodology to calculate the controlled toxics inventory as the baseline
inventory. We lack data on how the toxic fractions of the hydrocarbons may change for engines
designed to meet the new standards; therefore, we assume for the sake of analysis that the toxic
fractions do not change. In other words, we assume the same percent reductions in air toxics as
we calculate for hydrocarbons.
c. Controlled Emissions Inventory (Updated Control Case)
The main purpose of the updated control case is to consider changes between the
standards proposed in the NPRM and the standards finalized today. For these calculations, we
consider the heavy-duty vehicle standards as presented in Table n.B-1 and the standards phase-in
dates presented in Table II.B-2. All HDDEs are engine-certified, however; most heavy-duty
gasoline vehicles are chassis-certified. We refer to gasoline engines sold as part of a chassis as
"completes" and require these engines to be certified on a chassis-based test provided that the
vehicle does not have a gross vehicle weight rating more than 14,000 pounds. Other gasoline
engines are tested on an engine dynamometer and we refer to these as "incompletes."
H-117
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.B-2. Heavy-Duty Vehicle Exhaust Emissions Standards
Class
HDDE
HDGV, 2b Completes
HDGV, 3 Completes
HDGV Incompletes
Units
g/bhp-hr
g/mile
g/mile
g/bhp-hr
NMHC
0.14
0.195
0.230
0.14
NOx
0.20
0.2
0.4
0.20
PM
0.01
0.02
0.02
0.01
Table II.B-3. Heavy-Duty Vehicle Standards Phase-In (percent of production)
Model
Year
2007
2008
2009
2010+
HDDE (NMHC & NOx)
NPRM FRM
25% 50%
50% 50%
75% 50%
100% 100%
HDDE (PM/
NPRM FRM
100% 100%
100% 100%
100% 100%
100% 100%
HDGV3
NPRM FRM
100% 0%
100% 50%
100% 100%
100% 100%
A This applies to the closed crankcase requirement as well.
B This applies to evaporative emission standards as well.
As shown above, the actual values of the standards have not changed since proposal.
However, the implementation dates have changed somewhat. One other change is that we
assume that diesel engine manufacturers will design their engines to meet the NMHC with a
small compliance margin. In the NPRM, we assumed that particulate traps would result in a 90
percent reduction in NMHC. This is discussed in more detail in Chapter HI. Other than for
NMHC, the net effect of the changes in the FRM from the NPRM for HDDEs is small.
However, the FRM implementation dates essentially delay the HDGV standards by a year and a
half. The Updated Control Case calculates the reductions from HDVs using the same
methodology as the Air Quality Analysis Case except that the new HDDE NMHC assumptions
and FRM implementation dates are used.
Also, we consider the low sulfur diesel temporary compliance flexibilities and hardship
provisions in our calculations. These provisions allow as little as 75 percent of highway diesel
fuel sales to be 15 ppm sulfur beginning in 2006; increasing to 100% in 2010. In the NPRM, we
H-118
-------
Chapter II: Health and Welfare
proposed to require that all highway diesel fuel meet the standard in 2006. This delay in
production affects SOx and DSPM benefits from HDDEs.
2. HDDE Emission Reductions
a. Anticipated Reductions due to the New HDDE Standards
This section looks at tons/year emission inventories of NOx, PM, and NMHC from
HDDEs. These are the emissions that we are directly regulating from HDDEs. We present our
projected baseline and controlled emissions inventories in addition to our anticipated benefits.
Where there is a difference, we present both the results from the Air Quality Analysis Case
(AQAC) and the Updated Control Case (Updated). In addition, this section presents the total
production of highway diesel fuel which will be required to meet the low sulfur standard set
today.
/'. NOx Reductions
Today's standards should result in about a 90 percent reduction in NOx from new
engines. Table n.B-4 presents these projections with the estimated NOx benefits for selected
years.
Table II.B-4. Nationwide NOx Emissions from HDDEs
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Baseline
2,650
2,440
2,310
2,350
2,770
Controlled
AQAC Updated
2,620 2,600
2,020 2,040
1,080 1,090
582 587
291 292
Reduction
AQAC Updated
29 57
416 403
1,230 1,220
1,770 1,760
2,480 2,480
/'/'. PM Reductions from 2007 Model Year and Later
This section just looks at exhaust emission PM benefits that are directly the result of the
2007 standards. DSPM benefits are presented later. For engines meeting the new standards, we
consider low sulfur fuel to be necessary to enable the PM control technology. In other words, we
H-119
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
don't claim additional emissions reductions beyond the standard due to reductions in direct
sulfate PM except for the difference between certification and average in-use fuel sulfur levels as
discussed above.
The new standards should result in about a 90 percent reduction in exhaust PM from new
engines. This translates to a 16 percent reduction in total PM10 when brake and tire wear are
considered. Table II.B-5 presents these projections with the estimated PM benefits for selected
years. This table includes brake and tire wear, but does not include the direct sulfate benefits
from the existing fleet. These results do not change between the AQAC and Updated analyses.
Table II.B-5. Nationwide PM10 Exhaust and Break/Tire Wear Emissions from HDDEs
Without Existing Fleet Reductions (thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
Baseline Exhaust
96
84
80
86
104
Control Exhaust
91
57
28
15
8
Reduction
5
27
51
71
96
Brake/Tire Wear
13
15
17
19
23
Hi. NMHC Reductions
Although the standards give manufacturers the same phase-in for NMHC as for NOx,
we model the NMHC reductions to be fully in place for diesel engines in 2007. As discussed
earlier, we believe the use of aftertreatment for PM control will cause the NMHC levels to meet
the standard as soon as the PM standard goes into effect in 2007, but in the Updated Control
Case, no longer assume a 90 percent reduction due to the particulate trap. This standard will
result in about a 30 percent reduction in NMHC from new engines. Table n.B-6 presents these
projections with the estimated NMHC reductions for selected years.
n-120
-------
Chapter II: Health and Welfare
Table II.B-6. Nationwide NMHC Exhaust Emissions from HDDEs
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Baseline
184
185
191
206
240
Controlled
AQAC Updated
175 182
132 172
74 156
40 152
25 167
Reduction
AQAC Updated
9 2
53 13
117 35
166 54
217 74
iv. Fuel Consumption Estimates
Table II.B-7 presents national fuel consumption estimates for HDDEs. Table HB-8
presents our estimates of low sulfur fuel consumption. This total consumption includes on-
highway fuel used by light duty diesel vehicles and spillover into sources other than on-highway.
Our total consumption estimates are similar to EIA's production estimates and our highway fuel
consumption estimates are consistent with Federal Highway Association estimates of taxed
highway diesel fuel use.160
Table II.B-7. HDDE Fuel Consumption Estimates by Calendar Year (billion gallons)
Calendar Year
2007
2010
2015
2020
2030
LHDDE
4.26
4.52
4.93
5.30
5.95
MHDDE
5.57
5.94
6.53
7.06
8.02
HHDDE
26.4
27.9
30.4
32.6
36.5
Urban Bus
0.86
0.91
0.99
1.06
1.18
n-i2i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.B-8. Consumption of Highway Diesel Fuel Including Spillover (billion gallons)
Calendar Year
2007
2010
2015
2020
2030
Light-duty
0.37
0.39
0.43
0.46
0.53
Heavy-duty
37.1
39.3
42.8
46.0
51.6
Spillover
4.09
4.25
4.51
4.78
5.30
Total
41.5
44.0
47.8
51.2
57.5
V.
DSPM Reductions from Existing Fleet
Figure HB-1 shows our national projections (using the Updated Control Case) of direct
sulfate PM emissions from the pre-2007 engines using HD highway diesel fuel with and without
the low sulfur fuel. The low sulfur fuel should result in about a 95 percent reduction in direct
sulfate PM from pre-2007 engines. Table II.B-9 presents the estimated DSPM benefits from
HDDEs and other engines using the same fuel for selected years.
H-122
-------
Chapter II: Health and Welfare
7,000
• base fuel
• low S fuel
2000 2005 2010 2015 2020 2025 2030
Figure II.B-1. Projected DSPM from Pre-2007 Engines Using Highway Diesel Fuel
Table II.B-9. Existing Fleet PM Reductions From Low Sulfur Fuel
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
HDDEs
AQAC Updated
6.07 4.64
4.19 4.19
2.16 2.16
1.00 1.00
0.15 0.15
Other
AQAC Updated
0.73 0.56
0.50 0.50
0.25 0.25
0.11 0.11
0.02 0.02
Total Reductions
AQAC Updated
6.80 5.20
4.69 4.69
2.41 2.41
1.12 1.12
0.17 0.17
VI.
Crankcase Emission Reductions
H-123
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.B-10 presents our estimates of the baseline crankcase emissions from HDDEs.
As described earlier, we assume that the crankcase emissions would be zero for the controlled
case. These calculations do not differ between the AQAC and Updated analyses.
Table II.B-10. Crankcase Emissions from Uncontrolled HDDEs
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
NOx
0.7
3.7
7.1
9.5
12.8
PM
0.7
3.7
7.1
9.5
12.8
NMHC
0.7
3.7
7.1
9.5
12.8
vii. Sum of NOx, PM, and NMHC Reductions
As discussed above, we are anticipating large emission reductions in NOx, PM, and
NMHC from HDDEs as a result of the new exhaust emission standards. In addition, we are
anticipating reductions in PM from the existing fleet due to the low sulfur fuel and reductions
from 2007 and later MY engines due to the closed crankcase requirements. Table n.B-11
presents the total projected reductions from HDDEs for this rule for selected years.
Table II.B-11. Total Reductions from HDDEs for this Rule
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
NOx
AQAC Updated
29 58
419 406
1,240 1,230
1,780 1,770
2,490 2,490
PM
AQAC Updated
13 11
35 35
61 61
82 82
109 109
NMHC
AQAC Updated
10 2
57 17
124 43
175 64
229 87
n-124
-------
Chapter II: Health and Welfare
This action is the second of two rules requiring large reductions in NOx emissions from
HDDEs. The 2004 standards reduce NOx from 4 g/bhp-hr to about 2.3 g/bhp-hr. The new
standards will reduce NOx again by another 2.1 g/bhp-hr in 2007. This is a 95 percent reduction
in NOx from new engines. Figure n.B-2 presents (using the Updated Control Case) the
combined effects of the two standards on national HDDE NOx emissions. This figure also
includes crankcase emissions.
,000
re
o
.E
x
O
-1998stds
-2004stds
• 2007stds
2000 2005
2010
2015
2020
2025 2030
Figure II.B-2. Projected HDDE NOx Emissions Due to 2004 and 2007 Standards
Figure II.B-3 shows (using the Updated Control Case) our national projections of total
PM emissions with and without the new engine controls. This figure includes brake and tire
wear, crankcase emissions, and the direct sulfate PM benefits due to the use of low sulfur fuel by
the existing fleet.
H-125
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
160,000
140,000
^-120,000
(0
a)
• baseline
• controlled
• brake/tire wear
2000
2005
2010
2015
2020
2025
2030
Figure II.B-3. Projected Nationwide PM Emissions from HDDEs
n-126
-------
Chapter II: Health and Welfare
Figure II.B-4 shows (using the Updated Control Case) our national projections of total
NMHC crankcase and exhaust emissions from HDDEs with and without the new engine controls.
300,000
250,000 -
•Z1
• baseline
• controlled
0
T 1 1 r
2000 2005 2010 2015 2020 2025 2030
b.
Figure II.B-4. Projected Nationwide NMHC Emissions from HDDEs
Additional Reductions due to the New HDDE Standards
This section looks at tons/year emission inventories of carbon monoxide (CO), oxides of
sulfur (SOx), and air toxics from HDDEs. Although we are not including explicit new standards
for these pollutants in today's action, we believe the new standards will result in reductions in
CO, SOx, and air toxics. Here we present our anticipated benefits.
/'. CO Reductions
Although the CO standard for HDDEs remains at 37.1 g/bhp-hr, CO emission levels from
certified HDDEs are much lower. According to the emission factor report161 we use for baseline
EFs and DFs, baseline emissions for CO range from 1.0 to 1.3 g/bhp-hr for HDDEs. We believe
that the exhaust emission control technology that would be used to meet the standards would
result in excess of a 90 percent reduction in CO from baseline levels. This is because PM traps
have very high oxidation capabilities. We use 90 percent here to be conservative. Using this
assumption, Table n.B-12 presents projected reductions in CO from HDDEs. These results do
not change between the AQAC and Updated analyses.
H-127
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.B-12. Reductions in CO from HDDEs
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
CO Reduction
56
317
691
982
1,290
/'/'. SOx Reductions
We assume that all of the sulfur in the fuel not converted to direct sulfate PM is converted
to sulfur dioxide. For pre-2007 engines, we assume that 98 percent of the sulfur is converted to
SO2; for 2007 and later engines, we assume that 70 percent of the sulfur is converted to SO2.
Because we are converting from S to SO2, we use a molecular weight ratio of 64/32. Table HB-
13 presents our estimates of SOx reductions from HDDEs corresponding with the use of low
sulfur fuel. Table II.B-13 also presents SOx benefits from other sources using highway diesel
fuel as discussed earlier in this chapter.
Table II.B-13. Reductions in SOx from Low Sulfur Fuel
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
HDDE
AQAC
90
96
105
113
127
SOx Reduction
Updated
70
96
105
113
127
Other
AQAC
11
12
13
13
14
SOx Reduction
Updated
8
11
12
13
14
n-128
-------
Chapter II: Health and Welfare
Hi. Air Toxic Reductions
Table II.B-14 shows the estimated air toxics reductions associated with the anticipated
reductions in hydrocarbons. The difference between the toxics reductions from the Air Quality
Analysis Case and the Updated Control Case are due to the revised assumption about NMHC
reductions due to PM traps.
Table II.B-14. Reductions in Air Toxics from HDDEs
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Benzene
AQAC Updated
0.14 0.02
0.87 0.22
2.01 0.61
2.81 0.92
3.80 1.30
Formaldehyde
AQAC Updated
1.02 0.18
6.47 1.60
15.0 4.52
20.1 6.83
28.3 9.70
Acetaldehyde
AQAC Updated
0.38 0.07
2.38 0.59
5.52 1.66
7.71 2.52
10.4 3.57
1, 3-Butadiene
AQAC Updated
0.08 0.01
0.50 0.12
1.17 0.35
1.63 0.53
2.21 0.76
3. HDGV Emission Reductions
This section presents reductions in NOx, exhaust and evaporative NMHC, and air toxics
from HDGVs that we anticipate from this rule. Although, medium-duty passenger vehicles
(MDPV) are technically part of the HDGV class, they are not included in the standards finalized
today. Therefore, emissions from MDPVs are not included in the inventories presented here.
MDPVs were recently regulated under the Tier 2 light-duty vehicle rule.
Also, we do not claim benefits for reductions in California for HDGVs due to
California's comparably stringent LEV2 standards for these vehicles. However, the charts
presented below will include national inventories. In the tables, we will only present emissions
reductions that we are claiming for this rule.
The Air Quality Analysis Case includes emissions from MDPVs in its baseline, and
includes 50-state emissions reductions from the HDGVs regulated under this rule. Therefore, we
also present an adjusted AQAC inventory which only includes HDGVs covered by this rule and
distinguishes between 49-state and 50-state emission reductions. The Updated Control Case not
only accounts for the difference between the proposed and final standards, but accounts for
MDPVs and California reductions. All of the charts in this section are based on the Updated
H-129
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Control Case; however, the tables present the AQAC (with and without MDPV emissions) and
Updated (excludes MDPV emissions) results.
a.
NOx Reductions
Figure II.B-5 presents the projected NOx inventory with and without the new standards.
We believe the NOx standards will result in more than a 60 percent reduction in NOx from new
heavy-duty gasoline vehicles. Tables n.B-15.a and II.B-15.b present these projections with the
estimated NOx reductions for selected years for the AQAC and Updated inventories respectively.
Table n.B-15a distinguishes between the inventory with and without medium duty passenger
vehicles (MDPV). Although these vehicles are classified as HDGVs, they were included in the
Tier 2 standards and therefore are not included in today's standards. Table II.B-15b presents the
Updated control case which considers a delay in the standards compared to the AQAC inventory.
MDPV emissions are excluded from this table.
• baseline
• controlled
2000
2005
2010
2015
2020
2025
2030
Figure II.B-5. Projected Nationwide Exhaust NOx Emissions from HDGVs
n-iso
-------
Chapter II: Health and Welfare
Table II.B-15.a. Estimated Nationwide NOx Emissions from HDGVs Based on the Air
Quality Analysis Case (thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Baseline
AQAC w/oMDPVs
381 310
316 257
236 192
200 163
175 143
Controlled
AQAC w/oMDPVs
377 306
292 233
187 144
133 96
81 49
Reduction
50-state 49-state
4 3
24 21
48 43
67 60
94 84
Table ILB-15.b. Estimated Nationwide NOx Emissions from HDGVs
Based on the Updated Control Case (thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
Baseline
310
257
192
163
143
Controlled
310
244
154
105
54
Reduction
50-state 49-state
0 0
13 12
38 34
58 52
88 79
b.
Exhaust NMHC Reductions
Figure II.B-6 presents the projected exhaust NMHC inventory with and without the new
standards. We believe the NMHC standard will result in about a 30 percent reduction in exhaust
NMHC from new heavy-duty gasoline vehicles. Tables II.B-16.a and n.B-16.b present these
projections with the estimated exhaust NMHC reductions for selected years for the AQAC and
Updated inventories.
H-131
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
120,000
100,000
T
- baseline
• controlled
0
2000 2005 2010 2015 2020 2025 2030
Figure II.B-6. Projected Nationwide Exhaust NMHC Emissions from HDGVs
Table ILB-16.a. Estimated Nationwide Exhaust NMHC Emissions from HDGVs
Based on the Air Quality Analysis Case (thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Baseline
AQAC w/oMDPVs
93 76
71 58
60 49
61 50
64 52
Controlled
AQAC w/oMDPVs
93 75
68 55
55 44
51 39
49 37
Reduction
50-state 49-state
0.4 0.4
2.4 2.1
5.4 4.8
10.5 9.4
15.6 14.0
n-132
-------
Chapter II: Health and Welfare
Table II.B-16.b. Estimated Nationwide Exhaust NMHC Emissions from HDGVs
Based on the Updated Control Case (thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
Baseline
76
58
49
50
52
Controlled
76
56
45
41
37
Reduction
50-state 49-state
0 0
1.3 1.1
4.4 3.9
9.0 8.1
15.0 13.5
c. Evaporative Emission Reductions
Evaporative HC emissions include diurnal, resting loss, refueling, and running loss
emissions. To estimate evaporative emissions reductions from HDGVs, we used MOBILESb to
calculate percent reductions. We generated average national emission factors giving
consideration to northern and southern regions of the country, fuel programs,
inspection/maintenance programs, and time of year. This analysis uses the same methodology as
was used in the inventory analysis for the Tier 2 light-duty vehicle standards.162
Figure II.B-7 presents the projected nonexhaust HC inventory with and without the new
standards. We believe the new evaporative emissions standards would result in about a 12
percent reduction in nonexhaust HC from new heavy-duty gasoline vehicles. Tables U.B-17.a
and n.B-17.b present these projections with the estimated evaporative emission reductions for
selected years for the AQAC and Updated inventories.
H-133
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
160,000
«J 140,000
1 120,000
O
•c 100,000
" 80,000
O
X
0) 60,000
_>
5 40,000
o
Q.
> 20,000
0
-baseline
• controlled
2000 2005
2010 2015
2020
2025
2030
Figure II.B-7. Projected Nationwide Evaporative Emissions from HDGVs
Table II.B-17.a. Estimated Nationwide Evaporative Emissions from HDGVs
Based on the Air Quality Analysis Case (thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Baseline
AQAC w/oMDPVs
134 109
132 107
137 112
146 119
178 145
Controlled
AQAC w/oMDPVs
133 108
126 102
127 102
133 106
160 127
Reduction
50-state 49-state
1 1
6 5
10 9
13 12
17 16
n-134
-------
Chapter II: Health and Welfare
Table ILB-17.b. Estimated Nationwide Evaporative Emissions from HDGVs
Based on the Updated Control Case (thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
Baseline
109
107
112
119
145
Controlled
109
104
103
107
128
Reduction
50-state 49-state
0.0 0.0
3.4 3.0
8.9 8.0
12.2 10.9
17.1 15.3
d.
Air Toxics Reductions
The air toxics 49-state reductions for HDGVs are presented in Table HB-18 for the Air
Quality Analysis Case and the Updated Control Case.
Table II.B-18. Estimated 49-State Reductions in Air Toxics from HDGVs
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
Benzene
AQAC Updated
0.03 0.00
0.24 0.14
0.41 0.36
0.46 0.42
0.68 0.66
Formaldehyde
AQAC Updated
0.02 0.00
0.12 0.06
0.25 0.20
0.29 0.25
0.50 0.48
Acetaldehyde
AQAC Updated
0.00 0.00
0.03 0.02
0.07 0.06
0.09 0.08
0.16 0.15
1, 3-Butadiene
AQAC Updated
0.00 0.00
0.02 0.01
0.04 0.03
0.04 0.03
0.07 0.07
4. Total Emission Reductions
Figures HB-8 through II.B-10 present the total projected emissions of NOx, PM, and
NMHC from heavy-duty engines with and without the new exhaust, evaporative, crankcase, and
fuel sulfur standards. No reductions are assumed for HDGV PM. Tables II.B-19 through II.B-21
H-135
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
present the total NOx, PM, and NMHC benefits from heavy-duty engines that we anticipate from
this rule. Evaporative emission reductions are included in the NMHC benefits. Table HB-22
presents the total air toxics reductions. All of these projections are based on the Updated Control
Case. Reductions in California are not included in the tables.
re
o
.E
x
O
5,000,000
4,500,000
4,000,000
3,500,000
3,000,000
2,500,000
2,000,000
1,500,000
1,000,000
500,000
2000
-baseline
-controlled
2005
2010
2015
2020
2025
2030
Figure II.B-8. Projected NOx Inventory for Heavy-Duty Highway Vehicles
n-ise
-------
Chapter II: Health and Welfare
5
c
o
+J
r
o
.E
XL
O
X
500,000
450,000
400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000
0
• baseline
• controlled
2000
2005
2010
2015
2020
2025
2030
Figure II.B-10. Projected NMHC Inventory for Heavy-Duty Highway Vehicles
,000
• baseline
• controlled
2000
2005
2010
2015
2020
2025
2030
Figure II.B-9. Projected PM Inventory for Heavy-Duty Highway Vehicles
n-is?
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table II.B-19. Total NOx Emissions and Benefits for This Rule
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
HDV Baseline
2,970
2,710
2,520
2,520
2,930
HDV Controlled
2,910
2,290
1,250
692
346
Reduction
58
420
1,260
1,820
2,570
Table II.B-20. Total PM Emissions and Reductions for This Rule
(thousand short tons per year)
Calendar
Year
2007
2010
2015
2020
2030
HDV
Baseline
125
115
114
124
150
HDV
Controlled
114
79
53
42
41
HDV
Reduction
11
35
61
82
109
Other DSPM
Reduction*
0.6
0.6
0.6
0.7
0.7
Total
Reduction
12
36
61
82
109
* From sources other than HDDEs using on-highway low sulfur fuel.
H-138
-------
Chapter II: Health and Welfare
Table II.B-21. Total NMHC Emissions and Reductions for This Rule
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
HDVBaseline
376
358
361
386
451
HDV Controlled
374
337
305
301
332
Reduction
2
21
54
83
115
Table II.B-22. Total Reductions in Air Toxics for This Rule
(thousand short tons per year)
Calendar Year
2007
2010
2015
2020
2030
Benzene
0.02
0.36
0.96
1.34
1.96
Formaldehyde
0.18
1.67
4.72
7.08
10.2
Acetaldehyde
0.07
0.61
1.72
2.60
3.73
1,3-Butadiene
0.01
0.14
0.38
0.57
0.82
5. Differences from NPRM Inventory
For the NPRM we used a simplified analysis to calculate emissions inventories and
reductions from heavy-duty vehicles. For FtDDEs we took a top-down approach to modeling
emission inventories using a spreadsheet model. For FtDGVs we used emission factors
generated using a modified MOBILES model with the inputs of an average speed, average fuel,
and summertime average temperature. Neither of these approaches were sophisticated enough to
include county-by-county or hour-by-hour effects on the emission inventories. With that being
said, these inventories have proven to be similar to the FRM inventories.
The FRM inventories (as discussed above) are based on complex and time consuming
calculations in which emissions were summed in every county in the U.S. on an hourly basis.
These inventories were developed for 1996, 2007, 2020, and 2030. Mathematically, this bottom-
H-139
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
up approach gives us more precise results than using national averages. In addition, it allows us
to account for more specific effects of ambient conditions, roadway types, fuel parameters, and
scrappage rates on HDV emissions.
In this section, we compare the baseline emissions inventories from the NPRM to those
presented in the FRM. By looking at baseline inventories we can focus on the calculation
methodology separate from the differences caused by the changes in the FRM standards from the
proposed standards. The effects of these changes are described in detail earlier in this chapter.
We believe that the small changes between the NPRM and FRM inventories reflect better
analysis in the FRM inventory.
a. Heavy-Duty Diesel Engines
i. MOBILE Model (NOx, NMHC)
As discussed above, MOBILES was used with adjustment factors to calculate HDDE
NOx and NMHC in the final rule. The primary difference between the NPRM and FRM
analyses is that we only considered operation at a single average speed in the NPRM spreadsheet
analysis; therefore, no speed correction was made. However, for the FRM analysis, emissions
were calculated for twelve different roadway types. These roadway type distributions differed for
each county. Based on the national average distribution of roadway types reported in the FRM
analysis, we can roughly calculate the effect of speed on NOx and NMHC inventories. Using the
MOBILE speed correction factors, we approximate that the weighted average national speed
correction is -5 percent for NOx and -17 percent for NMHC.
Table II.B-23 compares the NPRM and FRM inventories for exhaust NOx and NMHC.
We look at exhaust emissions only, because crankcase emissions are still calculated the same
way in the FRM analysis as they were in the NPRM analysis. As shown in this table, most of the
change from the NPRM to the FRM inventory is due to the application of speed correction
factors. We also would expect variation between the two inventories due to the top-down versus
bottom-up methodology as discussed earlier.
n-140
-------
Chapter II: Health and Welfare
Table II.B-23. Comparison of NPRM and FRM HDDE Baseline Inventories
for NOx and NMHC (thousand short tons per year)
Calendar
Year
2007
2020
2030
NPRM
2,860
2,600
3,000
Exhaust NOx
FRM
2,650
2,350
2,770
% Change
-7%
-10%
-8%
NPRM
218
249
292
Exhaust NMHC
FRM
184
206
240
% Change
-16%
-17%
-18%
77. PART Model (PM, SOx)
As discussed above, PARTS was used with adjustment factors to calculate HDDE PM
and SOx in the final rule. Table II.B-24 shows very good correlation in exhaust PM and SOx
between the NPRM and FRM inventories. In this case, the FRM results are less sensitive to
roadway distribution because PART does not apply speed correction factors to PM and SOx. We
look at exhaust PM because the NPRM inventory did not include brake and tire wear. In
addition, we use the same analysis methodology to calculate crankcase emissions in the NPRM
and FRM.
Table II.B-24. Comparison of NPRM and FRM HDDE Baseline Inventories
for Exhaust PM and SOx (thousand short tons per year)
Calendar
Year
2007
2020
2030
NPRM
92
88
106
Exhaust PM
FRM
96
86
104
% Change
4%
-2%
-2%
NPRM
91
112
126
Exhaust SOx
FRM
92
116
130
% Change
2%
4%
4%
b. Heavy-Duty Gasoline Vehicles
As with the HDDE final analysis, the FRM bases the HDGV NOx and NMHC emissions
inventories on the MOBILES model with adjustment factors. In this case, the NPRM was also
based on MOBILES model runs with adjustments to the model year emission factors entered into
the model. However, the NPRM analysis was run for a typical summer day and for a single
speed of 20 mph. In addition, the NPRM did not consider the effects of inspection/maintenance
H-141
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
or reformulated fuel programs. As a result, we saw similar results as with the Tier 2 Inventory
Analysis in which NMHC decreased noticeably with the county-by-county, hour-by-hour analysis
compared to the more simple top down analysis.
Table II.B-25 compares the FRM and NPRM baseline inventories for NOx and NMHC.
To make a direct comparison, we adjusted the NPRM inventory to be a national inventory rather
than just a 49-state inventory. We believe that the differences in the inventories reflect more
precise calculations in the county level analysis which results in a better inventory for the FRM.
Table II.B-25. Comparison of NPRM and FRM HDGV Baseline Inventories
for NOx and NMHC (thousand short tons per year)
Calendar
Year
2007
2020
2030
NPRM
307
159
138
Exhaust NOx
FRM
310
163
143
% Change
1%
2%
3%
Exhaust and Evaporative NMHC
NPRM FRM % Change
216 185 -15%
196 169 -14%
230 197 -14%
6. Sensitivity Analysis for In-Use PM Deterioration
In our analysis of the HDDE emissions inventory, we may underestimate emissions,
especially PM, due to engine deterioration in-use. We believe that current modeling represents
properly maintained engines, but may not be representative of in-use tampering or
malmaintenance. However, data related to this issue is extremely limited and inconclusive and
we are in the process of collecting more data on in-use emission deterioration. Once this has
been completed we will be better able to decide whether or not we need to update our
deterioration rates. If we do update our deterioration rates, we will do so through a similar public
process as we are using to create the MOBILE6 model.
Although a substantial amount of work remains before we can update our deterioration
factors, we believe it is valuable to get a feel for the potential effects of in-use tampering and
malmaintenance on the PM emissions inventory. In this section, we present a sensitivity analysis
of these effects.
a. Methodology
Engine, Fuel, and Emissions Engineering, Inc. recently performed a study which suggests
that tampering and malmaintenance result in large increases in in-use PM emissions from heavy-
H-142
-------
Chapter II: Health and Welfare
duty diesel engines.163 The California Air Resources Board (ARE) uses the underlying data in
this report in developing its in-use deterioration rates for the EMFAC2000 emission model.164
The ARE HDDE deterioration rates are presented in Table II.B-26 and are compared to the
deterioration factors used in our inventory analysis. No deterioration is assumed for urban buses.
To perform our sensitivity analysis, we use the ARB deterioration rates and the NPRM
spreadsheet model to determine the increases in the HDDE exhaust PM inventory due to
tampering and malmaintenance. We then applied these increases to the exhaust PM inventory
presented above. For 2007 and later model year engines, we assumed that the ratio of the
deterioration rate to the emission standard is the same as for 2004 model year engines.
Table II.B-26. Comparison of EMFAC2000 and EPA PM Deterioration Rates
for HDDEs (grams per mile per 10,000 miles)
Model Year
Pre 1976
1977-79
1980-83
1984-86
1987
1988-99
1990
1991-93
1994-97
1998
1999-02
2003
2004
LHDDE
0.003
0.003
0.004
0.004
0.005
0.005
0.005
0.002
0.003
0.003
0.001
0.001
0.001
EMFAC2000
MHDDE
0.016
0.016
0.016
0.021
0.017
0.017
0.017
0.022
0.018
0.012
0.012
0.009
0.009
HHDDE
0.016
0.017
0.018
0.012
0.008
0.008
0.008
0.009
0.010
0.007
0.003
0.003
0.003
LHDDE
0.000
0.000
0.000
0.000
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
EPA Analysis
MHDDE
0.000
0.000
0.000
0.000
0.000
0.004
0.000
0.002
0.002
0.000
0.000
0.000
0.000
HHDDE
0.000
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
b.
Results
If we consider the EMFAC2000 deterioration rates presented in Table II.B-26, we see an
increase of over 50 percent in the HDDE exhaust PM emission inventory compared to the results
from the Updated Control Case. In 2030, we see an increase in the baseline PM inventory of 48
percent and an increase in the controlled PM inventory of 63 percent. This translates to an
exhaust PM reduction of 141,000 tons in 2030 due to the new standards compared to the 96,000
ton PM reduction when tampering and malmaintenance were not considered. Figure II.B-11
presents the exhaust PM inventory with ("High") and without ("FRM") considering tampering
and malmaintenance.
H-143
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
200,000
180,000
160,000
140,000
120,000
100,000
80,000
60,000
40,000
20,000
0
• High base
• High control
. FRM base
• FRM control
2000
2005
2010
2015
2020
2025
2030
Figure II.B-11. Projected HDDE Exhaust PM Inventory with and without
Consideration of Tampering and Malmaintenance
7. Contribution of HDVs to National Inventory
Nationwide, heavy-duty vehicles are projected to contribute about 15 percent of the total
NOx inventory, and 28 percent of the mobile source inventory in 2007. Heavy-duty NOx
emissions also contribute to fine particulate concentrations in ambient air due to the
transformation in the atmosphere to nitrates. The NOx reductions resulting from today's
standards will therefore have a considerable impact on the national NOx inventory. All highway
vehicles account for 34 percent and heavy-duty highway vehicles account for 20 percent of the
mobile source portion of national PM10 emissions in 2007. These inventories are based on the
analysis performed by Pechan used for the air quality modeling analysis.165166 Because this
inventory analysis does not include stationary source emissions from Alaska and Hawaii, Tables
II.B-27 through II.B-29 present emissions inventories for the 48 contiguous states.
H-144
-------
Chapter II: Health and Welfare
Table II.B-27. 2007 Baseline Emissions Inventories for 48 Contiguous States
(thousand short tons)
Source
Heavy-Duty Vehicles
Light-Duty Vehicles
Nonroad
Other
Total
VOC
415 (3%)
2,600 (18%)
2,120(15%)
9,140 (64%)
14,300
NOx
3,030 (15%)
2,950 (14%)
4,710(23%)
9,890 (48%)
20,600
SO 2
94(1%)
25 (0%)
1,040 (6%)
15,900 (93%)
17,100
PM*
126 (4%)
82 (3%)
408 (14%)
2,210 (78%)
2,830
CO
3,850 (4%)
39,300 (42%)
27,200 (29%)
22,600 (24%)
92,900
* excludes natural and miscellaneous sources
Table II.B-28. 2020 Baseline Emissions Inventories for 48 Contiguous States
(thousand short tons)
Source
Heavy-Duty Vehicles
Light-Duty Vehicles
Nonroad
Other
Total
VOC
420 (3%)
1,800 (13%)
2,000(14%)
10,100(70%)
14,300
2,
1
4,
NOx
550 (14%)
,310(7%)
040 (23%)
9,980 (56%)
17,900
SO 2
118(1%)
31(0%)
1,310(8%)
14,500 (91%)
16,000
PM*
126 (4%)
100 (3%)
450(15%)
2,380 (78%)
3,060
4
CO
,720 (4%)
44,600 (42%)
33
23
,900 (32%)
,700 (22%)
107,000
: excludes natural and miscellaneous sources
Table II.B-29. 2030 Baseline Emissions Inventories for 48 Contiguous States
(thousand short tons)
Source
Heavy-Duty Vehicles
Light-Duty Vehicles
Nonroad
Other
Total
VOC
491 (3%)
1,950 (12%)
2,230(14%)
11,000(70%)
15,700
NOx
2,940 (16%)
1,250 (7%)
4,320 (23%)
10,200 (55%)
18,700
S02
133 (1%)
35 (0%)
1,490 (9%)
14,800 (90%)
16,400
PM*
152 (5%)
114(3%)
511(15%)
2,560 (77%)
3,340
CO
5,730 (5%)
51,200(42%)
39,200 (33%)
24,400 (20%)
120,000
: excludes natural and miscellaneous sources
E-145
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
References on Ozone Health Effects
These documents are all publicly available, and are also available in EPA Air Docket A-95-58.
Devlin, R. B.; McDonnell, W. F.; Koren, H. S.; Becker, S. (1990) Prolonged exposure of humans
to 0.10 and 0.08 ppm ozone results in inflammation in the lung. Presented at: 83rd annual
meeting of the Air & Waste Management Association; June; Pittsburgh, PA. Pittsburgh,
PA: Air & Waste Management Association; paper no. 90-150.2.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.;
Koren, H. S. (1991) Exposure of humans to ambient levels of ozone for 6.6 hours causes
cellullar and biochemical changes in the lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Fernandes, A. L. G.; Molfino, N. A.; McClean, P. A.; Silverman, F.; Tarlo, S.; Raizenne, M.;
Slutsky, A. S.; Zamel, N. (1994) The effect of pre-exposure to 0.12 ppm of ozone on
exercise-induced asthma. Chest: 106: 1077-1082.
Folinsbee, L. J.; McDonnell, W. F.; Horstman, D. H. (1988) Pulmonary function and symptom
responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA 38:
28-35.
Gong, H., Jr.; Bradley, P. W.; Simmons, M. S.; Tashkin, D. P. (1986) Impaired exercise
performance and pulmonary function in elite cyclists during low-level ozone exposure in
a hot environment. Am. Rev. Respir. Dis. 134: 726-733.
Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone
concentration profiles on pulmonary function. Am. Rev. Respir. Dis. 146: 1487-1493.
Horstman, D. H.; Folinsbee, L. J.; Ives, P. J.; Abdul-Salaam, S.; McDonnell, W. F. (1990) Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-
1163.
Horvath, S. M.; Bedi, J. F.; Drechsler-Parks, D. M. (1986) Effects of peroxyacetyl nitrate alone
and in combination with ozone in healthy young women. J. Air Pollut. Control Assoc. 36:
265-270.
Koenig, J. Q.; Covert, D. S.; Morgan, M. S.; Horike, M.; Horike, N.; Marshall, S. G.; Pierson,
W. E. (1985) Acute effects of 0.12 ppm ozone or 0.12 ppm nitrogen dioxide on
pulmonary function in healthy and asthmatic adolescents. Am. Rev. Respir. Dis. 132:
648-651.
H-146
-------
Chapter II: Health and Welfare
Koenig, J. Q.; Covert, D. S.; Marshall, S. G.; Van Belle, G.; Pierson, W. E. (1987) The effects of
ozone and nitrogen dioxide on pulmonary function in healthy and in asthmatic
adolescents. Am. Rev. Respir. Dis. 136: 1152-1157.
Koenig, J. Q.; Covert, D. S.; Smith, M. S.; Van Belle, G.; Pierson, W. E. (1988) The pulmonary
effects of ozone and nitrogen dioxide alone and combined in healthy and asthmatic
adolescent subjects. Toxicol. Ind. Health 4: 521-532.
Koenig, J. Q.; Covert, D. S.; Hanley, Q. S.; Van Belle, G.; Pierson, W. E. (1990) Prior exposure
to ozone potentiates subsequent response to sulfur dioxide in adolescent asthmatic
subjects. Am. Rev. Respir. Dis. 141: 377-380.
Konig, G.; Rommelt, H.; Kienle, H.; Dirnagl, K.; Polke, H.; Fruhmann, G. (1980) Anderung der
bronchomotorischen Reagibilitat des Menschen durch Einwirkung von Ozon [Changes in
the bronchial reactivity of humans caused by the influence of ozone]. Arbeitsmed.
Sozialmed. Praeventivmed. 151: 261-263.
Koren, H. S.; Devlin, R. B.; Becker, S.; Perez, R.; McDonnell, W. F. (1991) Time-dependent
changes of markers associated with inflammation in the lungs of humans exposed to
ambient levels of ozone. Toxicol. Pathol. 19: 406-411.
Kulle, T. J.; Sauder, L. R.; Hebel, J. R.; Chatham, M. D. (1985) Ozone response relationships in
healthy nonsmokers. Am. Rev. Respir. Dis. 132: 36-41.
Larsen, R. I; McDonnell, W. F.; Horstman, D. H.; Folinsbee, L. J. (1991) An air quality data
analysis system for interrelating effects, standards, and needed source reductions: part 11.
a lognormal model relating human lung function decrease to O3 exposure. J. Air Waste
Manage. Assoc. 41: 455-459.
Linn, W. S.; Fischer, D. A.; Medway, D. A.; Anzar, U. T.; Spier, C. E.; Valencia, L. M.; Venet,
T. G.; Hackney, J. D. (1982a) Short-term respiratory effects of 0.12 ppm ozone exposure
in volunteers with chronic obstructive pulmonary disease. Am. Rev. Respir. Dis. 125:
658-663.
Linn, W. S.; Avol, E. L.; Shamoo, D. A.; Spier, C. E.; Valencia, L. M.; Venet, T. G.; Fischer, D.
A.; Hackney, J. D. (1986) A dose-response study of healthy, heavily exercising men
exposed to ozone at concentrations near the ambient air quality standard. Toxicol. Ind.
Health 2: 99-112.
Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D. (1994)
Effects of prolonged, repeated exposure to ozone, sulfuric acid, and their combination in
healthy and asthmatic volunteers. Am. J. Respir. Crit. Care Med. 150: 431-440.
H-147
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
McBride, D. E.; Koenig, J. Q.; Luchtel, D. L.; Williams, P. V.; Henderson, W. R., Jr. (1994)
Inflammatory effects of ozone in the upper airways of subjects with asthma. Am. J.
Respir. Crit. CareMed. 149: 1192-1197.
McDonnell, W. F.; Horstman, D. H.; Hazucha, M. J.; Seal, E., Jr.; Haak, E. D.; Salaam, S. A.;
House, D. E. (1983) Pulmonary effects of ozone exposure during exercise: dose-response
characteristics. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 54: 1345-1352.
McDonnell, W. F., IE; Horstman, D. H.; Abdul-Salaam, S.; House, D. E. (1985b)
Reproducibility of individual responses to ozone exposure. Am. Rev. Respir. Dis. 131:
36-40.
McDonnell, W. F.; Kehrl, H. R.; Abdul-Salaam, S.; Ives, P. J.; Folinsbee, L. J.; Devlin, R. B.;
O'Neil, J. J.; Horstman, D. H. (1991) Respiratory response of humans exposed to low
levels of ozone for 6.6 hours. Arch. Environ. Health 46: 145-150.
McDonnell, W. F.; Muller, K. E.; Bromberg, P. A.; Shy, C. M. (1993) Predictors of individual
differences in acute response to ozone exposure. Am. Rev. Respir. Dis. 147: 818-825.
Molfino, N. A.; Wright, S. C.; Katz, I; Tarlo, S.; Silverman, F.; McClean, P. A.; Szalai, J. P.;
Raizenne, M.; Slutsky, A. S.; Zamel, N. (1991) Effect of low concentrations of ozone on
inhaled allergen responses in asthmatic subjects. Lancet 338(8761): 199-203.
Schelegle, E. S.; Adams, W. C. (1986) Reduced exercise time in competitive simulations
consequent to low level ozone exposure. Med. Sci. Sports Exercise 18: 408-414.
Seal, E., Jr.; McDonnell, W. F.; House, D. E.; Salaam, S. A.; Dewitt, P. J.; Butler, S. O.; Green,
J.; Raggio, L. (1993) The pulmonary response of white and black adults to six
concentrations of ozone. Am. Rev. Respir. Dis. 147: 804-810.
Weymer, A. R.; Gong, H., Jr.; Lyness, A.; Linn, W. S. (1994) Pre-exposure to ozone does not
enhance or produce exercise-induced asthma. Am. J. Respir. Crit. Care Med. 149: 1413-
1419.
H-148
-------
Chapter II: Health and Welfare
References on PM Health Effects
These documents are all publicly available, and are also available in EPA Air Docket A-95-54.
Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.;
Raad-Young, M.; Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and
sulfates on the frequency of respiratory admissions to Ontario hospitals. Environ. Res. 65:
172-194.
Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality:
associations with particulates and acid aerosols. Environ. Res. 59: 362-373.
Dockery, D. W.; Pope, C. A., HI; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G.,
Jr.; Speizer, F. E. (1993) An association between air pollution and mortality in six U.S.
cities. N. Engl. J. Med. 329: 1753-1759.
Ito., K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk
sub-populations. J. Exposure Anal. Environ. Epidemiol.: in press.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10
associations in Los Angeles. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the
colloquium on particulate air pollution and human mortality and morbidity; January 1994;
Irvine, CA. Inhalation Toxicol. 7: 59-69.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The
association of ambient air pollution with twice daily peak expiratory flow rate
measurements in children. Am. J. Epidemiol. 141: 111-122.
Ostro, B. D.; Lipsett, M. J.; Wiener, M. B.; Seiner, J. C. (1991) Asthmatic responses to airborne
acid aerosols. Am. J. Public Health 81: 694-702.
Ostro, B. D.; Lipsett, M. J.; Mann, J. K.; Krupnick, A.; Harrington, W. (1993) Air pollution and
respiratory morbidity among adults in Southern California. Am. J. Epidemiol. 137:
691-700.
Pope, C. A., IE; Dockery, D. W. (1992) Acute health effects of PM10 pollution on symptomatic
and asymptomatic children. Am. Rev. Respir. Dis. 145: 1123-1128.
Pope, C. A., IE; Kalkstein, L. S. (1996) Synoptic weather modeling and estimates of the
exposure-response relationship between daily mortality and paniculate air pollution.
Environ. Health Perspect. 104: in press.
H-149
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Pope, C. A., IE; Kanner, R. E. (1993) Acute effects of PM10 pollution on pulmonary function of
smokers with mild to moderate chronic obstructive pulmonary disease. Am. Rev. Respir.
Dis. 147: 1336-1340.
Pope, C. A., IE; Dockery, D. W.; Spengler, J. D.; Raizenne, M. E. (1991) Respiratory health and
PM10 pollution: a daily time series analysis. Am. Rev. Respir. Dis. 144: 668-674.
Samet, J. M.; Zeger, S. L.; Berhane, K. (1995) The association of mortality and particulate air
pollution. In: Particulate air pollution and daily mortality: replication and validation of
selected studies, the phase I report of the particle epidemiology evaluation project
[preprint]. Cambridge, MA: Health Effects Institute; pp. 1-104.
Samet, J. M.; Zeger, S. L.; Kelsall, J. E.; Xu, J. (1996a) Air pollution and mortality in
Philadelphia, 1974-1988, report to the Health Effects Institute on phase IB: Particle
Epidemiology Evaluation Project. Cambridge, MA: Health Effects Institute; accepted.
Samet, J. M.; Zeger, S. L.; Kelsall, J. E.; Xu, J.; Kalkstein, L. S. (1996b) Weather, air pollution
and mortality in Philadelphia, 1973-1980, report to the Health Effects Institute on phase
IB, Particle Epidemiology Evaluation Project. Cambridge, MA: Health Effects Institute;
review draft.
Schwartz, J. (1993) Air pollution and daily mortality in Birmingham, Alabama. Am. J.
Epidemiol. 137: 1136-1147.
Schwartz, J. (1994a) Total suspended particulate matter and daily mortality in Cincinnati, Ohio.
Environ. Health Perspect. 102: 186-189.
Schwartz, J. (1994b) Air pollution and daily mortality: a review and meta analysis. Environ. Res.
64: 36-52.
Schwartz, J. (1994c) What are people dying of on high air pollution days? Environ. Res. 64:
26-35.
Schwartz, J. (1994d) Air pollution and hospital admissions for the elderly in Detroit, Michigan.
Am. J. Respir. Crit. CareMed. 150: 648-655.
Schwartz, J. (1994e) Air pollution and hospital admissions for the elderly in Birmingham,
Alabama. Am. J. Epidemiol. 139: 589-598.
Schwartz, J. (1994f) PM10, ozone, and hospital admissions for the elderly in Minneapolis, MN.
Arch. Environ. Health 49: 366-374.
n-150
-------
Chapter II: Health and Welfare
Schwartz, J. (1994g) Nonparametric smoothing in the analysis of air pollution and respiratory
illness. Can. J. Stat. 22: 1-17.
Schwartz, J. (1995) Short term fluctuations in air pollution and hospital admissions of the elderly
for respiratory disease. Thorax 50: 531-538.
Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology
7: 20-28.
Schwartz, J.; Dockery, D. W. (1992a) Increased mortality in Philadelphia associated with daily
air pollution concentrations. Am. Rev. Respir. Dis. 145: 600-604.
Schwartz, J.; Dockery, D. W. (1992b) Particulate air pollution and daily mortality in
Steubenville, Ohio. Am. J. Epidemiol. 135: 12-19.
Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease
in Detroit, Michigan. Am. J. Epidemiol. 142: 23-35.
Schwartz, J.; Slater, D.; Larson, T. V.; Pierson, W. E.; Koenig, J. Q. (1993) Particulate air
pollution and hospital emergency room visits for asthma in Seattle. Am. Rev. Respir. Dis.
147: 826-831.
Schwartz, J.; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.; Koutrakis,
P.; Speizer, F. E.; Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on
respiratory symptom reporting in children. Am. J. Respir. Crit. Care Med. 150:
1234-1242.
Schwartz, D. A.; Thorne, P. S.; Yagla, S. J.; Burmeister, L. F.; Olenchock, S. A.; Watt, J. L.;
Quinn, T. J. (1995) The role of endotoxin in grain dust-induced lung disease. Am. J.
Respir. Crit. Care Med. 152: 603-608.
Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996a) Is daily mortality associated specifically with
fine particles? J. Air Waste Manage. Assoc.: accepted.
Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; Le Tertre, A.;
Piekarksi, T.; Ponce de Leon, A.; Ponka, A.; Rossi, G.; Saez, M.; Shouten, J. P. (1996b)
Methodological issues in studies of air pollution and daily counts of deaths or hospital
admissions. J. Epidemiol. Commun. Health: in press.
Styer, P.; McMillan, N.; Gao, F.; Davis, J.; Sacks, J. (1995) The effect of airborne particulate
matter on daily death counts. Environ. Health Perspect. 103: 490-497.
H-151
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution
and respiratory hospital admissions in three New York State metropolitan areas: results
for 1988 and 1989 summers. J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital
admissions and summertime haze air pollution in Toronto, Ontario: consideration of the
role of acid aerosols. Environ. Res. 65: 271-290.
U. S. EPA Air Quality Criteria for Particulate Matter, Office of Research and Development,
EPA/600/P-95/001, April 1996.
H-152
-------
Chapter II: Health and Welfare
Chapter II. References
1. Clean Air Act, Section 202(a)(3)(B).
2. 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.
5. U.S. EPA, 1996, Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF.
4. U.S. EPA, 1995, Review of National Ambient Air Quality Standards for Nitrogen
Dioxide, Assessment of Scientific and Technical Information, OAQPS Staff Paper,
EPA-452/R-95-005.
5. U.S.EPA, 1993, Air Quality Criteria for Oxides of Nitrogen, EP A/600/8-9 l/049aF.
6. Vitousek, Pert M., John Aber, Robert W. Howarth, Gene E. Likens, et al. 1997. Human
Alteration of the Global Nitrogen Cycle: Causes and Consequences. Issues in Ecology.
Published by Ecological Society of America, Number 1, Spring 1997.
7. Technical Memorandum to EPA Air Docket A-99-06, April 20, 2000, Cincinnati
Redesignation in Attainment and Approval of 10-Year Maintenance Plan.
8. 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").
9. 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.
10. 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.
11. Air Quality Criteria Document for Ozone and Related Photochemical Oxidants, EPA
National Center for Environmental Assessment, July 1996, Report No. EPA/600/P-
93/004cF. The document is available on the internet at
H-153
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
http://www.epa.gov/ncea/ozone.htm.
12. U.S. EPA (1998) National Air Quality and Emissions Trends Report, 1997. Office of Air
Quality Planning and Standards. EPA 454/R-99-016.
13. U.S. EPA (1998) National Air Pollutant Emission Trends Update, 1970-1997. EPA-
454/E-98-007.
14. Gray, H.A.,Cass, G.R., Huntzicker, J.J., Heyerdahl, E.K. and Rau, J.A. (1986)
Characterization of Atmospheric Organic and Elemental Carbon Particle Concentrations
in Los Angeles. Environ. Sci. & Technol. 20:580-589.
15. Kleeman, M.J., Hughes, L.S., Allen, J.O., Cass, G.R. (1999) Source Contributions to the
Size and Composition Distribution of Atmospheric Particles: Southern California in
September 1996. Environ. Sci. Technol. 33:4331-4341.
16. U.S. EPA (1998) National Air Quality and Emissions Trends Report, 1997. Office of Air
Quality Planning and Standards. EPA 454/R-99-016.
17. U.S. EPA (2000) Health Assessment Document for Diesel Exhaust: SAB Review Draft.
EPA/600/8-90/057E Office of Research and Development, Washington, D.C. The
document is available electronically at www.epa.gov/ncea/dieslexh.htm.
18. EPA (2000) Review of EPA's Health Assessment Document for Diesel Exhaust (EPA
600/8-90/057E). Review by the Clean Air Scientific Advisory Committee (CASAC)
December 2000 EPA-SAB-CASAC-01-003.
19. Health Effects Institute (1995) Diesel Exhaust: A Critical Analysis of Emissions,
Exposure, and Health Effects pp. 253-292. April 1995.
20. National Institute for Occupational Safety and Health (NIOSH) (1988) Carcinogenic
effects of exposure to diesel exhaust. NIOSH Current Intelligence Bulletin 50. DHHS,
Publication No. 88-116. Centers for Disease Control, Atlanta, GA.
21. International Agency for Research on Cancer (1989) Diesel and gasoline engine exhausts
and some nitroarenes, Vol. 46. Monographs on the evaluation of carcinogenic risks to
humans. World Heath Organization, International Agency for Research on Cancer, Lyon,
France.
22. World Health Organization (1996) Diesel fuel and exhaust emissions: International
program on chemical safety. World Health Organization, Geneva, Switzerland.
23. Office of Environmental Health Hazard Assessment (1998) Health risk assessment for
diesel exhaust, April 1998. California Environmental Protection Agency, Sacramento,
CA.
n-154
-------
Chapter II: Health and Welfare
24. U.S. Department of Health and Human Services (2000) Ninth report on carcinogens.
National Toxicology Program, Research Triangle Park, NC.
ehis.niehs.nih.gov/roc/toc9.html.
25. Winer, A.M. and Busby, W.F.Jr. (1995) Atmospheric Transport and Transformation of
Diesel Emissions. In: Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and
Health Effects. Cambridge, MA. Health Effects Institute, pp.84-105.
26. U.S. EPA (1999) Health Assessment Document for Diesel Emissions: SAB Review
Draft. PEA/600/8-90/057D Office of Research and Development, Washington, D.C. The
document is available electronically at www.epa.gov/ncea/diesel.htm.
27. Kittelson, D.B., Arnold, M., Watts, W.F. (1999) Review of Diesel Particulate Matter
Sampling Methods. Final Report. University of Minnesota.
www.me.umn.edu/centers/cdr/Proj _EPA.html.
28. Zaebst, D.D., Clapp D.E., Blake L.M., Marlow D.A., Steenland K., Hornung R.W.,
Scheutzle D. and J. Butler (1991) Quantitative Determination of Trucking Industry
Workers Exposures to Diesel Exhaust Particles. Am. Ind. Hyg. Assoc. J.., 52:529-541.
29. Graboski, M. S., McCormick, R.L., Yanowitz, J., and L.B.A. Ryan (1998) Heavy-Duty
Diesel Testing for the Northern Front Range Air Quality Study. Colorado Institute for
Fuels and Engine Research, http://nfraqs.cira.colostate.edu/index2.htm.
30. Pierson, W. R.; Brachazek, W. W. (1983) Particulate Matter Associated with Vehicles on
the Road. Aerosol Sci. & Tech. 2:1-40.
31. Warner-Selph, M. A., Dietzmann, H.E. (1984) Characterization of Heavy-Duty Motor
Vehicle Emissions Under Transient Driving Conditions. Southwest Research Institute.
EPA-600/3-84-104.
32. U. S. EPA (2000) Health Assessment Document for Diesel Exhaust: SAB Review Draft.
EPA/600/8-90/057E Office of Research and Development, Washington, D.C. The
document is available electronically at www.epa.gov/ncea/dieslexh.htm.
33. Kleeman, M.J., Cass, G.R. (1999) Identifying the Effect of Individual Emissions Sources
on Particulate Air Quality Within a Photochemical Aerosol Processes Trajectory Model.
Atmos. Eviron. 33:4597-4613.
34. Kleeman, M.J., Hughes, L.S., Allen, J.O., Cass, G.R. (1999) Source Contributions to the
Size and Composition Distribution of Atmospheric Particles: Southern California in
September 1996. Environ. Sci. Technol. 33:4331-4341.
H-155
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
35. Cass, G.R. and H.A. Gray (1995) Regional Emissions and Atmospheric Concentrations of
Diesel Engine Particulate Matter: Los Angeles as a Case Study. In: Diesel Exhaust: A
Critical Analysis of Emissions, Exposure, and Health Effects. A Special Report of the
Institute's Diesel Working Group. Health Effects Institute, Cambridge, MA, pp. 125-137.
36. Schauer, J.J., Rogge, W.F., Hildemann, L.M., Mazureik, M.A., Cass, G.R., and B.R.T.
Simoneit (1996) Source Apportionment of Airborne Parti culate Matter Using Organic
Compounds as Tracers. Atmos. Environ. 30(22):3837-3855.
37. 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.
38. Maricopa Association of Governments (1999) The 1999 Brown Cloud Project for the
Maricopa Association of Governments Area, Final Report, December 1999.
39. 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.
40. 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. Air a& Waste
Management Association 87th Annual Meeting, Cincinnati, OH, June 19-24, 1994.
41. Fujita, E., Watson, J.G, Chow, J.C., Robinson, N.F., Richards, L.W., Kumar, N. (1998)
The Northern Front Rage Air Quality Study Final Report Volume C: Source
Apportionment and Simulation Methods and Evaluation.
http://nfiraqs.cira.colostate.edu/index2.html.
42. Cass, G.R. and H.A. Gray (1995) Regional Emissions and Atmospheric Concentrations of
Diesel Engine Particulate Matter: Los Angeles as a Case Study. In: Diesel Exhaust: A
Critical Analysis of Emissions, Exposure, and Health Effects. A Special Report of the
Institute's Diesel Working Group. Health Effects Institute, Cambridge, MA, pp. 125-137.
43. Kleeman, M.J., Cass, G.R. (1999) Identifying the Effect of Individual Emissions Sources
on Particulate Air Quality Within a Photochemical Aerosol Processes Trajectory Model.
Atmos. Eviron. 33:4597-4613.
44. Kleeman, M.J., Hughes, L.S., Allen, J.O., Cass, G.R. (1999) Source Contributions to the
Size and Composition Distribution of Atmospheric Particles: Southern California in
September 1996. Environ. Sci. Technol. 33:4331-4341.
H-156
-------
Chapter II: Health and Welfare
45. Salmon, L.G., Cass, G.R., Pedersen, D.U., Durant, J.L., Gibb, R., Lunts, A., and M. Utell
(1997) Determination of fine particle concentration and chemical composition in the
northeastern United States, 1995. Progress Report to Northeast States for Coordinated
Air Use Management (NESCAUM), September, 1999. www.nescaum.org.
46. 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 . ISSN: 0737-5352-32.
47. South Coast Air Quality Management District (2000) Multiple Air Toxics Exposure
Study in the South Coast Air Basin (MATES-II), Final Report and Appendices, March
2000.
48. Wittorff, D.N., Gertler, A.W., Chow, J.C., Barnard, W.R., Johngedyk, H.A. The Impact
of Diesel Paniculate Emission on Ambient Particulate Loadings. Air & Waste
Management Association 87th Annual Meeting, Cincinnati, OH, June 19-24, 1994.
49. California EPA (1998) Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant Appendix in Part A: Exposure Assessment. California Environmental
Protection Agency. California Air Resources Board April 22, 1998. Available at
http://www.arb.ca.gov/toxics/diesel/diesel.htm.
50. Brunekreef, B. (1999) Environmental diesel exhaust exposure and respiratory health of
children in the Netherlands. Presented at the Health Effects Institute Diesel Workshop:
Building a Research Strategy to Improve Risk Assessment. Stone Mountain, GA, March
7-9, 1999.
51. South Coast Air Quality Management District (2000) Multiple Air Toxics Exposure
Study in the South Coast Air Basin (MATES-II), Final Report and Appendices, March
2000.
52. California EPA Air Resources Board (1998) Measuring Concentration of Selected Air
Pollutants Inside California Vehicles. Final Report. Contract No. 95-339.
53. Watts, F.W., Jr. (1995) Assessment of Occupational Exposure to Diesel Emissions. In:
Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. Health
Effects Institute, Cambridge, MA, pp. 107-123.
54. Saverin, R., Braunlich, A., Dahman, D. et al. (1999) Diesel Exhaust and Lung Cancer
Mortality in Potash Mining. Am J. Ind. Med. 36:415-422.
55. Woskie, S.R., Smith, T.J., Hammond, S.K., Schenker, M.B., Garshick, E., Speizer, F.E.
(1988) Estimation of the Diesel Exhaust Exposures of Railroad Workers: I. Current
H-157
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Exposures. AmJ.Ind.Med. 13:381-394.
56. Birch, M.E. and Gary, R.A. (1996) Elemental Carbon-based Method for Monitoring
Occupational Exposures to Particulate Diesel Exhaust. Aerosol Sci. Technol. 25:221-241.
57. Watts, F.W., Jr. (1995) Assessment of Occupational Exposure to Diesel Emissions. In:
Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. Health
Effects Institute, Cambridge, MA, pp. 107-123.
58. Friones, J.R., Hinds, W.C., Duffy, R.M., Lafuente, E.J., Liu, W-C.V. (1987) Exposure of
Firefighters to Diesel Emissions in Fire Stations. Am. Ind. Hyg. Assoc. J. 48:202-207.
59. National Institute for Occupational Safety and Health (1990) Health Hazard Evaluation
Report: Yellow Freight Systems, Inc., Columbus, OH, Maybrook, NY. NIOSH Report
NO. HHE HETA 90-088-2110. U.S. DHHS, Public Health Service, Center for Disease
Control.
60. Zaebst, D.D., Clapp D.E., Blake L.M., Marlow D.A., Steenland K., Hornung R.W.,
Scheutzle D. and J. Butler (1991) Quantitative Determination of Trucking Industry
Workers Exposures to Diesel Exhaust Particles. Am. Ind. Hyg. Assoc. J., 52:529-541.
61. U.S. EPA. (1999) 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.
http://www.epa.gov/otaq/toxics.htm.
62. EPA (2000) National Air Toxics Assessment National-Scale Analysis.
http://www.epa.gov/ttn/uatw/nata
63. California EPA (1998) Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant Appendix in Part A: Exposure Assessment. California Environmental
Protection Agency. California Air Resources Board April 22, 1998. Available at
http://www.arb.ca.gov/toxics/diesel/diesel.htm.
64. Law, P.L., Lioy, P.J., Zelenka, M.P., Huber, A.H., and T.R. McCurdy (1997) Evaluation
of a Probabilistic Exposure model Applied to Carbon Monoxide (pNEM/CO) Using
Denver Personal Exposure Monitoring Data. J. Air & Waste Mange. Assoc. 47:491-500.
65. Johnson, T., M. McCoy and J. E. Capel. 1993. Enhancements to the Hazardous Air
Pollutant Exposure Model (HAPEM) as Applied to Mobile Source Pollutants. Prepared
for U. S. EPA, Office of Air Quality Planning and Standards and Office of Research and
Development by International Technology Corporation, September, 1993.
H-158
-------
Chapter II: Health and Welfare
66. Palma, T., M. Riley, and J. Capel. 1996. Development and Evaluation of Enhancements
to the Hazardous Air Pollutant Exposure Model (HAPEM-MS3). Prepared for U. S.
EPA, Office of Air Quality Planning and Standards and Office of Research and
Development by International Technology Corporation, September, 1996.
67. 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.
68. 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.
69. U.S. EPA (1993) Motor Vehicle Air Toxics Study Peer Review Comments. Available on
the web at: http://www.epa.gov/oms/regs/toxics.
70. U.S. EPA. (1999) 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.
http://www.epa.gov/otaq/toxics.htm.
71. U.S. EPA. (1999) 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.
http://www.epa.gov/otaq/toxics.htm.
72. Yanowitz, J., McCormick, RL., Graboski, MS (2000) Critical review: in-use emissions
from HD diesel vehicles. Environ Sci Technol. 34:729-740.
73. Glen, G. and Shadwick, D. (1998) 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.
74. U.S. EPA (1993) Motor Vehicle Air Toxics Study Peer Review Comments. Available on
the web at: http://www.epa.gov/oms/regs/toxics/toxprsum.txt.
75. DeLucchi, M. A. (1999) Analysis of Particulate Matter Emission Factors in the PARTS
Model. Submitted to EPA.
H-159
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
76. Northeastern States for Coordinated Air Use Management (NESCAUM) (1999) Peer
Review: U. S. EPA's Draft Estimation of Motor Vehicle Toxic Emissions and Exposure
in Selected Urban Areas. October, 1999.
77. Johnson, T. R. (1999) Comments on Exposure Assessment Methodology Described in
Three Documents. Submitted to EPA, August, 1999.
78. Brodowicz, P. through Lorang, P. to Ris, C., memorandum, March 1999. U.S.
Environmental Protection Agency, National Fuel and Vehicle Emissions Laboratory.
Determination of the Demographic Groups with the Highest Annual Average Modeled
Diesel particulate matter Exposure.
79. U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I and U. 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.
http://www.epa.gov/otaq/toxics.htm.
80. Brodowicz, P. through Lorang, P. to Ris, C., memorandum, March 1999. U.S.
Environmental Protection Agency, National Fuel and Vehicle Emissions Laboratory.
Determination of the Demographic Groups with the Highest Annual Average Modeled
Diesel particulate matter Exposure.
81. California EPA (1998) Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant Appendix in Part A: Exposure Assessment. California Environmental
Protection Agency. California Air Resources Board April 22, 1998. Available at
http://www.arb.ca.gov/toxics/diesel/diesel.htm.
82. U.S. EPA (1998) National Air Pollutant Emission Trends Update, 1970-1997. EPA-
454/E-98-007.
83. California Environmental Protection Agency, Office of Health Hazard Assessment (CAL-
EPA, OEHHA) (1998) Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant. Appendix in Part B Health Risk Assessment for Diesel Exhaust. April 22,
1998.
84. Steenland, K., Deddens, J., Stayner, L. (1998) Diesel Exhaust and Lung Cancer in the
Trucking Industry: Exposure-Response Analyses and Risk Assessment. Am. J Indus.
Medicine 34:220-228.
85. Harris, I.E. (1983) Diesel emissions and Lung Cancer. Risk Anal. 3:83-100.
86. U. S. EPA (2000) Health Assessment Document for Diesel Exhaust: SAB Review Draft.
EPA/600/8-90/057E Office of Research and Development, Washington, D.C. The
n-160
-------
Chapter II: Health and Welfare
document is available electronically at www.epa.gov/ncea/dieslexh.htm.
87. Bhatia, R., Lopipero, P., Smith, A. (1998) Diesel Exhaust Exposure and Lung Cancer.
Epidemiol. 9:84-91.
88. Lipsett, M., Campleman, S. (1999) Occupational Exposure to Diesel Exhaust and Lung
Cancer: AMeta-analysis. Am J Public Health 80:1009-1017.
89. U.S. EPA (2000) 1996 National Toxics Inventory, http://www.epa.gov/ttn/uatw/nata.
90. U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA
by Sierra Research, Inc. and Radian International Corporation/Eastern Research Group,
November 30, 1999. Report No. EPA420-R-99-029.
http://www.epa.gov/otaq/toxics.htm.
91. U.S. EPA (1998) Environmental Protection Agency, Carcinogenic Effects of Benzene:
An Update, National Center for Environmental Assessment, Washington, DC. 1998.
EPA/600/P-97/001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html.
92. U.S. 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.
93. Clement Associates, Inc. (1991) Motor vehicle air toxics health information, for U.S.
EPA Office of Mobile Sources, Ann Arbor, MI, September 1991.
94. International Agency for Research on Cancer (IARC) (1982) IARC monographs on the
evaluation of carcinogenic risk of chemicals to humans, Volume 29, Some industrial
chemicals and dyestuffs, International Agency for Research on Cancer, World Health
Organization, Lyon, France, p. 345-389.
95. Irons, R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry (1992) Synergistic action
of the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci.
89:3691-3695.
96. Lumley, M., H. Barker, and J.A. Murray (1990) Benzene in petrol, Lancet 336:1318-
1319.
97. U.S. EPA (1998) Environmental Protection Agency, Carcinogenic Effects of Benzene:
An Update, National Center for Environmental Assessment, Washington, DC. 1998.
EPA/600/P-97/001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html.
H-161
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
98. U.S. EPA (1993) 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.
99. Aksoy, M (1991) Hematotoxicity, leukemogenicity and carcinogenicity of chronic
exposure to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds. Molecular
Aspects of Monooxygenases and Bioactivation of Toxic Compounds. New York:
Plenum Press, pp. 415-434.
100. Goldstein, B.D. (1988) Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.
101. Aksoy, M., S. Erdem, and G. Dincol. (1974) Leukemia in shoe-workers exposed
chronically to benzene. Blood 44:837.
102. 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.
103. 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.
104. U.S. EPA (1999) 1990 Emissions Inventory of Forty Potential Section 112 (K)
Pollutants: Supporting Data for EPA's Section 112(k) Regulatory Strategy — Final
Report. Emission Factors and Inventory Group, Office of Air Quality Planning and
Standards, May, 1999.
105. U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA
by Sierra Research, Inc. and Radian International Corporation/Eastern Research Group,
November 30, 1999. Report No. EPA420-R-99-029.
http://www.epa.gov/otaq/toxics.htm.
106. U.S. 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.
107. U.S. EPA (1998) Draft Health Risk Assessment of 1,3-Butadiene, National Center for
Environmental Assessment, Office of Research and Development, U.S. EPA,
EPA/600/P-98/001A, February 1998.
H-162
-------
Chapter II: Health and Welfare
108. Scientific Advisory Board. 1998. An SAB Report: Review of the Health Risk Assessment
of 1,3-Butadiene. EPA-SAB-EHC-98, August, 1998.
109. EPA 1996. Proposed guidelines for carcinogen risk assessment. Federal Register
61(79):17960-18011.
110. Delzell, 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.
111. U.S. EPA (1999) Memo from Dr. AparnaKoppikar, ORD-NCEA to Laura McKlevey,
OAQPS and Pamela Brodowicz, QMS. Slope Factor for 1,3-Butadiene, April 26, 1999.
112. U.S. EPA. (2000) Personal communication from Dr. Aparna Koppikar, ORD-NCEA to
Pamela Brodowicz, EPA OTAQ.
113. U.S. 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.
http://www.epa.gov/ngispgm3/iris/subst/0139.htm.
114. U.S. 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.
http://www.epa.gov/ngispgm3/iris/subst/0139.htm.
115. U.S. EPA (1999) 1990 Emissions Inventory of Forty Potential Section 112 (K)
Pollutants: Supporting Data for EPA's Section 112(k) Regulatory Strategy - Final
Report. Emission Factors and Inventory Group, Office of Air Quality Planning and
Standards, May, 1999.
116. Ligocki, M.P., G.Z. Whitten, R.R. Schulhof, M.C. Causley, and G.M. Smylie (1991)
Atmospheric transformation of air toxics: benzene, 1,3-butadiene, and formaldehyde,
Systems Applications International, San Rafael, CA (SYSAPP-91/106).
117. U.S. 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.
118. Clement Associates, Inc. (1991) Motor vehicle air toxics health information, for U.S.
EPA Office of Mobile Sources, Ann Arbor, MI, September 1991.
119. U.S. EPA (1993) Motor Vehicle-Related Air Toxics Study, U.S. Environmental
Protection Agency, Office of Mobile Sources, Ann Arbor, MI, EPA Report No. EPA 420-
H-163
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
R-93-005, April 1993. http://www.epa.gov/otaq/toxics.htm.
120. Wilhelmsson, B. and M. Holmstrom. (1987) Positive formaldehyde PAST after
prolonged formaldehyde exposure by inhalation. TheLancet:l64.
121. 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.
122. 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.
123. Nordman, H., H. Keskinen, and M. Tuppurainen. (1985) Formaldehyde asthma - rare or
overlooked? J. Allergy Clin. Immunol. 75:91-99.
124. U.S. EPA (1999) 1990 Emissions Inventory of Forty Potential Section 112 (K)
Pollutants: Supporting Data for EPA's Section 112(k) Regulatory Strategy - Final
Report. Emission Factors and Inventory Group, Office of Air Quality Planning and
Standards, May, 1999.
125. U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA
by Sierra Research, Inc. and Radian International Corporation/Eastern Research Group,
November 30, 1999. Report No. EPA420-R-99-029.
http://www.epa.gov/otaq/toxics.htm.
126. Ligocki, M.P., G.Z. Whitten (1991) Atmospheric transformation of air toxics:
acetaldehyde and polycyclic organic matter, Systems Applications International, San
Rafael, CA, (SYSAPP-91/113).
127. U.S. 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.
128. California Air Resources Board (CARB) (1992) Preliminary Draft: Proposed
identification of acetaldehyde as a toxic air contaminant, Part B Health assessment,
California Air Resources Board, Stationary Source Division, August, 1992.
129. U.S. EPA (1997) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH, 1997. http://www.epa.gov/iris/subst/0290.htm.
130. U.S. EPA (1999) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH. http://www.epa.gov/iris/subst/0139.htm.
H-164
-------
Chapter II: Health and Welfare
131. U.S. EPA (1993) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH. http://www.epa.gov/ngispgm3/iris/subst/0364.htm.
132. U.S. EPA (2000) Exposure and Human Health Reassessment of 2,3,7,8-
Tetrachlorodibenzo-^-Dioxin (TCDD) and Related Compounds. Part HI: Integrated
Summary and Risk Characterization for 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD)
and Related Compounds. External Review Draft. EPA/600/P-00/001Ag.
133. U.S. EPA. (1999) 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.
http://www.epa.gov/otaq/toxics.htm.
134. Mathpro, Inc. (1999) Costs for Meeting 40 ppm Sulfur Content Standard for Gasoline in
PADDs 1-3, via MOBILE and CD TECH Desulfurization Processes. A Study performed
for the American Petroleum Institute, February 26, 1999.
135. 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.
136. 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.
137. U.S. EPA (1993) Motor Vehicle Air Toxics Study Peer Review Comments. Available on
the web at: http://www.epa.gov/oms/regs/toxics.
138. Systems Applications International (1994) Projected Emission Trends and Exposure
Issues for 1,3-Butadiene. Prepared for the American Automobile Manufacturers
Association, March, 1994.
139. "National Air Quality and Emissions Trends Report, 1996", EPA Document Number
454/R-97-013.
140. Letter from Governor Michael Leavitt of Utah, on behalf of the Western Governors'
Association, to EPA Administrator Carol Browner, dated June 29, 1998.
141. "Report of the Grand Canyon Visibility Transport Commission to the United States
Environmental Protection Agency", June 1996.
H-165
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
142. Source Contributions to Atmospheric Fine Carbon Particle Concentrations, Gray and
Cass, Atmospheric Environment, Vol. 32, No. 22, pp. 3805-3825 (1998).
143. Memorandum to Docket A-99-06 from Drew Kodjak on Current Nonattainment Areas
and Population, January 12, 2000.
144. "Procedures for Developing Base Year and Future Year Mass and Modeling Inventories
for the Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel
(HDD) Rulemaking," EPA420-R-00-020, October 2000, Prepared for EPA by E.H.
Pechan & Associates, Inc.
145. "Data Summaries of Base Year and Future Year Mass and Modeling Inventories for the
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel (HDD)
Rulemaking," EPA420-R-00-019, October 2000, Prepared for EPA by E.H. Pechan &
Associates, Inc.
146. Memorandum from Michael Samulski and John Koupal to Docket A-99-06, "Heavy-
Duty Vehicle Emission Factors and Adjustment Factors for the Final 2007 Heavy-Duty
Rule Inventory Analysis," May 26, 2000.
147. Memorandum from Michael Samulski to Docket A-99-06, "Revisions to Heavy-Duty
Vehicle Emission Factors and Adjustment Factors for the Final 2007 Heavy-Duty Rule
Inventory Analysis," November 2, 2000.
148. "Procedures for Developing Base Year and Future Year Mass and Modeling Inventories
for the Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel
(HDD) Rulemaking," EPA420-R-00-020, October 2000, Prepared for EPA by E.H.
Pechan & Associates, Inc.
149. Memorandum from Michael Samulski and John Koupal to Docket A-99-06, "Heavy-
Duty Vehicle Emission Factors and Adjustment Factors for the Final 2007 Heavy-Duty
Rule Inventory Analysis," May 26, 2000.
150. "Update Heavy-Duty Engine Emission Conversion Factors for MOBILE6: Analysis of
Fuel Economy, Non-Engine Fuel Economy Improvements, and Fuel Densities," U.S.
Environmental Protection Agency, EPA-420-P-98-014, May 1998.
151. Regulatory Impact Analysis for the Tier 2 final rulemaking, Air Docket A-97-10.
152. Draft MOBILE6 fuel economy estimates.
153. Petroleum Supply Annual, 1998, Volume #1, Energy Information Administration (EIA),
Department of Energy (DOE), DOE/EIA-0340(98)/1.
H-166
-------
Chapter II: Health and Welfare
154. Pagan, Jaime, "Investigation on Crankcase Emissions from a Heavy-Duty Diesel
Engine," U.S. Environmental Protection Agency, March, 1997.
155. "Analysis of the Impacts of Control Programs on Motor Vehicle Toxics Emissions and
Exposure in Urban Areas and Nationwide." Prepared for the Office of Mobile Sources by
Sierra Research, Inc. November 30, 1999.
156. "Procedures for Developing Base Year and Future Year Mass and Modeling Inventories
for the Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel
(HDD) Rulemaking," EPA420-R-00-020, October 2000, Prepared for EPA by E.H.
Pechan & Associates, Inc.
157. "Data Summaries of Base Year and Future Year Mass and Modeling Inventories for the
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel (HDD)
Rulemaking," EPA420-R-00-019, October 2000, Prepared for EPA by E.H. Pechan &
Associates, Inc.
158. Memorandum from Michael Samulski and John Koupal to Docket A-99-06, "Heavy-
Duty Vehicle Emission Factors and Adjustment Factors for the Final 2007 Heavy-Duty
Rule Inventory Analysis," May 26, 2000.
159. Memorandum from Michael Samulski to Docket A-99-06, "Revisions to Heavy-Duty
Vehicle Emission Factors and Adjustment Factors for the Final 2007 Heavy-Duty Rule
Inventory Analysis," November 2, 2000.
160. Fuel Oil and Kerosene Sales 1998, Energy Information Administration (EIA),
Department of Energy (DOE), DOE/EIA-0535(98).
161. "Update of Heavy-Duty Emission Levels (Model Years 1988-2004+) for Use in
MOBILE6," U.S. Environmental Agency, EPA420-R-99-010, April 1999.
162. Memorandum from Dave Brzezinski to the Tier 2 Docket (A-97-10), "A Modified
Version of MOBILES for Evaluation of Proposed Tier 2 Evaporative Emission
Standards," II-B-11, March 3, 1999.
163. "Modeling Deterioration in Heavy-Duty Diesel Particulate Emissions," Engine, Fuel,
and Emissions Engineering, Incorporated, prepared for the U.S. Environmental Protection
Agency, September 30, 1998.
164. "Public Meeting to Consider Approval of the Revisions to the State's On-Road Motor
Vehicle Emissions Inventory: Technical Support Document," California Air Resources
Board, Chapter 10, May 2000.
H-167
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
165. "Procedures for Developing Base Year and Future Year Mass and Modeling Inventories
for the Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel
(HDD) Rulemaking," EPA420-R-00-020, October 2000, Prepared for EPA by E.H.
Pechan & Associates, Inc.
166. "Data Summaries of Base Year and Future Year Mass and Modeling Inventories for the
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Diesel (HDD)
Rulemaking," EPA420-R-00-019, October 2000, Prepared for EPA by E.H. Pechan &
Associates, Inc.
H-168
-------
Chapter III: Emissions Standards Feasibility
Chapter III: Emissions Standards Feasibility
For the past 30 or more years, emission control development for gasoline vehicles and
engines has concentrated most aggressively on exhaust emission control devices. These devices
currently provide as much as or more than 95 percent of the emission control on a gasoline
vehicle. In contrast, the emission control development work for diesels has concentrated on
improvements to the engine itself to limit the emissions leaving the combustion chamber.3
However, during the past 15 years, more development effort has been put into diesel
exhaust emission control devices, particularly in the area of particulate matter (PM) control.
Those developments, and recent developments in diesel NOx exhaust emission control devices,
make the widespread commercial use of diesel exhaust emission controls feasible. Through use
of these devices, we believe emissions control similar to that attained by gasoline applications
will be possible with diesel applications. However, without low sulfur diesel fuel, these
technologies cannot be implemented on heavy-duty diesel applications. Low sulfur diesel fuel
will at the same time also allow these technologies to be implemented on light-duty diesel
applications.
Several exhaust emission control devices have been developed to control harmful diesel
PM constituents-the diesel oxidation catalyst (DOC), and the many forms of diesel parti culate
filters, or traps. DOCs have been shown to be durable in-use, but they control only a relatively
small fraction of the total PM and, consequently, do not address our PM concerns sufficiently.
Uncatalyzed diesel paniculate filters demonstrated high efficiencies many years ago, but the level
of the PM standard was such that it could be met through less costly "in-cylinder" control
techniques. Catalyzed diesel particulate filters (CDPF, also referred to as catalyzed filters or
catalyzed traps, along with the very similar continuously regenerating DPF or CR-DPF) have the
potential to provide major reductions in diesel PM emissions and provide the durability and
dependability required for diesel applications. Throughout this document we will use the
acronym CDPF to refer to both catalyzed diesel particulate filters and the similar continuously
regenerating diesel particulate filter. Because of the significant PM reductions that they enable
and their proven durability on low sulfur diesel fuel, we believe the CDPF will be the control
a Note that throughout this document we refer to diesel and gasoline vehicles and engines. We tend to use
those terms given the preponderance of vehicles using diesel fuel or gasoline fuel in the U.S. heavy-duty highway
market. However, when we refer to a diesel engine, we mean any engine using the diesel cycle. When we refer to a
gasoline engine or vehicle, we mean any Otto-cycle vehicle or engine. Therefore, the emission standards discussed
throughout this preamble apply equally to engines and vehicles fueled by alternative fuels, unless otherwise
specified in the regulatory text accompanying today's rule.
m-i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
technology of choice for the future control of diesel PM emissions. However, CDPFs cannot be
brought to market on diesel applications without low sulfur diesel fuel.
In addition to the diesel PM control devices, several exhaust emission control devices
also have been developed to control diesel NOx. For reasons discussed in this chapter, we
believe that the NOx adsorber is the most likely candidate to be used to meet future low diesel
exhaust emission standards on the variety of applications in the heavy-duty diesel market. While
other technologies exist that have the potential to provide significant emission reductions, such
as selective catalytic reduction systems for NOx control, we believe that the NOx adsorber will
likely be the only broadly applicable technology choice by the makers of engines and vehicles for
the national fleet in the 2007-2010 time frame. Neither of the technologies for NOx control
discussed here, the NOx adsorber or the compact SCR, can meet the Phase 2 standards without
low sulfur diesel fuel.b
As for gasoline engines and vehicles, improvement continues to be made to gasoline
emissions control technology. This includes improvement to catalyst designs in the form of
improved washcoats and improved precious metal dispersion. Much effort has also been put into
improved cold start strategies that allow for more rapid catalyst light-off. This can be done by
retarding the spark timing to increase the temperature of the exhaust gases, and by using air-gap
manifolds, exhaust pipes, and catalytic converter shells to decrease heat loss from the system.
These improvements to gasoline emission controls will be made in response to the
California LEV-II standards and the federal Tier 2 standards. These improvements should
transfer well to the heavy-duty gasoline segment of the fleet. With such migration of light-duty
technology to heavy-duty vehicles and engines, we believe that considerable improvements to
heavy-duty gasoline emissions can be realized, thus allowing vehicles to meet much more
stringent standards than currently required.
The following discussion provides more detail on the technologies which are capable of
achieving the emission standards contained in this final rule. The purpose of this Chapter is to
discuss the emission reduction capability of these emission control technologies, as well as
discuss their sensitivity to sulfur in diesel fuel. We start with diesel applications, the technology
expected and its need for low sulfur diesel fuel, and finish with gasoline applications.
b The Phase 1 heavy-duty diesel standards are those required in 65 FR 59896, October 6, 2000, and are
often referred to as the "2004 standards." The standards required for 2007 and later model year engines,
promulgated in this rule represent Phase 2.
m-2
-------
Chapter III: Emissions Standards Feasibility
A. Feasibility of the Heavy-Duty Diesel Standards
1. Engine Out Improvements
Diesel engines have made great progress in lowering engine out emissions from 6.0
g/bhp-hr NOx and 0.6 g/bhp-hr PM in 1990 to 4.0 g/bhp-hr NOx and 0.1 g/bhp-hr PM in 1998.
These reductions came initially with improvements to combustion and fuel systems. Introduction
of electronic fuel systems in the early 1990s allowed lower NOx and PM levels without
sacrificing fuel economy. This, combined with increasing fuel injection pressures, have been the
primary technologies that have allowed emission levels to be reduced to current levels.
Further engine out NOx reductions to the levels necessary to comply with the Phase 1
standard of 2.5 g/bhp-hr NOx+NMHC will come primarily from the addition of cooled exhaust
gas recirculation (EGR). EGR recirculates a portion of the exhaust back to the intake manifold
where it is drawn into the combustion chamber. The earliest EGR systems were uncooled. The
recirculation of the hot exhaust gases to the intake manifold had the undesirable result of
increasing the intake manifold temperature. This temperature rise reduces the NOx reduction
potential of uncooled EGR since the NOx formation rate is proportional to temperature.1 The
intake manifold temperature rise also reduces the density of the fresh intake air, thus reducing the
mass of fresh air drawn into the combustion chamber which lowers the air-fuel ratio. Lower air-
fuel ratios generally increase PM emissions because there is less available oxygen to fully
combust the carbon.
To overcome the lower air-fuel ratio and intake air density, the EGR gas can be cooled
and/or the turbocharger can be adjusted to a higher pressure ratio. Both of these will almost
certainly be done, but the latter tends to increase the work required of the turbocharger which
increases fuel consumption. The former, cooling the recirculated exhaust gases, is limited by
condensation concerns since a significant portion of the exhaust gases is water vapor.2 The water
vapor generally contains sulfuric acid as a result of the combustion of fuel-borne sulfur. This
combustion results in SO2 in the exhaust gas which can react with excess oxygen and water to
form sulfuric acid (H2SO4). The level of EGR cooling is thus limited by the desire to prevent
condensation of this corrosive water and sulfuric acid mixture. Therefore, the intake manifold
temperature in an EGR equipped engine, even a cooled EGR engine, is usually higher than that
found in a non-EGR engine.
More sophisticated electronic control systems will be necessary to control the EGR
system and turbo machinery. 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 turbo chargers will also require a sophisticated control
algorithm to take advantage of their transient response, EGR pumping, and air flow control
m-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
characteristics. In addition, the turbo machinery 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. Consequently, the
Phase 1 heavy-duty diesel standards are expected to dramatically increase the capabilities of
future ECMs compared to current non-EGR equipped ECMs.
We believe that reductions in engine out emissions beyond the Phase 1 levels may be
attainable with low sulfur diesel fuel and more experience with cooled EGR systems. Low sulfur
fuel will allow more EGR to be used at lower temperatures because of the reduced threat of
sulfuric acid formation. In addition, recirculating the exhaust gases from downstream of a CDPF
may allow different EGR pumping configurations to be feasible. Current EGR systems draw
exhaust gases from the exhaust manifold upstream of the turbocharger turbine and recirculate
them through the EGR cooler and into the intake manifold downstream of the turbo compressor
and aftercooler. Such a system is called a high pressure loop EGR system because the gases are
drawn from high pressure upstream of the turbo turbine and recirculated to high pressure
downstream of the turbo compressor and aftercooler.
By contrast, a low pressure loop EGR system could draw some exhaust gases from the
exhaust downstream of the turbo turbine and any exhaust emission control devices and
recirculate those gases through the aftercooler and into the air intake system upstream of the
compressor. The low pressure loop approach increases the effectiveness of the EGR system
because it eliminates the high pressure loop EGR system's dependency on the pressure variations
that exist between the intake and exhaust manifolds. To date, low pressure loop EGR has not
been considered viable for HD applications because of the potential durability concerns
associated with recirculating exhaust gas containing particles and sulfuric acid through the
turbocharger compressor and aftercooler. The particles and acid accumulate in the aftercooler
(typically made of aluminum) plugging and corroding it. The turbocharger compressor is also
subject to particulate buildup and corrosion. But, by adding a CDPF and low sulfur fuel, the
particles and acid would be reduced significantly and these durability concerns would be
minimized.
Low pressure loop EGR systems provide many advantages over high pressure loop EGR
systems. For example, low pressure loop EGR allows more EGR to be pumped across a wider
engine operating range than with some other EGR configurations. As already pointed out, the
EGR does not have to be pumped against changing turbocharger pressure differentials found in
high pressure loop EGR systems that pump exhaust from the exhaust manifold upstream of the
turbocharger turbine through an EGR cooler to the intake manifold. For promoting EGR, the
pressure differential between the exhaust and intake manifolds can vary from very favorable at
rated speed to very unfavorable near torque peak. The unfavorable pressure differential near
torque peak requires work (i.e., pumping work) to be done to provide EGR during such engine
operation, thereby causing a fuel economy penalty. However, the low pressure loop EGR system
m-4
-------
Chapter III: Emissions Standards Feasibility
is not dependent on the pressure differential characteristics of the turbocharger, meaning that it
does not require this additional EGR pumping work. Therefore, we believe low pressure loop
systems may allow increased EGR rates, improved fuel economy, and perhaps even further
reductions of engine out emissions. However, these potential engine out emission reductions are
expected to be modest and are not expected to be sufficient to meet the emission standards
required by this final rule.
2. Meeting the PM Standard
Diesel PM consists of three primary constituents: elemental carbon particles from
incomplete combustion, which make up the largest portion of the total PM; the soluble organic
fraction (SOF), which consists of unburned hydrocarbons that have condensed into liquid
droplets or have adsorbed onto the surfaces of the elemental carbon particles; and sulfates with
associated water, which result from oxidation of fuel-borne sulfur in the engine's exhaust.
Several exhaust emission control devices have been developed to control diesel PM
constituents — the diesel oxidation catalyst (DOC), and the many forms of PM filters, CDPFs, or
PM traps. DOCs have been shown to be durable in use, but they effectively control only the SOF
portion of the total PM which, especially on today's engines, constitutes only around 10 to 30
percent of the total PM. Therefore, the DOC alone is not capable of meeting the FTP 0.01 g/bhp-
hr PM standard set in this final rule.
Only the catalyzed diesel particulate filter (CDPF) is capable of providing the level of
control required to meet the Phase 2 standards. In the past, the CDPF has demonstrated high
trapping efficiency, but regeneration of the collected PM has been a serious challenge, when
operating on anything other than low sulfur diesel fuel. The CDPF works by passing the exhaust
through a ceramic or metallic filter to collect the PM. The collected PM, mostly elemental
carbon particles but also a fraction of the SOF portion of PM at some low temperature
conditions, must then be burned off the CDPF before the filter becomes plugged. This burning
off of collected PM is referred to as "regeneration." The CDPF demonstrated high PM trapping
efficiencies many years ago, but the level of the applicable PM standard was such that it could be
met through less costly "in-cylinder" control techniques. As a result, the CDPF found little use
in the diesel market. The un-catalyzed diesel particulate filter is unlikely to be able to meet the
0.01 g/bhp-hr PM standard as they are only moderately effective at controlling the SOF fraction
of the PM. In addition, they require active regeneration technology which must be engaged
frequently making the systems expensive to operate (i.e., increasing fuel consumption) and less
reliable.
m-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
a. Catalyzed Diesel Particulate Filters
We believe the kind of diesel particulate filter that will be able to meet the PM standard
in a reliable, durable, cost effective manner, and the type of diesel parti culate filter that will
prove to the be the industry's technology of choice, is one capable of regenerating on an
essentially continuous basis such as the catalyzed diesel particulate filter or CDPF. These
CDPFs will be able to achieve very low PM emissions because:
they are highly efficient at trapping all forms of diesel PM;
• they employ precious metals to reduce the temperature at which regeneration
occurs, thereby allowing for passive regeneration under normal operating
conditions typical of a diesel engine;0
they have lower average back-pressure thereby reducing potential fuel economy
impacts, because they regenerate on a continuous basis; and,
they need no extra burners or heaters as would be required by an active
regeneration system, thereby reducing potential fuel economy impacts.
These CDPFs are able to provide in excess of 90 percent control of diesel PM, provided
they are operated on diesel fuel with sulfur levels at or below 15 ppm. High sulfur level diesel
fuel creates two problems for CDPFs. First, the CDPF cannot regenerate properly with current
fuel sulfur levels as such sulfur levels poison the catalyst inhibiting the NO to NO2 reaction
severely limiting regeneration of the trapped PM.3 Second, because SO2 is readily oxidized to
SO3 across the precious metals catalysts used in CDPFs, the Supplemental Emission Test (SET)
0.01 g/bhp-hr PM standard contained in today's final rule cannot be achieved with diesel fuel
sulfur levels greater than 15 ppm because of the resultant increase in sulfate PM emissions
("sulfate make"). Table in.A-1 shows that at a 15 ppm sulfur level, a CDPF can achieve 0.009
g/bhp-hr over the Supplemental Emission Test (SET), but at 30 ppm sulfur, the SET results are
0.017 g/bhp-hr, which would exceed the standard. See the discussion later in this chapter for
further information on CDPFs and sulfur, in particular section IH.A.T.a.ii, "Loss of PM Control
Effectiveness."
More than one emission control manufacturer is developing these precious metal
catalyzed, passively regenerating CDPFs. In field trials, they have demonstrated highly efficient
PM control and promising durability. A recent publication documents results from a sample of
these field test engines after years of use in real world applications.4 The sampled CDPFs had on
0 For CDPF regeneration without precious metals, temperatures in excess of 650°C must be obtained. At
such high temperatures, elemental carbon will burn provided sufficient oxygen is present. However, diesel engines
rarely if ever operate with such high exhaust temperatures. For example, exhaust temperatures on the HDE Federal
Test Procedure cycle typically range from 100°C to 450°C. Precious metal CDPFs use platinum to oxidize NO in
the exhaust to NO2, which is capable of oxidizing carbon at temperatures as low as 250°C to 300°C.
m-6
-------
Chapter III: Emissions Standards Feasibility
average four years of use covering more than 225,000 miles in applications ranging from city
buses to garbage trucks to intercity trains, with some units accumulating more than 360,000
miles. When tested on the US Heavy-Duty Federal Test Procedure (HD FTP), they continued to
demonstrate PM reductions in excess of 90 percent.
The experience gained in these field tests also helps to clarify the need for very low sulfur
diesel fuel. In Sweden and some European city centers where below 10 ppm diesel fuel sulfur is
readily available, more than 3,000 CDPFs have been introduced into retrofit applications without
a single failure. This success on 10 ppm sulfur fuel is all the more impressive as some of these
units have been in operation for more than six years. The field experience in areas where sulfur
is capped at 50 ppm has been less definitive. In regions without extended periods of cold
ambient conditions (such as the United Kingdom) field tests on 50 ppm cap sulfur fuel have been
positive, with no reported durability issues. Of course, it should be mentioned that an HDDE
equipped with a CDPF operating on 50 ppm sulfur would not meet the PM emission standards
contained in this final rule, as discussed below. These good results in the UK are contrasted with
field tests in Finland where colder winter conditions are sometimes encountered (similar to many
northern regions of the United States). The testing in Finland revealed a failure rate of 10 percent
(14 failures in the test program) when operated on fuel with a sulfur cap of 50 ppm. This 10
percent failure rate has been attributed to insufficient CDPF regeneration due to fuel sulfur in
combination with low ambient temperatures.5 Other possible reasons for the high failure rate in
Finland when contrasted with the Swedish experience appear to be unlikely. The Finnish and
Swedish fleets were substantially similar, with both fleets consisting of transit buses powered by
Volvo and Scania engines in the 10 to 11 liter range. Further, the buses were operated in city
areas and none of the vehicles were operated in northern extremes such as north of the Arctic
Circle.6 Given that the fleets in Sweden and Finland were substantially similar, and given that
ambient conditions in Sweden are expected to be similar to those in Finland, we believe that the
increased failure rates noted here are due to the higher fuel sulfur level in a 50 ppm cap fuel
versus a 10 ppm cap fuel.d Testing on an even higher fuel sulfur level of 200 ppm was conducted
in Denmark on a fleet of 9 vehicles. In less than six months, all of the vehicles in the Danish
fleet had failed due to plugging of the CDPFs.7 We believe that this real world testing clearly
indicates that increasing diesel fuel sulfur levels limit CDPF regeneration, leading to plugging of
the CDPF even at fuel sulfur levels as low as 50 ppm.
d The average temperature in Helsinki, Finland, for the month of January is 21°F. The average
temperature in Stockholm, Sweden, for the month of January is 26°F. The average temperature at the University of
Michigan in Ann Arbor, Michigan, for the month of January is 24°F. The temperatures reported here are from
www.worldclimate.com based upon the Global Historical Climatology Network (GHCN) produced jointly by the
National Climatic Data Center and Carbon Dioxide Information Analysis Center at Oak Ridge National Laboratory
(ORNL).
m-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Another program evaluating CDPFs in the field is the ARCO Emission Control Diesel
(EC-D) program.6 In that program, a one-year technology validation is being run to evaluate EC-
D and CDPFs using diesel vehicles operating in southern California. The fuel's performance,
impact on engine durabiltiy and vehicle performance, and emission characteristics are being
evaluated in several fleets in various applications. The program is still ongoing, but interim
results have been made available.8 These interim results have shown that vehicles retrofitted
with CDPFs and fueled with EC-D (7.4 ppm sulfur) emitted 91 percent to 99 percent less PM
compared to the vehicles fueled with California diesel fuel (121 ppm sulfur) having no exhaust
filter equipment. Further, the test vehicles equipped with the CDPFs and fueled with EC-D have
operated reliably during the program start-up period and no significant maintenance issues have
been reported for the school bus, tanker truck and grocery truck fleets that have been operating
for over six months (approximately 50,000 miles).9
From these results, we can further conclude that lighter applications (such as large pick-
up trucks and other light- and medium-heavy-duty applications), having lower exhaust
temperatures than heavier applications, may experience similar failure rates even in more
temperate climates and would, therefore, need lower sulfur fuel even in the United Kingdom.
These results are understood to be due to the effect of sulfur on the CDPF's ability to create
sufficient NO2 to carry out proper filter regeneration. Without the NO2, the CDPF continues to
trap PM at high efficiency, but it is unable to oxidize, or regenerate, the trapped PM. The
possible result is a plugged CDPF.
Much development effort is underway worldwide to bring PM exhaust emission control
devices to market. One of the drivers is the Euro IV PM standard set to become effective in
2005.f This standard sets a PM emission target that forces CDPF usage. In anticipation of the
2005 introduction date, field testing is already underway in several countries with CDPFs. We
believe the experience gained in Europe with these technologies will coincide well with the 2007
standards contained in this final rule. The timing of the new standards harmonizes the heavy-
duty highway PM technologies with those expected to be used to meet the light-duty highway
Tier 2 standards. With this level of development already under way, we are confident that the
PM standard will be met provided low sulfur fuel is made available.
The data currently available show that CDPFs can provide significant reductions in PM,
and are capable of meeting a standard of 0.01 g/bhp-hr PM. CDPFs with precious metal
catalysts, in conjunction with low sulfur fuel, have been shown to be more than 90 percent
e EC-D is a diesel fuel developed recently by ARCO (Atlantic Richfield Company) from typical crude oil
using a conventional refining process and having a fuel sulfur content less than 15 ppm.
f The Euro IV standards are 2.6 g/hp-hr NOx and 0.015 g/hp-hr PM over the European Stationary Cycle
and European Transient Cycle.
m-8
-------
Chapter III: Emissions Standards Feasibility
efficient over the FTP and across the NTE zone.10 Figure in.A-lg shows representative FID FTP
CDPF efficiencies with fuel sulfur levels near 15 ppm.11121314 Since the total PM removal
efficiency of the CDPF is roughly linear with fuel sulfur (as shown in the DECSE program
results15), it can be seen from Figure IHA-1 that even at the 15 ppm sulfur cap, current CDPFs
can produce greater than 90 percent trapping efficiency over the HD FTP. Based on the current
FTP standard of 0.10 g/bhp-hr, these CDPFs will easily achieve the FTP 0.01 g/bhp-hr PM
standard contained in today's final rule.
100%
95%
90%
70%
15
20 25 30
Fuel Sulfur (ppm)
35
40
45
50
Navistar
Figure III.A-1. HD CDPF PM Removal Efficiency Over the Federal Test
Procedure
As part of the EPA National Vehicle and Fuel Emissions Laboratory (NVFEL) test
program to evaluate CDPFs and NOx adsorbers, we performed CDPF testing over the hot-start
FIDDE FTP. This testing included CDPF evaluation with three ppm sulfur fuel, which produced
greater than a 95 percent reduction in PM, with a post-CDPF emission rate of 0.004 g/bhp-hr
PM.16 In addition, we performed testing of a complete system which included CDPFs, NOx
adsorbers, and a clean-up diesel oxidation catalyst, which also produced on average greater than
a 95 percent reduction during triplicate hot-start UDDE FTP testing, with an average post-CDPF
g Figure III.A-1 includes a Navistar data point at 200 ppm sulfur and 61 percent PM reduction; this data
point does not appear in the figure so that the data from 0 ppm sulfur to 50 ppm sulfur can be more easily viewed.
m-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
emission rate of 0.002 g/bhp-hr PM (with a 95 percent confidence interval of ± 0.001 g/bhp-hr),
using six ppm sulfur fuel (additional discussion of this NVFEL test program can be found in
section in.A.S.b of this RIA, as well as in the docket for this final rule).h As indicated by Figure
in.A-1, when typical particulate sulfate conversion rates for the HDDE FTP are considered, the
EPA NVFEL test program would be projected to produce greater than a 90 percent PM reduction
at a fuel sulfur level of 15 ppm.
The engine operating conditions have little impact on the particulate trapping efficiency
of carbon particles by CDPFs, so the greater than 90 percent efficiency for elemental carbon
particulate matter will apply to engine operation within the NTE zone, as well as the test modes
which comprise the SET. This is supported by a study by Johnson Matthey which showed
greater than 95 percent reduction in elemental carbon PM over a broad range of operating
modes.17 This same paper also shows large reductions in the soluble organic fraction of the PM
across the engine operating map. However, engine operation will affect the CDPF regeneration
and oxidation of SO2 to sulfate PM (i.e., "sulfate-make"). Sulfate-make will reduce the
measured PM removal efficiency at some NTE operating conditions and supplemental steady-
state modes, even at the 15 ppm fuel sulfur cap. Figure UI.A-2 shows PM removal efficiency as
a function of fuel sulfur for a CDPF when tested over the SET.18 From the graph, it can be seen
that fuel sulfur level has a stronger effect on PM removal efficiency over the SET than over the
HD FTP. This increased sensitivity to fuel sulfur is caused by the higher temperatures that are
found at some of the steady-state modes. High exhaust temperatures promote the oxidation of
SO2 to SO3 (which then combines with water in the exhaust, forming a hydrated sulfate) across
the precious metals found in CDPFs. The sulfate emissions condense in the atmosphere (as well
as in the CFR mandated dilution tunnel used for PM testing) forming PM. Figure IU-A-2 shows
PM reductions of 85 percent or greater are achievable with 15 ppm sulfur fuel. Engine-out PM
emission rates over the SET test are typically 50 percent or less of the FTP PM emission rates,
primarily because carbonaceous PM formation is greater under transient engine operation as
compared to steady-state operation. For example, model year 2000 certification data for a
number of HDDE families shows SET PM emission rates between 0.02 and 0.05 g/bhp-hr.19
Therefore, an 85 percent reduction in PM over the SET test is sufficient to comply with the 2007
SET PM standard contained in this final rule.
h The NVFEL emission measurement test equipment used to evaluate the performance of a HDDE
equipped only with a CDPF (no NOx adsorbers) utilized PM sample equipment and procedures consistent with the
existing 40 CFR part 86, subpart N provisions (including dual 70 mm filters). The NVFEL test equipment used to
evaluate the performance of a complete system (CDPFs, NOx adsorbers, clean-up DOC) utilized PM sample
equipment and techniques consistent with the new PM measurement regulations contained in this rule, including a
single, high efficiency 47 mm filter.
HI-10
-------
Chapter III: Emissions Standards Feasibility
100%
95%
90%
85%
80%
75%
70%
Catalyzed PM Traps
10 15 20 25 30
Fuel Sulfur (ppm)
35
40
45
50
"DECSE-Catalyzed Diesel Participate Filter
"DECSE-Continously Regenerating Diesel Participate Filter
Figure III.A-2. HD PM Removal Efficiency for a CDPF Over the Supplemental
Emission Test (SET)
Nonetheless, as shown in Table ni.A-1, a HDDE equipped with CDPFs available today is
capable of meeting the SET standard (equal to l.Ox FTP standard, or 0.01 g/bhp-hr) with 15 ppm
fuel. Table in.A-1 shows data from the Diesel Emission Control Sulfur Effects (DECSE) test
program, a program conducted by the US Department of Energy in cooperation with industry to
provide insight into the relationship between advanced emission control technologies and diesel
fuel sulfur levels. Interim report number four of this program gives the total PM emissions from
a heavy-duty diesel engine operated with a CDPF on several different fuel sulfur levels. Table
in.A-1 also shows interpolated points representing a straight line fit through the DECSE data
illustrating the expected total PM emissions from a heavy-duty diesel engine on the SET at
various fuel sulfur levels. As shown, the PM emissions at a 15 ppm sulfur level would be 0.009
g/bhp-hr, ten percent below the 0.01 g/bhp-hr standard set in this final rule, which demonstrates
the feasibility of the standard at 15 ppm sulfur.
m-ii
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table III.A-1. PM Emissions from a Heavy-Duty Diesel Engine at the
Indicated Fuel Sulfur Levels
Fuel Sulfur Level
3
7*
15*
30
150
Supplemental Steady State
Tailpipe PM
[g/bhp-hr]
0.003
0.006
0.009
0.017
0.071
Relative to the Standard
(%)
-70
-40
-10
70
610
* The PM emissions at these sulfur levels are based on a straight-line fit to the DECSE
program data;20 PM emissions at other sulfur levels are actual DECSE data.
The NTE requirement, unlike the HD FTP or SET standard, is not a composite test (i.e.,
the 20 minute transient HD FTP cycle or the 13-steady-state SET modes). In fact, a number of
the individual modes within the SET test fall within the NTE engine control zone. As discussed
above, CDPFs are very efficient at capturing elemental carbon PM (>95 percent), but sulfate-
make under certain operating conditions may exceed the FTP or SET standard, which is part of
the reason the PM NTE standard is greater than the FTP and SET PM standards. In addition to
the composite SET results, the DECSE test program also reported PM performance results at two
individual steady-state modes, the peak-torque condition and a "road-load" condition. The peak-
torque test mode produces very high exhaust gas temperature (and would therefore be
representative of the highest sulfate particulate generating conditions) and the road-load
condition is intended to be representative of a typical HD diesel engine line-haul cruise operation
(75 percent load, Euro B speed). A linear fit of the DECSE PM emission results for the road-
load and peak-torque conditions between the three ppm sulfur test fuel and the 30 ppm sulfur test
fuel point shows that the two CDPFs which were tested produced an 88-94 percent or greater
reduction at a linear interpolated sulfur level of seven ppm (near the expected in-use average) and
an 83-90 percent reduction at a linear interpolated sulfur level of 15 ppm (capped level), for both
test operating points. For both CDPFs, the road-load condition resulted in lower sulfate make
and higher overall PM reduction than the peak-torque condition. Based on this information,
under very high parti culate sulfate formation conditions, at 15 ppm sulfur a CDPF can produce at
least an 83 percent reduction, and at the expected refinery average sulfur level of seven ppm,
when operated at very high sulfate conversion engine conditions a CDPF can produce at least an
88 percent reduction. It should be noted that a prolonged steady-state test condition at the peak-
torque mode for a HDDE is representative of the highest exhaust gas temperature producing
engine operating conditions. The DECSE testing conditions for these two steady-state points
HI-12
-------
Chapter III: Emissions Standards Feasibility
prescribed a five minute warm-up and a 20 minute sample collection, for a total of 25 minutes of
operation at peak-torque. The peak-torque test data described above could be considered
representative of the worst case paniculate sulfate generating conditions. The data show that
even under these test conditions, an 83 percent reduction would be sufficient to comply with the
NTE provisions, even considering PM sulfate make, when tested on 15 ppm sulfur fuel. Under
the Phase 1 rule, a HDDE could emit PM emissions subject to the NTE requirements as high as
0.13 g/bhp-hr (1.25 x 0.10 g/bhp-hr). An 83 percent reduction from this engine would result in a
PM emission rate of 0.022 g/bhp-hr, which would comply with the 2007 NTE PM standards
contained in this final rule.1
Under the Phase 1 HDDE provisions (which includes the Phase 1 FTP standards and the
2007 NTE provisions as they apply to the Phase 1 FTP standards), emission "carve-out" zones of
the NTE control zone were defined. These carve-out zones are areas within the defined NTE
control zone which are excluded from meeting the NTE standards for specific emissions. The
Phase 1 rule defined two types of carve-out zones, one which applied to all regulated emissions
(gaseous emissions and PM), and one carve-out zone which only applied to PM. The PM only
carve-out zone was specified to exclude low load, high speed engine operation from the NTE
requirements. During these conditions, FtDDEs not equipped with CDPFs can produce higher
PM emission rates, and it was decided within the Phase 1 rule to exclude FtDDEs from
complying with the PM NTE requirements when operated within the defined PM carve-out
zones. With the application of CDPFs to HDDEs, these PM only carve-out zones are not needed.
As discussed previously, CDPFs are very effective at reducing engine-out PM. During the recent
NVFEL test program which evaluated the effectiveness of diesel CDPFs on low sulfur diesel
fuel, we included one steady-state test point which was within the Phase 1 rule's PM carve-out
zone. Specifically, the test point was 228 ft-lbs torque and 2,415 rpm (listed as Mode 6 in Figure
in.A-6 of this Chapter), which places the test condition near the center of the Phase 1 rule's PM
carve-out zone for this engine. At this operating condition, one of the CDPFs reduced engine-out
PM by more than 95 percent, from 0.068 g/bhp-hr to 0.003 g/bhp-hr, the second CDPF produced
similar results. Based on the high PM reduction capability of CDPFs when operated on low
sulfur diesel fuel, and their demonstrated ability to achieve >90 percent reductions when operated
inside the Phase 1 PM carve-out zones, we have eliminated the PM-only carve-out zones from
the NTE requirements.
1 The PM NTE standard contained in this final rule is 1.5 x FTP standard, or 1.5 x 0.01 g/bhp-hr. 40 CFR
86.007-1 l(a)(4)(v) specifies that the rounding procedures in ASTM E29-90 should be applied to the NTE emission
standard, therefore, the NTE standard is rounded to the same number of significant digits as the FTP standard, i.e.,
1.5 x O.Olg/bhp-hr is rounded to 0.02 g/bhp-hr. An engine with a measured NTE PM emission rate of 0.022 g/bhp-
hr would also be rounded using ASTM rounding provisions, and would be rounded to the same number of
significant digits as the standard, so 0.022 g/bhp-hr would round to 0.02 g/bhp-hr, and would meet the NTE PM
standard.
m-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The NTE requirements apply not only during standard laboratory conditions, but also
during the expanded ambient temperature, humidity, and altitude limits defined in the
regulations. We believe the NTE PM standard is technologically feasible across this range of
ambient conditions. As discussed above, CDPFs are mechanical filtration devices, and ambient
temperature changes will have minimal effect on CDPF performance. Ambient altitude will also
have minimal, if any, effects on CDPF filtration efficiencies, and ambient humidity should have
no effect on CDPF performance. As discussed above, particulate sulfate make is sensitive to
high exhaust gas temperatures, however, at sea-level conditions, the NTE requirements apply up
to ambient temperatures which are only 14°F greater than standard test cell conditions (100°F
under the NTE, versus 86°F for FID FTP laboratory conditions). At an altitude of 5,500 feet
above sea-level, the NTE applies only up to an ambient temperature within the range of standard
laboratory conditions (i.e., 86°F). These small or non-existent differences in ambient
temperature should have little effect on the sulfate make of CDPFs, and as discussed above, even
when tested under at an engine operating test mode representative of the highest particulate
sulfate generating conditions (25 minutes at peak-torque operation) with 15 ppm sulfur diesel
fuel, we predict the engine would comply with the PM NTE standard. Based on the available test
data and the expected impact of the expanded, but constrained, ambient conditions under which
engines must comply with the NTE, we conclude that the PM NTE standard is technologically
feasible by 2007, provided low sulfur diesel fuel (<15 ppm) is available.
There may be a need to remove, clean, and reverse these CDPFs at regular intervals to
remove ash build-up resulting from engine oil. Small amounts of oil can enter the exhaust via
the combustion chamber (past the pistons rings and valve seals), and via the crankcase
ventilation system. This can lead to ash build-up, primarily as a result of the metallic oil
additives used to provide pH control. This pH control is necessary, in part, to neutralize sulfuric
acid produced as a byproduct of burning fuel containing sulfur. However, with reduced fuel
sulfur, these oil additives could be reduced, thereby reducing the rate of ash build-up and
lengthening any potential cleaning intervals. It may also be possible to use oil additives that are
less prone to ash formation to reduce the need for periodic maintenance to at least those specified
in CFR 86.004-25 (100,000 miles or 3,000 hours for light heavy-duty vehicles, and 150,000
miles or 4,500 hours for medium- and heavy-duty engines). Periodic maintenance would consist
of reversing the CDPF and/or washing it out with compressed air or water. Consequently, we
conclude that CDPFs will be able to meet the required emission life with minimal maintenance.
b. Control of Ultra-Fine PM
CDPFs reduce PM by capturing and burning particles. Ninety percent of the PM mass
resides in particle sizes that are less than 1000 nanometers (nm) in diameter, and half of these
particles are less than 200 nm.2122 23 24 25 Fortunately, CDPFs have very high particle capture
efficiencies. PM less than 200 nm is captured efficiently by diffusion onto surfaces within the
CDPF walls. Larger particles are captured primarily by inertial impaction onto surfaces due to the
HI-14
-------
Chapter III: Emissions Standards Feasibility
tortuous path that exhaust gas must take to pass through the porous CDPF walls. Capture
efficiency for elemental carbon (soot) and metallic ash is nearly 100 percent; therefore,
significant PM can only form downstream of the CDPF. Volatile PM forms from sulfate or
organic vapors via nucleation, condensation, and/or adsorption during initial dilution of raw
exhaust into the atmosphere. Kleeman, et al., and Kittelson, et al., independently demonstrated
that these volatile particles reside in the ultra-fine PM range (i.e., <100 nm range).26 27 Thus
ultra-fine PM is comprised primarily of semi-volatile PM. The organic portion of semi-volatile,
ultra-fine PM can be controlled via oxidation over a PGM catalyst. The sulfate portion of semi-
volatile, ultra-fine PM can be reduced by eliminating sulfur from the fuel. Furthermore, the work
of Kittelson et al. suggests that reducing sulfate PM in this manner will reduce the number of
nucleation sites available for the nucleation of ultra-fine organic PM, forcing more of the organic
material to adsorb onto the much larger soot agglomerates and thus reducing the number of ultra-
fine organic particles.
Modern CDPFs have been shown to be very effective at reducing PM mass. In addition,
recent data shows that they are also very effective at reducing the overall number of emitted
particles when operated on low sulfur fuel. Hawker, et. al., found that a modern CDPF reduced
particle count by over 95 percent, including some of the smallest measurable particles (< 50 nm),
at most of the tested conditions. The lowest observed efficiency in reducing particle number was
86 percent. No generation of particles by the CDPF was observed under any tested conditions.28
Kittelson, et al., confirmed that ultra-fine particles can be reduced by a factor often by oxidizing
volatile organics, and by an additional factor often by reducing sulfur in the fuel. CDPFs
efficiently oxidize nearly all of the volatile organic PM precursors, and elimination of as much
fuel sulfur as possible will substantially reduce the number of ultra-fine PM emitted from diesel
engines. The combination of CDPFs with low sulfur fuel is expected to result in very large
reductions in both PM mass and the number of ultra-fine particles.
3. Meeting the NOx Standard
NOx emissions from gasoline powered vehicles are controlled to extremely low levels
through the use of the three-way catalyst technology first introduced in the 1970s. Historically,
reduction of NOx emissions in the oxygen-rich environment typical of diesel exhaust has been
significantly more difficult because known catalytic NOx reduction mechanisms like the gasoline
three-way catalyst work only when the oxygen content of the exhaust is very low. Nevertheless,
significant progress has been made in developing catalytic emission control technologies that
reduce the NOx to form harmless oxygen and nitrogen in the oxygen rich (lean burn) exhaust
environment typical of diesel engines. These devices are the lean NOx catalyst, the NOx
adsorber, selective catalytic reduction (SCR), and non-thermal plasma.
m-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
a. Lean NOx Catalysts
Lean NOx catalysts have been under development for some time, and two methods have
been developed for using a lean NOx catalyst depending on the level of NOx reduction desired
though neither method can produce more than a 30 percent NOx reduction. The "active" lean
NOx catalyst injects a reductant that serves to reduce NOx to N2 and O2 (typically diesel fuel is
used as the reductant). The reductant is introduced upstream of, or into, the catalyst. The
presence of the reductant provides locally oxygen poor conditions which allows the NOx
emissions to be reduced by the catalyst.
The lean NOx catalyst washcoat incorporates a zeolite catalyst that acts to adsorb
hydrocarbons from the exhaust stream. Once adsorbed on the zeolite, the hydrocarbons will
oxidize and create a locally oxygen poor region that is more conducive to reducing NOx. To
promote hydrocarbon oxidation at lower temperatures, the washcoat can incorporate platinum or
other precious metals. The platinum also helps to eliminate the emission of unburned
hydrocarbons that can occur if too much reductant is injected, referred to as "hydrocarbon slip."
With platinum, the NOx conversion can take place at the low exhaust temperatures that are
typical of diesel engines. However, the presence of the precious metals can lead to production of
sulfate PM, as already discussed for PM control technologies.
Active lean NOx catalysts have been shown to provide up to 30 percent NOx reduction
under limited steady-state conditions. However, this NOx control is achieved with a fuel
economy penalty upwards of 7 percent due to the need to inject fuel into the exhaust stream.29
NOx reductions over the HD transient FTP are only on the order of 12 percent due to excursions
outside the optimum NOx reduction efficiency temperature range for these devices.30
Consequently, the active lean NOx catalyst does not appear to be capable of enabling the
significantly lower NOx emissions required by the NOx standard.
The "passive" lean NOx catalyst uses no reductant injection. Therefore, the passive lean
NOx catalyst is even more limited in its ability to reduce NOx because the exhaust gases
normally contain very few hydrocarbons. For that reason, today's passive lean NOx catalyst is
capable of best steady state NOx reductions of less than 10 percent. Neither approach to lean
NOx catalysis listed here can provide the significant NOx reductions required to satisfy the air
quality needs discussed in chapter n.
b. NOx Adsorbers
NOx emissions from gasoline powered vehicles are controlled to extremely low levels
through the use of the three-way catalyst technology first introduced in the 1970s. Today, an
advancement upon this well developed three-way catalyst technology, the NOx adsorber, has
shown that it too can make possible extremely low NOx emissions from lean burn engines such
HI-16
-------
Chapter III: Emissions Standards Feasibility
as diesel engines. The potential of the NOx adsorber catalyst is limited only by its need for
careful integration with the total vehicle system (as was done for three-way catalyst equipped
passenger cars in the 1980s and 1990s) and by poisoning of the catalyst from sulfur in the fuel.
The following subsections describe the function, design and technical challenges remaining for
application of the NOx adsorber catalyst to heavy-duty diesel vehicles.
/'. How do NOx Adsorbers Work?
The NOx adsorber catalyst is a further development of the three-way catalyst technology
developed for gasoline powered vehicles more than twenty years ago. The NOx adsorber
enhances the three-way catalyst function through the addition of storage materials on the catalyst
surface which can adsorb NOx under oxygen rich conditions. This enhancement means that a
NOx adsorber can allow for control of NOx emissions under lean burn (oxygen rich) operating
conditions typical of diesel engines.
Three-way catalysts reduce NOx emissions as well as HC and CO emissions (hence the
name three-way) by promoting oxidation of HC and CO to water and CO2 using the oxidation
potential of the NOx pollutant and in the process reducing the NOx emissions to atomic nitrogen,
N2. Said another way, three-way catalysts work with exhaust conditions where the net oxidizing
and reducing chemistry of the exhaust is approximately equal, allowing the catalyst to promote
complete oxidation/reduction reactions to the desired exhaust components, carbon dioxide (CO2),
water (H2O) and nitrogen (N2). The oxidizing potential in the exhaust comes from NOx
emissions and some oxygen (O2) which is not consumed during combustion. The reducing
potential in the exhaust comes from HC and CO emissions, which represent products of
incomplete combustion. Operation of the engine to ensure that the oxidizing and reducing
potential of the combustion and exhaust conditions is precisely balanced is referred to as
stoichiometric engine operation.
If the exhaust chemistry varies from stoichiometric conditions emission control is
decreased. If the exhaust chemistry is net "fuel rich," meaning there is an excess of HC and CO
emissions in comparison to the oxidation potential of the NOx and O2 present in the exhaust, the
excess HC and CO pollutants are emitted from the vehicle. Conversely, if the exhaust chemistry
is net "oxygen rich" (lean burn), meaning there is an excess of NOx and O2 in comparison to the
reducing potential of the HC and CO present in the exhaust, the excess NOx pollutants are
emitted from the vehicle. It is this oxygen rich operating condition that typifies diesel engine
operation. Because of this, diesel engines equipped with three-way catalysts (or simpler
oxidation catalysts) have very low HC and CO emissions while NOx (and O2) emissions remain
almost unchanged from the high engine out levels. For this reason, when diesel engines are
equipped with catalysts (diesel oxidation catalysts, or DOCs) they have HC and CO emissions
that are typically lower, but have NOx emissions that are an order of magnitude higher, than for
gasoline engines equipped with three-way catalysts.
HI-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The NOx adsorber catalyst works to overcome this situation by storing NOx emissions
when the exhaust conditions are oxygen rich. Unfortunately the storage capacity of the NOx
adsorber is limited, requiring that the stored NOx be periodically purged from the storage
component. If the exhaust chemistry is controlled such that when the stored NOx emissions are
released the net exhaust chemistry is at stoichiometric or net fuel rich conditions, then the three-
way catalyst portion of the catalyst can reduce the NOx emissions in the same way as for a
gasoline three-way catalyst equipped engine. Simply put, the NOx adsorber works to control
NOx emissions by storing NOx on the catalyst surface under lean burn conditions typical of
diesel engines and then by reducing the NOx emissions with a three-way catalyst function by
periodically operating under stoichiometric or fuel rich conditions.
The NOx storage process can be further broken down into two steps. First the NO in the
exhaust is oxidized to NO2 across an oxidation promoting catalyst, typically platinum. Then the
NO2 is further oxidized and stored on the surface of the catalyst as a metallic nitrate (MNO3).
The storage components are typically alkali or alkaline earth metals that can form stable metallic
nitrates. The most common storage component is barium carbonate (BaCO3) which can store
NO2 as barium nitrate (Ba(NO3)2) while releasing CO2. In order for the NOx storage function to
work, the NOx must be oxidized to NO2 prior to storage and a storage site must be available (the
device cannot be "full"). During this oxygen rich portion of operation, NOx is stored while HC
and CO emissions are oxidized across the three-way catalyst components by oxygen in the
exhaust. This can result in near zero emissions of NOx, HCs, and CO under the net oxygen rich
operating conditions typical of diesel engines.
The NOx adsorber releases and reduces NOx emissions under fuel rich operating
conditions through a similar two step process, referred to here as NOx adsorber regeneration.
The metallic nitrate becomes unstable under net fuel rich operating conditions, decomposing and
releasing the stored NOx. Then the NOx is reduced by reducing agents in the exhaust (CO and
HCs) across a three-way catalyst system, typically containing platinum and rhodium. Typically
this NOx regeneration step occurs at a significantly faster rate than the period of lean NOx
storage such that the fuel rich operation constitutes only a small fraction of the total operating
time. Since this release and reduction step, NOx adsorber regeneration, occurs under net fuel
rich operating conditions, NOx emissions can be almost completely eliminated. But for some of
the HC and CO emissions, "slip"(failure to remove all of the HC and CO) may occur during this
process. The HC and CO slip can be controlled with a downstream "clean-up" catalyst that
promotes their oxidation or potentially by controlling the exhaust constituents such that the
excess amount of the HC and CO pollutants at the fuel rich operating condition is as low as
possible, that is, as close to stoichiometric conditions as possible.
The difference between stoichiometric three-way catalyst function and the newly
developed NOx adsorber technology can be summarized as follows. Stoichiometric three-way
catalysts work to reduce NOx, HCs and CO by maintaining a careful balance between oxidizing
HI-18
-------
Chapter III: Emissions Standards Feasibility
(NOx and O2) and reducing (HCs and CO) constituents and then promoting their mutual
destruction across the catalyst on a continuous basis. The newly developed NOx adsorber
technology works to reduce the pollutants by balancing the oxidation and reduction chemistry on
a discontinuous basis, alternating between net oxygen rich and net fuel rich operation in order to
control the pollutants. This approach allows lean-burn engines (oxygen rich operating), like
diesel engines, to operate under their normal operating mode most of the time, provided that they
can periodically switch and operate such that the exhaust conditions are net fuel rich for brief
periods. If the engine/emission control system can be made to operate in this manner, NOx
adsorbers offer the potential to employ the highly effective three-way catalyst chemistry to lean
burn engines.
/'/'. Where are NOx Adsorbers used Today?
NOx adsorber catalysts were first introduced in the power generation market less than
five years ago. Since then, NOx adsorber systems in stationary source applications have enjoyed
considerable success. In 1997, the South Coast Air Quality Management District of California
determined that a NOx adsorber system provided the "Best Available Control Technology" NOx
limit for gas turbine power systems.31 Average NOx control for these power generation facilities
is in excess of 92 percent.32 A NOx adsorber catalyst applied to a natural gas fired powerplant
has demonstrated better than 99 percent reliability for more than 21,000 hours of operation while
controlling NOx by more than 90 percent.33 The experience with NOx adsorbers in these
stationary power applications shows that NOx adsorbers can be highly effective for controlling
NOx emissions for extended periods of operation with high reliability.
The NOx adsorber's ability to control NOx under oxygen rich (fuel lean) operating
conditions has lead industry to begin applying NOx adsorber technology to lean burn engines in
mobile source applications. NOx adsorber catalysts have been developed and are now in
production for lean burn gasoline vehicles in Japan, including several vehicle models sold by
Toyota Motor Corporation.j The 2000 model year saw the first U.S. application of this
technology with the introduction of the Honda Insight, certified to the California LEV-IULEV
category standard. These lean burn gasoline applications are of particular interest because they
are similar to diesel vehicle applications in terms of lean NOx storage and the need for periodic
NOx regeneration under transient driving conditions. The fact that they have been successfully
applied to these mobile source applications shows clearly that NOx adsorbers can work under
transient conditions provided that engineering solutions can be found to periodically cause
normally lean-burn exhaust conditions to operate in a rich regeneration mode.
1 Toyota requires that their lean burn gasoline engines equipped with NOx adsorbers are fueled on
premium gasoline in Japan, which has an average sulfur content of six ppm.
HI-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Hi. Can NOx Adsorbers be applied to Diesel Engines?
NOx adsorbers work to control NOx emissions by storing the NOx pollutants on the
catalyst surface during oxygen rich engine operation (lean burn engine operation) and then by
periodically releasing and reducing the NOx emissions under fuel rich exhaust conditions. This
approach to controlling NOx emissions can work for a diesel engine provided that the engine and
emission control system can be designed to work in concert, with relatively long periods of
oxygen rich operation (typical diesel engine operation) followed by brief periods of fuel rich
exhaust operation. The ability to control the NOx emissions in this manner is the production
basis for lean burn NOx emission control in stationary power systems and for lean burn gasoline
engines. As outlined below we believe that there are several approaches to accomplish the
required periodic operation on a diesel engine.
(a) With In-Cylinder Control Systems
The most frequently mentioned approach for controlling the exhaust chemistry of a diesel
engine is through in-cylinder changes to the combustion process. This approach roughly mimics
the way in which lean-burn gasoline engines function with NOx adsorbers. That is the engine
itself changes in operation periodically between "normal" lean burn (oxygen rich) combustion
and stoichiometric or even fuel rich combustion in order to promote NOx control with the NOx
adsorber catalyst. For diesel engines this approach typically requires the use of common rail fuel
systems which allow for multiple fuel injection events along with an air handling system which
includes exhaust gas recirculation (EGR).
The normal lean burn engine operation can last from as little time as 15 seconds to more
than three minutes as the exhaust NOx emissions are stored on the surface of the NOx adsorber
catalyst. The period of fuel lean, oxygen rich, operation is determined by the NOx emission rate
from the engine and the storage capacity of the NOx adsorber. Once the NOx adsorber catalyst is
full (once an unacceptable amount of NOx is slipping through the catalyst without storage) the
engine must switch to fuel rich operation in order to regenerate the NOx adsorber.
The engine typically changes to fuel rich operation by increasing the EGR rate, by
throttling the fresh air intake, and by introducing an additional fuel injection event late in the
combustion cycle. The increased EGR rate works to decrease the oxygen content of the intake
air by displacing fresh air that has a high oxygen content with exhaust gases that have a much
lower oxygen content. Intake air throttling further decreases the amount of fresh air in the intake
gases again lowering the amount of oxygen entering the combustion chamber. The combination
of these first two steps serves to lower the oxygen concentration in the combustion chamber,
decreasing the amount of fuel required in order to reach a fuel rich condition. The fuel is
metered then into the combustion chamber in two steps under this mode of operation. The first,
or primary, injection event meters a precise amount of fuel in order to deliver the amount of
m-20
-------
Chapter III: Emissions Standards Feasibility
torque (energy) required by the operator demand (accelerator pedal input). The second injection
event is designed to meter the amount of fuel necessary in order to achieve a net fuel rich
operating condition. That is, the primary plus secondary injection events introduce an excess of
fuel when compared to the amount of oxygen in the combustion chamber. The secondary
injection event occurs very late in the combustion cycle so that no torque is derived from its
introduction. This is necessary so that the switching between the normal lean burn operation and
this periodic fuel rich operation is transparent to the user.
Additional ECM capability will be necessary to monitor the NOx adsorber and determine
when the NOx regeneration events are necessary. This could be done in a variety of ways,
though they fall into two general categories: predictive and reactive. The predictive method
would estimate or measure the NOx flow into the adsorber in conjunction with the predicted
adsorber performance to determine when the adsorber is near capacity. Then, upon entering
optimal engine operating conditions, a NOx regeneration would be performed. This particular
step is similar to an on-board diagnostic (OBD) algorithm waiting for proper conditions to
perform a functionality check. During the NOx regeneration, sensors would determine how
accurately the predictive algorithm performed, and adjust it accordingly. The reactive method is
envisioned to monitor NOx downstream of the NOx adsorber and, if NOx slippage is detected, a
regeneration event would be triggered. This method is dependent on good NOx sensor
technology. This method would also depend on the ability to regenerate under any given engine
operating condition, since the algorithm would be reacting to indications that the adsorber had
reached its NOx storage capacity. In either case, we believe these algorithms are not far removed
from those used today for other purposes. When used in combination with the sophisticated
control systems that will be available, we expect that NOx regeneration events can be seamlessly
integrated into engine operation such that the driver may not be aware that the events are taking
place.
Using this approach of periodic switching between normal lean burn operation and brief
periods of fuel rich operation all accomplished within the combustion chamber of a diesel engine
is one way in which an emission control system for a diesel engine can be optimized to work
with the NOx adsorber catalyst. This approach requires no new engine hardware beyond the air
handling and advanced common rail fuel systems that many advanced diesel engines will have
already applied in order to meet the Phase 1 EGR based NOx standard. For this reason an in-
cylinder approach is likely to appeal to engine manufacturers for product lines where initial
purchase cost is the most important factor in determining engine purchases. A Department of
Energy (DOE) research program has already demonstrated that this approach can work.34
This in-cylinder approach is not without some drawbacks. The high EGR rates and very
low oxygen content of the intake air supply during rich operation can lead to poor combustion
quality, increased fuel consumption and increased PM formation. Since the all of the exhaust
gases must be made rich under this approach, the amount of fuel added from the secondary
HI-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
injection event can be substantial leading to an even greater increase in fuel consumption.
Further the secondary injection event which occurs very late in the combustion cycle has the
potential to lead to dilution of the engine lubricating oil with diesel fuel. This can occur when
the fuel injection spray "over-penetrates" and impinges on the cylinder walls. The fuel on the
cylinder walls is "scraped" into the engine crankcase by the same piston ring technology that is
designed to control oil consumption. Dilution of the lubricating oil can lead to increased engine
wear rates. Fortunately the period of fuel rich operation is typically very brief in relation to the
period of normal lean burn operation. As an example lean burn operation may have a one minute
duration while the fuel rich operation can be as brief as two seconds. The fact that the fuel rich
operation occurs for only brief periods helps to alleviate concerns about this operating mode.
Further, the complete emission control system can be designed to address concerns about the
very brief increase in PM emissions through the use of a CDPF.
(b) With External Control Systems
The in-cylinder approach to optimizing a diesel engine NOx emission control system to
work with a NOx adsorber has several drawbacks which may make it a less desirable solution for
heavy heavy-duty diesel engines which can have an extremely long engine life and for which fuel
economy is a greater concern than initial purchase price. For these applications it would be
desirable to develop a system which could function outside of the engine's combustion system
independent of engine operating mode. This would allow the diesel engine itself to continue to
be designed for maximum durability and minimum fuel consumption while always operating in
an oxygen rich environment as is typical of today's diesel engines. This is precisely what is done
today for NOx adsorber systems applied to stationary power sources.
One approach to accomplish this goal is through the use of a so called "dual-bed" or
"multiple-bed" NOx adsorber catalyst system. Such a system is designed so that the exhaust
flow can be partitioned and routed through two or more catalyst "beds" which operate in parallel.
Multiple-bed NOx adsorber catalysts restrict exhaust flow to part of the catalyst during its
regeneration. By doing so, only a portion of the exhaust flow need be made rich, reducing
dramatically the amount of oxygen needing to be depleted and thus the fuel required to be
injected in order to generate a rich exhaust stream. One simple example of a multiple bed NOx
adsorber is the dual-bed system in Figure in.A-3. In this example, the top half of the adsorption
catalyst system is regenerating under a low exhaust flow condition (exhaust control valve nearly
closed), while the remainder of the exhaust flow is bypassed to a lower half of the system. A
system of this type would have the following characteristics:
• Half of the system would operate with a major flow in an "adsorption mode",
where most of the exhaust is well lean of stoichiometric (A > 1 or »1, typical
HI-22
-------
Chapter III: Emissions Standards Feasibility
diesel exhaust), NO is converted to NO2 over a Pt-catalyst, and stored as a
metallic nitrate within the NOx adsorbent material.1"
The other half of the system would have its exhaust flow restricted to just a small
fraction (~5 percent) of the total flow and would operate in a regeneration mode.
- While the flow is restricted for regeneration, a small quantity of fuel is
sprayed into the regenerating exhaust flow at the beginning of the
regeneration event.
- The fuel is oxidized by the oxygen in the exhaust until sufficient oxygen
is depleted for the stored NOx to be released. This occurs at exhaust
conditions of A, < 1.
- At these conditions, NOx can also be very efficiently reduced to N2 and
O2 over a precious metal catalyst.
At the completion of regeneration, the majority of the flow can then be
reintroduced into the regenerated half of the system by opening the flow control
valve.
Simultaneously, flow is restricted to the other half of the system to allow it to
regenerate.
NOx
Adsorber
NOx
Adsorber
4
Flow-
Partially Closed
Exhaust-flow
Control Valve
Fully Open
Exhaust-flow
Control Valve
Secondary
Fuel Injector
(on)
Exhaust Flow
/ / Secondary
Injector
(off)
Diesel Engine
Figure III.A-3. Schematic Representation of the Operation of a Dual-Bed NOx
Adsorption Catalyst
k A condition of 1 = 1 means that there are precisely the needed quantity of reactants for complete reaction
at equilibrium. 1 < 1 means that there is insufficient oxygen, 1 > 1 means that there is exess oxygen.
m-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The primary advantage of such a system is to significantly reduce fuel consumption
compared to single-bed approaches to NOx adsorber catalysts. Since oxygen must be depleted
from the exhaust during regeneration of the NOx adsorber, depleting oxygen from a minor flow
requires much less fuel than depleting oxygen from the entire exhaust as would be required with
a single-bed NOx adsorber approach. Control of the system is also somewhat less complicated
due to the segregation of the exhaust control external to the engine. This avoids some of the
issues highlighted above with a secondary injection event used for an in-cylinder approach. The
disadvantage is the need for additional hardware and a somewhat more complicated exhaust
system.
Although the schematic shows two separate systems, the diversion of exhaust flow can
occur within a single catalyst housing, and with a single catalyst monolith. Toyota has already
demonstrated flow-switching within a single device with their new combination CDPF/NOx
adsorber.35 There may also be advantages to using more than one partition for the NOx adsorber
system, for example:
Multiple bed NOx adsorbers increase adsorption capacity by allowing more
complete regeneration than is typically possible with a single bed.
• Use of multiple beds allows desulfation of one bed while normal NOx adsorption
and regeneration events occur in other beds.
The NOx adsorber performance can be enhanced by incorporating a catalyzed diesel
particulate filter (CDPF) into the system. A number of synergies exist between NOx adsorber
systems and CDPFs. Both systems rely on conversion of NO to NO2 over a Pt catalyst for part of
their functioning. Partial oxidation reforming of diesel fuel to hydrogen and CO over a Pt-
catalyst has been demonstrated for fuel-cell applications. A similar reaction to reform the fuel
upstream of the NOx adsorber during regeneration would provide a more reactive reductant for
desorption and reduction of NOx. Heavier fuel hydrocarbons are known to inhibit NOx
reduction on the NOx adsorption catalyst since competitive adsorption by hydrocarbons on the
precious metal sites inhibits NOx reduction during adsorber regeneration.36 Partial oxidation of
the secondary fuel injected into the exhaust during regeneration could lead to sooting of the fuel.
Using a CDPF upstream of the NOx adsorber, but downstream of the secondary fuel injection,
allows partial oxidation of the fuel hydrocarbons to occur over the Pt catalyst on the surface of
the CDPF. The wall-flow design of the CDPF efficiently captures any soot formed during partial
oxidation of the fuel injected into the exhaust, preventing any increase in soot emissions. The
partial oxidation reaction over the CDPF is exothermic, which could be used increase the rate of
temperature rise for the NOx adsorber catalyst after cold starts, similar to the use of light-off
catalysts with cascade three-way catalyst systems.37 The schematic in Figure IE. A-4 shows the
integrated dual-bed NOx adsorber and CDPF system developed for testing at EPA-NVFEL,
along with a potential second generation of this type of emission control system having an
HI-24
-------
Chapter III: Emissions Standards Feasibility
additional partition for NOx adsorption/regeneration and further integration of the components.
The use of a DOC for HC and H2S control may be necessary downstream of the NOx adsorbers
partitions. The system tested at NVFEL is described in more detail in a memo to the docket for
this final rule.38
m-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Secondary
Fuel Injector
(on)
CDPF
Partially Closed
Exhaust-flow
Control Valve
Fully Open
Exhaust-flow
Control Valve
Diesel Engine
Exhaust Flow
/ I Secondary
Flow fcT Fuel Injector
(off)
Integrated,
partitioned,
wall-flow
CDPF and
low-temp
NOx
Three flow paths
with independent
injectors and flow
control valves
Partitioned,
flow-through
high-temp
NOx
DOC for HC
and H2S
(not
partitioned)
Air Gaps
Figure III.A-4. A functional schematic representation of the PM and NOx exhaust
emission control system tested at NVFEL shown together with one possible approach
having the same functionality, but with further integration of components
m-26
-------
Chapter III: Emissions Standards Feasibility
A multiple-bed device of this type could be manufactured using a single, wall-flow
monolith within a single housing with an internal valve for flow diversion. One possible
configuration of such a system is shown in Figure IE. A-4 above. Toyota has already
demonstrated a similar concept, catalyst can with internal valving that uses a single wall-flow
monolith. Components washcoated onto the surfaces of the wall-flow monolith of the Toyota
system are used to provide oxidation catalysis for efficient PM regeneration, and also provide
alkaline-metal adsorption sites for NOx storage during lean operation. The internal valving is
currently used by Toyota for flow reversal, and the device was originally configured for the NOx
adsorber function to regenerate as a single bed during full rich operation. A similar device could
be reconfigured to allow diversion of the exhaust gases through alternating portions of the wall-
flow monolith for a dual-bed approach. Integration of NOx and PM control components into this
sort of dual-bed system would:
• allow the low-fuel consumption benefits of a multiple-bed NOx adsorber
approach,
• provide a wall-flow CDPF for partial oxidation of the secondary fuel needed for
regeneration
• reduce the mass of the entire system for improved performance
reduce the size of the system for better integration of the system into applications
with tight packaging constraints (as with some light- and medium-heavy-duty
diesel engine applications)
• reduce the price of the system by allowing the use of a single monolith and
housing instead of four or more separate devices.
iv. How Efficient are Diesel NOx Adsorbers ?
Research into applying the NOx adsorber catalyst to diesel exhaust is only a few years old
but benefits from the larger body of experience with stationary power sources and with lean burn
gasoline systems. In simplest terms the question is how well does the NOx adsorber store NOx
under normal lean burn diesel engine operation, and then how well does the control system
perform the NOx regeneration function. Both of these functions are affected by the temperature
of the exhaust and of the catalyst surface. For this reason efficiency is often discussed as a
function of exhaust temperature under steady-state conditions. This is the approach used in this
section. The potential for both NOx storage and reduction to operate at very high efficiencies
can be seen through careful emission control system design as described below.
(a) At Storing NOx Under Oxygen Rich (fuel lean) Conditions?
The NOx storage function as described in section IH.A.S.b.i., above, consists of oxidation
of NO to NO2 and then storage of the NOx as a metallic nitrate on the catalyst surface. The
effectiveness of the catalyst at accomplishing these tasks is dependent upon exhaust temperature,
HI-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
catalyst temperature, precious metal dispersion, NO storage volume, and transport time (mass
flow rates through the catalyst). Taken as a whole these factors determine how effective a NOx
adsorber based control system can store NOx under lean burn diesel engine operation.
Catalyst and exhaust temperature are important because the rate at which the desirable
chemical reactions occur is a function of the local temperature where the reaction occurs. The
reaction rate for NO to NO2 oxidation and for NOx storage increases with increasing
temperature. Beginning at temperatures as low as 100°C NO oxidation to NO2 can be promoted
across a platinum catalyst at a rate high enough to allow for NOx storage to occur. Below 100°C
the reaction can still occur (as it does in the atmosphere) however the reaction rate is so slow as
to make NOx storage ineffective below this temperature in a mobile source application. At
higher exhaust temperatures, above 400°C, two additional mechanisms affect the ability of the
NOx adsorber to store NOx. First the NO to NO2 reaction products are determined by an
equilibrium reaction which favors NO rather than NO2. That is across the oxidation catalyst, NO
is oxidizing to form NO2 and NO2 is decaying to form NO at a rate which favors a larger fraction
of the gas being NO rather than NO2. As this is an equilibrium reaction when the NO2 is
removed from the gas stream by storage on the catalyst surface, the NOx gases quickly "re-
equilibrate" forming more NO2. This removal of NO2 from the gas stream and the rapid
oxidation of NO to NO2 means that in spite of the NO2 fraction of the NOx gases in the catalyst
being low at elevated conditions (30 percent at 400°C) the storage of NOx can continue to occur
with high efficiencies, near 100 percent.
Unfortunately the other limitation of high temperature operation is not so easily
overcome. The metallic nitrates that are formed on the catalyst surface and that serve to store the
NOx emissions under fuel lean operating conditions can become unstable at elevated
temperatures. That is, the metallic nitrates thermally decompose releasing the stored NOx under
lean operating conditions allowing the NOx to exit the exhaust system "untreated." The
temperature at which the storage metals begin to thermally release the stored NOx emissions
varies dependent upon the storage metal or metals used, the relative ratio of the storage metals,
and the washcoat design. Changes to catalyst formulations can change the upper temperature
threshold for thermal NOx desorption by as much as 100°C.39 Thermal stability is the primary
factor determining the NOx control efficiency of the NOx adsorber at temperatures higher than
400-500°C.
(b) At Reducing NOx Under Fuel Rich Conditions?
The NOx adsorber catalyst releases stored NOx emissions under fuel rich operating
conditions and then reduces the NOx over a three-way catalyst function. While the NOx storage
function determines the NOx control efficiency during lean operation it is the NOx release and
reduction function that determines the NOx control efficiency during NOx regeneration. Since
NOx storage can approach near 100 percent effectiveness for much of the temperature range of
HI-28
-------
Chapter III: Emissions Standards Feasibility
the diesel engine, the NOx reduction function often determines the overall NOx control
efficiency.
NOx release can occur under relatively cool exhaust temperatures even below 200°C for
current NOx adsorber formulations. Unfortunately the three-way NOx reduction function is not
operative at such cool exhaust temperatures. The lowest temperature at which a chemical
reaction is promoted at a defined efficiency (often 50 percent) is referred to as the "light-off
temperature. The 80 percent light-off temperature for the three-way catalytic NOx reduction
function of current NOx adsorbers is between 200°C and 250°C. Therefore, even though NOx
storage and release can occur at cooler temperatures, NOx control is limited under steady-state
conditions to temperatures greater than this light-off temperature.
Under transient operation however, NOx control can be accomplished at temperatures
below this NOx reduction light-off temperature provided that the period of operation at the lower
temperature is preceded by operation at higher temperatures and provided that the low
temperature operation does not continue for an extended period. This NOx control is possible
due to two characteristics of the system specific to transient operation. First, NOx control can be
continued below the light-off temperature because storage can continue below that temperature.
If the exhaust temperature again rises above the NOx reduction light-off temperature before the
NOx adsorber storage function is full the NOx reduction can then precede at high efficiency.
Said another way, if the excursions to very low temperatures are brief enough, NOx storage can
precede under this mode of operation followed by NOx reduction when the exhaust temperatures
are above the light-off temperature. Although this sounds like a limited benefit because NOx
storage volume is limited, in fact it can be significant, because the NOx emission rate from the
engine is low at low temperatures. While the NOx storage rate may be limited such that at high
load conditions the lean NOx storage period would be as short as 30 seconds, at the very low
NOx rates typical of low temperature operation (operation below the NOx reduction light-off
temperature) this storage period can increase dramatically. This is due to the NOx mass flow rate
from the engine changing by several orders of magnitude between idle conditions and full load
conditions. The period of lean NOx storage would be expected to increase in inverse proportion
to the NOx emission rate from the engine. Therefore the period of NOx storage under light load
conditions could likewise be expected to increase by orders of magnitude as well.
Transient operation can further allow for NOx control below the NOx reduction light-off
temperature due to the thermal inertia of the emission control system itself. The thermal inertia
of the emission control system can work to warm the exhaust gases to a local temperature high
enough to promote the NOx reduction reaction even though the inlet exhaust temperatures are
below the light-off temperature for the catalyst. In testing at NVFEL (discussed below in section
IH.A.S.b.v.c) exhaust temperatures were above the NOx reduction light-off temperature for
testing at engine loads as low as 25 percent of full load. In as much as heavy-duty diesel engines
are expected to operate under some load for most operating conditions, the exhaust temperature
HI-29
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
will be expected to be above this threshold on average. Therefore the NOx emission control
system temperature is expected to be above the light-off temperature for almost all operation,
even when the engine exhaust temperature drops below this level due to the thermal inertia of the
exhaust control system.
The combination of these two effects was observed during testing of NOx adsorbers at
NVFEL especially with regards to NOx control under idle conditions. It was observed that when
idle conditions followed loaded operation, for example when cooling the engine down after a
completing an emission test, that the NOx emissions were effectively zero (below background
levels) for extended periods of idle operation (for more than 10 minutes). Additionally it was
discovered that the stored NOx could be released and reduced in this operating mode even
though the exhaust temperatures were well below 250°C provided that the regeneration event
was triggered within the first 10 minutes of idle operation (before the catalyst temperature
decreased significantly). However, if the idle mode was continued for extended periods (longer
than 15 minutes) NOx control eventually diminished. The loss of NOx control at extended idle
conditions appeared to be due to the inability to reduce the stored NOx leading to high NOx
emissions during NOx regeneration cycles.
NOx control efficiency with the NOx adsorber technology under steady-state operating
conditions can be seen to be limited by the light-off temperature threshold of the three-way
catalyst NOx reduction function. Further a mechanism for extending control below this
temperature is described for transient operation and is observed in testing of NOx adsorber based
catalyst systems.
(c) For Overall Diesel NOx Control?
Overall NOx adsorber efficiency reflects the composite effectiveness of the NOx adsorber
in storing, releasing and reducing NOx over repeated lean/rich cycles. As detailed above,
exhaust temperatures play a critical role in determining the relative effectiveness of each of these
catalyst functions. These limits on the individual catalyst functions can explain the observed
overall NOx control efficiency of the NOx adsorber, and can be used to guide future research to
improve overall NOx adsorber efficiency and the design of an integrated NOx emission control
system.
At low exhaust temperatures overall NOx control is limited by the light-off temperature
threshold of the three-way NOx reduction function in the range from 200°C to 250°C. At high
temperatures (above 400° to 500°C) overall NOx control is limited by the thermal stability of the
NOx storage function. For exhaust temperatures between these two extremes NOx control can
occur at virtually 100 percent effectiveness.
HI-30
-------
Chapter III: Emissions Standards Feasibility
The ability of the complete system including the engine and the emission control system
to control NOx emissions consistently well in excess of 90 percent is therefore dependent upon
the careful management of temperatures within the system. Figure IHA-5 provides a pictural
representation of these constraints and indicates how well a diesel engine can match the
capabilities of a NOx adsorber based NOx control system. The figure shows accumulated NOx
emission (grams) over the heavy-duty FTP test for both a light heavy-duty (LFID) and a heavy
heavy-duty (F£HD) engine. The engine-out NOx emissions are shown as the dark bars on the
graphs. The accumulated NOx emissions shown here, divided by the integrated work over the
test cycle gives a NOx emission rate of 4 g/bhp-hr (the 1998 HD emission standard) for each of
these engines. Also shown on the graph as a solid line is the steady-state NOx conversion
efficiency for a NOx adsorber, MECA "B", used in testing at NVFEL (see section A.v.c below
for more details on testing at NVFEL). The line has been annotated to show the constraint under
low temperature operation (three-way catalyst light-off). The white bars on the graph represent
an estimate of the tailpipe NOx emissions that could be realized from the application of the NOx
adsorber based upon the steady-state efficiency curve for adsorber MECA "B". These estimated
tailpipe emissions are highest in the temperature range below 250°C even though the engine out
NOx emissions are the lowest in this region. This is due to the light-off temperature threshold
for the NOx three-way reduction function.
ni-3i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
LHD Diesel Estimated NOx Adsorber Effectiveness over HD FTP
.100
90
- MECA "B" NOx Adsorber (%)
^| Engine Out NOx FTP (4g NOx Engine)
I | Projected FTP Tailpipe NOx
250
300
350
400
450
—«- 0
500
E
ro
D)
O
0_
LL
Q
I
(D
o
E
LLJ
X
O
z
T3
-t-«
(0
O
O
Catalyst Inlet Temperature (°C)
HMD Diesel Estimated NOx Adsorber Effectiveness over HD FTP
100
90
— MECA "B" NOx Adsorber (%)
^f Engine Out NOx FTP (4g NOx Engine)
I | Projected FTP Tailpipe NOx
300
350
400
450
10
500
o
O
CL
Q
I
OJ
o
(/)
c
g
I
LLJ
X
O
jo
^
E
^
o
o
Catalyst Inlet Temperature (°C)
Figure III.A-5. NOx Adsorber Efficiency Characteristics versus Exhaust Temperature
m-32
-------
Chapter III: Emissions Standards Feasibility
Since the conversion efficiencies are based upon steady-state operation it is likely that the
low temperature performance could be better than estimated here due to catalyst's ability to store
the NOx emissions at these low temperatures and then to reduce them when transient operation
raises the exhaust temperatures above the three-way light-off temperature. This assertion
provides one explanation for differences noted between this approximation to the FTP NOx
efficiency for the LHD engine shown in Figure HI. A-5 above and actual NOx adsorber efficiency
demonstrated with this engine in the NVFEL test program. Based upon the figure above (using
the steady-state conversion estimate) the NOx adsorber catalyst should have provided less than
an 84 percent reduction in NOx emissions over the FTP. However testing at NVFEL (detailed in
section in.A.v.c) has already demonstrated a 90 percent reduction in NOx emissions with this
same engine and catalyst pair without significant optimization of the system. Clearly then
steady-state NOx adsorber performance estimates can underestimate the real NOx reductions
realized in transient vehicle operation as typified by the HD FTP.
The tailpipe NOx emissions are the lowest in the range from 250°C to 450°C, even
though this is where the majority of the engine out NOx emissions are created, because of the
high overall NOx reduction efficiency of the NOx adsorber system under these conditions. At
temperatures above 500°C the NOx conversion efficiency of the NOx adsorber can be seen to
decrease. However since exhaust temperatures over the FTP for both of these engines remained
below 450°C this loss of NOx control at high temperatures did not affect the overall NOx
conversion efficiency. As detailed in section HI. A.v.c below, this loss in NOx efficiency at high
temperatures is a more important consideration for the SET test where higher exhaust
temperatures at some test points are possible.
Figure IE. A-5 shows that the temperature window of a current technology NOx adsorber
catalyst is well matched to the exhaust temperature profiles of a light heavy-duty and a heavy
heavy-duty diesel engine operated over the heavy-duty FTP driving cycle. Testing at NVFEL on
the same light heavy-duty engine operated over the SET, shows that even for extended high load
operation, as typified by the 100 percent load test points in the procedure, NOx conversion
efficiencies remained near or above 90 percent (See discussion of the NVFEL test program in
section in.A.S.b.v.c, below).
The discussion above makes it clear that when the engine and NOx adsorber based
emission control system are well matched that NOx reductions can be far in excess of 90 percent.
Conversely it can be inferred that if exhaust temperatures are well in excess of 500°C or well
below 200°C for significant periods of engine operation then NOx control efficiency may be
reduced. Fortunately the temperature window for NOx adsorber and diesel engines are inherently
well matched as shown in Figure ni.A-5. Researchers are known to be developing and testing
new NOx adsorber formulations designed to increase the high temperature stability of the NOx
adsorber.40 The unique characteristics of the NOx adsorber will mean that integrated total
HI-33
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
systems approaches will be needed in order to ensure compliance with the NOx standards under a
wide range of conditions.
v. Progress in NOx Adsorber Development for Diesel Engines
(a) Industry Progress
The rapid development of the NOx adsorber technology is not limited to stationary power
and gasoline applications, but includes markets where low sulfur diesel fuel is already available
or has been mandated to coincide with future emission standards. In Japan, Toyota Motor
Corporation has recently announced that it will begin introducing vehicles using its Diesel
Particulate - NOx Reduction (DPNR) system in 2003. This system uses a NOx adsorber catalyst
applied on the surface of a CDPF, providing greater than 80 percent reductions in both PM and
NOx. This system is being designed to operate with fuel in Japan that will have a 50 ppm sulfur
cap but only with a regulated useful life of 50,000 miles (for heavy heavy-duty diesel engines in
the US, the regulated useful life is 435,000 miles). Toyota notes, however, that the DPNR
system requires fuel with low sulfur content in order to maintain high efficiency and good fuel
economy for a long duration.41 In Europe, both Daimler Chrysler and Volkswagen, driven by a
need to meet stringent Euro IV emission standards, have published results showing how they
would apply the NOx adsorber technology to their diesel-powered passenger cars. Volkswagen
reports that it has already demonstrated NOx emissions of 0.137 g/km (0.22 g/mi), a 71 percent
reduction, on a diesel powered Passat passenger car equipped with a NOx adsorber catalyst.42
Likewise, in the United States, heavy-duty engine manufacturers have begun investigating
the use of NOx adsorber technologies as a more cost effective means to control NOx emissions
when compared to more traditional in-cylinder approaches. For example, Cummins Engine
Company reported at DOE's 1999 Diesel Engine Emissions Reduction workshop, that they had
demonstrated an 80 percent reduction in NOx emissions over the Supplemental Steady State test
using a NOx adsorber catalyst.43 In a separate presentation to members of the oil and engine
industries, Cummins reported using a NOx adsorber catalyst to demonstrate 98 percent NOx
control over the heavy-duty FTP, resulting in NOx emissions of 0.055 g/bhp-hr from an engine
out level of approximately 3 g/bhp-hr.44
(b) DOE'sDECSE Programs
The U.S. Department of Energy (DOE) has funded several test programs at national
laboratories and in partnership with industry to investigate the NOx adsorber technology. Most
of these test programs are part of the Advanced Petroleum Based Fuel (APBF) program of
DOE's Office of Transportation Technology (OTT). These programs are often referred to as the
Diesel Emission Control Sulfur Effects (DECSE) program which is itself one of the APBF
HI-34
-------
Chapter III: Emissions Standards Feasibility
programs. Five reports documenting the DECSE program are available from the DOE OTT
website (www.ott.doe.gov/decse) and were used extensively throughout our analysis.45 46 47 48 49
At Oak Ridge National Laboratory DOE researchers have been working to demonstrate
the application of a NOx adsorber catalyst to a light-duty diesel passenger car. For this testing a
Mercedes A170 diesel vehicle was evaluated on the light-duty chassis dynamometer driving
cycle tests. The original equipment manufacturer (OEM) supplied catalysts were removed from
the car and were replaced with a light-off catalyst, a NOx adsorber catalyst and a synthesis gas
reductant system (bottled gases). The synthesis gas reductant system was designed to simulate
exhaust constituents that could be produced by in-cylinder late cycle injection. Overall, NOx
emissions were reduced by more than 90 percent when operating on three ppm sulfur diesel fuel.
NOx reductions of 89 percent were realized over the USO6 test cycle while 96 percent reductions
were realized over the SC03 test cycle. Subsequent testing of the vehicle revealed that poisoning
of the catalyst with sulfur had substantially reduced the NOx adsorber performance of the car in a
little as 600 miles of driving on 150 ppm sulfur fuel.50
The researchers concluded that NOx adsorbers show promise for enabling significant
reductions in diesel NOx emissions based upon their demonstrated FTP and US06 emission
results using a synthesis gas injection system to simulate late cycle, in-cylinder injection of diesel
fuel. They further concluded that sulfur loading equivalent to 3,000 miles of operation on 30
ppm sulfur fuel caused a marked decrease in NOx conversion and that commercial use of the
NOx adsorber will require effective desulfation, sulfur traps, or another solution to the sulfur
poisoning problem.51
In the DECSE program, an advanced diesel engine equipped with common rail fuel
injection and exhaust gas recirculation (EGR) was combined with a NOx adsorber catalyst to
control NOx emissions. The system used an approach similar to the in-cylinder control approach
described in section in.A.3.b.iii.a, above. Rich regeneration conditions are created for the NOx
adsorber catalyst regeneration through increased EGR rates and a secondary injection event
designed to occur late enough in the engine cycle so as not to change engine torque output.
Using this approach, the DECSE program has shown NOx conversion efficiencies exceeding 90
percent over a catalyst inlet operating temperature window of 300°C to 450°C. This performance
level was achieved while staying within the four percent fuel economy penalty target defined for
regeneration calibration.52
(c) NVFEL' s NOx Adsorber Evaluation Program
As part of an effort to evaluate the rapidly developing state of this technology, the
Manufacturers of Emission Control Association (MECA) provided four different NOx adsorber
catalyst formulations to EPA for evaluation. Testing of these catalysts at the National Vehicle
and Fuel Emission Laboratory (NVFEL) revealed that all four formulations were capable of
HI-35
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
reducing NOx emissions by more than 90 percent over the broad range of operation in the SET
procedure (sometimes called the EURO in test). At operating conditions representative of "road-
load" operation for a heavy duty on-highway truck, the catalysts showed NOx reductions as high
as 99 percent resulting in NOx emissions well below 0.1 g/bhp-hr from an engine out level of
nearly 5 g/bhp-hr. Testing on the FTP has shown similarly good results, with hot start FTP NOx
emissions reduced by more than 90 percent. These results demonstrate that significant NOx
reductions are possible over a broad range of operating conditions with current NOx adsorber
technology, as typified by the FTP and the SET procedures.
The test program at NVFEL can be divided into phases. The first phase began with an
adsorber screening process using a single leg of the planned dual leg system. The goals of this
screening process, a description of the test approach, and the results are described below. The
next phase of the test program consisted of testing the dual leg system using a NOx adsorber
chosen during the first phase in each of two legs.
Testing Goals — Single Leg NOx Adsorber System
The goal of the NOx adsorber screening process was to evaluate available NOx adsorber
formulations from different manufacturers with the objective of choosing an adsorber with 90
percent or better NOx reduction for continued evaluation. To this end, four different adsorber
formulations were provided from three different suppliers. Since this was a screening process
and since a large number of each adsorber formulation would be required for a full dual leg
system, it was decided to run half of a dual leg system (a single leg system) and mathematically
correct the emissions and fuel economy impact to reflect a full dual leg system. The trade-off
was that the single leg system would only be able to run steady state modes, as the emissions
could not be corrected over a transient cycle. The configuration used for this test was similar to
that shown in Figure in.A-4, but with a catalyst installed on only one side of the system.
Test Approach — Single Leg NOx Adsorber System
The single leg system consisted of an exhaust brake, a fuel injector, CDPF, and a NOx
adsorber in one test leg. The other leg, the "bypass leg," consisted of an exhaust brake that
opened when the test leg brake was closed; this vented the remainder of the exhaust out of the
test cell. Under this set up, the test leg, i.e., the leg with the adsorber, was directed into the
dilution tunnel where the emissions were measured and then compensated to account for
emissions from the bypass leg. The restriction in the bypass leg was set to duplicate the
backpressure of the test leg so that, while bypassing the test leg to conduct a NOx regeneration,
the backpressure of the bypass leg simulated the presence of a NOx adsorber system. A clean-up
diesel oxidation catalyst (DOC) downstream of the NOx adsorber was not used for this testing.
HI-36
-------
Chapter III: Emissions Standards Feasibility
The measured emissions had to be adjusted to account for the lack of any NOx adsorber
in the bypass leg. For this correction, it was assumed that the bypass leg's missing (virtual)
adsorber would adsorb only while the actual leg was regenerating. It was also assumed the
virtual adsorber would have regeneration fuel requirements in proportion to its adsorbing time.
The emissions performance of the virtual adsorber was assumed to be the same as the
performance of the actual adsorber. With these assumptions, the gaseous emissions could be
adjusted as detailed in a memo to the docket describing this test program further.53
Test Results — Single Leg NOx Adsorber System
Two sets of steady-state modes were run with each adsorber formulation. These modes
consisted of the SET modes and the AVL 8 mode composite FTP prediction.1 The modes are
illustrated in Figure IE. A-6 and are numbered sequentially one through 20 to include both the
eight AVL modes and the 13 SET modes (the idle mode is repeated in both tests). The mode
numbers shown in the figure are denoted as "EPA" modes in the subsequent tables to
differentiate between the AVL and SET modes which have duplicate mode numbers. The NTE
zones are also shown in Figure in.A-6 to show that these two sets of modes give comprehensive
coverage of the NTE zone. The modes were run with varying levels of automation, with the
general strategy being to inject sufficient fuel during regeneration to obtain a lambda at or
slightly fuel rich of stoichiometric (A, < 1). The NOx regenerations were then timed to achieve
the desired NOx reduction performance. The adsorber formulations were identified as A, B, D,
and E. Prior to testing, each set of adsorbers were aged at 2500 rpm, 150 Ib-ft for 40 minutes,
then 2500 rpm full load for 20 minutes, repeated for a total of 10 hours.
1 The AVL 8 mode test procedure is a steady-state test procedure developed by Anstalt fur
Verbrennungskraftmaschinen, Prof. Dr. Hans List (or Institute for Internal Combustion
Engines) to approximate the transient FTP.
HI-37
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
700
600
500
=- 400
300
200
100
800
1000
1200
1400
1600
1800
Speed (rpm)
2000
2200
2400
2600
2800
Figure III.A-6. Modal Definitions
(the mode numbers here correspond to the "EPA" modes given in the subsequent tables)
The SET and AVL Composite emission results, along with the NOx reduction
performance vs. adsorber inlet temperature, are shown in Figures IHA-7 through ni.A-10 for
each of the tested NOx adsorber formulations. The SET composites for all four adsorber
formulations had NOx reductions in excess of 90 percent with under a three percent FE impact.
The HC emissions varied most widely, most likely due to differences in regeneration strategies,
and to some extent, adsorber formulation. The HC emissions with the exception of adsorber "A"
were very good, less than 0.1 g/hp-hr over the SET and less than 0.2 g/hp-hr over the AVL
composite. It should be noted that no DOC was used to clean up the HC emissions.
Another point to note is that the EPA mode 1 data for each composite is the same. This is
because EPA mode 1, low idle, is too cold for effective steady-state regeneration, but efficient
NOx adsorption can occur for extended periods of time. For either of these composite tests, a
regeneration would not be needed under such conditions and, therefore, the idle mode was
HI-38
-------
Chapter III: Emissions Standards Feasibility
considered to have no FE impact (See discussion in section HI.A.S.b.iv of this chapter). EPA
mode 1 has very little impact on either composite in any case because of the low power and
emission rate. EPA mode 2 also had very low steady-state temperatures, and the difficulty
regenerating at this mode can be seen in the HC and FE impacts. But, like EPA mode 1, EPA
mode 2 would adsorb for extended periods of time without need for regeneration.
The AVL composite showed greater differences between the adsorber formulations than
the SET. Three of the adsorbers achieved greater than 90 percent NOx reduction over the AVL
composites with the other adsorber at 84 percent NOx reduction. The greater spread in NOx
reduction performance was, in part, due to this composite's emphasis on EPA mode 8, which was
at the upper end of the NOx reduction efficiency temperature window. Adsorber E had an EPA
mode 8 NOx reduction of 66 percent, and the NOx reduction efficiency vs. inlet temperature
graph clearly shows that this formulation's performance falls off quickly above 450°C. In
contrast, the other formulations do not show such an early, steep loss in performance. The FE
impacts vary more widely also, partly due to the test engineers' regeneration strategies,
particularly with the low temperature modes, and to the general inability to regenerate at very low
temperature modes at steady-state. It should be noted that none of the regeneration strategies
here can be considered fully optimized, as they reflect the product of trial and error
experimentation by the test engineers. With further testing and understanding of the technology a
more systematic means for optimization should be possible. In spite of the trial and error
approach the results shown here are quite promising.
The AVL composite was developed as a steady state engine-out emission prediction of
the HDDE transient cycle. With exhaust emission control devices, it loses some of its accuracy
because of the inability of the emission control devices to be regenerated at the low temperature
modes (EPA modes 1, 2, 5). In real world conditions, the HDDE does not come to steady-state
temperatures at any of these modes, and the adsorber temperatures will be higher at EPA modes
1, 2, and 5 than the stabilized steady-state values used for this modal testing. Consequently, the
actual HDDE transient cycle performance is expected to be much better than the composites
would suggest (See discussion of transient testing below).
Based on the composite data and the temperature performance charts, amongst other
factors, adsorber formulation B was chosen for further dual leg performance work. Both
composites for this formulation were well above 90 percent. The NOx vs. temperature graph,
Figure in.A-8, also shows that this formulation was a very good match for this engine.
HI-39
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.0
4.6
4.7
5.0
5.0
5.0
5.5
4.0
5.0
4.0
4.8
5.0
4.8
Composite Results 4.6
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
461
357
411
384
427
287
498
293
515
282
404
357
BSNOx
(g/hp-hr)
0.16
0.11
0.07
0.06
0.13
0.24
0.25
0.89
0.14
0.48
0.42
0.08
0.14
NOx Red
100%
98%
98%
99%
97%
95%
95%
78%
97%
88%
91%
98%
97%
HC*
(g/hp-hr)
0.00
0.92
1.02
1.35
0.11
0.81
1.39
0.36
1.88
1.12
0.68
0.62
0.70
FE Impact
*
0.0%
2.4%
2.0%
2.6%
1 .3%
1 .6%
3.3%
1 .9%
4.1%
3.8%
3.5%
3.0%
2.8%
0.31 93% 0.91 * 2.6% *
Adsorber
Inlet T
(C)
144
172
346
430
286
325
386
505
BSNOx
(g/hp-hr)
0.16
0.83
0.36
0.20
0.37
0.08
0.10
1.06
NOx Red
100%
91%
96%
97%
93%
98%
98%
74%
HC*
(g/hp-hr)
0.00
0.75
1.10
2.16
4.93
2.30
2.38
0.03
FE Impact
*
0.0%
7.7%
3.1%
3.0%
3.6%
3.6%
3.1%
1 .9%
0.44 91% 1.69* 2.9%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens
than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
100%
B-
-------
Chapter III: Emissions Standards Feasibility
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.0
4.6
4.7
5.0
5.0
5.0
5.5
4.0
5.0
4.0
4.8
5.0
4.8
Composite Results 4.6
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
498
366
446
375
420
296
524
293
537
280
426
357
BSNOx
(g/hp-hr)
0.16
0.18
0.07
0.14
0.06
0.07
0.18
0.46
0.36
0.56
0.29
0.24
0.11
NOx Red
100%
96%
98%
97%
99%
98%
97%
89%
93%
86%
94%
95%
98%
HC*
(g/hp-hr)
0.00
0.01
0.04
0.01
0.08
0.10
0.10
0.01
0.05
0.04
0.03
0.04
0.02
FE Impact
*
0.0%
1.2%
0.5%
1 .5%
0.7%
2.3%
0.3%
3.2%
0.4%
4.3%
0.4%
4.3%
0.9%
0.27 94% 0.03* 2.2%*
Adsorber
Inlet T
(C)
144
162
355
446
263
346
403
544
BSNOx
(g/hp-hr)
0.16
0.56
0.30
0.09
0.66
0.11
0.05
0.73
NOx Red
100%
94%
96%
98%
88%
98%
99%
82%
HC*
(g/hp-hr)
0.00
2.11
0.16
0.23
0.25
0.03
0.02
0.35
FE Impact
*
0.0%
1 .8%
0.3%
0.9%
1 .6%
0.4%
1 .4%
4.0%
0.33 93% 0.19* 2%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens
than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
100%
— 80%
60%
C
.o
"u
3
1!
OL
X
O
40%
20%
0%
200 250 300 350 400 450
Adsorber Inlet Temperature (C)
500
550
Figure III.A-8. SET & AVL Composites, and Temperature vs.
NOx Chart for Adsorber B
ni-4i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.00
4.60
4.70
5.00
5.00
5.00
5.50
4.00
5.00
4.00
4.80
5.00
4.80
Adsorber
Inlet T
(C)
144
451
356
400
377
431
305
501
303
489
278
391
330
BSNOx
(g/hp-hr)
0.16
0.18
0.14
0.09
0.07
0.11
0.23
0.16
0.15
0.93
0.57
0.12
0.21
NOx Red
100%
96%
97%
98%
99%
98%
96%
96%
97%
93%
88%
98%
96%
HC*
(g/hp-hr)
0.00
0.07
0.15
0.05
0.01
0.02
0.14
0.04
0.14
0.09
0.18
0.10
0.09
FE Impact
*
0.0%
1 .3%
1 .7%
1 .6%
1.2%
1 .6%
2.3%
2.1%
3.1%
1 .7%
3.5%
1 .8%
2.9%
Composite Results
4.6 I
0.28
94%
0.08'
1.9%*
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
162
359
427
273
301
363
493
BSNOx
(g/hp-hr)
0.16
0.56
0.08
0.14
1.25
0.52
0.66
0.31
NOx Red
100%
94%
99%
98%
77%
89%
87%
92%
HC*
(g/hp-hr)
0.00
2.11
0.30
0.19
0.26
0.13
0.04
0.08
FE Impact
*
0.0%
1 .8%
3.1%
1 .7%
6.4%
1 .9%
1 .4%
1 .6%
0.51 90% 0.14* 1.9%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens
than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
100%
80%
.
It
UJ
C
o
1!
OL
X
O
60%
40%
20%
0%
200 250 300 350 400 450
Adsorber Inlet Temperature (C)
500
550
Figure III.A-9. SET & AVL Composites, and Temperature vs.
NOx Chart for Adsorber D
m-42
-------
Chapter III: Emissions Standards Feasibility
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.00
4.60
4.70
5.00
5.00
5.00
5.50
4.00
5.00
4.00
4.80
5.00
4.80
Composite Results 4.6
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
455
343
442
377
419
412
392
294
492
388
391
327
BSNOx
(g/hp-hr)
0.16
0.47
0.07
0.36
0.08
0.29
0.14
0.05
0.09
0.95
0.11
0.12
0.22
NOx Red
100%
89%
98%
93%
98%
94%
98%
99%
98%
76%
98%
98%
95%
HC*
(g/hp-hr)
0.00
0.02
0.05
0.07
0.01
0.03
0.05
0.02
0.26
0.03
0.03
0.10
0.02
FE Impact
*
0.0%
2.1%
0.9%
9.0%
1 .5%
1 .6%
1 .7%
2.1%
4.4%
2.0%
2.4%
1 .8%**
1 .4%
** Md 19 data from Adsorber D
0.33 93% 0.05* 2.9%*
Adsorber
Inlet T
(C)
144
166
339
449
256
313
372
508
BSNOx
(g/hp-hr)
0.16
7.39
0.09
0.65
1.36
0.35
0.12
1.39
NOx Red
100%
16%
99%
89%
75%
92%
97%
66%
HC*
(g/hp-hr)
0.00
1.02
0.05
0.01
0.91
0.21
0.10
0.04
FE Impact
*
0.0%
71 .9%
2.3%
2.1%
15.8%
5.6%
2.6%
3.3%
0.80 84% 0.16* 5.4%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens
than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
100%
80%
C
.o
"u
3
1!
OL
X
O
60%
40%
20%
0%
200 250 300 350 400 450
Adsorber Inlet Temperature (C)
500
550
Figure III.A-10. SET & AVL Composites, and Temperature
vs. NOx Chart for Adsorber E
m-43
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Testing Goals — Dual Leg NOx Adsorber System
After completing the screening process and selecting NOx adsorber "B," the dual leg
system was developed. The testing goal for the dual leg system was to demonstrate that NOx
adsorbers are capable of 90 percent NOx reductions over the HD FTP and SET tests with a
current production engine. Once the capability of the devices to achieve the NOx reductions is
established, testing will be done to evaluate the impact of higher fuel sulfur levels (15 ppm) and
aging effects on adsorber performance.
Testing Approach — Dual Leg NOx Adsorber System
The steady state SET testing was conducted in a manner similar to that used in the
screening process described above. The modes were run with varying levels of automation, with
the general strategy being to inject sufficient fuel during regeneration to obtain a lambda at or
slightly fuel rich of stoichiometric (A, < 1). The NOx regenerations were then timed to achieve
the targeted 90 percent NOx reduction. Each mode was run twice by different calibrators to
investigate the adsorber's emission and fuel usage sensitivity to different combinations of
regeneration frequency and fuel injection rates. The regeneration control and optimization
strategies are described in more detail in a memo to the docket for this rule.54 The engine and the
integrated dual-bed NOx adsorber/CDPF system are also described in more detail in section
m.A.S.b.iii.
The transient FIDDE FTP regeneration control was accomplished using a time-based
regeneration schedule. This control regenerated on a prescribed schedule of time and fuel
quantities so that regenerations occurred at predetermined engine conditions during the transient
cycle. This control represents a simplified control strategy that was used due to the lack of time
to develop a true, non-time based control algorithm.
The transient FIDDE FTP results presented here are for hot-start cycles only. The
adsorber system was not optimized for cold start performance and would not provide a
meaningful assessment of adsorber warmup performance. In order to better simulate the "cold-
soak-hot" procedure called for in the HDDE FTP, a preconditioning mode was chosen to provide
adsorber temperatures at the start of the "hot" cycle that would be similar to those found
following the "cold-soak" portion of the test. The mode chosen was EPA mode 10 (1947 rpm,
328 Ib-ft) which resulted in adsorber inlet temperatures (i.e., at the outlet of the CDPF) at the
start of the hot cycle of about 280°C. Another purpose for the preconditioning was to ensure the
adsorbers were in the same condition at the start of each test. Given that our regeneration control
system did not automatically take into account the starting condition of the NOx adsorbers, this
preconditioning was necessary to provide repeatable transient test results. We expect more
realistic control systems would not need such preconditioning. For this preconditioning, the
HI-44
-------
Chapter III: Emissions Standards Feasibility
adsorbers were regenerated frequently in order to ensure a consistent, relatively clean adsorber
state at the start of the transient cycle. The preconditioning consisted of 30 seconds of
regeneration followed by 30 seconds of adsorption as shown below in Figure ni.A-11.
700 i
600
500
•a
Q.
3 400
o
2 300
Q.
X
200
100
- Adsorber Leg '
Adsorber Leg 0
Adsorber Leg 1
Adsorber Leg 0
T 0.7
*tjw/\rvw-vvT^
1— M~ 0.6
Nox Adsorber In (raw)
- NOx Adsorber Out (raw) -
i
HC Analyzer (dilute)
yv
Regeneration Fuel
0.5
0.4 E
ro
OL
+ 03
0.2
- 0.1
160
180
200
220
240
260
280
Time (s)
Figure IILA-11. NOx Adsorber FTP Preconditioning Cycle used in NVFEL Testing
Test Results — Dual Leg NOx Adsorber System
Supplemental Emission Test (SET) Results
The SET is made up of the 13 Euro in modes. Several modes were run twice by different
engineers, and the best calibration was chosen for the SET composite. Table in. A-2 shows the
SET composite test results. These data show that 90 percent NOx reductions were possible over
the SET composite, with a modal NOx reduction range from 77 percent to 98 percent. The
adsorber NOx and HC reduction performance varied primarily as a function of exhaust
temperature. Modes with high temperatures (>500°C) tended to have lower NOx reduction
performance with this adsorber formulation. High temperature EPA modes 9, 15 and 17 had
HI-45
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
lower performance due to the conflict between the high NOx mass flow rate from the engine, the
reduced storage capacity of the NOx adsorber due to ongoing sulfur poisoning and the reduction
in storage capacity near 500°C observed for this adsorber. The combination of these three factors
resulted in higher NOx slippage during adsorption.
Table III.A-2. SET Composite Test Results with the Dual Leg NOx Adsorber System
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.0
4.6
4.7
5.0
5.0
5.0
5.5
4.0
5.0
4.0
4.8
5.0
4.8
Composite Results 4.6
Adsorber
Inlet T
(C)
144
493
373
444
404
456
304
521
343
510
283
409
361
BSNOx
(g/hp-hr)
0.16
0.71
0.09
0.17
0.07
0.51
0.28
0.56
0.34
0.91
0.22
0.41
0.12
NOx Red
100%
84%
98%
96%
98%
90%
95%
86%
93%
77%
95%
92%
98%
HC
(g/hp-hr)
0.00
0.16
0.28
0.24
0.14
0.11
0.11
0.31
0.09
0.54
0.56
0.13
0.10
FE Impacl
0.0%
1.8%
2.3%
2.8%
2.6%
1.9%
2.5%
2.2%
1.9%
1.8%
3.0%
1.8%
2.0%
0.45 90% 0.27 2.1%
Conversely, low exhaust temperatures did not seem to impact the NOx storage capability
of the adsorber. The ability to store at low temperatures is reflected in the low idle, EPA mode 1
performance. Nearly 100 percent NOx reduction could be realized for several minutes until the
adsorber's storage sites filled up with NOx, particularly when coming down to low idle from
higher temperature modes. NOx regeneration at low idle after the adsorber had cooled to a low
steady-state temperature was not possible with this adsorber in our testing.
The FE impact was defined as the percent increase in fuel consumption caused by the
adsorber regeneration fuel, or the mass of fuel used for regeneration, divided by the mass of fuel
consumed by the engine during one regeneration and adsorption cycle. The FE impact varied
from 1.8 to 3.0 percent depending on the mode. The low temperature modes (EPA modes 14,
and 18) tended to have higher FE impacts. This was caused by the combination of low engine
fuel consumption and low HC utilization efficiency seen with this catalyst at these temperatures.
Given the short time spent calibrating the regeneration events, and the relatively early stage of
catalyst development, we anticipate significant improvements in regeneration strategies will be
possible.
Finally, at the time these SET emission tests were conducted, the NOx adsorber system
had accumulated 172 hours of operation. During that time, 530 gallons of five ppm equivalent
HI-46
-------
Chapter III: Emissions Standards Feasibility
(some three ppm and some six ppm) sulfur fuel was consumed by the engine and the NOx
adsorber regenerations. For a light heavy-duty diesel truck averaging 20 miles per gallon of fuel,
the 530 gallons of fuel consumed here would be equivalent to more than 10,000 miles of driving.
No desulfations were performed during any of the testing, though it is expected that a NOx
adsorber system in-use would have been desulfated at least twice, and more likely three times,
during this amount of driving. Consequently, the adsorbers' performance would likely have been
even better had they been desulfated as anticipated.
HOPE Transient FTP Test Results
The transient cycle data were taken with a DOC downstream of the NOx adsorbers and
insulation on the exhaust from the engine to the CDPF. Table ILL A-3 shows a timed regeneration
schedule that was developed to switch between adsorbers, and to control when and how much
fuel was injected for NOx regeneration. The first column in the table, "Time," represents the
time from the start of the FTP and is used as a trigger to switch which leg is adsorbing. That is,
at 30 seconds, and 60 seconds, etc., the adsorbing leg is bypassed and the previously bypassed leg
starts adsorbing. The second column, "Fuel Rate," is the rate that fuel is injected into the
bypassed (regenerating) adsorber. The final column is the time for which the fuel is injected at
the rate specified in the previous column.
Figure III.A-3. Timed Regeneration Schedule for Switching between NOx Adsorber Legs
Time
(s)
20
40
60
80
120
180
240
260
320
410
430
470
500
530
580
630
655
680
Fuel Rate
(Ib/min)
0
0
0
0
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
Injection Time
(s)
0
0
0
0
2.0
2.0
1.4
2.0
2.0
1.0
1.5
1.5
1.2
1.5
1.8
1.0
1.0
1.0
Time
(s)
705
725
750
775
800
825
850
875
900
945
975
995
1070
1145
1165
1185
1200
Fuel Rate
(Ib/min)
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0
0
0.25
0.25
0
0
0
0
Injection Time
(s)
1.0
1.0
1.0
1.5
1.0
1.0
1.3
1.2
1.3
0
0
2.0
2.0
0
0
0
0
m-47
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
This regeneration strategy resulted in an average, over three HDDE FTPs, of 0.25 g/hp-hr
NOx, 0.002 g/hp-hr PM and virtually no CO (See Table ffl.A-4). These represent greater than
90 percent reductions from the engine out emission levels. HC emissions decreased slightly
compared to the baseline results. The relatively small HC emission increase was the result of HC
slippage during NOx regeneration. Two factors contributed to the HC slippage. One factor was
the relatively low HC oxidation efficiency of the DOC used downstream of the adsorbers. Back-
to-back testing with a raw gas analyzer at several steady state modes revealed that the lightly
catalyzed DOC (-10 g/ft3 Pt) had an oxidation efficiency of less then 60 percent, where more
effective DOCs are capable of 90 percent HC reductions. The second factor was that more fuel
was injected than was absolutely necessary to release and reduce the stored NOx. The excess HC
then contributed to HC emissions and FE Impact. Determining the best strategy for injecting the
fuel so that it is most efficiently utilized will be important future work.
Table III.A-4. HDDE FTP Emissions from NVFEL Test Program
Emission
NOx (g/hp-hr)
HC (g/hp-hr)
CO (g/hp-hr)
PM (g/hp-hr)
FE Impact (%)
Run# 1
0.26
0.28
0.00
0.003
2.4
Run #2
0.25
0.30
0.00
0.002
2.3
Run #3
0.40
0.19
0.04
0.002
2.3
Average
0.25
0.28
0.00
0.002
2.3
Engine Out
3.66
0.29
1.46
0.089
-
At the time these FTP emission tests were conducted, the NOx adsorber system had
accumulated 190 hours of operation. During that time, 653 gallons of five ppm equivalent (some
three ppm and some six ppm) sulfur fuel was consumed by the engine and the NOx adsorber
regenerations. For a light heavy-duty diesel truck averaging 20 miles per gallon of fuel, the 653
gallons of fuel consumed here would be equivalent to more than 13,000 miles of driving. No
desulfations were performed during any of the testing, though it is expected that a NOx adsorber
system in-use would have been desulfated at least twice, and more likely three or four times,
during this amount of driving. Consequently, the adsorbers' performance would likely have been
even better had they been desulfated as anticipated.
HI-48
-------
Chapter III: Emissions Standards Feasibility
vi. Can a NOx Adsorber Equipped Diesel Engine Meet the NOx Standards?
(a) The FTP Standard
As discussed in section HI.A.S.b.v.c, above, we have demonstrated in our laboratory that
a NOx adsorber can produce greater than 90 percent reduction in NOx emissions over the hot-
start HDDE transient FTP. The results of this test program lead us to believe NOx adsorbers will
be capable of meeting the Phase 2 FTP NOx emission standard of 0.20 g/bhp-hr. The test
program discussed under section IH.A.S.b.v.c utilized a non-EGR equipped engine certified to
the 1999 FID standards (i.e., the 4.0 g/bhp-hr NOx standard). As discussed in more detail in the
test report in the docket which documents that test program, the regeneration strategy we used to
produce the greater than 90 percent reduction had significant room for improvement.55 For
example, the raw data collected during the FTP runs indicated numerous regeneration events
which were either unnecessary (because the adsorber bed was not slipping NOx) or regeneration
events which released NOx but did not reduce the released NOx because the adsorber did not
achieve a locally rich condition. We conclude from this initial test program that NOx reductions
greater than 90 percent are achievable.
As mentioned, the data presented in Table in.A-4 is from an HDDE certified to the 4.0
g/bhp-hr NOx FTP standard. If the engine-out levels are reduced to achieve the 2004 FTP
standards through the use of cooled EGR, our understanding of how the NOx adsorber
technology works leads us to believe NOx adsorbers would also be capable of achieving 90
percent or greater emission reduction on a 2004 technology engine. Therefore, a NOx adsorber
used on an HDDE certified to the Phase 1 FTP standard of 2.5 g/bhp-hr NMHC+NOx would
enable a HDDE to achieve the Phase 2 FTP NOx standard. In addition, as discussed in section
in.A.S.b.v.a of this chapter, one HDDE manufacturer has also demonstrated greater than a 98
percent NOx reduction over the HD FTP using a NOx adsorber.56
The results discussed above have all been demonstrated over the hot-start portion of the
HDDE FTP, but the HD FTP also includes a cold-start test. A complete HDDE FTP involves
three test sequences. First, the 20 minute duty-cycle test is run with the engine at the same
ambient temperature as the test cell (between 68°F and 86°F). This can be achieved with a long
soak period, or a forced engine cool-down. Second, following the cold-start run the engine
undergoes a hot-soak which lasts 10 minutes. Finally, the 20 minute duty-cycle test is run a
second time. The HDDE FTP emission level for the engine is determined by weighting the cold-
start emissions by 1/7 (-14 percent), and weighting the hot-start emission results by 6/7 (-86
percent). Historically, for a HDDE not equipped with an exhaust emission control device, the
cold-start and hot-start emissions from a HDDE have been nearly identical. However, with the
application of exhaust emission control devices, such as a NOx adsorber, the cold-start test will
become a design challenge for diesel manufacturers, just as it has been a design challenge for
light-duty gasoline vehicle manufacturers for more than 20 years. As discussed above, NOx
HI-49
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
adsorbers do have optimal temperature operating windows, and thus will represent a design
challenge.
Manufacturers have a number of tools available to them to overcome this challenge:
• The volume, shape, and substrate material have a significant effect on the warm-
up time of a NOx adsorber (just as they do for a light-duty three-way catalysts).
Manufactures will optimize the make-up of the adsorber for best light-off
characteristics, such as the thin-walled ceramic monolith catalysts typical of
modern low emission light-duty gasoline applications.
• The packaging of the exhaust emission control devices, including the use of
insulating material and air-gap exhaust systems, will also decrease light-off time,
and we expect manufacturers to explore those opportunities.
• The location of the adsorber, with respect to it's proximity to the exhaust
manifold, will have a significant impact on the light-off characteristics.
As discussed in more detail in section HI.A.S.b.iv.a, NOx adsorbers have the
ability to store NOx at temperatures much less than the three-way catalyst
function temperature operating window, on the order of 100°C. This is unlike the
performance of light-duty gasoline catalysts, and it would allow the NOx adsorber
to store NOx for some period of time prior to the light-off time of the three-way
function of its catalyst, resulting in an overall lower effective temperature for the
device.
These first four tools available to manufacturers all deal with system design opportunities
to improve the cold-start performance of the NOx adsorber system. In addition, manufactures
have a number of active tools which can be used to enhance the cold-start performance of the
system. These include the use of engine start-up routines which have a primary purpose of
adding heat to the exhaust to enhance NOx adsorber light-off For example:
retarded injection timing;
intake air throttling;
post-injection addition of fuel; or
or increasing back-pressure with an exhaust brake or a VGT system.
ni-so
-------
Chapter III: Emissions Standards Feasibility
We anticipate manufacturers will explore all of these tools in order to choose the best
combination necessary to minimize light-off time and improve the cold-start FTP performance.
Considering that the cold-start test is weighted approximately 14 percent, a manufacturer could
achieve the composite FTP standard with considerably less than 90 percent reduction over the
cold-start test, provided the hot-start test achieves greater than 90 percent reduction. Considering
the tools available to manufacturers, and the several years of lead time, we conclude the cold-
start FTP challenges will not be a barrier to the achievement of the FTP NOx standard
established in this rule.
Based on the data and discussion provided in this section, we conclude that the Phase 2
FTP NOx standard is technologically feasible.
(b) The Supplemental Emission Test (SET) NOx Standard
The SET NOx requirements established in the Phase 1 rule for 2007 sets an NMHC+NOx
standard equal to 1.0 x the Phase 1 FTP standard of 2.5 g/bhp-hr NMHC+NOx. Based on current
certification data for HDDE's which indicate HC values on the order of 0.2 to 0.3 g/bhp-hr, we
anticipate that Phase 1 technology engines will achieve the SET standard with NOx emissions on
the order of 2.2 to 2.3 g/bhp-hr. The SET standard promulgated with this Phase 2 rule sets a
NOx standard of 1.0 x the Phase 2 FTP NOx standard of 0.20 g/bhp-hr. This requires a NOx
reduction on the order of 90 percent or more from Phase 1 technology engines, which we have
demonstrated is achievable with NOx adsorber technology, as discussed below.
Section ni.A.3.b.iv.c discusses the NOx adsorber NOx reduction efficiencies as a
function of exhaust gas temperature. The majority of these data show emission reductions of
greater than 90 percent are achievable across the range of exhaust gas temperatures typical of a
HDDE during the SET test procedure.
Section ni.A.3.b.v.c contains SET emissions data from four different NOx adsorbers
using a single-bed exhaust configuration tested at NVFEL. This test data shows NOx reductions
for the SET between 93 and 94 percent. Section in.A.3.b.v.c also contains SET test data
collected using a dual-bed exhaust configuration which achieved a 90 percent NOx reduction.
Based on the information presented in this Chapter, and summarized above, we conclude
that the SET NOx standard will be technologically feasible by model year 2007.
(c) The Not-to-Exceed NOx Standard
Under the Phase 1HDDE rule, NTE emission requirements for NMHC+NOx specify the
NTE standard as 1.25 x FTP standard. The Phase 1 FTP standard for NMHC+NOx is 2.5 g/bhp-
hr, therefore the Phase 1 NMHC+NOx NTE standard is 3.1 g/bhp-hr. As discussed in the Phase
ni-5i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1 final rule, we would expect the break-down between NMHC and NOx emissions for the Phase
1 NTE standard to mostly NOx emissions, on the order of 3.0 g/bhp-hr NOx, with the remainder
being NMHC. In this rule, we have promulgated the Phase 2 engine NOx NTE standard as 1.5 x
FTP standard, i.e., 1.5 x 0.20 g/bhp-hr, which is 0.30 g/bhp-hr NOx. Therefore, a 90 percent
reduction in NOx emissions is necessary from Phase 1 engines in order to achieve the Phase 2
NTE NOx standard in this final rule. As discussed below, this 90 percent reduction is
technologically feasible by model year 2007 across the range of engine operating conditions and
ambient conditions subject to the NTE standards specified in this rule. Also as discussed below,
some modifications to the NTE provisions to address technical issues which arise from the
application of advanced NOx catalyst systems have been included in this final rule.
Section ni.A.3.b.v.c ("NVFEL's NOx Adsorber Evaluation Program"), contains a
description of the NOx adsorber evaluation test program run by our EPA laboratory. Included in
that section is test data on four different NOx adsorbers for which extensive steady-state mapping
was performed in order to calculate the SET and AVL composites (See Figures HI. A-7 through
IE. A-10). Several of the test modes presented in these figure are not within the NTE NOx
control zone, and would not be subject to the NTE standard. The following modes listed in these
four figures are within the NTE NOx control zone, EPA modes 6 - 13, 15, 17, 19, 20. For all of
the adsorbers, efficiencies of 90 percent or greater were achieved across the majority of the NTE
zone. The region of the NTE zone for which efficiencies less than 90 percent were achieved
were concentrated on or near the torque curve (EPA modes 8, 9, 15 and 17) with the exception of
Adsorber D, for which EPA modes 6 and 7 achieved 87 percent and 89 percent NOx reduction
respectively. However, Adsorber D was able to achieve NOx reductions greater than 90 percent
along the torque curve. The test modes along the torque curve represent the highest exhaust gas
temperature conditions for this test engine, on the order of 500°C. As discussed in Section
in.A.3.b.iv.c, 500°C is near the current upper temperature limit of the peak NOx reduction
efficiency range for NOx adsorbers, therefore it is not unexpected that the NOx reductions along
the torque curve for the test engine are not as high as in other regions of the NTE zone. We
would expect manufacturers to choose a NOx adsorber formulation which matches the exhaust
gas temperature operating range of the engine. In addition, the steady-state mode data in section
in.A.3.b.v.c were collected under stabilized conditions. In reality, actual in-use operation of a
heavy-duty diesel vehicle would likely not see periods of sustained operation along the torque
curve, and therefore the likelihood the NOx adsorber bed itself would achieve temperatures in
excess of 500°C would be diminished. Regardless, as discussed in Section ni.A.3.b.iv.a & c, we
expect incremental improvements in the high temperature NOx reduction capabilities of NOx
adsorbers between now and model year 2007 will be achieved through improvements in NOx
adsorber formulations.57'58 As discussed above, only small improvements in the current
characteristics are necessary in order to achieve 90 percent NOx reductions or greater across the
NTE control zone.
HI-52
-------
Chapter III: Emissions Standards Feasibility
As discussed in section HI.A.S.b.vi.a, the use of advanced NOx reduction catalyst systems
on HDDEs will present cold-start challenges for HDDEs similar to what light-duty gasoline
manufacturers have faced in the past, due to the light-off characteristics of the NOx adsorber.
We have previously discussed the tools available to HDDE manufacturers to overcome these
challenges in order to achieve the Phase 2 FTP NOx standard. The majority of engine operation
which occurs within the NTE control zone will occur at exhaust gas temperatures well above the
light-off requirement of the NOx adsorbers. Figures ni.A-7 through ILL A-10 in section
lU.A.S.b.v.c ("NVFEL's NOx Adsorber Evaluation Program") show that all test modes which are
within the NTE control zone have exhaust gas temperatures greater than 300°C which, as
discussed in section ni.A.S.b.iv, is well within the peak NOx reduction efficiency range of
current generation NOx adsorbers. However, though the NTE does not include engine start-up
conditions, it is conceivable that a HDDE vehicle which has not been warmed up could be started
and very quickly be operated under conditions which are subject to the NTE standard; for
example, within a minute or less of vehicle operation after the vehicle has left an idle state. The
NTE regulations specify a minimum emissions sampling period of 30 seconds. Conceivably the
vehicle emissions could be measured against the NTE provisions during that first minute of
operation, and in all likelihood it would not meet the NTE NOx standard set in this final rule.
Given that the FTP standards will require control of cold-start emissions, manufacturers will be
required to pay close attention to cold start to comply with the FTP. As discussed above,
operation with the NTE will be at exhaust gas temperatures within the optimum NOx reduction
operating window of the NOx adsorbers. In addition, the NOx adsorber is capable of adsorbing
NOx at temperatures on the order of 100°C. Figures ni.A-7 through II.A.-10 all show NOx
emission reductions on the order of 70 - 80 percent are achieved at temperatures as low as 250°C.
Therefore, we have established a low temperature exhaust gas threshold of 250°C, below which
specified NTE requirements do not apply.
The minimum emissions sample time established under the Phase 1 rule for NTE testing
is 30 seconds. This testing requirement was premised on the use of Phase 1 HDDE emission
control technology such as EGR and fuel injection timing. These emission control devices tend
to produce brake-specific mass emission rates of exhaust pollutants which do not have periodic,
orders of magnitude changes in brake-specific emission rates within the NTE control zone when
averaged over a 30 second sample time. However, this is not the case for the NOx adsorber
catalysts. As discussed throughout this Chapter, NOx adsorbers require active regeneration
events, which can produce near zero mass emission rates during the adsorption phase, followed
by relatively large spikes in NOx and HC emissions during the regeneration phase. This is
illustrated in Figure in.A-11, above, which shows that engine out NOx under steady-state
conditions on the order of 640 ±15 ppm, which is fairly continuous. However, the NOx
emissions downstream of the NOx adsorber are both much lower and are characterized by
periodic, orders of magnitude changes in emissions. The NOx concentration downstream of the
adsorber shows periods of near zero ppm NOx lasting approximately 10 seconds, followed by a
NOx peak with a maximum concentration of approximately 40 ppm, with the spike lasting
m-53
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
approximately four seconds. A similar phenomenon can be seen in Figure IHA-11 for
hydrocarbon emissions. Because of this unique periodic nature of the NOx adsorber system, we
have modified the NTE sample time provisions in the regulations, to assure that the emission
spikes described above are not measured in isolation during NTE testing. The regulations
specify that for any emission control system which requires discreet regeneration events, if a
regeneration event occurs within the emissions sample, the emissions averaging time must be at
least as long at the time between regenerations events (i.e., a regeneration period), multiplied by
the number of full regeneration events within the sample period. This provision to account for
regeneration events ensures that the unique operation of the NOx adsorber system will not cause
an inappropriate exceedance of the NTE limits.
The NTE requirements apply not only during laboratory conditions applicable to the
transient FTP and the SET tests, but also under the wider range of ambient conditions for
altitude, temperature and humidity specified in the regulations. These expanded conditions will
have minimal impact on the emission control systems expected to be used to meet the NTE NOx
standard contained in this final rule. Under the Phase 1 rule, NTE emissions under the expanded
NTE testing conditions can be as high as 3.1 g/bhp-hr NMHC+NOx (1.25 x 2004 FTP standard).
Therefore, we assume here that engines in the 2007 time frame are capable of achieving 3.1
g/bhp-hr NMHC+NOx over the NTE without the use of the NOx control devices needed to
achieve the standards contained in this rule. Thus, we analyze the impact of the NTE expanded
testing conditions on the NOx adsorber, not on the base engine which is capable of achieving the
Phase 1 NTE requirements. In general, it can be said that the performance of the NOx adsorbers
are only effected by the exhaust gas stream to which the adsorbers are exposed. Therefore, the
impact of ambient humidity, temperature, and altitude will only effect the performance of the
adsorber to the extent these ambient conditions change the exhaust gas conditions (i.e., exhaust
gas temperature and gas constituents). The ambient humidity conditions subject to the NTE
requirement will have minimal, if any, impact on the performance of the NOx adsorbers. The
exhaust gas itself, independent of the ambient humidity, contains a very high concentration of
water vapor, and the impact of the ambient humidity on top of the products of dry air and fuel
combustion are minimal. The effect of altitude on NOx adsorber performance should also be
minimal, if any. The NTE test procedure regulations specify an upper bound on NTE testing for
altitude at 5,500 feet above sea-level. The Phase 1 regulations require compliance with an NTE
NMHC+NOx limit of 1.25 x the Phase 1 FTP standard up to this altitude. As discussed above, a
90 percent reduction in NOx emissions from the Phase 1 technology engines is, therefore,
necessary to comply with the NTE standard established in this rule. The decrease in atmospheric
pressure at 5,500 feet should have minimal impact on the NOx adsorber performance. Increasing
altitude can decrease the air-fuel ratio for HDDEs which can in turn increase exhaust gas
temperatures; however, as discussed in the Phase 1 final rule, Phase 1 technology HDDEs can be
designed to target air-fuel ratios at altitude which will maintain appropriate exhaust gas
temperatures, as well as maintain engine-out PM levels near the 0.1 g/bhp-hr level, within the
ambient conditions specified by the NTE test procedure. Finally, the NTE regulations specify
HI-54
-------
Chapter III: Emissions Standards Feasibility
ambient temperatures which are broader than the FTP temperature range of 68-86°F. The NTE
test procedure specifies no lower ambient temperature bounds. However, as discussed above, we
have limited NTE requirements on NOx (and HC) for engines equipped with NOx (and/or HC)
catalysts to include only engine operation with exhaust gas temperatures greater than 250°C.
Therefore, low ambient temperatures will not present any difficulties for NTE NOx compliance.
The NTE also applies under ambient temperatures which are higher than the FTP laboratory
conditions. The NTE applies up to a temperature of 100°F at sea-level, and up to 86°F at 5,500
feet above sea-level. At altitudes in between, the upper NTE ambient temperature requirement is
a linear fit between these two conditions. At 5,500 feet, the NTE ambient temperature
requirement is the same as the upper end of the FTP temperature range (86°F), and therefore will
have no impact on the performance of the NOx adsorbers, considering that majority of the test
data described throughout this chapter was collected under laboratory conditions. The NTE
upper temperature limits at sea-level is 100°F, which is 14°F. (7.7°C) greater than the FTP range.
This increase is relatively minor, and while it will increase the exhaust gas temperature, in
practice the increase should be passed through the engine to the exhaust gas, and the exhaust gas
would be on the order of 8°C higher. Within the exhaust gas temperature range for a HDDE
during NTE operation, an 8°C increase is very small. As discussed above, we expect
manufacturer to choose an adsorber formulation which is matched to a particular engine design,
and we would expect the small increase in exhaust gas temperature which can occur from the
expanded ambient temperature requirements for the NTE will be taken into account by the
manufacturer when designing the complete emission control system.
To summarize, based on the information presented in this Chapter, and the analysis and
discussion presented in this section, we conclude the NTE NOx requirement (1.5 x FTP
standard) contained in this final rule will be feasible by model year 2007.
vii. Are Diesel NOx Adsorbers Durable ?
The considerable success in demonstrating NOx adsorbers, as outlined above, makes us
confident that the technology is capable of providing the level of conversion efficiency needed to
meet the NOx standard. However, there are several engineering challenges that will need to be
addressed in going from this level of demonstration to implementation of durable and effective
emission control systems on production vehicles. In addition to the generic need to optimize
engine operation to match the NOx adsorber performance, industry will further need to address
issues of system and catalyst durability. The nature of these issues are understood well today.
The hurdles that must be overcome have direct analogues in technology issues that have been
addressed previously in other automotive applications and are expected to be overcome with
many of the same solutions. In this section we will describe the major technical hurdles to
address in order to ensure that the significant emission reductions enabled through the
application of NOx adsorbers is realized throughout the life of heavy-duty diesel vehicles. The
m-55
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
section is organized into separate durability discussions for the system components (hardware)
and various near and long term durability issues for the NOx adsorber catalyst itself.
(a) NOx Adsorber Regeneration Hardware Durability
The system we have described in Figure IE. A-4 represents but one possible approach for
generating the necessary exhaust conditions to allow for NOx adsorber regeneration and
desulfation. The system consists of three catalyst substrates (for a CDPF/Low Temperature NOx
Adsorber, a High Temperature NOx Adsorber and an Oxidation Catalyst), a support can that
partitions the exhaust flow through the first two catalyst elements, three fuel injectors, and a
means to divert exhaust flow through one or more of the catalyst partitions. Although not shown
in the figure, a NOx /O2 sensor is also likely to be needed for control feedback and on-board
diagnostics(OBD). All of these elements have already been applied in one form or another to
either diesel or gasoline engines in high volume long life applications.
The environment in an automotive exhaust system is extremely harsh with high
temperatures, high humidity and high levels of mechanical vibration. For all of these reasons
care is taken to design components to function over the life of a vehicle. Despite these
challenging conditions, technologies have been developed over the last 30 or more years that are
well suited to exhaust conditions. One of the most ubiquitous components on a modern
passenger car is the three-way catalyst. Its design has evolved over the years so that today it is
highly efficient, reliable and durable.
The NOx adsorber system we described earlier borrows several components from the
gasoline three-way catalyst systems and benefits from the years of development on three way
catalysts. The catalyst substrates (the ceramic support elements on which a catalyst coating is
applied) have developed through the years to address concerns with cracking due to thermal
cycling and abrasive damage from vehicle vibration. The substrates applied for diesel NOx
adsorbers will be virtually identical to the ones used for today's passenger cars in every way but
size. They are expected to be equally durable when applied to diesel applications as has already
been shown in the successful application of diesel oxidation catalysts (DOCs) on some diesel
engines over the last 10 years.
The NOx/O2 sensor needed for regeneration control and OBD is another component
originally designed and developed for gasoline powered vehicles (in this case lean-burn gasoline
vehicles) that are already well developed and can be applied with confidence in long life for NOx
adsorber based diesel emission control. The NOx/O2 sensor is an evolutionary technology based
largely on the current Oxygen (O2) sensor technology developed for gasoline three-way catalyst
based systems. Oxygen sensors have proven to be extremely reliable and long lived in passenger
car applications, which see significantly higher temperatures than would normally be
encountered on a diesel engine. Diesel engines do have one characteristic that makes the
HI-56
-------
Chapter III: Emissions Standards Feasibility
application of NOx/O2 sensors more difficult. Soot in diesel exhaust can cause fouling of the
NOx/O2 sensor damaging its performance. However this issue can be addressed through the
application of a catalyzed diesel particulate filter (CDPF) in front of the sensor. The CDPF then
provides a protection for the sensor from PM while not hindering its operation. Since the NOx
adsorber is expected to be located downstream of a CDPF in each of the potential technology
scenarios we have considered this solution to the issue of PM sooting is readily addressed.
The catalyst can, the metal frame that holds the catalyst substrates in place and in the
exhaust system, is a well developed technology in its own right. The catalyst can must be able to
secure the catalyst substrates while not chipping or cracking the substrates ceramic material.
Further, in the system described in Figure HI. A-4, the can must also partition the catalysts into
two or more regions. While this is a departure from the way today's three-way catalysts are
made, it is not a significant technical challenge when compared to the complex internal
geometries of some muffler designs. Corrosion and weld durability are two other important
considerations in can design. Advances in material science and manufacturing processes being
made for gasoline catalysts designs to meet the stringent Tier 2 standards are expected address
these issues.
Fuel is metered into a modern gasoline engine with relatively low pressure pulse-width-
modulated fuel injection valves. These valves are designed to cycle well over a million times
over the life of a vehicle while continuing to accurately meter fuel. Applying this technology to
provide diesel fuel as a reductant for a NOx adsorber system is a relatively straightforward
extension of the technology. A NOx adsorber system would expect to cycle far fewer times over
its life when compared to the current long life of gasoline injectors. However, these gasoline fuel
injectors designed to meter fuel into the relatively cool intake of a car can not be directly applied
to the exhaust of a diesel engine. In the testing done at NVFEL, a similar valve design was used
that had been modified in material properties to allow application in the exhaust of an engine.
Thus while benefitting from the extensive experience with gasoline based injectors a designer
can, in a relatively straightforward manner, improve the characteristics of the injector to allow
application for exhaust reductant regeneration.
The NOx adsorber system we describe in Figure HI. A-4 requires a means to partition the
exhaust during regeneration and to control the relative amounts of exhaust flow between two or
more regions of the exhaust system. Modern diesel engines already employ a valve designed to
carry out this very task. Most modern turbochargers employ a wastegate valve that allows some
amount of the exhaust flow to bypass the exhaust turbine in order to control maximum engine
boost and limit turbocharger speed. These valves can be designed to be proportional, bypassing a
specific fraction of the exhaust flow in order to track a specified boost pressure for the system.
Turbocharger wastegate valves applied to heavy-duty diesel engines typically last the life of the
engine in spite of the extremely harsh environment within the turbocharger. This same valve
approach could be applied in order to accomplish the flow diversion required for diesel NOx
HI-57
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
adsorber regeneration and desulfation. Since temperatures will be typically cooler at the NOx
adsorber compared to the inlet to the exhaust turbine on a turbocharger, the control valve would
be expected to be equally reliable when applied in this application.
Toyota has announced its intention to manufacture a CDPF/NOx adsorber based catalyst
system (called Diesel Particulate NOx Reduction (DPNR)) for application to diesel trucks for the
2003 model year. Schematics of Toyota's prototype suggest that they are applying a wastegate
type valve to accomplish flow diversion.59 Also the catalyst can has been designed so that the
exhaust flow-path can be redirected during the NOx regeneration step.60 Toyota's intent to
introduce this type of system in such a short time frame indicates that the technologies needed to
apply the NOx adsorber catalyst are likely to be extensions of existing technologies with which
Toyota is already well familiar.
Therefore, the system components needed to implement a NOx adsorber catalyst system
reflect relatively straight-forward extensions of existing automotive hardware which has already
demonstrated long life and high levels of reliability.
(b) NOx Adsorber Catalyst Durability
In many ways a NOx adsorber like other motor vehicle catalysts, acts like a small
chemical process plant. It has specific chemical processes that it promotes under specific
conditions with different elements of the catalyst materials. There is often an important sequence
to the needed reactions and a need to match process rates in order to keep this sequence of
reactions going. Because of this need to promote specific reactions under the right conditions
early catalysts were often easily damaged. This damage prevents or slows one or more the
reactions causing a loss in emission control.
For example, contaminants from engine oil, like phosphorous or zinc, could attach to
catalysts sites partially blocking the site from the exhaust constituents and slowing reactions.
Similarly, lead added to gasoline in order to increase octane levels bonds to the catalyst sites
causing poisoning as well. Likewise, sulfur which occurs naturally in petroleum products like
gasoline and diesel fuel can poison many catalyst functions preventing or slowing the desired
reactions. High exhaust temperatures experienced under some driving conditions can cause the
catalyst materials to sinter (thermally degrade) decreasing the surface area available for reactions
to decrease.
All of these problems have been addressed over time for the gasoline three-way catalysts,
resulting in the high efficiency and long life durability now typical of modern vehicles. In order
to accomplish this changes were made to fuels and oils used in vehicles (e.g., lead additives
banned from gasoline, sulfur levels reduced in gasoline distillates, specific oil formulations for
HI-58
-------
Chapter III: Emissions Standards Feasibility
aftertreatment equipped cars), and advances in catalysts designs were needed to promote
sintering resistant catalyst formulations with high precious metal dispersion.
The wealth of experience gained and technological advancements made over the last 30
years of gasoline catalyst development can now be applied to the development of the NOx
adsorber catalyst. The NOx adsorber is itself an incremental advancement from current three-
way catalyst technology. It adds one important additional component not currently used on three-
way catalysts, NOx storage catalyst sites. The NOx storage sites (normally alkali or alkaline
earth metals) allow the catalyst to store NOx emissions with extremely high efficiency under the
lean burn conditions typical of the diesel exhaust. It also adds a new durability concern due to
sulfur storage on the catalyst.
This section will explore the durability issues of the NOx adsorber catalyst applied to
diesel vehicles. It describes the effect of sulfur in diesel fuel on catalyst performance, the
methods to remove the sulfur from the catalyst through active control processes, and the
implications for durability of these methods. It then discusses these durability issues relative to
similar issues for existing gasoline three-way catalysts and the engineering paths to solve these
issues. This discussion shows that the NOx adsorber is an incremental improvement upon the
existing three-way catalyst, with many of the same solutions for the expected durability issues.
(b. 1) Sulfur Poisoning of the NOx Storage Sites
The NOx adsorber technology is extremely efficient at storing NOx as a nitrate on the
surface of the catalyst, or adsorber (storage) bed, during lean operation. Because of the
similarities in chemical properties of SOx and NOx, the SO2 present in the exhaust is also stored
on the catalyst surface as a sulfate. The sulfate compound that is formed is significantly more
stable than the nitrate compound and is typically not released during the NOx release and
reduction step (NOx regeneration step). Since the NOx adsorber is virtually 100 percent
effective at capturing SO2 in the adsorber bed, sulfate compounds quickly occupy the NOx
storage sites on the catalyst thereby reducing and eventually rendering the catalyst ineffective for
NOx reduction (poisoning the catalyst).
Figure IH.A-12 shows the effect of sulfur poisoning of a NOx adsorber catalyst as
reported by the DOE DECSE program. The graph shows the NOx adsorber efficiency versus
exhaust inlet temperature under steady-state conditions for a diesel engine based system. The
three dashed lines that overlap each other show the NOx conversion efficiency of the catalyst
when sulfur has been removed from the catalyst. The three solid lines show the effect of sulfur
poisoning on the catalyst at three different fuel sulfur levels over different periods of extended
aging (up to 250 hours). From the figure, it can be seen that even with three ppm sulfur fuel a
significant loss in NOx efficiency can occur in as little as 250 hours. Further, it can be seen that
quite severe sulfur poisoning can occur with elevated fuel sulfur levels. Catalyst performance
HI-59
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
was degraded by more than 70 percent over only 150 hours of operation when 30 ppm sulfur fuel
was used.61
100
c
o
80
I 60
> 40
o
O
20
250
300 350 400
Catalyst Inlet Temperature [C]
450
500
• -x— DECSE II after desulfation (3-ppm)
• -•- - DECSE II after desulfation (16-ppm)
• -•- - DECSE II after desulfation (30-ppm)
:—DECSE II before desulfation (3-ppm, 250 hrs aging)
-DECSE II before desulfation (16-ppm, 200 hrs aging)
-DECSE II before desulfation (30-ppm, 150 hrs aging)
Figure III.A-12. Comparison of NOx Conversion Efficiency before and after Desulfation
The DECSE researchers drew three important conclusions from Figure IHA-12:
• Fuel sulfur, even at very low levels like three ppm, can limit the performance of
the NOx adsorber catalyst significantly.
Higher fuel sulfur levels, like 30 ppm, dramatically increase the poisoning rate,
further limiting NOx adsorber performance.
• Most importantly though, the figure shows that if the sulfur can be removed from
the catalyst through a desulfation (or desulfurization) event, the NOx adsorber can
provide high NOx control even after exposure to sulfur in diesel fuel. This is
HI-60
-------
Chapter III: Emissions Standards Feasibility
evidenced by the sequence of the data presented in the figure. The three high
conversion efficiency lines show the NOx conversion efficiencies after a
desulfation event which was preceded by the sulfur poisoning and degradation
shown in the solid lines.
The increase in sulfur poisoning rate is important to understand in order to look at the
means to overcome the dramatic sulfur poisoning shown here. Sulfur accumulates in the NOx
storage sites preventing their use for NOx storage. In other words, they decrease the storage
volume of the catalyst. The rate at which sulfur fills NOx storage sites is expected to be directly
proportional to the amount of sulfur that enters the catalyst. Therefore, for a doubling in fuel
sulfur levels a corresponding doubling in the SOx poisoning rate would be predicted. In the case
of the two most commonly discussed fuel sulfur levels, our proposed 15 ppm sulfur cap with an
expected in use average less than 10 ppm and a 50 ppm sulfur cap with a regulated average of 30
ppm, the difference in average sulfur levels would indicate at least a three-fold increase in sulfur
poisoning rate (<10 versus 30).
The design of a NOx adsorber will need to address accommodating an expected volume
of sulfur before experiencing unacceptable penalties in either lost NOx control efficiency or
increased fuel consumption due to more frequent NOx regenerations. The amount of operation
allowed before that limit is realized for a specific adsorber design will be inversely proportional
to fuel sulfur quantity. In the theoretical case of zero sulfur, the period of time before the sulfur
poisoning degraded performance excessively would be infinite. For a more practical fuel sulfur
level like the < 10 ppm average expected with a 15 ppm fuel sulfur cap the period of operation
before unacceptable poisoning levels have been reached is expected to be less than 40 hours
(with today's NOx adsorber formulations).62 In the case of a 30 ppm average fuel sulfur level,
this period would be reduced by a factor of three or more.
Future improvements in the NOx adsorber technology are expected due to its relatively
early state of development. Some of these improvements are likely to include improvements in
the kinds of materials used in NOx adsorbers to increase the means and ease of removing stored
sulfur from the catalyst bed. However, because the stored sulfate species are inherently more
stable than the stored nitrate compounds (from stored NOx emissions), we expect that future
NOx adsorbers will continue to be poisoned by sulfur in the exhaust. Therefore a separate sulfur
release and reduction cycle (desulfation cycle) will always be needed in order to remove the
stored sulfur.
(b.2) NOx Adsorber Desulfation
Numerous test programs have shown that sulfur can be removed from the catalyst surface
through a sulfur regeneration step (desulfation step) not dissimilar from the NOx regeneration
function.63^64'65'66'67'68 The stored sulfur compounds are removed by exposing the catalyst to hot
HI-61
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
and rich (air-fuel ratio below the stoichiometric ratio of 14.5 to 1) conditions for a brief period.
Under these conditions, the stored sulfate is released and reduced in the catalyst. This sulfur
removal process, called desulfation or desulfurization in this document, can restore the
performance of the NOx adsorber to near new operation.
Most of the information in the public domain on NOx adsorber desulfation is based upon
research done either in controlled bench reactors using synthetic gas compositions or on
advanced lean burn gasoline engine vehicles. As outlined above these programs have shown that
desulfation of NOx adsorber catalysts can be accomplished under certain conditions but the work
does not directly answer whether NOx adsorber desulfation is practical for diesel engine exhaust
conditions. The DECSE Phase II program answers that question.
Phase II of the DECSE program developed and demonstrated a desulfurization
(desulfation) process to restore NOx conversion efficiency lost to sulfur contamination. The
engine used in the testing was a high speed direct injection diesel selected to provide a
representative source of diesel exhaust and various exhaust temperature profiles to challenge the
emission control devices. The desulfation process developed in the DECSE Phase II program
controlled the air to fuel ratio and catalyst inlet temperatures to achieve the high temperatures
required to release the sulfur from the device. Air to fuel ratio control was accomplished in the
program with exhaust gas recirculation (EGR) and a post injection of fuel to provide additional
reductants. (See discussion in section IH.A.S.b.iii.a, which describes this approach for NOx
regeneration.) Using this approach the researchers showed that a desulfation procedure could be
developed for a diesel engine with the potential to meet in-service engine operating conditions
and acceptable driveability conditions. The NOx efficiency recovery accomplished in DECSE
Phase II using this approach is shown in Figure HI. A-12, above.
The effectiveness of NOx adsorber desulfation appears to be closely related to the
temperature of the exhaust gases during desulfation, the exhaust chemistry (relative air to fuel
ratio), and to the NOx adsorber catalyst formulation.69'70 Lower air to fuel ratios (more available
reductant) works to promote the release of sulfur from the surface, promoting faster and more
effective desulfation. Figure in. A-13 shows results from Ford testing on NOx adsorber
conversion efficiency with periodic aging and desulfation events in a control flow reactor test.71
The control flow reactor test uses controlled gas constituents that are meant to represent the
potential exhaust gas constituents from a lean burn engine. The solid line with the open triangles
labeled "w/o regen" shows the loss of NOx control over thirteen hours of testing without a
desulfation event and with eight ppm sulfur in the test gas (this is roughly equivalent to 240 ppm
fuel sulfur, assuming an air to fuel ratio for diesels of 30:1).72 From the figure it can be seen that
without a desulfation event, sulfur rapidly degrades the performance of the NOx adsorber
catalyst. The remaining two lines show the NOx adsorber performance with periodic sulfur
regeneration events timed at one hour intervals and lasting for 10 minutes (a one hour increment
on 240 ppm fuel sulfur would be approximately equivalent to 34 hours of operation on seven
HI-62
-------
Chapter III: Emissions Standards Feasibility
ppm fuel). The desulfation events were identical to the NOx regeneration events, except that the
desulfation events occurred at elevated temperatures. The base NOx regeneration temperature
for the testing was 350°C. The sulfur regeneration, or desulfation, event was conducted at two
different gas temperatures of 550°C and 600°C to show the effect of exhaust gas temperature on
desulfation effectiveness, and thus NOx adsorber efficiency. From Figure HI.A-13 it can be seen
that, for this NOx adsorber formulation, the NOx recovery after desulfation is higher for the
desulfation event at 600°C than at 550°C.
95%
60%
# of SOx and DeSOx events (1hr periods)
Figure IILA-13. Flow Reactor Testing of a NOx Adsorber with Periodic
Desulfations
As suggested by Figure HI. A-13, it is well known that the rate of sulfur release (also
called sulfur decomposition) in a NOx adsorber increases with temperature.73 74 However, while
elevated temperatures directionally promote more rapid sulfur release, they also can directionally
promote sintering of the precious metals in the NOx adsorber washcoat. The loss of conversion
efficiency due to exposure of the catalyst to elevated temperatures is referred to as thermal
degradation in this document.
(b.3) Thermal Degradation
HI-63
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The catalytic metals that make up most exhaust emission control technologies, including
NOx adsorbers, are designed to be dispersed throughout the catalyst into as many small catalyst
"sites" as possible. By spreading the catalytic metals into many small catalyst sites, rather than
into a fewer number large sites, catalyst efficiency is improved. This is because smaller catalyst
sites have more surface area per mass, or volume, of catalyst when compared to larger catalyst
sites. Since most of the reactions being promoted by the catalyst occur on the surface, increasing
surface area increases catalyst availability and thus conversion efficiency. While high dispersion
(many small catalyst sites) is in general good for most catalysts, it is even more beneficial to the
NOx adsorber catalyst because of the need for the catalytic metal sites to perform multiple tasks.
NOx adsorber catalysts typically rely on platinum to oxidize NO to NO2 prior to adsorption of the
NO2 on an adjacent NOx storage site. Under rich operating conditions, the NOx is released from
the adsorption site, and the adjacent platinum (or platinum + rhodium) catalyst site can serve to
reduce the NOx emissions into N2 and O2. High dispersion, combined with NO oxidation, NOx
storage and NOx reduction catalyst sites being located in close proximity, provide the ideal
catalyst design for a NOx adsorber catalyst. High temperatures, especially under oxidizing
conditions, can promote sintering of the platinum and other PGM catalyst sites, permanently
decreasing NOx adsorber performance.
Catalyst sintering is a process by which adjacent catalyst sites can "melt" and regrow into
a single larger catalyst site (crystal growth). The single larger catalyst site has less surface area
available to promote catalytic activity than the original two or more catalyst sites that were
sintered to form it. This loss in surface area decreases the efficiency of the catalyst.75 High
temperatures promote sintering of platinum catalysts especially under oxidizing conditions.76
Therefore, it is important to limit the exposure of platinum based catalysts to high exhaust
temperatures especially during periods of lean operation. Consequently, the desire to promote
rapid desulfation of the NOx adsorber catalyst technology by maximizing the desulfation
temperature and the need to limit the exposure of the catalyst to the high temperatures that
promote catalyst sintering must be carefully balanced. An example of this tradeoff can be seen in
Figure in. A-14 below, which shows the NOx conversion efficiency of three NOx adsorber
catalysts evaluated after extended periods of sulfur poisoning followed by sulfur regeneration
periods.77 The three catalysts (labeled A, B, and C) are identical in formulation and size but were
located at three different positions in the exhaust system of the gasoline direct injection engine
used for this testing. Catalyst A was located 1.2 meters from the exhaust manifold, catalyst B 1.8
meters from the exhaust manifold and catalyst C was located 2.5 meters from the exhaust
manifold. Locating the catalysts further from the engine lowered the maximum exhaust
temperature and thus catalyst bed temperature experienced during the programmed sulfur
regeneration cycle. Catalyst A experienced the highest catalyst bed temperature of 800°C, while
catalyst C experienced the lowest catalyst bed temperature of 650°C. Catalyst B experienced a
maximum catalyst bed temperature of 730°C. Figure HI. A-14 shows that an optimum
desulfation temperature exists which balances the tradeoffs between rapid sulfur regeneration and
thermal degradation (thermal sintering) at high temperatures.
HI-64
-------
Chapter III: Emissions Standards Feasibility
Influence of Maximum Temperature in Durability Cycle on Engine Bench
100%
o
o
O
IO
fO
+J
re
o
'55
I
o
O
x
O
80% --
60% --
40% --
20% --
0%
Lower Temperatures Decrease
Sulfur Deconposition Rates and
the Effective Period of Desulfation
B
Aging Time:
100 hours
High Temperatures Promote
Sulfate Decomposition but
Increase Precious Metal Sintering
600 650 700 750 800
Maximum Catalyst Bed Temperature (°C) in Durability Cycle
From SAE 1999-01-3501 Figure 7
850
Figure III.A-14. Influence of Maximum Catalyst Bed Temperature During
Desulfation
The DECSE Phase II program, in addition to investigating the ability of a diesel engine /
NOx adsorber based emission control system to desulfate, provides a preliminary assessment of
catalyst durability when exposed to repeated aging and desulfurization cycles. Two sets of tests
were completed using two different fuel sulfur levels (three ppm and 78 ppm) to investigate these
durability aspects. The first involved a series of aging, performance mapping, desulfurization
and performance mapping cycles. An example of this testing is shown below in Figure in.A-15.
The graph shows a characteristic "sawtooth" pattern of gradual sulfur poisoning followed by an
abrupt improvement in performance after desulfation. The results shown in Figure in. A-15 are
for two identical catalysts one operated on 3 ppm sulfur fuel (catalyst S5) and the other operated
on 78 ppm sulfur fuel (catalyst S7). For the catalyst operated on 3 ppm sulfur fuel the loss in
performance over the ten hours of poisoning is noted to be very gradual. There appears to be
little need to desulfate that catalyst at the ten hour interval set in the experiment. In fact it can be
seen that in several cases the performance after desulfation is worse than prior to desulfation.
This would suggest as discussed above, that the desulfation cycle can itself be damaging to the
catalyst. In actual use we would expect that an engine operating on 3 ppm sulfur fuel would not
desulfate until well beyond a ten hour interval and would be engineered to better withstand the
damage caused by desulfation, as discussed later in this section. For the catalyst operated on 78
HI-65
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
ppm sulfur fuel the loss in performance over the ten hour poisoning period is dramatic. In order
to ensure continued high performance when operating on 78 ppm sulfur fuel the catalyst would
require frequent desulfations. From the figure it can be inferred that the desulfation events would
need to be spaced at intervals as short as one to two hours in order to maintain acceptable
performance.
100%
90% <
80%
70% -
§ 60%-|
50% -
X
O
40% -
30% -
20% -
10% -
0% -
Aging 10 hrs
of 3 ppm fuel
-
^:
Aging 10 hrs
of 78 ppm fuel =
• DECSE Catalyst S7
- Aged on 78 ppm S
• DECSE Catalyst S5
- Aged on 3 ppm S
0 10 20 30 40 50
Time (hours) cycle of 10 hrs Sulfur Aging / 6 min Desulfation
Figure IILA-15. Integrated NOx Conversion Efficiency following Aging and Desulfation
As a follow on to the work shown in Figure UI.A-15, the desulfation events were repeated
an additional 60 times without sulfur aging between desulfation events. This was done to
investigate the possibility of deleterious affects from the desulfation event itself even without
additional sulfur poisoning. As can be seen in Figure HI.A-16, the investigation did reveal that
repeated desulfation events even without additional sulfur aging can cause catalyst deterioration.
As described previously, high temperatures can lead to a loss in catalyst efficiency due to thermal
degradation (sintering of the catalytic metals). This appears to be the most likely explanation for
the loss in catalyst efficiency shown here. For this testing, the catalyst inlet temperature was
controlled to approximately 700°C, however the catalyst bed temperatures could have been
higher.78
m-66
-------
Chapter III: Emissions Standards Feasibility
Based on the work in DECSE Phase n, the researchers concluded that
• The desulfurization procedure developed has the potential to meet in-service engine
operating conditions and to provide acceptable driveability conditions.
• Although aging with 78 ppm sulfur fuel reduced NOx conversion efficiency more than
aging with three ppm sulfur fuel as a result of sulfur contamination, the desulfurization
events restored the conversion efficiency to nearly the same level of performance.
However, repeatedly exposing the catalyst to the desulfurization procedure developed in
the program caused a continued decline in the catalyst's desulfated performance.
• The rate of sulfur contamination during aging with 78 ppm sulfur fuel increased with
repeated aging / desulfurization cycles (from 10 percent per ten hours to 18 percent per
ten hours). This was not observed with the three ppm sulfur fuel, where the rate of
decline during aging was fairly constant at approximately two percent per ten hours.
100%
80% -
90% -
C
O
i
Jo *c 70%i
0) O
O 'i=
,_ O 60% -
O X
50% -
40% -
30% -
LU
X 20% -
O
10% -
0%
DECSE Catalyst S7
DECSE Catalyst S8
10
20
30
40
50
60
70
Desulfation Events (# of desulfation cycles)
Figure III.A-16. Integrated NOx Conversion Efficiency after Repeated Desulfation
The data available today on current NOx adsorber formulations shows clearly that sulfur
can be removed from the surface of the NOx adsorber catalyst. The initial high performance
HI-67
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
after a desulfation event is then degraded over time by the presence of sulfur until the next
desulfation event. The resulting characteristic NOx adsorber performance level over time
exhibits a saw-tooth pattern with declining performance followed by rapid recovery of
performance following desulfation. The rate of this decline increases substantially with higher
fuel sulfur levels. In order to ensure a gradual and controllable decline in performance fuel sulfur
levels must be minimized. However, even given very low fuel sulfur levels, gradual decline in
performance must be periodically overcome. The development experience so far shows that
diesel engines can accomplish the required desulfation event. The circumstances that effectively
promote rapid desulfation also promote thermal degradation. It will therefore be important to
limit thermal degradation.
(b.4) Limiting Thermal Degradation
The issue of thermal degradation of NOx adsorber catalyst components is similar to the
thermal sintering issues faced by light-duty three-way catalysts for vehicles developed to meet
current California LEV and future Federal Tier 2 standards using platinum+rhodium (Pt+Rh)
catalysts. Initial designs were marked by unacceptable levels of platinum sintering which limited
the effectiveness of Pt+Rh catalysts. This problem has been overcome through modifications to
the catalyst supports and surface structures that stabilize the precious metals at high temperatures
(>900 °C). Stabilization of ceria components using Zirconium (Zr) has pushed the upper
temperature limits of ceria migration to well over 1000 °C.79'80 Stabilization components can
function in a number of ways. Some are used to "fill" structural vacancies, for example "open"
locations within a crystalline lattice, thus strengthening the lattice structure. Such strengthening
of crystalline lattice structures is particularly important at high temperatures. Other types of
stabilizing components can act as obstructions within a matrix to prevent migration of
components, or can enhance the mobility of other molecules or atoms, such as oxygen. An
approach to the stabilization of NOx adsorber catalyst components that is similar to the
approaches taken with LEV three-way catalyst designs should help to minimize thermal sintering
of components during desulfation.
In many ways, limiting the thermal degradation of the NOx adsorber catalyst should be
easier than for the gasoline three-way catalyst. Typical exhaust gas temperatures for a heavy
light-duty gasoline truck (e.g., a Ford Expedition) commonly range from 450°C to more than
800°C during normal operation.81 A heavy-duty diesel engine in contrast rarely has exhaust gas
temperatures in excess of 500°C. Further, even during the desulfation event, exhaust
temperatures are expected to be controlled well below 800°C. Therefore the NOx adsorber when
applied to diesel engines is expected to see both lower average temperatures and lower peak
temperatures when compared to an equivalent gasoline engine. Once thermal degradation
improvements are made to NOx adsorber catalysts, thermal degradation will reasonably be
expected to be less than the level predicted for future Tier 2 gasoline applications.
HI-68
-------
Chapter III: Emissions Standards Feasibility
In addition to the means to improve the thermal stability of the NOx adsorber by applying
many of the same techniques being perfected for the Tier 2 gasoline three-way catalyst
applications, an additional possibility exists that the desulfation process itself can be improved to
give both high sulfur removal and to limit thermal degradation. The means to do this might
include careful control of the maximum temperature during desulfation in order to limit the
exposure to high temperatures. Also, improvements in how the regeneration process occurs may
provide avenues for improvement. Low air to fuel ratios (high levels of reductant) are known to
improve the desulfation process. The high level of reductant may also help to suppress oxygen
content in the exhaust to further limit thermal degradation.
Researchers at Ford Scientific Research Labs have investigated NOx adsorber catalyst
desulfation (called DeSOx in their work) to answer the question: "if a regeneration process
(sulfur regeneration) is required periodically, will the high temperatures required for the
regeneration have deleterious, irreversible effects on NOx efficiency?" To explore the issue of
NOx adsorber durability after repeated desulfation events, Ford conducted repeated sequential
sulfur poisoning and desulfation cycles with a NOx adsorber catalyst. The results of their
experiment are shown in Figure in.A-17.82 As shown in Figure UI.A-17, the NOx adsorber
sample underwent more than 90 poisoning and desulfation cycles with 12 hours occurring
between the end of one desulfation to the end of the next desulfation without a measurable loss in
post-desulfation performance. This testing was done using a laboratory tool called a pulsator,
used to study ceramic monolith catalyst samples. The ceramic test samples were heated to
between 700°C and 750°C. These results indicate that for some combinations of temperatures
and reductant chemistries the NOx adsorber can be repeatedly desulfated without a significant
loss in NOx reduction efficiency. This work indicates that it is possible to optimize the
desulfation process to allow for adequate sulfur removal without a significant decrease in NOx
reduction efficiency.
HI-69
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
100
10
20
30
40 50
DeSOx event
60
70
Figure III.A-17. Repeated Sulfur Poisoning and Desulfation on a Bench Pulsator
These results indicate that, with further improvements to the NOx adsorber catalyst
design incorporating the experience gained on gasoline three-way catalysts and continuing
improvements in the control of the desulfation, degradation of the NOx adsorber catalyst with
each desulfation event can be limited. However, the expectation remains that there will be some
level of deterioration with desulfation that must be managed to ensure long term high efficiency
of the NOx adsorber. This means that the number and frequency of desulfation events must be
kept to a minimum. The key to this is to limit the amount of sulfur to which the catalyst is
exposed over its life. In this way, the deterioration in performance between desulfation events is
controlled at a gradual rate and the period between desulfations can be maximized to limit
thermal degradation. There appears to be a general consensus among the technology developers
in the emission control industry that, when sulfur levels are controlled with an average level
below 10 ppm, the frequency and number of desulfation events will be at a level such that long
term durability can be assured. This is evidenced by comments provided on our NPRM. Some
examples of those comments are:
1. The Department (the U. S. Department of Energy) is confident that, assuming a
reasonable rate of technology development before 2006, dieselfuel sulfur levels
averaging 10 ppm or less will enable emission control devices (NOx adsorbers and
m-70
-------
Chapter III: Emissions Standards Feasibility
CDPFs) to operate effectively over the full useful vehicle life and, therefore, allow the
vehicles/engines to meet the future standards (the Phase 2 standards)." - letter from DOE
to Bob Perciascepe 6 September 2000, EPA Docket A-99-06 Item IV-G-28.
2. ... the technological challenges posed by the proposed 2007 HDD standards are
achievable. .. with the surety of specific standards at a known date, along with a
concerted effort by the engine manufacturers and the emission control technology
industry, once again we will "make it happen " with technology and integrated systems
that meet the standards and are durable, letter from Martin Lassen - Johnson Matthey
Catalytic Systems division 19 October 2000, EPA Docket A-99-06 Item IV-G-55.
3. We believe all NOx adsorber development issues have been identified and the
technology is proceeding according to schedule. Letter from John Mooney Director,
Technology Development and Business Systems Engelhard Corporation to Margo Oge,
U.S. EPA 3 October 2000, EPA Docket A-99-06 Item IV-G-38.
EPA agrees and expects this progress is likely to occur along the developmental paths
discussed above.
(d) Overall System Durability
NOx emission control with a NOx adsorber catalyst based systems is an extension of the
very successful three-way catalyst technology. NOx adsorber technology is most accurately
described as incremental and evolutionary with system components that are straightforward
extensions of existing technologies. Therefore, the technology benefits substantially from the
considerable experience gained over the past 30 years with the highly reliable and durable three-
way catalyst systems of today.
The following observations can be made from the data provided in the preceding sections
on NOx adsorber durability:
• NOx adsorber catalysts are poisoned by sulfur in diesel fuel, even at fuel sulfur
levels as low as three ppm.
A sulfur regeneration event (desulfation) can restore NOx adsorber performance.
A diesel engine can produce exhaust conditions that are conducive to desulfation.
Desulfation events which require high catalyst temperatures can cause sintering of
the catalytic metals in the NOx adsorber thereby reducing NOx control efficiency.
HI-71
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The means exist from the development of gasoline three-way catalysts to improve
the NOx adsorber's thermal durability.
• In carefully controlled experiments, NOx adsorbers can be desulfated repeatedly
without an unacceptable loss in performance.
The number and frequency of desulfation events must be limited in order to
ensure any gradual thermal degradation over time does not excessively deteriorate
the catalyst.
Based on these observations, we are confident that NOx adsorber technology for MY2007
and later engines will be durable over the life of heavy-duty diesel vehicles, provided fuel with a
15 ppm sulfur cap is used. Without the use of this low sulfur fuel, we can no longer be confident
that the increased number of desulfation cycles that will be required to address the impact of
sulfur on efficiency can be accomplished without unrecoverable thermal degradation and thus
loss of NOx adsorber efficiency. Limiting the number and frequency of these deleterious
desulfation events through the use of diesel fuel with sulfur content less than 15 ppm allows us to
conclude with confidence that NOx adsorber catalysts will be developed that are durable
throughout a vehicle's life.
viii. Will NOx Adsorber Desulfation Lead to Undesirable H2S Emissions?
The desulfation process for a NOx adsorber catalyst is directly analogous to the NOx
regeneration process. The stored sulfur emissions are released and reduced over the catalyst
before exiting the tailpipe. The final reduction state for the sulfur emissions can be any one of
several sulfur products including hydrogen sulfide (H2S) and sulfur dioxide (SO2). Hydrogen
sulfide has a strong unpleasant odor which is often described as smelling like a rotten egg. For
this reason it is preferable to emit the stored sulfur as SO2 rather than H2S.
There are several possible ways to control H2S emissions from an automotive catalyst.
For gasoline three-way catalysts which initially had problems with H2S emissions, the solution
has been the inclusion of nickel oxide to the catalyst formulation. Nickel oxide will store sulfur
under reducing conditions as nickel sulfide and will release sulfur under oxidizing conditions as
SO2. This storage and release of sulfur emissions in the three-way catalyst has proved effective
in controlling H2S emissions. This approach to controlling sulfur emissions may be possible for
NOx adsorber catalysts as well, although because NOx adsorbers store sulfur under lean
conditions it may prove to be more difficult to use this approach. A catalyst formulated to store
sulfur under both rich and lean conditions would presumably be hindered in its ability to
completely remove sulfur under either condition.
HI-72
-------
Chapter III: Emissions Standards Feasibility
Another solution to address H2S emissions would be the inclusion of a downstream
"clean-up" catalyst which could oxidize H2S emissions, released under rich conditions, to a more
desirable sulfur product, namely SO2. The oxidizing potential for this reaction could come from
either oxygen storage elements in the clean-up catalyst or from a second oxygen rich exhaust
stream which would join the reducing stream from the desulfating NOx adsorber prior to entering
the clean-up catalyst. The use of a clean-up catalyst is identified as a means to control NMHC
emissions during NOx regeneration as described in sections in.A.S.b.iii and UI.A.5. The same
approach described in those sections will work well to control H2S emissions. In our cost
analysis described in Chapter V, we have included the cost of a diesel oxidation catalyst in the
total system cost in order to address both NMHC emission concerns under NOx regeneration
conditions and H2S concerns under desulfation conditions.
ix. Can SOx Traps Protect NOx Adsorbers from High Sulfur Diesel Fuel?
The preceding discussion of NOx adsorbers assumes that SOx (SO2 and SO3) emissions
will be "trapped" on the surface of the catalyst, effectively poisoning the adsorber and requiring a
"desulfation" (sulfur removal event) to recover catalyst efficiency. We believe that, at the 15
ppm cap fuel sulfur level, this strategy will allow effective NOx control over the life of heavy-
duty vehicles. As an alternative to desulfation of the NOx adsorber itself, some researchers are
investigating the use of an adsorber catalyst (SOx adsorber) designed to preferentially store sulfur
emissions in order to serve as a protective catalyst for sulfur sensitive technologies such as NOx
adsorbers. The device would then either require replacement or its own desulfation event in
order allow the SOx adsorber to last for the life of a heavy-duty diesel vehicle.
Replacement of the SOx adsorber on a periodic basis appears to be a workable solution to
the problem of sulfur in diesel fuel only for fuel sulfur levels well below 15 ppm. Analysis
provided by Cummins Engine Company estimates that a sulfur trap could store approximately
one pound of SO2 per cubic foot of catalyst volume.83 Based on this assumption an estimate was
made of the package volume needed to store one year's worth of sulfur from diesel fuel for a
heavy heavy-duty application that traveled 250,000 miles in one year while averaging seven
miles per gallon:
For five ppm sulfur fuel, the "disposable" SOx adsorber volume would be 2.5 ft3.
For 15 ppm sulfur fuel, the "disposable SOx adsorber volume would be 7.5 ft3.
For 50 ppm sulfur fuel, the "disposable" SOx adsorber volume would be 25 ft3.
While it may be possible to imagine packaging a removable SOx adsorber with a volume
of approximately 2.5 cubic feet (-71 liters) for a heavy heavy-duty vehicle, packaging a 25 cubic
HI-73
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
foot (710 liter) catalyst is probably impossible. For comparison typical heavy heavy-duty diesel
engine displacements range from 10 to 15 liters in volume.
An alternative approach is to use a SOx adsorber in series with the NOx adsorber and
then to either divert the effluent around the NOx adsorber during the SOx adsorber desulfation
event or allow the effluent to go through the NOx adsorber under rich conditions which are less
conducive to sulfur storage on the catalyst. The first approach of diverting the effluent around
the NOx adsorber is theoretically possible, but has several practical limitations. Although SOx
adsorbers are highly efficient at capturing sulfur in the exhaust they are not 100 percent
effective.84 This means that some fugitive sulfur will invariably be deposited on the NOx
adsorber catalyst. In other words this approach only defers the needed desulfation event to a later
time, although it would allow for less frequent desulfation events which is helpful in controlling
thermal degradation. Additionally, diverting the sulfur laden gas around the NOx adsorber adds
complexity to the system which is not necessary, since the NOx adsorber itself must also be able
to be desulfated. If the sulfur emissions are allowed to travel through the NOx adsorber catalyst
(rather than being diverted around) under rich conditions some significant fraction of the sulfur
will still be trapped on the NOx adsorber catalyst.85 Again, this approach may allow for
somewhat less frequent NOx adsorber desulfations but will not eliminate the need for periodic
desulfation of the NOx adsorber catalyst.
Based on the eventual need for NOx adsorber desulfation under any of the scenarios
described here and the impractical package size of a replacement SOx trap, we believe the only
viable approach is to have 15 ppm sulfur diesel fuel.
c. Selective Catalytic Reduction (SCR)
Diesel Selective Catalytic Reduction (SCR) is an adaptation of stationary technology that
has been in use for some time. Ammonia (NH3) is injected into the exhaust upstream of a
vanadium/titanium (V2O5/TiO2) catalyst to reduce NOx. The following reactions occur:
4NH3 + 4 NO + O2 - 4N2 + 6H2O
2NH3 + NO + NO2 - 2N2 + 3H2O
4NH3 + 3NO2 - 7N2 + 6H2O
The ammonia is typically stored onboard the vehicle as a urea solution ((NH2)2CO) since
ammonia is hazardous in its raw form. The urea solution is then injected upstream of the catalyst
which breaks down the urea into ammonia and carbon dioxide. The ammonia must be injected in
proportion to the NOx produced by the engine. If too much ammonia is injected for the amount
of NOx present, the excess ammonia can pass unreacted through the SCR catalyst; this is referred
to as "ammonia slip." In a mobile transient application, controlling the urea injection to prevent
ammonia slippage is key. A diesel oxidation catalyst (DOC) containing platinum can be used
HI-74
-------
Chapter III: Emissions Standards Feasibility
downstream of the SCR system to control ammonia slippage by oxidizing any slipped ammonia
to N2 and H2O. A DOC can also be used upstream of the SCR system to improve NOx reduction
performance by converting NO to NO2. Optimum NOx reduction occurs when the NOx has a
significant NO2 fraction (note that diesel engine out NOx typically has only a small fraction of
NO2). Systems that use a DOC to improve cold temperature performance are called "compact
SCR" systems due to their relatively small size when compared to conventional SCR systems.
The urea SCR has been developed for stationary applications and is currently being
refined for the transient operation found in mobile applications. The reduction efficiency
window for this device is similar to the NOx adsorber, with greater than 80 percent efficiency at
exhaust temperatures as low as 250 C.86 Testing has shown HD FTP cycle NOx reductions of 77
percent.87 Such efficiencies would allow NOx levels of 0.5 g/hp-hr to be possible with today's
technology starting with a 2.0 g/hp-hr cooled EGR engine. Lower levels would be possible with
engine out emission reductions. Over the NTE zone the SCR has been shown to have 65-99
percent efficiency.88
Implementation of SCR poses unique difficulties due to the need to create a new supply
chain for the urea. A SCR system consumes urea at a rate proportional to the NOx emission rate.
SCR testing today has shown urea consumption rates ranging from three to six percent of the
amount of fuel burned. Therefore, a line haul truck with a 300 gallon fuel tank would need 9 to
18 gallons of urea for every fill-up. Likewise, a large sport utility vehicle (SUV) with a 50 gallon
fuel tank would need 1.5 to 3 gallons of urea for every fill-up. If the urea were distributed in
liquid form, this would mean an additional on-board tank for urea that would probably have to be
replenished at each refueling. Further, without an adequate urea supply onboard, whether by
accident or by user intent, the SCR system would become useless, converting none of the NOx.
Since the urea is expected to cost approximately 80 cents per gallon, there would be some
incentive for the user not to refill the urea tank.89 Since driving performance of the engine is not
normally affected by the absence of urea, manufacturers would have to provide incentive for the
users to continue refilling the urea tank so that the in-use benefit of the SCR system would be
fully realized. What form such a refilling incentive would take is not known. A standardized
ammonia distribution format (liquid, solid, etc.), delivery infrastructure, and anti-tampering
measures are all issues that would need to be addressed to make this technology viable.
Urea SCR catalysts, like NOx adsorbers, need low sulfur diesel fuel to achieve high NOx
conversion efficiencies and to control sulfate PM emissions. If low sulfur fuel is required, SCR
NOx control may be possible in some applications by 2007. For a further discussion of SCR
system sensitivity to sulfur in diesel fuel, and of its need for low sulfur diesel fuel, refer to
section in.A.7. However we believe there are significant barriers to its general use for meeting
the 2007 standards.
HI-75
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
There would need to be adequate safeguards in place to ensure the urea is used
throughout the life of the vehicle since, given the added cost of urea and the fact that urea
depletion would not normally affect driveability, there would be an incentive not to refill the urea
tank. This could lead to considerable uncertainties regarding the effectiveness of SCR, even if
EPA were to promulgate the regulations that likely would be needed to require the regular
replenishment of urea. Some would argue that this is the key issue with regard to urea SCR
systems, but that this issue could be addressed by designing engines with on-board diagnostic
systems utilizing a NOx sensor that would observe a loss of NOx control. When observed, the
engine would be designed to reduce power gradually until a 50 percent loss of power was
realized. This power loss would serve to encourage the user to replenish the urea tank.90 While
such an approach may be possible, it raises concerns for public safety as poor engine
performance could lead to inadequate power for safe merging onto highways and other related
driving situations. We remain hesitant to base a national program on such technology when
important issues such as driver training on the need to refill the urea tank and the consequences
of failure to do so cannot be appropriately controlled. This approach would seem to suggest a
need for EPA mandated spot checks of individual vehicles to ensure compliance with the NOx
standard. How such a program would work and the burden that it might place on small business
entities was not addressed in the comments. The California Trucking Association has raised
concerns about the appropriateness of putting this regulatory burden on truckers when a simpler
technology such as a diesel NOx adsorber was available instead.91 Without measures similar to
these we doubt that users would consistently remember to fill their urea tanks. Since failure to
provide urea would lead to a total loss of NOx control, we would need to model the loss of NOx
control to be expected from an SCR based program.
While SCR systems are capable of limited operation on current sulfur level fuels, their
efficiency is reduced at the low temperatures typical of much of diesel engine operation and they
run the risk of ammonia slip. Consequently, to achieve the NOx standard, a SCR system would
likely need platinum-containing oxidation catalysts upstream and downstream. The presence of
any platinum in the system, whether for conversion of ammonia slip or for conversion of NO to
NO2, would lead to the production of sulfate PM and loss of NOx reduction efficiency.
Therefore, like every exhaust emission control technology discussed so far, the elimination of
fuel sulfur is imperative for this technology to be effective.
d. Non-Thermal Plasma Assisted Catalysts
Another approach to NOx reduction is the non-thermal plasma assisted catalyst. This
system works by applying a high voltage across two metal plates in the exhaust stream to form
ions that serve as oxidizers. Essentially, the plasma would displace a conventional platinum
based oxidation catalyst in function. Once oxidized to NO2, NOx can be more readily reduced
over a precious metal catalyst or used as an oxidizer, as in CDPFs. A potential drawback of this
technology is the high voltage and power requirement. Generation of this power is expected to
HI-76
-------
Chapter III: Emissions Standards Feasibility
entail a two to three percent fuel economy penalty.92 We expect that, if and when the non-
thermal plasma approach to NOx control becomes viable, it will also require the use of low
sulfur diesel fuel due to its reliance on a precious metal catalyst to reduce the NO2.93
4. Meeting the NMHC Standard
Meeting the NMHC standards under the lean operating conditions typical of the biggest
portion of NOx adsorber operation should not present any special challenges to diesel
manufacturers. Since all of the devices discussed above - CDPFs, NOx adsorbers, and SCR -
contain platinum and other precious metals to oxidize NO to NO2, they are also very efficient
oxidizers of hydrocarbons. NMHC reductions of greater than 95 percent have been shown over
transient FTP and the SET modes.94 Given that typical engine out HC is expected to be in the
0.20 g/bhp-hr range for engines meeting the Phase 1 standards, this level of NMHC reduction
will mean that under lean conditions emission levels will be well below the standard.
The NOx regeneration strategies for the NOx adsorber technology may prove difficult to
control precisely, leading to a possible increase in HC emissions under the rich operating
conditions required for NOx regeneration. Even with precise control of the regeneration cycle,
HC slip may prove to be a difficult problem due to the need to regenerate the NOx adsorber
under net rich conditions (excess fuel) rather than the stoichiometric (fuel and air precisely
balanced) operating conditions typical of a gasoline three-way catalyst. It seems likely therefore,
that in order to meet the HC standards we have set, an additional clean up catalyst may be
required. A diesel oxidation catalyst, like those applied historically for HC and partial PM
control, can reduce HC emissions (including Toxic HCs) by more than 90 percent.95 This amount
of additional control along with optimized NOx regeneration strategies will ensure very low HC
emissions.
During the NVFEL NOx adsorber test evaluation program, we performed extensive
testing of a system which included CDPFs, NOx adsorbers, and a clean-up diesel oxidation
catalyst with low platinum loading. As discussed in section HI.A.S.b of this RIA, and in more
detail in the technical memorandum to the docket detailing this test program, without the use of a
clean-up DOC we encountered test conditions which resulted in high HC emissions from the
NOx adsorber regeneration events. As discussed in section HI.A.S.b, this complete system, when
tested over the hot-start HDDE FTP, resulted in HC emissions of 0.25 g/bp-hr, a 13 percent
reduction from the baseline values for the test engine. However, the clean-up DOC we used for
this evaluation program had a relatively light precious metal loading (-10 g/ft3) and a relatively
low cell density (300 cpsi). Emissions sampling upstream and downstream of this DOC
indicated it's oxidation efficiency was less than 60 percent. More effective DOC formulations
have been shown to produce greater than an 90 percent reduction in hydrocarbons.96 In addition,
as discussed in section ni.A.3.b, the NVFEL evaluation program did not optimize the
hydrocarbon reductant injection strategy in the short time available to the test program. During
HI-77
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
the testing we saw opportunities for optimizing the use of the injected diesel fuel to achieve the
desired NOx reduction which we were not able to pursue. With additional time, we would
expect to both decrease the HC slip from the emission control system, as well as employ a much
more effective DOC, capable of HC reductions on the order of 80 percent. When combined, we
expect the NMHC standard will be achieved over the FTP and SET tests, as well as during NTE
testing. With a more effective downstream clean-up DOC to control HC slip during the periodic
NOx regeneration event, the HC standard we have set here can be met.
As discussed in section ni.A.S.b of this RIA regarding NOx emissions, the minimum
emission sample time provisions for the NTE test have been changed to reflect the potential for
short-duration high HC emissions which can occur following a regeneration event. This change
to the NTE minimum sample time approach will address any feasibility concerns which could
arise because of the short-term increase in HC emissions immediately following a regeneration
event, by increasing the sample time to include the time period until the next regeneration. In
addition, the NMHC NTE provisions do not apply until the hydrocarbon emission control device
(e.g., DOC) has achieved a warmed up exhaust gas temperature of at least 250°C on the outlet of
the device. This same provision applies to the NOx NTE standard. With these additional
constraints placed on NTE testing, we conclude the NTE provisions can be achieved.
5. Meeting the Crankcase Emissions Requirements
The most common way to eliminate crankcase emissions has been to vent the blow-by
gases into the engine air intake system, so that the gases can be re-combusted. Until today's
rulemaking, we have required that crankcase emissions be controlled only on naturally aspirated
diesel engines. We have made an exception for turbocharged heavy-duty diesel engines in the
past because of concerns regarding fouling that could occur from diesel PM and engine oil,
which are included in the crankcase emissions, when routing the crankcase blow-by into the
turbocharger and aftercooler. However, this is an environmentally significant exception since
most heavy-duty diesel trucks use turbocharged engines, and a single engine can emit over 100
pounds of NOx, NMHC, and PM from the crankcase over its lifetime. Over the past several
years technology has become available which allows us to eliminate the exception for
turbocharged diesel engines, as discussed below.
We anticipate that the heavy-duty diesel engine manufacturers will be able to close the
crankcase using one of two methods. First, by using closed crankcase filtration systems. We are
aware of at least two companies which produce closed crankcase filtration systems for the heavy-
duty diesel market today, as described in more detail below.97'98 Second, the blow-by gases could
be routed directly into the exhaust system upstream of the emission control equipment. Finally,
if the manufacturer chooses not to close the crankcase, the manufacturer must add the emission
from the open crankcase ventilation system to the emissions from the engine downstream of any
emission control equipment, e.g., the open-crankcase emissions would be added into the FTP
HI-78
-------
Chapter III: Emissions Standards Feasibility
emission results. Thus, the regulatory provision has been written such that if adequate control
can be had without "closing" the crankcase then the crankcase can remain "open."
We expect that in order to meet the stringent tailpipe emission standards set in this rule
manufacturers will have to utilize closed crankcase approaches as described here. Closed
crankcase filtration systems work by separating oil and particulate matter from the blow-by
gases through single or dual stage filtration approaches, routing the blow-by gases into the
engine's intake manifold and returning the filtered oil to the oil sump. Closed crankcases are
required for new heavy-duty diesel vehicles in Europe starting in 2000. Oil separation
efficiencies in excess of 80 percent have been demonstrated with production ready prototypes of
two stage filtration systems after more than 500 hours of testing." 10° By eliminating 80 percent
or more of the oil that would normally be vented to the atmosphere, the system works to reduce
oil consumption and to eliminate concerns over fouling of the intake system when the gases are
routed through the turbocharger. Mercedes-Benz currently utilizes this type of system on
virtually all of its heavy-duty diesel engines sold in Europe, and Mercedes-Benz has certified at
least one on-highway HDDE in the U.S. equipped with such a system since at least 1999.101
An alternative approach could be to route the blow-by gases into the exhaust system
upstream of the CDPF which would be expected to effectively trap and oxidize the engine oil and
diesel PM. This approach may require the use of low sulfur engine oil to ensure that oil carried
in the blow-by gases does not compromise the performance of the sulfur-sensitive emission
control equipment. Further this approach would likely require some means to generate a
favorable pressure differential in order to allow the blow-by gases to flow into the exhaust.
Given the available means to control crankcase emissions, we have eliminated this
exception.
6. The Complete System
We expect that the technologies described above would be integrated into a complete
emission control system optimized for cost, reliability and package size. The engine-out
emissions will be balanced with the exhaust emission control package in such a way that the
result is the most beneficial from a cost, fuel economy and emissions standpoint. The engine-out
exhaust characteristics will also have a role in assisting the exhaust emission control devices
used. The NOx adsorber, for instance, will require periods of oxygen-depleted exhaust flow in
order to accomplish NOx regeneration and to allow for sulfur control using desulfation events.
This may be most efficiently done by reducing the air-fuel ratio that the engine is operating under
during the regeneration to reduce the oxygen content of the exhaust, or alternatively by
partitioning the exhaust flow such that only a small portion of the exhaust flow is used for NOx
regeneration, thereby reducing the amount of oxygen needing to be depleted through fuel
addition. Further, it is envisioned that the PM device will be integrated into the exhaust system
HI-79
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
upstream of the NOx reduction device. This placement would allow the PM trap to take
advantage of the engine-out NOx as an oxidant for the particulate, while removing the particulate
so that the NOx exhaust emission control device will not have to deal with large PM deposits
which may cause a deterioration in performance. Further it allows the NOx adsorber to make use
of the upstream PM filter as a pre-catalyst to oxidize some NO to NO2 and to partially oxidize the
reductant (diesel fuel or exhaust hydrocarbons) to a more desirable reductant form such as CO
before entering the NOx adsorber. Of course, there is also the possibility of integrating the PM
and NOx exhaust emission control devices into a single unit to replace a muffler and save space
(Toyota's DNPR system being an example of this approach).102 The final component in any of
these system configurations is likely to be some form of clean up catalyst which can provide
control of HC slip during NOx regeneration as well as H2S slip during SOx regeneration.
Particulate free exhaust may also allow for new options in EGR system design to optimize its
efficiency.
We expect that the emission reduction efficiency of the exhaust emission control system
will vary across the NTE zone as a function of exhaust temperature and space velocity."1
Consequently, to maintain the NTE emission cap, the engine-out emissions would have to be
calibrated with exhaust emission control system performance characteristics in mind. This
would be accomplished by lowering engine-out emissions where the exhaust emission control
system was less efficient, for example by retarding fuel injection timing or increasing the EGR
rate. Conversely, where the exhaust emission control system is very efficient at reducing
emissions, the engine-out emissions could be tuned for higher emissions and better fuel
economy. These trade-offs between engine-out emissions and exhaust emission control system
performance characteristics are similar to those of gasoline engines with three-way catalysts in
today's light-duty vehicles and can be overcome through similar system based engineering
solutions. Managing and optimizing these trade-offs will be crucial to effective implementation
of exhaust emission control devices on diesel applications.
In considering how these technologies might be integrated into a complete system, we
have carefully considered the safety aspects of the above system. Based on our understanding of
these technologies and the fact that we received no substantive comments on safety, we are
confident that there are no undue safety concerns associated with the system. Given the proper
diesel fuel sulfur level, actual field data have shown that PM traps function properly in-use
without plugging. As for the NOx adsorber system, there is nothing about the expected system
that causes concern for safety. Injection of diesel fuel upstream of the NOx adsorber for the
purpose of NOx regeneration or desulfation presents no safety concern given the low volatility of
diesel fuel.
The term, "space velocity," is a measure of the volume of exhaust gas that flows through a device.
ni-so
-------
Chapter III: Emissions Standards Feasibility
7. The Need for Low Sulfur Diesel Fuel
In discussing in the preceding sections the technologies that we expect to be needed in
order to meet the stringent emissions standards set in this rulemaking, we have described in some
detail the impact that sulfur has on these technologies. Because of the importance of the fuel
sulfur control portion of this rulemaking, this section will provide a comprehensive overview of
the need for low sulfur diesel fuel to enable the technologies capable of achieving the heavy-duty
vehicle emission standards.
In order to evaluate the effect of sulfur on diesel exhaust control technologies we
identified three key factors which we used to categorize the impact of sulfur in fuel on emission
control function. These factors were efficiency, reliability, and fuel economy. Taken together
these three factors lead us to believe that diesel fuel sulfur levels of 15 ppm will be required in
order to make feasible the heavy-duty vehicle emission standards. Brief summaries of our
analyses for each of these factors are provided below.
Efficiency: The efficiency of emission control technologies to reduce harmful pollutants is
directly affected by sulfur in diesel fuel. Initial and long term conversion efficiencies for NOx,
HC, CO and diesel PM emissions are significantly reduced by sulfur poisoning of the catalyst.
NOx conversion efficiencies with the NOx adsorber technology in particular are dramatically
reduced in a very short time due to sulfur poisoning of the NOx storage bed. In addition, total
PM control efficiency is adversely impacted by the formation of sulfate PM. As explained in
detail in the following sections, all of the advanced NOx and PM technologies described here
have the potential to make significant amounts of sulfate PM under operating conditions typical
of heavy-duty vehicles. The formation of sulfate PM is likely to be in excess of the total PM
standard for diesel fuel sulfur levels above 15 ppm. Based on the strong negative impact of
sulfur on emission control efficiencies for all of the technologies evaluated, we believe that 15
ppm represents an upper threshold for diesel fuel sulfur.
Reliability: Reliability refers to the expectation that emission control technologies must continue
to function as required under all operating conditions for the life of the vehicle. As discussed in
the following sections, sulfur in diesel fuel can prevent proper operation of both NOx and PM
control technologies. This can lead to permanent loss in emission control effectiveness and even
catastrophic failure of the systems. Sulfur in diesel fuel impacts reliability by decreasing catalyst
efficiency (poisoning of the catalyst), increasing diesel PM loading on CDPFs, and by negatively
impacting system regeneration functions. Among the most serious reliability concerns with
sulfur levels greater than 15 ppm are those associated with failure to properly regenerate. In the
case of the NOx adsorber, failure to regenerate will lead to rapid loss of NOx emission control as
a result of sulfur poisoning of the NOx adsorber bed. In the case of the CDPF, sulfur in the fuel
reduces the reliability of the regeneration function. If regeneration does not occur, catastrophic
failure of the CDPF could occur . It is only by the availability of very low sulfur diesel fuels that
HI-81
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
these technologies become feasible. The analysis given in the following section indicates that
diesel fuel sulfur levels of 15 ppm are needed in order to ensure robust operation of the
technologies we believe will be needed to meet the standards under the variety of operating
conditions anticipated to be experienced in the field.
Fuel Economy: Fuel economy impacts due to sulfur in diesel fuel are associated with both NOx
and PM control technologies. The NOx adsorber sulfur regeneration cycle (desulfurization or
desulfation cycle) can consume significant amounts of fuel unless fuel sulfur levels are very low.
The larger the amount of sulfur in diesel fuel, the greater this impact on fuel economy. As sulfur
levels increase above 15 ppm, our projected fuel economy impact quickly transitions above one
percent and doubles with each doubling of fuel sulfur level. Likewise, CDPF regeneration is
inhibited by sulfur in diesel fuel. This leads to increased PM loading in the CDPF and increased
work to pump exhaust across this restriction. With very low sulfur diesel fuel, CDPF
regeneration can be optimized to give a lower (on average) exhaust backpressure and thus better
fuel economy. Thus, for both NOx and PM technologies the lower the fuel sulfur level the
better.
a. Catalyzed Diesel Particulate Filters and the Need for Low Sulfur Fuel
As discussed earlier in this section, un-catalyzed diesel particulate filters require exhaust
temperatures in excess of 650°C in order for the collected PM to be oxidized by the oxygen
available in diesel exhaust. That temperature threshold for oxidation of PM by exhaust oxygen
can be decreased to 450°C through the use of base metal catalytic technologies. Unfortunately,
for a broad range of operating conditions diesel exhaust is significantly cooler than 400°C. If
oxidation of the trapped PM could be assured to occur at exhaust temperatures lower than 300°C,
then diesel particulate filters will be expected to be robust for most applications and operating
regimes. The only means that we are aware of to ensure oxidation of PM (regeneration of the
CDPF) at such low exhaust temperatures is by using oxidants which are more readily reduced
than oxygen. One such oxidant is NO2.
NO2 can be produced in diesel exhaust through the oxidation of the nitrogen monoxide
(NO), created in the engine combustion process, across a catalyst. The resulting NO2-rich
exhaust is highly oxidizing in nature and can oxidize trapped diesel PM at temperatures as cool
as 250°C.103 Platinum is the primary catalyst used to promote the oxidation of NO to NO2.
Therefore in order to ensure passive regeneration of the diesel particulate filters, significant
amounts of platinum are being used in the wash-coat formulations of advanced catalyzed diesel
particulate filters (CDPFs). The use of platinum to promote the oxidation of NO to NO2
introduces several system vulnerabilities affecting both the durability and the effectiveness of the
CDPF when sulfur is present in diesel exhaust. The two primary mechanisms by which sulfur in
diesel fuel limits the robustness and effectiveness of CDPFs are inhibition of regeneration (as a
result of inhibition of the oxidation of NO to NO2) and a dramatic loss in total PM control
HI-82
-------
Chapter III: Emissions Standards Feasibility
effectiveness due to the formation of sulfate PM. Unfortunately these two mechanisms tradeoff
against one another in the design of CDPFs. Changes to improve the reliability of regeneration
by increasing catalyst loadings lead to increased sulfate emissions and thus loss of PM control
effectiveness. Conversely, changes to improve PM control by reducing the use of platinum
group metals and, therefore, limiting sulfate make leads to less reliable regeneration. In our
view, the only means of achieving good PM emission control and reliable operation is to reduce
sulfur in diesel fuel to 15 ppm , as shown in the following subsections.
/'. Inhibition of CDPF Regeneration Due to Sulfur
The passively regenerating CDPF technologies rely on the generation of a very strong
oxidant, NO2, to ensure that the elemental carbon captured by the CDPF's filtering media is
oxidized under normal operating conditions. NO2 is produced through the oxidation of NO in the
exhaust across a platinum catalyst. This oxidation is inhibited by sulfur poisoning of the
catalytic metals.104 This inhibition limits the total amount of NO2 available for oxidation of the
trapped diesel PM, thereby raising the minimum exhaust temperature required to ensure CDPF
regeneration. The balance point temperature is the temperature at which PM accumulation
matches the PM oxidation rate in a CDPF. In other words, the lowest temperature at which the
CDPF would never plug due to PM buildup. Figure ni.A-18 shows that going from three ppm
sulfur fuel to 30 ppm sulfur fuel significantly increases the balance point of these CDPFs"
through inhibition of the NO2 conversion process.105 This seemingly small change in balance
point temperature (approximately 10 percent) is significant because temperatures in the range
shown here are representative of likely exhaust temperatures for many diesel vehicles under
normal driving cycles. Were typical exhaust temperatures in excess of 400°C for most engine
operating conditions, this change would be less important. Without sufficient NO2, the amount
of PM trapped in the CDPF will continue to increase and can lead to excessive exhaust back
pressure, low engine power, and even catastrophic failure of the CDPF itself.
n CR-DPF in the figure refers to a continuously regenerating diesel paniculate filter, CDPF refers to a
catalyzed diesel paniculate filter. Both devices are nearly functionally identical, and the term CDPF is used for
either device in the text.
m-83
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
450
400
U
n 3 PPM SULFUR
•30 PPM SULFUR
Figure IILA-18. Effect of Fuel Sulfur on Regeneration Temperature
Full field test evaluations and retrofit applications of these catalytic trap technologies are
occurring in parts of Europe where low sulfur diesel fuel is already available.0 The experience
gained in these field tests helps to clarify the need for low sulfur diesel fuel. In Sweden and
some European city centers where below 10 ppm diesel fuel sulfur is readily available, more than
3,000 catalyzed diesel particulate filters have been introduced into retrofit applications without a
single failure. Given the large number of vehicles participating in these test programs, the
diversity of the vehicle applications which included intercity trains, airport buses, mail trucks,
city buses and garbage trucks, and the extended time periods of operation (some vehicles have
been operating with traps for more than 5 years and in excess of 300,000 miles), there is a strong
indication of the robustness of this technology on 10 ppm low sulfur diesel fuel.106 The field
experience in areas where sulfur is capped at 50 ppm has been less definitive. In regions without
extended periods of cold ambient conditions, such as the United Kingdom, field tests on 50 ppm
cap low sulfur fuel have also been positive, matching the durability at 10 ppm, although sulfate
0 Through tax incentives 50 ppm cap sulfur fuel is widely available in the United Kingdom and 10 ppm
sulfur fuel is available in Sweden and in certain European city centers.
HI-84
-------
Chapter III: Emissions Standards Feasibility
PM emissions are much higher. However, field tests on 50 ppm fuel in Finland, where colder
winter conditions are sometimes encountered (similar to many parts of the United States),
showed a significant number of failures (-10 percent) due to trap plugging. This 10 percent
failure rate has been attributed to insufficient trap regeneration due to fuel sulfur in combination
with low ambient temperatures.107 Other possible reasons for the high failure rate in Finland
when contrasted with the Swedish experience appear to be unlikely. The Finnish and Swedish
fleets were substantially similar, with both fleets consisting of transit buses powered by Volvo
and Scania engines in the 10 to 11 liter range. Further, the buses were operated in city areas and
none of the vehicles were operated in northern extremes such as north of the Arctic Circle.108
Given that the fleets in Sweden and Finland were substantially similar, and given that ambient
conditions in Sweden are expected to be similar to those in Finland, we believe that the increased
failure rates noted here are due to the higher fuel sulfur level in a 50 ppm cap fuel versus a 10
ppm cap fuel.p Testing on an even higher fuel sulfur level of 200 ppm was conducted in
Denmark on a fleet of 9 vehicles. In less than six months all of the vehicles in the Danish fleet
had failed due to trap plugging.109 The failure of some fraction of the traps to regenerate when
operated on fuel with sulfur caps of 50 ppm and 200 ppm is believed to be primarily due to
inhibition of the NO to NO2 conversion as described here. Similarly the increasing frequency of
failure with higher fuel sulfur levels is believed to be due to the further suppression of NO2
formation when higher sulfur level diesel fuel is used.
The failure mechanisms experienced by CDPFs due to low NO2 availability vary
significantly in severity and long term consequences. In the most fundamental sense, the failure
is defined as an inability to oxidize the stored PM at a rate fast enough to prevent net
accumulation of the PM over time. The excessive accumulation of PM over time blocks the
passages through the filtering media, making it more restrictive to exhaust flow. The exhaust
pressure upstream of the CDPF must increase in order to continue to force the exhaust through
the now more restrictive filter. This increase in exhaust pressure is commonly referred to as
increasing "exhaust backpressure" on the engine.
The increased exhaust backpressure represents increased work being done by the engine
to force the exhaust gas through the increasingly restrictive CDPF. Unless the CDPF is
frequently cleansed of the trapped PM, this increased work can lead to reductions in engine
performance and increases in fuel consumption. This loss in performance may be noted by the
vehicle operator in terms of poor acceleration and generally poor driveability of the vehicle. This
p The average temperature in Helsinki, Finland, for the month of January is 21°F. The average
temperature in Stockholm, Sweden, for the month of January is 26°F. The average temperature at the University of
Michigan in Ann Arbor, Michigan, for the month of January is 24°F. The temperatures reported here are from
www.worldclimate.com based upon the Global Historical Climatology Network (GHCN) produced jointly by the
National Climatic Data Center and Carbon Dioxide Information Analysis Center at Oak Ridge National Laboratory
(ORNL).
m-85
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
progressive deterioration of engine performance as more and more PM is accumulated in the
filter media is often referred to as "trap plugging." Whether trap plugging occurs, and the speed
at which it occurs, will be a function of many variables in addition to the fuel sulfur level; these
variables include the vehicle application, its duty cycle, and ambient conditions. However, if the
fuel sulfur level is sufficient to prevent CDPF regeneration in some real world conditions
experienced, trap plugging could theoretically occur with just one fill-up.q This is not to imply
that any time a vehicle is refueled once with high sulfur fuel trap plugging will occur. In fact, we
believe the likelihood of a single misfueling event causing failure of the CDPF to be small,
because adverse driving conditions (low duty cycle and very cold ambient conditions) would also
have to occur while the fuel is in the vehicle. Rather it is important to know that the use of fuel
with sulfur levels higher than 15 ppm significantly increases the chances of CDPF failure.
Catastrophic failure of the CDPF can occur when excessive amounts of PM are trapped in
the CDPF due to a lack of NO2 for oxidation. This failure occurs when excessive amounts of
trapped PM begin to oxidize at high temperatures (i.e., CDPF regeneration temperatures of
>1000°C) leading to a "run-away" combustion of the PM. This can cause temperatures in the
filter media to increase in excess of that which can be tolerated by the CDPF itself. For the
cordierite material commonly used as the trapping media for CDPFs, the high thermal stresses
caused by the high temperatures can cause the material to crack or melt. This can allow
significant amounts of the diesel PM to pass through the CDPF without being captured during
the remainder of the vehicle's life. That is, the CDPF is destroyed and PM emission control is
lost.
As shown above, sulfur in diesel fuel inhibits NO oxidation leading to increased exhaust
backpressure, reduced fuel economy, and compromised reliability. We, therefore, believe that in
order to ensure reliable and economical operation over a wide range of expected operating
conditions a diesel fuel sulfur level of 15 ppm will be needed. With these very low sulfur levels
we believe, as demonstrated by experience in Europe, that CDPFs will prove to be both durable
and effective at controlling diesel PM emissions to the very low levels required by this standard.
/'/'. Loss of PM Control Effectiveness
In addition to inhibiting the oxidation of NO to NO2, the sulfur dioxide (SO2) in the
exhaust stream is itself oxidized to sulfur trioxide (SO3) at very high conversion efficiencies, by
q Assuming a 10 liter engine, that a CDPF is plugged when it accumulates 7 g/1 of CDPF volume, the
CDPF is two times the volume of the engine, the engine emits 0.1 g/hp-hrPM, the CDPF does not regenerate, and a
HD engine produces 3.013 hp-hr/mi (fromMOBILE6). ThenPM is emitted at a rate of 0.1 g/hp-hr times 3.013 hp-
hr/mi, or 0.3 g/mi. Given that the CDPF can contain 7g/l times 10 1 times 2, or 140 g of PM, then the CDPF will
plug in 140 g PM divided by 0.3 g/mi, or 462 miles. HD trucks typically have a cruising range of more than 500
miles, so it is conceivable that the CDPF could plug in as little as one tank of fuel.
HI-86
-------
Chapter III: Emissions Standards Feasibility
the precious metals in the CDPFs. The SO3 serves as a precursor to the formation of hydrated
sulfuric acid (H2SO4+H2O), or sulfate PM, as the exhaust leaves the vehicle tailpipe. Virtually all
of the SO3 is converted to sulfate under dilute exhaust conditions in the atmosphere as well in the
dilution tunnel used in heavy-duty engine testing. The sulfate formed in the dilution tunnel is
then collected and measured as part of the total PM. Since virtually all sulfur present in diesel
fuel is converted to SO2, the precursor to SO3, as part of the combustion process, the total sulfate
PM is directly proportional to the amount of sulfur present in diesel fuel. Therefore, even though
CDPFs are very effective at trapping and/or oxidizing the elemental carbon and the SOF portions
of the total PM, the overall PM reduction efficiency of CDPFs drops off rapidly with increasing
sulfur levels due to the production of sulfate PM (i.e., "sulfate make," see Figures HI. A-l and
IH.A-2).
SO2 oxidation is promoted across a catalyst in a manner very similar to the oxidation of
NO, except it is converted at higher rates (Figure ni.A-19r), with peak conversion rates in excess
of 50 percent (Table in.A-5).110 The SO2 oxidation rate for a platinum based oxidation catalyst
typical of the type which might be used in conjunction with, or as a washcoat on, a CDPF can
vary significantly with exhaust temperature. At the low temperatures typical of some urban
driving and the heavy-duty federal test procedure (FID-FTP), the oxidation rate is relatively low,
perhaps no higher than ten percent. However at the higher temperatures that might be more
typical of non-urban highway driving conditions and the supplemental emission test (SET, also
called the EURO HI or 13 mode test), the oxidation rate may increase to 50 percent or more.
These high levels of sulfate make across the catalyst are in contrast to the very low SO2 oxidation
rate typical of diesel engines (less than 2 percent). This variation in expected diesel exhaust
temperatures means that there will be a corresponding range of sulfate production expected
across a CDPF.
r In Figure III.A-19, the legend shows values of 42,000 hr"1 and 373,000 hr"1. These values refer to "space
velocity," which is a measure of the volume of exhaust gas that flows through a device; these can be taken to mean
"low flow rate" at 42,000 hf1 and "high flow rate" at 373,000 hf'.
HI-87
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
100
90
80
—*— NO Conversion 42,000 hr-1
-•-SO2 Conversion 42,000 hr-1
-A-NO Conversion, 373,000 hr-1
-X-SO2 Conversion, 373,000 hr-1
250
300
350
Temperature (C)
400
450
500
Figure III.A-19. NO and SO2 Conversion Rates Over Platinum
m-88
-------
Chapter III: Emissions Standards Feasibility
Table III.A-5. SO2 Oxidation Rates for a Platinum Oxidation Catalyst
at the Indicated Catalyst Inlet Temperatures
Catalyst
Temperature
200°C
250°C
300°C
350°C
400°C
450°C
SO 2 Oxidation*
Rate
1-3%
4-11%
10-45%
20-80%
30-90%
40-90%
Operation Represented
Idle, very low load
HD-FTP some Urban Driving
EURO IE some Rural Driving
EURO IE some Rural Driving
EURO IE some Rural Driving
Peak Torque and Rated Conditions
* Range in oxidation rates accounts for variations in exhaust flow through the CDPF, at very
high flow rates SO2 oxidation is minimized and at low flow rates SO2 oxidation is maximized.
The US Department of Energy in cooperation with industry conducted a study entitled
Diesel Emission Control Sulfur Effects (DECSE) to provide insight into the relationship between
advanced emission control technologies and diesel fuel sulfur levels. Interim report number four
of this program gives the total PM emissions from a heavy-duty diesel engine operated with a
CDPF on several different fuel sulfur levels. A straight line fit through this data is presented in
Table in.A-6 below showing the expected total direct PM emissions from a heavy-duty diesel
engine on the supplemental steady state test cycle.8
s Note that direct emissions are those pollutants emitted directly from the engine or from the tailpipe
depending on the context in which the term is used, and indirect emissions are those pollutants formed in the
atmosphere through the combination of direct emissions and atmospheric constituents.
HI-89
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table III.A-6. Estimated PM Emissions from a Heavy-Duty Diesel Engine
at the Indicated Average Fuel Sulfur Levels
Fuel Sulfur
[ppm]
O
7*
15*
30
150
Supplemental Steady State *
Tailpipe PM
[g/bhp-hr]
0.003
0.006
0.009
0.017
0.071
Total PM Increase
Relative to 3 ppm Sulfur
Test Point
--
100%
200 %
470 %
2300 %
* The PM emissions at these sulfur levels are estimated based on a straight-line fit to the
DECSE program data; PM emissions at other sulfur levels are actual DECSE data.
in
Table IHA-6 makes it clear that there are significant PM emission reductions possible
with the application of CDPFs and low sulfur diesel fuel. At the observed sulfate PM
conversion rates, the DECSE program results show that the PM standard is feasible for CDPF-
equipped engines operated on fuel with a sulfur level at or below 15 ppm. The results also show
that CDPF control effectiveness is rapidly degraded at higher diesel fuel sulfur levels due to the
high sulfate PM make observed with this technology. It is clear that PM reduction efficiencies
are limited by sulfur in diesel fuel and that, in order to realize the PM emissions benefits sought
in this rule, diesel fuel sulfur levels must be very low.
Hi. Increased Maintenance Cost for Catalyzed Diesel Particulate Filters Due to
Sulfur
In addition to the direct performance and durability concerns caused by sulfur in diesel
fuel, it is also known that sulfur can lead to increased maintenance costs, shortened maintenance
intervals, and poorer fuel economy for CDPFs. CDPFs are highly effective at capturing the
inorganic ash produced from metallic additives in engine oil. This ash is accumulated in the
CDPF and is not removed through oxidation, unlike the trapped carbonaceous PM. Periodically
the ash must be removed by mechanical cleaning of the CDPF with compressed air or water.
This maintenance step is anticipated to occur on intervals of well over one hundred thousand
miles. However, sulfur in diesel fuel increases this ash accumulation rate through the formation
of metallic sulfates in the CDPF, which increases both the size and mass of the trapped ash. By
increasing the ash accumulation rate the sulfur shortens the time interval between the required
maintenance of the CDPF and negatively impacts fuel economy.
HI-90
-------
Chapter III: Emissions Standards Feasibility
b. Diesel NOx Catalysts and the Need for Low Sulfur Fuel
All of the NOx aftertreatment technologies discussed previously in chapter in are
expected to utilize platinum to oxidize NO to NO2 to improve the NOx reduction efficiency of
the catalysts at low temperatures or as in the case of the NOx adsorber, as an essential part of the
process of NOx storage. This reliance on NO2 as an integral part of the reduction process means
that the NOx aftertreatment technologies, like the PM aftertreatment technologies, will have
problems with sulfur in diesel fuel. In addition NOx adsorbers have the added constraint that the
adsorption function itself is blocked by the presence of sulfur. These limitations due to sulfur in
the fuel affect both overall performance of the technologies and, in fact, the very feasibility of the
NOx adsorber technology.
/'. Sulfur Poisoning (Sulfate Storage) on NOx Adsorbers
The NOx adsorber technology relies on the ability of the catalyst to store NOx as a nitrate
on the surface of the catalyst, or adsorber (storage) bed, during lean operation. Because of the
similarities in chemical properties of SOx and NOx, the SO2 present in the exhaust is also stored
by the catalyst surface as a sulfate. The sulfate compound that is formed is significantly more
stable than the nitrate compound and is typically not released and reduced during the NOx release
and reduction step. Since the NOx adsorber is highly effective at capturing SO2 in the adsorber
bed, the poisoning of the catalyst occurs rapidly. As a result, sulfate compounds quickly occupy
all of the NOx storage sites on the catalyst thereby rendering the catalyst ineffective for NOx
reduction (poisoning the catalyst). Figure ni.A-20 clearly illustrates this effect at 3, 16, and 30
ppm fuel sulfur levels.112
HI-91
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
3 16 30
Fuel Sulfur Level (ppm)
Figure III.A-20. Diesel Fuel Sulfur Effect on NOx Adsorber
Performance after 150 hours
The stored sulfur compounds can be removed by exposing the catalyst to hot (>650°C)
and rich (air-fuel ratio below the stoichiometric ratio of 14.5 to one) conditions for a brief
period.113 Under these conditions, the stored sulfate is released and reduced in the catalyst.114115
While research to date on this procedure has been very favorable with regards to sulfur removal
from the catalyst, it has revealed a related vulnerability of the NOx adsorber catalyst. Under the
high temperatures used for desulfation, the metals that make up the storage bed can change in
physical structure. This leads to lower precious metal dispersion, or "metal sintering," (a less
even distribution of the catalyst sites) reducing the effectiveness of the catalyst.116 This
degradation of catalyst efficiency due to high temperatures is often referred to as thermal
degradation. Thermal degradation is known to be a cumulative effect. That is, with each
excursion to high temperature operation, some additional degradation of the catalyst occurs.
One of the best ways to limit thermal degradation is by limiting the accumulated number
of desulfation events over the life of the vehicle. Since the period of time between desulfation
events is expected to be determined by the amount of sulfur accumulated on the catalyst (the
higher the sulfur accumulation rate, the shorter the period between desulfation events) the
desulfation frequency is expected to be proportional to the fuel sulfur level. In other words for
each doubling in the average fuel sulfur level, the frequency and accumulated number of
desulfation events are expected to double. We believe, therefore, that the diesel fuel sulfur level
HI-92
-------
Chapter III: Emissions Standards Feasibility
must be set as low as possible in order to limit the frequency and duration of desulfation events.
Without control of fuel sulfur levels below 15 ppm, we can no longer conclude with sufficient
confidence that sulfur poisoning can be controlled without unrecoverable thermal degradation.
Some would argue that the NOx adsorber technology could meet the NOx standard using diesel
fuel with a 30 ppm average sulfur level. This would imply that the NOx adsorber could tolerate
more than a three to four fold increase in desulfation frequency (when compared to an expected
fuel sulfur level of 7 to 10 ppm with a 15 ppm cap) without any increase in thermal degradation.
This conclusion is inconsistent with our understanding of the technology that, with each
desulfation event, some thermal degradation occurs. Therefore, we believe that diesel fuel sulfur
levels must be at or below 15 ppm in order to limit the number and frequency of desulfation
events. Limiting the number and frequency of desulfation events will limit thermal degradation
and, thus, enable the NOx adsorber technology to meet the NOx standard. For additional
discussion of thermal degradation refer to the previous discussion in section ni.A.S.b.vii on NOx
adsorber durability.
Sulfur in diesel fuel for NOx adsorber equipped engines will also have an adverse effect
on fuel economy. The desulfation event requires controlled operation under hot and net fuel rich
exhaust conditions. These conditions, which are not part of a normal diesel engine operating
cycle, can be created through the addition of excess fuel to the exhaust. This addition of excess
fuel causes an increase in fuel consumption. We have developed a spreadsheet model that
estimates the frequency of desulfation cycles from published data and then estimates the fuel
economy impact from this event.117 Table ni.A-7 shows the estimated fuel economy impact for
desulfation of a NOx adsorber at different fuel sulfur levels assuming a desired 90 percent NOx
conversion efficiency. The estimates in the table are based on assumed average fuel sulfur levels
associated with different sulfur level caps. Note that, although we can estimate the fuel
consumption penalty of operation on diesel fuel sulfur levels higher than 15 ppm, this analysis
does not consider the higher degree of thermal degradation due to the more frequent desulfation
events which are required for operation on these higher sulfur levels.
HI-93
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table III.A-7. Estimated Fuel Economy Impact from
Desulfation of a 90 Percent Efficient NOx Adsorber
Fuel Sulfur Cap
[ppm]
500
50
25
15
5
Average Fuel Sulfur
[ppm]
350
30
15
7
2
Fuel Economy
Penalty [%]
27
2
1
<1
<«1
The table shows that the fuel economy penalty associated with sulfur in diesel fuel is
noticeable even at average sulfur levels as low as 15 ppm and increases rapidly with higher sulfur
levels. It also shows that the 15 ppm sulfur cap will be expected to result in a fuel economy
impact of less than one percent absent other changes in engine design.
As a consequence of requiring desulfation to occur before the NOx adsorber catalyst
degrades to a level below 90 percent, the fuel economy impacts at higher sulfur levels described
here are substantial. Therefore it would be logical to consider the possibility of allowing further
degradations in NOx performance (below 90 percent) before desulfation in order to reduce this
fuel economy impact. Recent results from industry contradict that position, however, indicating
that when deep poisoning of the catalyst occurs due to higher fuel sulfur levels (or presumably
extend periods of poisoning without desulfation) the ability of the catalyst to recover from the
sulfur poisoning is compromised.118 This data from a gasoline direct injection application
indicates that desulfation events sequenced on a fixed interval with only minimal poisoning
allowed for full recovery of NOx performance (eight ppm sulfur fuel, regenerated on a fixed
driving cycle with 32,000 km of vehicle operation). These good results are contrasted with
performance on 30 ppm sulfur fuel in which NOx adsorber desulfation occurred on the same
fixed interval (thus allowing greater levels of poisoning before desulfation). For this case NOx
control performance was never fully recovered at each desulfation step and, therefore, continued
to gradually decrease over time from an initial efficiency of 95 percent to 80 percent over the
same 32,000 km of vehicle operation.
Future improvements in the NOx adsorber technology are expected and needed if the
technology is to provide the environmental benefits we have projected today. Some of these
improvements are likely to include improvements in the means and ease to remove stored sulfur
from the catalyst bed. However, because the stored sulfate species are inherently more stable
HI-94
-------
Chapter III: Emissions Standards Feasibility
than the stored nitrate compounds (from stored NOx emissions), we expect that a separate release
and reduction cycle (desulfurization cycle) will always be needed in order to remove the stored
sulfur. Therefore, we believe that fuel with a sulfur level at or below 15 ppm sulfur will be
necessary in order to control thermal degradation of the NOx adsorber catalyst and to limit the
fuel economy impact of sulfur in diesel fuel.
/'/'. Paniculate Sulfate Production for NOx Control Technologies
The NOx adsorber technology relies on a platinum based oxidation function in order to
ensure high NOx control efficiencies. As discussed more fully in section HI.A.T.a, platinum
based oxidation catalysts form sulfate PM from sulfur in the exhaust gases significantly
increasing PM emissions when sulfur is present in the exhaust stream. The NOx adsorber
technology relies on the oxidation function to convert NO to NO2 over the catalyst bed. For the
NOx adsorber this is a fundamental step prior to the storage of NO2 in the catalyst bed as a
nitrate. Without this oxidation function the catalyst will only trap that small portion of NOx
emissions from a diesel engine which is NO2. This would reduce the NOx adsorber effectiveness
for NOx reduction from in excess of 90 percent to something well below 20 percent. The NOx
adsorber relies on platinum to provide this oxidation function due to the need for high NO
oxidation rates under the relatively cool exhaust temperatures typical of diesel engines. Because
of this fundamental need for a catalytic oxidation function, the NOx adsorber inherently forms
sulfate PM when sulfur is present in diesel fuel, since sulfur in fuel invariably leads to sulfur in
the exhaust stream.
The Compact-SCR technology, like the NOx adsorber technology, uses an oxidation
catalyst to promote the oxidation of NO to NO2 at the low temperatures typical of much of diesel
engine operation. As discussed above, there are substantial questions regarding the ability of
SCR systems to be implemented successfully to meet the requirements finalized today. By
converting a portion of the NOx emissions to NO2 upstream of the ammonia SCR reduction
catalyst, the overall NOx reductions are improved significantly at low temperatures. Without this
oxidation function, low temperature SCR NOx effectiveness is dramatically reduced making
compliance with the NOx standard impossible. As discussed previously in section HI. A.7,
platinum is known to be a good catalyst to promote NO oxidation, even at low temperatures.1
Therefore, future Compact-SCR systems would need to rely on a platinum oxidation catalyst in
order to provide the required NOx emission control. This use of an oxidation catalyst in order to
enable good NOx control means that Compact SCR systems will produce significant amounts of
sulfate PM when operated on anything but the lowest fuel sulfur levels due to the oxidation of
SO2 to sulfate PM promoted by the oxidation catalyst.
4 Platinum group metals include platinum, palladium, rhodium, and other precious metals.
HI-95
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Without the oxidation catalyst promoted conversion of NO to NO2, neither of these NOx
control technologies can meet the NOx standard set here. Therefore each of these technologies
will require low sulfur diesel fuel to control the sulfate PM emissions inherent in the use of
oxidation catalysts. The NOx adsorber technology may be able to limit its impact on sulfate PM
emissions by releasing stored sulfur as SO2 under rich operating conditions. The Compact-SCR
technology, on the other hand, has no means to limit sulfate emissions other than through lower
catalytic function or lowering sulfur in diesel fuel. The degree to which the NOx emission
control technologies increase the production of sulfate PM through oxidation of SO2 to SO3
varies somewhat from technology to technology, but it is expected to be similar in magnitude and
environmental impact to that for the PM control technologies discussed previously in section
lU.A.T.a.ii, since both the NOx and the PM control catalysts rely on precious metals to achieve
the required NO to NO2 oxidation reaction.
At fuel sulfur levels below 15 ppm this sulfate PM concern is greatly diminished.
Without this low sulfur fuel, the NOx control technologies are expected to create PM emissions
well in excess of the PM standard regardless of the engine-out PM levels.
c. Contribution of Sulfur from Engine Lubricating Oils
Current engine lubricating oils have sulfur contents which can range from 2,500 ppm to
as high as 8,000 ppm by weight. Since engine oil is consumed by heavy-duty diesel engines in
normal operation, it is important that we account for the contribution of oil derived sulfur in our
analysis of the need for low sulfur diesel fuel. One way to give a straightforward comparison of
this effect is to express the sulfur consumed by the engine as an equivalent fuel sulfur level. This
approach requires that we assume specific fuel and oil consumption rates for the engine.
Assuming that a heavy-duty diesel engine consumes one quart of engine oil in 2,000 miles of
operation, that engine oil sulfur levels range from 2,000 to 8,000 ppm, and that the engine
consumes fuel at a rate of one gallon per six miles of operation, the range of equivalent fuel
sulfur levels can be estimated. Using these assumptions, the estimated range is from two to
seven ppm diesel fuel sulfur equivalence.119 If values at the upper end of this range accurately
reflect the contribution of sulfur from engine oil to the exhaust this would be a concern as it
would represent as much as half of the total sulfur in the exhaust under a 15 ppm diesel fuel
sulfur cap (with an average sulfur level assumed to be approximately seven to 10 ppm).
However, we believe that this simplified analysis, while valuable in demonstrating the need to
investigate this issue further, overstates the likely sulfur contribution from engine oil by a
significant amount. Current heavy-duty diesel engines operate with open crankcase ventilation
systems which "consume" oil by carrying oil from the engine crankcase into the environment.
This consumed oil is correctly included in the total oil consumption estimates, but should not be
included in estimates of oil entering the exhaust system for this analysis, since as currently
applied this oil is not introduced into the exhaust. Thus the assumption of one quart of oil in
2,000 miles of operation being consumed and thus entering the exhaust system overstates the oil
HI-96
-------
Chapter III: Emissions Standards Feasibility
contribution by the fraction of oil that exits from today's open crankcase systems. In the future
we expect diesel engine manufacturers to rely on closed crankcase filtration systems to filter this
oil from the blow-by gases and return the oil to the engine's crankcase, thus lowering engine oil
consumption.
As an alternate approach to estimate the amount of oil and thus oil borne sulfur present in
the exhaust, projected emission rates for Phase 1 technology engines can be made. The Phase 1
HD emission standards set a 0.1 g/bhp-hr PM emission rate for all classes of heavy-duty diesel
vehicles. If we assume that virtually all oil consumed by the engine is emitted as diesel PM and
that this soluble organic fraction (SOF) makes up 30 percent of diesel PM we can estimate how
much oil is consumed. This estimate is made assuming that the engine oil has a sulfur content of
5,000 ppm, that 30 percent of PM emissions are from engine oil, and that the engine brake
specific fuel consumption rate is 0.300 Ibm/bhp-hr. The equivalent fuel sulfur level from engine
oil is then calculated as
sulfur level [ppm] = ( 5.000 [ppml x 0.1 [g/bhp-hrl x 30 F%1 ^
( 454 [g/lbm] x 0.300 [Ibm/bhp-hr] )
A higher fuel consumption rate decreases the relative amount of sulfur from engine oil in this
estimate. Using this approach we have estimated that the equivalent fuel sulfur level from engine
oil is approximately one ppm.
As a further attempt to better understand the amount of sulfur contributed from engine oil
in the exhaust we have looked at the results from the DECSE test program. The DECSE
program reports sulfate emissions from a heavy-duty diesel engine equipped with highly
catalyzed CDPFs and operated on diesel fuel at several fuel sulfur levels. A commonly used
motor oil with sulfur content of approximately 3,500 ppm was chosen for this testing. Since the
PM emission control technologies used in this testing are very sensitive to sulfur (converting
sulfur to sulfate PM emissions at a rate of approximately 40 percent) they should reveal
sensitivities to sulfur from lube oil. By taking the sulfate emission results reported by DECSE at
fuel sulfur levels of 3 and 30 ppm sulfur we can estimate the amount of sulfate emissions (and
thus sulfur contribution) from the engine oil. The intercept (the predicted sulfate emissions at 0
ppm sulfur fuel) of a straight-line fit through the two test points should reveal the amount of
sulfate produced from oil derived sulfur. Figure ni.A-21 shows the results of this analysis.120
The intercept value shown in the figure is slightly below zero indicating that in spite of the high
sulfur conversion rate typical of these emission control devices the amount of lube oil derived
sulfate emissions is unmeasurable. Although some amounts of sulfur from lubricating oils are
almost certainly present in the exhaust, this analysis seems to indicate that it will not be a
significant fraction of the total sulfur even for fuel sulfur levels as low as 15 ppm.
HI-97
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
• CR-DPF Sulfate
• CDPF Sulfate
— CDPF Linear Fit
— CR-DPF Linear Fit
15 20 25 30
Diesel Fuel Sulfur Level (ppm)
Figure III.A-21. Sulfate PM Emissions versus Diesel Fuel Sulfur Level
with 3,500 ppm Sulfur Engine Oil
B. Feasibility of the 2008 Standards for Heavy-Duty
Gasoline Vehicles & Engines
Gasoline emission control technology has evolved rapidly in recent years. 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
uncontrolled emissions. Since then, light-duty gasoline emission standards have undergone two
major reductions, our Tier 1 and Tier 2 standards, and heavy-duty gasoline emission standards
have undergone three changes toward ever lower levels. Despite that, some of today's heavy-
duty vehicle emissions are well below levels necessary to meet the current federal heavy-duty
gasoline standards, the Phase 1 heavy-duty gasoline standards to be implemented in the 2005
model year,121 and the California Low Emission Vehicle (LEV) standards for medium-duty
vehicles." The continuing emissions reductions have been brought about by ongoing
improvements in engine air-fuel management hardware and software plus improvements in
exhaust system and catalyst designs.
11 The Phase 1 heavy-duty program is a reference to the 2004 heavy-duty final rule which set the 2004
model year HD diesel standards and the 2005 model year HD gasoline standards. (See 65 FR 59896, October 6,
2000) The 2007 final rule represents Phase 2 of this heavy-duty standard setting effort.
HI-98
-------
Chapter III: Emissions Standards Feasibility
These improvements to gasoline emission controls have been made in response to the
California LEV-II standards and the new federal Tier 2 standards.122123 Some of this
development work was contributed by EPA in a very short timeframe and with very limited
resources in support of our Tier 2 rulemaking effort.124 These improvements should transfer well
to the heavy-duty gasoline segment of the fleet. Given the dramatic improvements in gasoline
emission control technology in recent years, it is clear that there is no need to invent new
technologies to meet emission levels below the 2005 heavy-duty gasoline standards. Instead,
existing technologies can be applied to heavy-duty gasoline engines more effectively and more
broadly; the focus being on the application and optimization of these existing technologies. With
the migration of light-duty technology to heavy-duty vehicles and engines, we believe that
considerable improvements to heavy-duty gasoline emissions can be realized, thus enabling the
stringent 2008 standards.
The most significant improvement facilitating the low emission levels of today's gasoline
vehicles has been to the traditional three-way catalyst, which now warms up and lights off very
rapidly and is substantially more durable than in the past. Dramatic improvements have been
realized also in fuel metering, which is now far more precise and accurate than previous systems.
Improvements have been made also 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. Equally important, if not more so, is that
emission control calibrations continue to become more refined and sophisticated as calibrators
become more skilled and computing power increases.
Fuel quality also plays an important role in improving vehicle emissions. In our Tier 2
rule for light-duty vehicles and trucks, we required that gasoline sulfur levels be reduced to a 30
ppm average with an 80 ppm cap. This sulfur level reduction is the primary enabler for the Tier
2 standards. Likewise for the 2008 heavy-duty gasoline standards. The Tier 2 gasoline sulfur
reduction that enables the technology needed to meet the Tier 2 standards (0.07 g/mi NOx, on
average) will enable that same technology on heavy-duty gasoline vehicles, thus enabling the
2008 heavy-duty gasoline standards.
1. Gasoline Exhaust Emission Control Technology Descriptions
Table in.B-1 below lists specific types of exhaust emission controls that we project may
be used on heavy-duty gasoline vehicles to meet the 2008 heavy-duty gasoline standards. We do
not believe that all of these technologies would be needed to meet the 2008 standards on every
vehicle. The choices manufacturers make and the 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, we believe that most, if not all, vehicles
will use the technology.
HI-99
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table III.B-1. Exhaust Emission Control Hardware and
Technologies That May be Used to Meet the 2008 Heavy-Duty
Gasoline Standards
Fast Light-Off Exhaust Gas
Oxygen Sensor
Retarded Spark Timing at Start-
Up
More Precise Fuel Control
32-bit Microprocessor
Manifold with Low Thermal
Capacity
Air- Assisted Fuel Injection
Engine Modifications
Secondary Air Injection Exhaust
Heat Optimized Exhaust Pipe
Leak-Free Exhaust System
Close-Coupled Catalyst
Improved Catalyst Washcoats
Increased Catalyst Volume and
Precious Metal Loading
Full Electronic Exhaust Gas
Recirculation
This section discusses in detail some of the technologies that may be used to meet the
2008 standards. The technology descriptions are divided into five categories:
base engine improvements;
• improved fuel control;
improved fuel atomization;
• improved exhaust and exhaust emission control systems; and
• improved engine calibrations.
a. 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
HCs 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"
m-ioo
-------
Chapter III: Emissions Standards Feasibility
combustion chamber designs, multiple valves with variable-valve timing, and exhaust gas
recirculation.
/'. 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."v 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 as an unburned HC. In addition, some components in lubricating oil can
poison the catalyst and reduce its effectiveness. To reduce oil consumption, vehicle
manufacturers are expected to 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. 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"). Further improvements can be realized by positioning 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 thereby reducing NOx formation.
v "Land height" is the distance between the top of the piston and the first piston ring.
m-ioi
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
/'/'. Improved EGR Design
One of the most effective means of reducing engine-out NOx emissions is exhaust gas
recirculation. By recirculating exhaust gases into the combustion chamber, the overall air-fuel
mixture is diluted, lowering peak combustion temperatures and reducing NOx. As discussed
above, the use of high swirl, high turbulence combustion chambers can allow the amount of EGR
to be increased from current levels of 15 to 17 percent to levels possibly as high as 20 to 25
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 may require more precise EGR control to
improve upon NOx emission control. Some manufacturers use a mechanical back-pressure
system that measures 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 heavy-duty gasoline applications
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.
Hi. 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
w 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.
m-102
-------
Chapter III: Emissions Standards Feasibility
intake and exhaust also increases. By doubling the number of intake and exhaust valves,
pumping losses are reduced, improving the volumetric efficiency and useful power output.
In addition to gains in breathing, the multiple-valve (typically 4-valve) design allows the
spark plug to be positioned closer to the center of the combustion chamber (as discussed above)
which decreases the distance the flame must travel inside the chamber. In addition, the two
streams of incoming gas can be used to achieve greater mixing of air and fuel, further increasing
combustion efficiency thereby lowering 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 LEV and ULEV standards, more vehicles will have to make improvements to the
induction system, including the use of variable valve timing.
iv. Leak-Free Exhaust Systems
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-
IE-103
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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.
b. 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 one 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 FIEGO 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. Because of this placement and the decreased susceptibility to aging, we
m-104
-------
Chapter III: Emissions Standards Feasibility
expect the downstream sensor to survive the full life of the vehicle without replacement. 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.
By 2008, most vehicle manufacturers are expected to use a dual oxygen sensor system for
monitoring the catalyst as part of the OBD system required by the Phase 1 heavy-duty rule. As
discussed above, most manufacturers also will use the secondary HEGO sensor for fuel trim (i.e.,
minor adjustments) of the fuel control system. We anticipate that all manufacturers will use this
secondary sensor for fuel trim.
/'/'. Universal Oxygen Sensors
The universal exhaust gas oxygen (UEGO) sensor, also called a "linear oxygen sensor",
could replace conventional HEGO sensors. Conventional HEGO sensors only determine if an
engine's A/F is richer or leaner than stoichiometric, providing no indication of the exact level of
the A/F. In contrast, UEGO's are capable of recognizing both the direction and magnitude of A/F
transients since the voltage output of the UEGO is "proportional" with changing A/F (i.e., each
voltage value corresponds to a certain A/F). Therefore, proportional A/F control is possible with
the use of UEGO sensors, facilitating faster response of the fuel feedback control system and
tighter control of A/F.
Although some gasoline applications currently use UEGO sensors, discussions with
various manufacturers suggest mixed opinions 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 HEGO sensors. Planar HEGO sensors (also known as "fast light-off HEGO sensors) have a
thimble design that is considerably lighter than conventional designs. The main benefits are
shorter heat-up time and faster sensor response.
/'/'/'. Individual Cylinder A/F Control
Another method for tightening fuel control is to control the A/F in each individual
cylinder. Current fuel control systems control the A/F for the entire engine or a bank of
cylinders. By controlling A/F for the entire engine or a bank of cylinders, any necessary
adjustments made to fuel delivery for the engine are applied to all cylinders simultaneously,
regardless of whether all cylinders need the adjustment. For example, there is usually some
deviation in A/F between cylinders. If a particular cylinder is rich, but the "bulk" A/F indication
for the engine is lean, the fuel control system will simultaneously increase the amount of fuel
delivered to all of the cylinders, including the rich cylinder. Thus, the rich cylinder becomes
even richer having a potentially negative effect on the net A/F.
m-105
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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.
iv. Adaptive Fuel Control Systems
The fuel control systems of virtually all current vehicles incorporate a feature known as
"adaptive memory" or "adaptive block learn." Adaptive fuel control systems automatically
adjust the amount of fuel delivered to compensate for component tolerances, component wear,
varying environmental conditions, varying fuel compositions, etc., to more closely maintain
proper fuel control under various operating conditions.
For most fuel control systems in use today, the adaption process affects only steady-state
operation conditions (i.e., constant or slowly changing throttle conditions). Because transient
operating conditions have always provided a challenge to maintaining precise fuel control, the
use of adaptive fuel control for transient operation would be extremely valuable. Accurate fuel
control during transient driving conditions has traditionally been difficult because of inaccuracies
in predicting the air and fuel flow under rapidly changing throttle conditions. Air and fuel
dynamics within the intake manifold (fuel evaporation and air flow behavior), and the time delay
between measurement of air flow and the injection of the calculated fuel mass, result in
temporarily lean A/F during transient operation. Variation in fuel properties, particularly
distillation characteristics, also increases the difficulty in predicting A/F during transients. These
can all lead to poor driveability and an increase in NOx emissions.
v. 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 gasoline applications, it is considered
expensive and those vehicles equipped with the feature are expensive, higher end vehicles.
m-106
-------
Chapter III: Emissions Standards Feasibility
Because of its high cost, it is not anticipated that drive-by-wire technology will become
commonplace in the near future.
c. Improvements in Fuel Atomization
In addition to maintaining a stoichiometric A/F ratio, it is also important that a
homogeneous air-fuel mixture be delivered at the proper time and that the mixture is finely
atomized to provide the best combustion characteristics and lowest emissions. Poorly prepared
air-fuel mixtures, especially after a cold start and during the warm-up phase of the engine, result
in significantly higher emissions of unburned HC since combustion of the mixture is less
complete. By providing better fuel atomization, more efficient combustion can be attained,
which should aid in improving fuel economy and reducing emissions. Sequential multi-point
fuel injection and air-assisted fuel injectors are examples of the most promising technologies
available for improving fuel atomization.
/'. Sequential Multi-Point
Typically, conventional multi-point fuel injection systems inject fuel into the intake
manifold by injector pairs. This means that rather than injecting fuel into each individual
cylinder, a pair of injectors (or even a whole bank of injectors) fires simultaneously, sending fuel
into several cylinders. Since only one of the cylinders is actually ready for fuel at the moment of
injection, the other cylinder(s) gets too much or too little fuel. With this less than optimum fuel
injection timing, fuel puddling and intake manifold wall wetting can occur, both of which can
hinder complete combustion. Sequential injection, on the other hand, delivers a more precise
amount of fuel that is required by each cylinder to each cylinder at the appropriate time. Because
of the emission reductions and other performance benefits "timed" fuel injection offers,
sequential fuel injection systems are very common on today's vehicles and are expected to be
incorporated in all vehicles soon.
/'/'. Air-Assisted Fuel Injectors
Another method used to further homogenize the air-fuel mixture is to use 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.
m-107
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
d. Improvements to Exhaust and Exhaust Emission Control Systems
Over the last five years or so, there have been tremendous advancements in exhaust
emission control systems. Catalyst manufacturers have been 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 adsorber technology. The advancements to
exhaust emission control systems are probably the single most important area of emission control
development.
/'. Catalysts
As previously mentioned, significant changes in catalyst formulation, size and design
have been made in recent years and additional advances in these areas are still possible.
Palladium (Pd) is likely to continue as the precious metal of choice for close-coupled
applications and may 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 gasoline applications. 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, that provide the
surface area support to which the precious metals adhere, and base components (rare earth
metals) such as lanthanum, ceria, and zirconia, that act as promoters and stabilizers while also
encouraging storage and reduction of oxygen.
m-108
-------
Chapter III: Emissions Standards Feasibility
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. Rhodium is primarily used for reducing NOx
emissions. Generally, and preferably, NOx emissions are reduced in the top washcoat layer
while CO and HC are still present. Then, the CO and HC can be oxidized in the bottom
washcoat layer. Figure in.B-1 illustrates the impact coating architecture (multi-layered
washcoat technology) can have on emission performance."
SAE 960802: 1.8 liter 4 cyl; 100 h aged; Pd/Rh=5/l @ 50 g/cu. ft.
• Single layer Pd/Rh
HTwo layer - Pdtop
fjTwo layer- Pd bottom
THC
NOx
Figure III.B-1. Impact of Coating Architecture on
HC and NOx Emissions
Manufacturers also have been developing catalysts with thinner walled substrates that
allow for a higher substrate cell density, and low thermal mass catalysts for close-coupled
applications. These developments improve mass transfer at high engine loads, increase catalyst
surface area, and speed up light-off time during cold starts. The greater the number of cells there
are, the more surface area that exists to which washcoat components and precious metals can
x Figure III.B-1 shows "% breakthrough in European driving cycle" on the y-axis; this can be defined as
the percentage of emissions that pass through the catalyst without being converted to H2O, CO2, and N2 during the
European test cycle.
m-109
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
adhere. This results 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 today's
conventional catalysts are 400 cpsi.
We also have projected that, in order to meet the 2008 heavy-duty gasoline emission
standards, catalyst volumes would have to increase. Current heavy-duty gasoline applications
have catalyst volumes slightly lower than the corresponding engine displacement. We believe
that most heavy-duty gasoline vehicles would likely need to increase catalyst volumes on the
order often percent. 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 also have projected that some level of increased precious metal loading (i.e., catalyst
loading) would be necessary to meet the 2008 heavy-duty gasoline standards. Typical catalyst
loadings for current heavy-duty gasoline applications are four grams/liter (g/L) of catalyst
volume. We believe that, based on input from catalyst suppliers and vehicle manufacturers,
catalysts meeting the 2008 standards would need loadings more on the order of five g/L.
However, catalyst suppliers have also indicated to us that they and vehicle manufacturers are
constantly working on ways to reduce the amount of precious metal loading (a process they refer
to as "thrifting"). Thrifting is achieved in several ways. One of the most common is matching
the catalyst to the attributes of the vehicle. By working in unison, vehicle manufacturers and
catalyst suppliers are able to thrift or reduce the amount of precious metal used in a given
application by attempting to optimize the vehicle fuel control strategy, exhaust mass flow rate,
and exhaust temperature with various catalyst parameters, such as catalyst location, substrate
design, cell density, oxygen storage capability, and precious metal and base metal dispersion, to
name a few. Other methods of thrifting are the constant improvements being made to washcoat
architecture - that is, constant improvement to the materials used in the washcoat formulation so
that the precious metals and other components better adhere to the substrate surface. Finally,
improvements to washcoat application processes also can improve significantly the catalyst
performance while still allowing for 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 precious metal
dispersion means that, rather than having relatively large "clumps" of precious metals unevenly
dispersed throughout the catalyst surface, many smaller precious metal sites are dispersed
uniformly throughout the catalyst surface. This type of dispersion increases the chance for
pollutants to come into contact with the precious metal and thus react into a harmless emission.
Therefore, as thrifting continues, precious metal loading may actually decrease rather than
increase, although this very likely outcome has not been incorporated into our cost estimates
presented in Chapter V of this RIA.
m-no
-------
Chapter III: Emissions Standards Feasibility
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. This results in more heat escaping into the exhaust manifold during the
exhaust stroke while having a negligible impact on fuel economy.125 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.
/'/'. 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 more quickly. 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 a detailed individual
application for each vehicle or engine design.
Hi. Heat Managed Exhaust Systems
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.
e. 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
m-iii
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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.
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 engine and powertrain control modules become 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 us 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 and complying with stringent light-duty LEV and
NLEV standards has not required the use of advanced hardware, such as electrically heated
catalysts or HC adsorbers as some had originally predicted they would need.
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.
2. The 2008 Heavy-Duty Gasoline Exhaust Emission Standards
The 2008 heavy-duty gasoline complete vehicle standards are equivalent to the California
LEV-II program LEV standards for vehicles in the same weight ranges, with the exception of the
PM standards. The 2008 NOx level for 8,500 to 10,000 pound vehicles is 0.2 g/mi and the NOx
level for 10,000 to 14,000 pound vehicles is 0.4 g/mile. The NMHC standards are 0.195 and
0.23 g/mile for the 8,500 to 10,000 pound and 10,000 to 14,000 pound vehicles, respectively.
The California LEV-II LEV standards for PM emissions are 0.12 g/mi, while the new federal
heavy-duty vehicle PM standard will be 0.02 g/mi. The California PM standards were originally
set with consideration given to diesel vehicles not equipped with emission control devices, hence
the much higher level than finalized for federal heavy-duty vehicles.
The 2008 heavy-duty gasoline incomplete vehicle standards, for which California has no
standards as all gasoline vehicles are required by California to certify on the chassis
IH-112
-------
Chapter III: Emissions Standards Feasibility
dynamometer, are 0.20 g/bhp-hrNOx, 0.14 g/bhp-hr NMHC, and 0.01 g/bhp-hr PM. Table ffl.B-
2 shows the new 2008 heavy-duty gasoline exhaust emission standards along with the California
emission standards for the equivalent vehicles. Also shown in Table ni.B-2 are some of the
federal Tier 2 bin levels for the purpose of comparison with the standards for light-duty trucks
and medium-duty passenger vehicles (MDPVs).y
Table III.B-2. Emission Standards for Select Federal and California Gasoline
Vehicles & Engines
Program
Phase 2 Heavy-
Duty Gasoline
Phase-in:
2008: 50%
2009: 100%
Cal LEV-II MDV
(LEV Only)
Phase-in:
2004-2006
2007: 40% B
Tier 2 Light-Duty
Phase-in:
2004-2009
MDPV Bin
available only
thru 2008
Weight Range or
Tier 2 Bin
Vehicles 8,500-10,000 Ibs
(g/mi)
Vehicles 10,000-14,000
Ibs
(g/mi)
Engines > 8,500 Ibs
(g/bhp-hr)
8,500-10,000 Ibs
(g/mi)
10,000-14,000 Ibs
(g/mi)
MDPV Interim Bin
(g/mi)
Final High Bin
(g/mi)
Average Bin
(g/mi)
Useful Life
(mi)
120,000
120,000
110,000
120,000
120,000
120,000
120,000
120,000
NOx
0.2
0.4
0.20
0.2
0.4
0.9
0.20
0.07
NMHCA
0.195
0.230
0.14
0.195
0.230
0.28
0.125
0.09
ECHO
0.032
0.040
N/A
0.032
0.040
0.032
0.018
0.018
PM
0.02
0.02
0.01
0.12
0.12
0.12
0.02
0.01
A Non-methane hydrocarbon (NMHC) or, for LEV-II and Tier 2, non-methane organic gas (NMOG).
B In 2007, the remaining 60% of California MD Vs must be certified to the more stringent ULEV levels.
y Medium-duty passenger vehicles are defined as any complete vehicle between 8,500 and 10,000 pounds
GVWR designed primarily for the transportation of persons. The definition specifically excludes any vehicle that
(1) has a capacity of more than 12 persons total or, (2) is designed to accommodate more than 9 persons in seating
rearward of the driver's seat or, (3) has a cargo box (e.g., pick-up box or bed) of six feet or more in interior length.
(See the Tier 2 final rulemaking, 65 FR 6698, February 10, 2000.)
m-113
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The standards shown in Table ni.B-2 for the 2008 heavy-duty gasoline vehicles and
engines are phased-in on 50 percent of vehicles and engines in the 2008 model year, and 100
percent in the 2009 model year. Under the California LEV-II program, by the 2007 model year
manufacturers are required to build at least 40 percent of their MDVs to the LEV category
standards shown in Table ni.B-2, and 60 percent of their MDVs to the tighter ULEV category
standards. The LEV-n ULEV category standards for MDVs are, for NOx and NMOG
respectively, 0.2 and 0.143 g/mi for 8,500 to 10,000 pounds, and 0.4 and 0.167 g/mi for 10,000
to 14,000 pounds.126
Under the federal Tier 2 program, the MDPVs are the vehicles most similar to the heavy-
duty gasoline vehicles required to meet the 2008 HD gasoline standards. During the 2008 model
year, half of the MDPVs can be certified in the MDPV interim bin shown in Table ni.B-2, while
the other half must be certified within the final Tier 2 bin structure and included in the
manufacturer's corporate average NOx standard of 0.07 g/mi NOx.127 The highest bin in the final
Tier 2 structure is the final high bin shown in Table ni.B-2. Then, in the 2009 model year, the
MDPV interim bin is no longer available and all MPDVs must be certified in the final Tier 2 bin
structure and included in the 0.07 g/mi NOx corporate average standard.
Therefore, the phase-in of the 2008 heavy-duty gasoline standards provides consistency
with the Tier 2 program because the standards affecting similarly sized vehicles are phased-in on
the same schedule. Further, those heavy-duty vehicles being phased-out (i.e., those 50 percent
not meeting the new standard), would be certified to the Phase 1 heavy-duty gasoline standards
which are equal to the MDPV interim bin for 8,500 to 10,000 pounds.128 Again, the standard
structure and implementation timing are consistent.
The engine standards shown in Table ni.B-2 are equal to the engine standards set for
diesel engines over 8,500 pound incomplete vehicles. For diesel engines, those standards are
phased-in on a 50/50/50/100 percent schedule beginning in the 2007 model year. For gasoline
engines, the phase-in schedule is consistent with the gasoline vehicle phase-in schedule of 50/100
percent beginning in the 2008 model year. This provides consistency and cost efficiency because
the engines certified to the engine standards are the same engines being certified to the vehicle
standards.
3. Current Exhaust Emission Certification Levels for Heavy-Duty
Gasoline Vehicles & Engines
Tables ni.B-3 and in.B-4 provide certification results from the 2000 model year for
heavy-duty gasoline vehicles and engines, respectively. The vehicle data is California medium-
duty vehicle (MDV) certification data and the engine data is EPA certification data. The tables
provide an indication of the emission levels that are being achieved through the application of
current emission control technologies.
IH-114
-------
Chapter III: Emissions Standards Feasibility
Table IHB-3 clearly shows that several vehicles have at least one emission constituent at
or below the Phase 2 standards, and six vehicles have both NOx and NMHC levels at or below
the future standards. We recognize that not all of these levels necessarily comply with the Phase
2 standards because they provide little or no compliance margin. Nonetheless such low levels
eight years prior to implementation suggest that the future standards are clearly within reach.
Table in.B-4 shows that current engines are being designed to be near the Phase 2 standards for
2008 despite being certified to the much higher current standards of 4.0 g/bhp-hr NOx and 0.9
g/bhp-hr NMHC. Based on industry input, we believe that manufacturers will continue the
process of replacing their old engines with more advanced engines over the next several years.
As new and more advanced engines are introduced, we anticipate that they will be capable of
achieving the 2008 standards.
IE-US
-------
Table III.B-3. 2000 Model Year Vehicle Certification Data (gram/mile)^
Mfr
Daimler
Chrysler
Ford
General
Motors
Same
Eng
Fam.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Mode?
Ram 3 500 Cab
Chassis 4WD
Ram 2500 P/U 4WD
Ram 2500 P/U 2WD
Ram 3 500 P/U 4 WD
Ram 3 500 P/U 2 WD
B3500Van2WD
Ram 2500 Cab
Chassis 4WD
Excursion 4WD
E350 2WD
F350 4WD
E250 Strip Chassis
2WD
E250 Econoline 2WD
K3500 P/U 4WD
K2500 Silverado
4WD
K2500 Suburban
4WD
Engine Size
(liters)
5.9
5.9
8.0
8.0
8.0
5.2
5.9
5.4
5.3
6.8
6.8
4.2
5.7
7.4
6.0
6.0
GVWR
(Ibs.)
11000
8800
8800
10500
11000
11000
8700
8800
8900
9300
9300
11000
8600
10000
8600
8600
NOx
(g/mi)
0.48
0.4
0.2
0.41
0.34
0.56
0.66
0.72
0.67
0.29
0.4
0.38
0.34
0.34
0.35
0.34
0.19
0.19
0.22
0.21
0.66
0.6
0.61
0.67
HC°
(g/mi)
0.16
0.097
0.084
0.2
0.19
0.22
0.26
0.24
0.23
0.14
0.15
0.1675
0.147
0.147
0.1615
0.1421
0.1003
0.1003
0.12
0.11
0.24
0.18
0.13
0.17
Stds
Tierl
LEV
Tierl
Tierl
Tierl
LEV
LEV
LEV
LEV
LEV
LEV
LEV
Tierl
Tierl
Tierl
Tierl
Tierl
Sales
Area0
CA
CA
CA
CA
FA
CA
CA
CF
CA
CA
CA
CF
FA
CA
CA
CA
CA
A Shaded entries are those at or below the Phase 2 emission standards.
B Some of these models may be Tier 2 medium-duty passenger vehicles.
IH-116
-------
Chapter III: Emissions Standards Feasibility
Table III.B-4. 2000 Model Year Engine Certification Data (g/bhp-hr)^
Mfr
DaimlerChrysler
Ford
General Motors
Same
Engine
Family
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Engine
Size
(liters)
5.9
8.0
5.4
6.8
4.3
5.7
6.0
7.4
Service Class
<14k
<14k
CFF/ULEV
<14k
CFF/ULEV
all
<14k
Fed<14k
Fed CFF/LEV
50 State <14k
Fed CFF/LEV
50 State <14k
Fed CFF/LEV
50 State <14k
50 State >14k
NOx (g/bhp-
hr)
1.291
1.14
0.66
0.66
0.48
0.48
0.48
0.9
2.7
2.0
1.7
0.52
1.7
3.7
0.8
3.7
HC
(g/bhp-hr)
0.18
0.13
0.10
0.10
0.13
0.13
0.12
0.2
0.3
0.2
0.3
0.2
0.6
0.6
0.5
0.6
NMHC
(g/bhp-hr)
0.2
n/a
0.2
n/a
0.5
Shaded entries are those at or below the Phase 2 emission standards.
4. Technological Feasibility of the 2008 Heavy-Duty Gasoline
Exhaust Emission Standards
We believe that the most promising overall emission control strategy for heavy-duty
gasoline engines is the combination of improved three-way catalysts and improved electronic
control of engine air-fuel ratio. Control of the air-fuel ratio is important because the three-way
catalyst is only effective if the air-fuel ratio is at a narrow band near stoichiometry. For example,
for an 80 percent conversion efficiency of HC, CO, and NOx with a typical three-way catalyst,
the air-fuel ratio must be maintained within a fraction of one percent of stoichiometry. During
transient operation, this minimal variation cannot be maintained with open-loop control. For
IH-117
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
closed-loop control, the air-fuel ratio in the exhaust is measured by an oxygen sensor and used in
a feedback loop. The throttle position, fuel injection, and spark timing can then be adjusted for
given operating conditions to result in the proper air-fuel ratio in the exhaust. Most, if not all,
engines have been equipped with closed loop controls. Some engines have been equipped with
catalysts that are achieving catalyst efficiencies in excess of 95 percent. This is one key reason
engine and vehicle certification levels are very low. In addition, electronic control can be used to
adjust the air-fuel ratio and spark timing to adapt to lower engine temperatures, thereby
controlling HC emissions during cold start operation.
All HD gasoline engines are equipped with three-way catalysts. Engines may be
equipped with a variety of different catalyst sizes and configurations. Manufacturers choose
catalysts to fit their needs for particular vehicles. Typically, current federal vehicle catalyst
systems contain either a single converter, or two converters in series or in parallel.2 A converter
is constructed of a substrate, a washcoat, and a catalytic material (e.g., precious metals). The
substrate may be metallic or ceramic with a flow-through design similar to a honeycomb.
Improvements in substrate and washcoat materials and technology have improved catalyst
performance significantly in recent years. A high surface area washcoat is used to provide a
suitable surface for the catalytic material. Under high temperatures, the catalytic material will
increase the rate of chemical reaction of the exhaust gas constituents. Current catalyst systems
on HD vehicles tend to have fairly low precious metal loading and total catalyst volumes
typically 80 to 90 percent of engine volumes. Current precious metal loadings tend to be in the
range of one to four g/L, and we expect most precious metal loadings to be up to four g/L for the
purpose of meeting the 2005 standards.
Significant changes in catalyst formulation have been made in recent years and additional
advances in these areas are still possible. Platinum, palladium and rhodium (Pt, Pd, and Rh) are
the precious metals typically used in catalysts.aa Historically, platinum has been widely used.
Today, palladium is being used much more widely due to its ability to withstand very high
exhaust temperatures. In fact, some HD vehicles currently are equipped with palladium-only
catalysts. Other catalysts contain all three metals or contain no platinum but both palladium and
rhodium. Some manufacturers have suggested that they will use Pd/Rh in lieu of tri-metal or
conventional Pt/Rh catalysts for underfloor applications. The underlying driver of which metals
are used, and in what proportion, is the price of those metals. As platinum prices rise, more
palladium is used; as palladium prices rise, more platinum is used. The same can be said of
rhodium.
z In contrast to some California LEV program medium-duty vehicles which have close-coupled catalysts.
aa Platinum, palladium, and rhodium are platinum-group metals, or PGM, which also includes indium,
osmium, and ruthenium.
IH-118
-------
Chapter III: Emissions Standards Feasibility
We project that the 2008 heavy-duty gasoline standards will require the application of
advanced engine and catalyst systems similar to those projected for their light-duty counterparts
to meet the Tier 2 standards. The technologies and emission control strategies that will be used
for medium-duty passenger vehicles (MDPVs), which have a GVWR greater than 8,500 pounds,
should also apply directly to heavy-duty gasoline vehicles. Historically, manufacturers have
introduced technology on light-duty gasoline vehicles and then applied those technologies to
their heavy-duty gasoline applications. We expect that manufacturers will take this same
approach to meeting the 2008 heavy-duty gasoline standards, through the application of
technology developed to meet light-duty Tier 2 standards beginning in the 2004 model year.
Improved calibration and systems management will be critical in optimizing the
performance of the engine with the advanced catalyst system. Precise air/fuel control must be
tailored for emissions performance and must be optimized. Calibration refinements may also be
needed for EGR system optimization and to reduce cold start emissions through methods such as
spark timing retard. We also project that electronic control modules with expanded capabilities
will be needed on some vehicles and engines.
We also expect increased use of other technologies in conjunction with those described
above. We expect some increased use of air injection to improve upon cold start emissions. We
may also see air-gap manifolds, exhaust pipes, and catalytic converter shells as a means of
improving upon catalyst light-off times, thereby reducing cold start emissions. Other, non-
catalyst related improvements to gasoline emission control technology include higher speed
computer processors which enable more sophisticated engine control algorithms and improved
fuel injectors providing better fuel atomization and improved fuel combustion.
For engine certified systems, the biggest concern will be the thermal durability of the
catalysts due to the heavier loads typical of the larger, more commonly engine certified, systems.
However, there is less emphasis on cold start emissions on the engine certification test procedure
than the chassis test procedure. As a result, there may be less use of close-coupled catalysts for
engine certified systems, although we have assumed the same implementation of that technology
for vehicles and engines.
Catalyst system durability is a key issue in the feasibility of the standards. Historically,
catalysts have deteriorated when exposed to very high temperatures and this has long been a
concern for heavy-duty work vehicles. Manufacturers have often taken steps to protect catalysts
by ensuring exhaust temperatures remain in an acceptable range. Catalyst technologies in use
currently are much improved over the catalysts used only a few years ago. The improvements
have come with the increased use of palladium, which has superior thermal stability, and through
much improved washcoat technology. The use of rhodium with palladium will also enhance
performance of the catalyst. The catalysts have been shown to withstand temperatures typically
experienced in heavy-duty applications. Nonetheless, as a worst-case assumption, we are
IH-119
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
assuming that 2008 model year heavy-duty gasoline vehicles and engines will require up to a 25
percent increase in precious metal loadings over the estimated 2005 loadings (i.e., 5 g/L versus
2005 levels of 4 or 4.5 g/L) to ensure acceptable catalyst durability characteristics while meeting
the standards. Further, we are assuming that catalyst volumes will increase to roughly coincide
with engine displacement volume. As a result of the greater catalyst volume, and the greater
PGM loading per unit volume, we are assuming considerably more precious metals for greater
durability. These costs are reflected in the cost estimates in Chapter V.
We believe that manufacturers will be able to achieve the 2008 standards by optimizing
all of these technologies. Current catalyst systems using some of these technologies have already
shown potential to reduce emissions to close to the required levels. Some current California
vehicles in the 8,500-10,000 pound range are certified to levels below 0.2 g/mile NOx. The
California Air Resources Board tested an advanced catalyst system on a vehicle loaded to a test
weight comparable to a heavy-duty vehicle test weight and achieved NOx and NMOG levels of
0.1 g/mile and 0.16 g/mile, respectively. Furthermore, the California vehicle with the advanced
catalyst had not been optimized as a system to take full advantage of the catalyst's capabilities.
In a light-duty truck technology demonstration program performed for our Tier 2
rulemaking effort, we found that a combination of calibration changes and improvements to the
catalyst system resulted in heavy light-duty truck (LDT4) NOx emission levels well below, and
NMHC/NMOG emissions slightly below, the Tier 2 intermediate useful life standards (0.05 g/mi
NOx and 0.075 g/mi NMOG).129 The catalyst improvements consisted of increases in volume
and precious metal loading, and higher cell-densities than those found in the original hardware.
Figures ni.B-2 and in.B-3 show the results of our testing on a Ford Expedition and a Chevrolet
Silverado, respectively.130 These figures demonstrate the dramatic improvements in emissions
that are possible with even the fairly simple enhancements that were done in a very short time as
part of this test program.
m-120
-------
Chapter III: Emissions Standards Feasibility
0.6
California LEV-1 MDV-3 Standard
0.5 -
0.4 -
O)
0.3 4
LJJ
0.24
0.1 -
• 1999 LEV Ford Expedition (certification
data)
fjEPA Expedition, stock/OEM configuration
QEPA Expedition, advanced catalysts,
calibration changes, air-gap exh. manifolds
OSwRI/CARB/EPA Expedition, advanced
catalysts, secondary air, calibration changes
Tier 2 Full-Life Stand
Intermediate Life Standard
0.05 0.1
NMHC/NMOG Emissions (g/mi)
0.15
Figure III.B-2. Emissions after an equivalent of 50,000 miles for various
tested configurations of Ford Expedition LDT4 SUVs with 5.4L V8 engines
m-121
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
0.4
California LEV-1 MDV-2 Standard
0.3 -
I
S
i
w 0.2 -
0.1 .
>1999 GM Chevrolet Silverado Certification
Data
QEPA Silverado, advanced catalysts, air-gap
manifold, minor calibration changes
<>SwRI/MECA Silverado, advanced catalysts
secondary air, calibration changes
Tier 2 Full-Life Standard
"ntermediate-LifeStlndard '"
0.05 0.1
NMOG/NMHC Emissions (g/mi)
0.15
Figure III.B-3. Emissions after an equivalent of 50,000 miles for various tested
configurations of 1999 GM Chevrolet Silverado LDT3 pickups with 5.3L V8
engines
The most significant difference between LDT4s (the heaviest of the light-duty
classification) and medium-duty passenger vehicles (MDPV), which had been heavy-duty
gasoline vehicles prior to our Tier 2 rule, 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 emissions than typical LDT4s.
These higher engine out emissions may be expected due to both the larger engine displacement,
and the 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 light-duty vehicles and trucks. The only difference will
likely be the need for larger catalysts with higher precious metal loadings than found in current
systems.
We believe that the test weight difference should not have a significant impact on the
emission levels to which these vehicles can be certified. We have tested a Ford Excursion and
IH-122
-------
Chapter III: Emissions Standards Feasibility
baseline results with a green (i.e., "new") catalyst indicate that emission levels are approximately
at, or slightly above, the 2008 heavy-duty standards. We tested the Excursion at loaded vehicle
test weight (curb weight + 300 Ib, or -7400 pounds) and again at adjusted loaded vehicle weight
(curb + half payload, or -8000 pounds) and found that the engine-out and tailpipe emission
results for NMHC and NOx were within ten percent for the two test weights. In other words, the
additional weight (approximately 600 Ibs) had no impact on emission performance. This is
borne out in the data shown in Table ni.B-3, which shows that the DaimlerChrysler 8.0L engine
used in the Ram 2500 Pickup (GVWR = 8,800 Ibs) and the Ram 3500 Pickup (GVWR = 10,500
Ibs) both have very similar emission levels despite having different payloads and, therefore,
different test weights. In fact, the heavier vehicle actually had slightly lower emissions. This is
also true with the Ford 6.8L engine used in the E350 (GVWR = 9300 Ibs) and in the F350
(GVWR = 11,000 Ibs); both of these vehicles have essentially the same emission levels. This is
significant because the majority of the heavy-duty vehicles falling under the 2008 heavy-duty
standards are large panel vans and pick-up trucks which typically weigh the same or less than
MPDVs.
Additionally, we believe that the 2008 standards will require manufacturers to focus some
effort on engine-out emissions control, and that engine-out NOx levels in the 6 to 8 g/bhp-hr are
reasonably achievable.bb Since some engines are already in this range, we believe that future
engines may even deliver lower engine out emissions. Current HD gasoline engines rely very
little on EGR. Recalibration of engine systems, including the EGR system and perhaps some
modest hardware changes to those systems, will be necessary. EGR plays a key role in reducing
engine-out NOx and system redesign may allow more effective use of this technology.
Lastly, the proposed averaging, banking, and trading (ABT) program can be an important
tool for manufacturers in implementing the new standards. The ABT program will allow
manufacturers to comply with the more stringent standards by introducing emissions controls
over a longer period of time, as opposed to doing so during one or two model years.
Manufacturers plan their product introductions well in advance. With ABT, manufacturers can
better manage their product lines so that the new standards do not interrupt their product
introduction plans. Also, the program allows manufacturers to focus on higher sales volume
vehicles first to earn credits and then use those credits for low sales volume vehicles. We believe
manufacturers have significant opportunity to earn credits in the pre-2008 time frame by selling
their California LEV II certified vehicles nationwide. Further, we are allowing manufacturers to
apply credits earned on vehicle sales to their engine sales, and vice versa, although a 20 percent
bb Note that the Phase 1 HD rule requires that diesel engines meet a standard of 2.5 g/bhp-hr
NOx+NMHC, of which we believe 2.2 to 2.3 grams will be NOx emissions. That emission level will not require
exhaust emission control technology (i.e., catalysts) and will instead be met through use of cooled EGR. In effect,
the 2.2 gram NOx level is an engine-out NOx level that will be achieved by diesel engines complying with that
standard.
IH-123
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
discount will apply.cc This ability to use credits across the vehicle/engine categories provides
even greater incentive to introduce LEV II vehicles earlier than 2008 to earn credits that can be
applied to the lower sales volume engine certified systems. This should provide attractive cost
efficiencies to manufacturers while having no negative effects on, and perhaps even improving,
air quality.
This discussion highlights our belief that there are numerous proven and existing
technologies available that will allow heavy-duty gasoline vehicles to meet our 2008 heavy-duty
gasoline exhaust emission standards. Therefore, we believe that these technologies, combined
with low sulfur gasoline, ABT, and considerable leadtime before the 2008 model year
implementation, will make the 2008 standards technologically feasible for heavy-duty gasoline
vehicles.
5. The 2008 Heavy-Duty Gasoline Evaporative Emission Standards
The new evaporative emission standards for heavy-duty vehicles and engines are shown
in Table ni.B-5. These standards will apply to heavy-duty gasoline-fueled vehicles and engines,
and methanol-fueled heavy-duty vehicles and engines. Consistent with existing standards, the
standard for the two day diurnal plus hot soak test sequence would not apply to liquid petroleum
gas (LPG) fueled and natural gas fueled HDVs.
Table III.B-5. New Heavy-Duty Evaporative Emission
StandardsA
(grams per test)
Category
8,500 - 14,000 Ibs
>14,000 Ibs
3 Day Diurnal +
Hot Soak
1.4
1.9
Supplemental 2 Day Diurnal +
Hot Soak8
1.75
2.3
A To be implemented on the same schedule as the gasoline engine and vehicle
exhaust emission standards shown in Table III.B-2. These new standards do n<
apply to medium-duty passenger vehicles, and do not apply to diesel fueled
vehicles and engines.
B Does not apply to LPG or natural gas fueled HDVs.
00 As explained in the preamble to this rule, we believe this 20 percent discount is necessary to account for
the uncertainty in converting between g/mi standards and g/bhp-hr standards.
IH-124
-------
Chapter III: Emissions Standards Feasibility
These new standards represent more than a 50 percent reduction in the numerical
standards as they exist today. The Phase 1 heavy-duty rule made no changes to the numerical
value of the standard, but it did put into place new evaporative emission test procedures for
heavy-duty complete gasoline vehicles.dd For establishing evaporative emission levels from
complete heavy-duty vehicles, the standards shown in Table in.B-5 presume the test procedures
required in the Phase 1 heavy-duty rule.
The new standards for 8,500 to 14,000 pound vehicles are consistent with the Tier 2
standards for medium-duty passenger vehicles (MDPV). MDPVs are of consistent size and have
essentially identical evaporative emission control systems as the remaining work-oriented HDVs
in the 8,500 to 10,000 pound weight range. Therefore, the evaporative emission standards should
be equivalent. We are requiring those same standards for the 10,000 to 14,000 pound HDVs
because, historically, the evaporative emission standards have been consistent throughout the
8,500 to 14,000 pound weight range. The HDVs in the 10,000 to 14,000 pound range are
essentially equivalent in evaporative emission control system design as the lighter HDVs;
therefore, continuing this historical approach is appropriate.
The evaporative emission standards for the over 14,000 pound HDVs are slightly higher
because of their slightly larger fuel tanks and for non-fuel emissions related to larger vehicle
sizes. This is consistent with past evaporative emission standards. The levels for the over
14,000 pound HDVs maintains the same ratio relative to the 8,500 to 14,000 pound HDVs as
exists with current evaporative standards. To clarify, the current standards for the 3 day diurnal
test are 3 and 4 grams/test for the 8,500 to 14,000 and the over 14,000 pound categories,
respectively. The ratio of 3:4 is maintained for the new 2008 standards, 1.4:1.9.
The new standard levels are slightly higher than the California LEV-II standard levels.
The California standard levels are 1.0 and 1.25 for the 3-day and the 2-day tests, respectively.
However, federal vehicles are certified using the higher-volatility federal test fuel.ee Arguably,
the federal and California evaporative emission standards are equivalent in stringency despite the
difference in standard levels.
dd The test procedure changes codify a commonly approved waiver allowing heavy-duty gasoline vehicles
to use the light-duty driving cycle for demonstrating evaporative emission compliance. The urban dynamometer
driving schedule (UDDS) used for heavy-duty vehicles is somewhat shorter than that used for light-duty vehicles,
both in terms of mileage covered and minutes driven. This results in considerably less time for canister purge under
the heavy-duty procedure than under the light-duty procedure. We recognize this discrepancy and have routinely
provided waivers under the enhanced evaporative program that allow the use of the light-duty procedures for heavy-
duty certification testing. This is consistent with CARB's treatment of equivalent vehicles. (See 65 FR 59896,
October 6, 2000.)
ee The federal test fuel specification for fuel volatility, the Reid Vapor Pressure, is 8.7 to 9.2 psi. The
California test fuel specification is 6.7 to 7.0 psi.
m-125
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The new evaporative emission standards are to be implemented on the same schedule as
the gasoline engine and vehicle exhaust standards shown in Table in.B-2. This will allow
manufacturers to plan any needed changes to new vehicles at the same time, although it is not
necessary that the exhaust and evaporative standards be phased-in on the same vehicles and
engines. Also, the revised durability provisions finalized in the Tier 2 rule will apply. These
provisions require a durability demonstration using fuel containing at least 10 percent alcohol.
Alcohol can break down the materials used in evaporative emission control systems. This
provision should not have an impact on the feasibility of the new standards.
6. Technological Feasibility of the 2008 Heavy-Duty Gasoline
Evaporative Emission Standards
The new evaporative emission standards appear to be feasible now. Many designs have
been certified that already meet these standards. A review of 1998 through 2000 model year
certification data indicates that nearly all evaporative system families in the 8,500 to 14,000
pound range comply with the proposed 1.4 g/test standard, while all evaporative system families
in the over 14,000 pound range comply with the proposed 1.9 g/test standard. Table ni.B-6
summarizes the 1998 through 2000 model year evaporative emission certification data.
IH-126
-------
Chapter III: Emissions Standards Feasibility
Table III.B-6. 1998-2000 Model Year Evaporative Emission
Certification Data
(grams/test)
Manufacturer
DaimlerChrysler
Ford
General Motors
Category
(GVWR)
<14k
<14k
>14k
<14k
>14k
3 Day Diurnal Emissions
(grams/test)
0.74
0.64
0.64
1.01
1.06
1.17
1.80
1.17
1.80
0.67
0.78
0.85
1.03
1.29
1.57
1.69
1.74
0.67
0.73
0.85
1.03
1.29
1.57
1.69
1.74
There are two approaches to reducing evaporative emissions for a given fuel. One is to
minimize the potential for permeation and leakage by reducing the number of hoses, fittings and
connections. The second is to use less permeable hoses and lower loss fittings and connections.
Manufacturers are already employing both approaches. 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. 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
IH-127
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 materials such as
those discussed above are necessary to ensure that the benefits are captured in-use. Rather than
requiring new application of these materials, our new evaporative standards will simply ensure
their consistent and continued use on most engines by discouraging manufacturers from
switching to cheaper materials or designs to take advantage of the large safety margins they have
under current standards.
Additionally, most manufacturers are moving to "returnless" fuel injection systems.
Through more precise fuel pumping and metering, these systems eliminate the return line in the
fuel injection system. The return line carries unneeded fuel from the fuel injectors back to the
fuel tank. Because the fuel injectors are in such close contact with the hot engine, the fuel
returned from the injectors to the fuel tank has been heated. This returned fuel is a significant
source of fuel tank heat and vapor generation. The elimination of the return line also reduces the
total length of hose on the vehicle though which vapors can permeate, and it reduces the number
of fittings and connections through which fuel can leak.
Steel fuel tanks and steel fuel lines have essentially zero losses due to permeation, but are
vulnerable to leakage at joints and interfaces. Manufacturers are moving toward plastic fuel
tanks for their lighter weight and greater ability to be molded to odd shapes. However, plastic
tanks are permeable and are also susceptible to seepage and higher permeability at areas where
connections and welds are made. Materials and manufacturing techniques exist to reduce these
losses.
IH-128
-------
Chapter III: Emissions Standards Feasibility
Chapter III. References
1. NOx formation rate increases rapidly with temperature, see Heywood J.B.: Internal
Combustion Engine Fundamentals, McGraw-Hill, Inc, New York, p. 586, 1988.
2. McKinley, T., Modeling Sulfuric Acid Condensation in Diesel Engine EGR Coolers,
SAE 970636, 1997.
3. Cooper, B., et al., "Role of NO in Diesel Paniculate Emission Control", SAE 890404.
Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
4. Allansson, et al, European Experience of High Mileage Durability of Continuously
Regenerating Diesel Particulate Filter Technology. SAE 2000-01-0480.
5. Letter from Dr. Barry Cooper to Don Kopinski US EPA, EPA Docket A-99-06.
6. Telephone conversation between Dr. Barry Cooper, Johnson Matthey, and Todd
Sherwood, EPA, EPA Docket A-99-06.
7. Letter from Dr. Barry Cooper to Don Kopinski US EPA, EPA Docket A-99-06.
8. LeTavec, Chuck, et al., "EC-Diesel Technology Validation Program Interim Report,"
SAE 2000-01-1854; Clark, Nigel N., et al., "Class 8 Trucks Operating On Ultra-Low
Sulfur Diesel With Particulate Filter Systems: Regulated Emissions," SAE 2000-01-
2815; Vertin, Keith, et al., "Class 8 Trucks Operating On Ultra-Low Sulfur Diesel With
Particulate Filter Systems: A Fleet Start-Up Experience," SAE 2000-01-2821.
9. Vertin, Keith, et al., "Class 8 Trucks Operating On Ultra-Low Sulfur Diesel With
Particulate Filter Systems: A Fleet Start-Up Experience," SAE 2000-01-2821.
10. Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels, Manufacturers of Emissions
Controls Association, June 1999 contained in Air Docket A-99-06 item JJ-G-139.
11. Schenk, Charles "Summary of EPA PM Efficiency Data" memo to EPA Air Docket A-
99-06 item JJ-B-15.
12. Navistar written comments dated July 13, 1999, in response to the Control of Diesel Fuel
Quality Advanced Notice of Proposed Rulemaking (64 FR 26142, May 13, 1999), which
can be found in Public Docket A-99-06.
13. Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels, Manufacturers of Emissions
IH-129
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Controls Association, June 1999 Air Docket A-99-06.
14. Schenk, Charles "Summary of EPA PM Efficiency Data" memo to EPA Air Docket A-
99-06. Data taken at EPA with a current production 6 1 HD engine with a CDPF and
Phillips 3 ppm (DECSE) fuel.
15. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase U Interim Data Report
No. 4, Diesel Paniculate Filters-Final Report, January 2000, Table Cl, Air Docket A-99-
06 also available at www.ott.doe.gov/decse.
16. Memorandum from Chuck Schenk, EPA, "Summary of EPA PM efficiency data.",
available in EPA Air Docket A-99-06, docket item II-B-15.
17. Hawker, P., et. al., Effect of Continuously Regenerating Diesel Particulate Filter on Non-
Regulated Emissions and Particle Size, SAE 980189, Figure 6.
18. Diesel Emission Control-Sulfur Effects Program, Phase I Interim Data Report No. 3,
October, 1999, Air Docket A-99-06 also available at www.ott.doe.gov/decse.
19. Memorandum from William Charmley, EPA, "Summary of Model Year 1999 and 2000
Federal On-highway Heavy-duty Diesel Engine Families Certified as Compliant with
Not-to-Exceed Requirements, Euro-3 Steady State Requirements, and Maximum
Allowable Emission Limits Requirements", available in EPA Air Docket A-98-32.
20. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 4, Diesel Particulate Filters-Final Report, January 2000, Table Cl, Air Docket A-99-
06 also available at www.ott.doe.gov/decse.
21. Dolan, D.F., Kittelson, D.B., Whitby, K.T., "Measurement of Diesel Exhaust Particle
Size Distributions", ASME Paper Number 75-WA/APC-5, 1975.
22. Cheng, Y.S., Yeh, C.H., Mauderly, J.L., Mokler, B.V., "Characterization of Diesel
Exhaust in a Chronic Inhalation Study", Am. Ind. Hyg. Assoc. J., 1984, 45:547-555.
23. Cantrell, B.K., Rubow, K.L., "Diesels in Underground Mines: Measurement and Control
of Particulate Emissions", U.S. Bureau of Mines Publication IC9324.
24. Kittelson, D.B., "Engines and Nanoparticles: A Review", J. Aerosol Sci 1998 29:575-
588.
25. Bagley, S.T., Gratz, L.D., Johnson, J.H., McDonald, J.F., Environ. Sci. Technol. 1998
32:1183-1191.
m-130
-------
Chapter III: Emissions Standards Feasibility
26. Kleeman, M.J., Schauer, J.J., Cass, G. R., 2000, Size and Composition Distribution of
Fine Particulate Matter Emitted From Motor Vehicles, Environmental Science and
Technology, Vol. 34, No. 7.
27. Kittelson, D. B., 2000, Presentation on Fuel and Lube Oil Sulfer and Oxidizing
Aftertreatment System Effects on Nano-particle Emissions from Diesel Engines.
Presented in United Kingdom April 12, 2000 Air Docket A-99-06 item H-G-149.
28. Hawker, P., et. al., Effect of a Continuously Regenerating Diesel Particulate Filter on
Non-Regulated Emissions and Particle Size Distribution, SAE 980189.
29. Diesel Emission Control-Sulfur Effects Program, Phase I Interim Data Report No. 1,
August, 1999, www.ott.doe.gov/decse EPA Docket A-99-06.
30. Kawanami, M., et. al., Advanced Catalyst Studies of Diesel NOx Reduction for Highway
Trucks, SAE 950154.
31. Letter from Barry Wallerstein, Acting Executive Officer, SCAQMD, to Robert Danziger,
Goal Line Environmental Technologies, dated December 8, 1997, www.glet.com Air
Docket A-99-06 item H-G-137.
32. Reyes and Cutshaw, SCONOx Catalytic Absorption System, December 8, 1998,
www.glet.com Air Docket A-99-06 item D-G-147.
33. Danziger, R. et. al. 21,000 Hour Performance Report on SCONOX, 15 September 2000
EPA Docket A-99-06 item IV-G-69.
34. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase n Summary Report:
NOx Adsorber Catalysts, October 2000 EPA Docket A-99-06.
35. Memo to Air Docket A-99-06 from Todd Sherwood, item number IV-E-31.
36. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
37. McDonald, J., et al., "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty
Trucks," SAE 2000-01-1957. Copy available in EPA Air Docket A-99-06, docket item
IV-G-131.
38. Schenk, Charles "Summary of NVFEL Testing of Advanced NOx and PM Emission
Control Technologies" memo to EPA Docket A-99-06.
IH-131
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
39. Brogan, M, et. al., Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions
of Lean Running Gasoline Engines, SAE 962045.
40. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
41. Revolutionary Diesel Aftertreatment System Simultaneously Reduces Diesel Particulate
Matter and Nitrogen Oxides, Toyota Motor Corporation press release, July 25, 2000 EPA
Docket A-99-06 item IV-G-36.
42. Pott, E., et. al., Potential of NOx-Trap Catalyst Application for DI-Diesel Engines EPA
Docket A-99-06.
43. Cunningham, M. J. and Huang S. C., presentation to 1999 DEER Conference, NOx
Adsorber Catalysts Applied to Heavy-Duty Diesel Test Cycles, EPA Docket A-99-06
itemIV-G-65.
44. Memo to EPA docket A-99-06 II-E.
45. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 1, August 1999, EPA Docket A-99-06.
46. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 2: NOx Adsorber Catalysts, October 1999, EPA Docket A-99-06.
47. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Date Report
No. 3: Diesel Fuel Sulfur Effects on Parti culate Matter Emissions, November 1999, EPA
Docket A-99-06.
48. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 4, Diesel Paniculate Filters-Final Report, January 2000, EPA Docket A-99-06.
49. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase U Summary Report:
NOx Adsorber Catalysts, October 2000 EPA Docket A-99-06.
50. West, B., et al., "NOx Adsorber Performance in A Light-Duty Diesel Vehicle", SAE
2000-01-2912. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
51. West, B., et al., "NOx Adsorber Performance in A Light-Duty Diesel Vehicle", SAE
2000-01-2912. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
52. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase U Summary Report:
NOx Adsorber Catalysts, October 2000 EPA Docket A-99-06.
IH-132
-------
Chapter III: Emissions Standards Feasibility
53. Schenk, Charles "Summary of NVFEL Testing of Advanced NOx and PM Emission
Control Technologies" memo to EPA Docket A-99-06 item IV-A-29.
54. Schenk, Charles "Summary of NVFEL Testing of Advanced NOx and PM Emission
Control Technologies" memo to EPA Docket A-99-06 item IV-A-29.
55. Schenk, Charles "Summary of NVFEL Testing of Advanced NOx and PM Emission
Control Technologies" memo to EPA Docket A-99-06 item IV-A-29.
56. Memo to EPA docket A-99-06 II-E-25.
57. Brogan, M, et. al., Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions
of Lean Running Gasoline Engines, SAE 962045.
58. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
59. "Toyota Develops Aftertreatment System," by Mike Osenga, Diesel Progress North
American Addition, October 2000 page 52, Air Docket A-99-06.
60. Memo to Air Docket A-99-06, 13 October 2000, from Todd Sherwood documenting a
meeting with representatives of Toyota. Item IV-E-31.
61. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase n Summary Report:
NOx Adsorber Catalysts, October 2000, Air Docket A-99-06.
62. Memo from Byron Bunker to Docket A-99-06, "Estimating Fuel Economy Impacts of
NOx Adsorber De-Sulfurization," December 10, 1999.
63. Jobson, E. et al, "Research Results and Progress in LeaNOx n - A Cooperation for Lean
NOx Abatement," SAE 2000-01-2909.
64. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst," SAE 1999-01-3501. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
65. Guy on, M. et al, "NOx-Trap System Development and Characterization for Diesel
Engines Emission Control," SAE 2000-01-2910.
66. Dou, Danan and Bailey, Owen, "Investigation of NOx Adsorber Catalyst Deactivation,"
SAE 982594. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
IH-133
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
67. Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.
68. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
69. Guyon, M. et al, "NOx-Trap System Development and Characterization for Diesel
Engines Emission Control," SAE 2000-01-2910.
70. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE
982594. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
71. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
72. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", Figure 5 SAE 982595. Copy available in EPA Air Docket A-99-
06, docket item IV-G-131.
73. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
74. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE
982594. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
75. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology,
page 64-65. 1995 Van Nostrand Reinhold Publishing.
76. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology,
Chapter 6. 1995 Van Nostrand Reinhold Publishing.
77. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst," SAE 1999-01-3501. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
78. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase n Summary Report:
NOx Adsorber Catalysts, October 2000, Air Docket A-99-06.
79. Tanaka, H., Yamamoto, M., "Improvement in Oxygen Storage Capacity," SAE 960794.
IH-134
-------
Chapter III: Emissions Standards Feasibility
80. Yamada, T., Kobayashi, T., Kayano, K., Funabiki M., "Development of Zr Containing
TWC Catalysts", SAE 970466.
81. McDonald, Joseph, and Lee Jones, U.S. EPA, "Demonstration of Tier 2 Emission Levels
for Heavy Light-Duty Trucks," SAE 2000-01-1957. Copy available in EPA Air Docket
A-99-06, docket item IV-G-131.
82. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
83. Handout materials provided by Cummins Engine Company in a meeting with EPA, OMB
and affected industries. EPA Docket A-99-06 item II-E-25.
84. Parks, J. et al., "Near-Zero NOx Control for Diesel Aftertreatment", SAE 1999-01-2890.
Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
85. Bailey, O. et al., "Sulfur Traps for NOx Adsorbers: Materials Development and
Maintenance Strategies for Their Application," SAE 2000-01-1205. Copy available in
EPA Air Docket A-99-06, docket item IV-G-131.
86. Klein, H., et al, NOx Reduction for Diesel Vehicles, Degussa-Huls AG, Corning Clean
Diesel Workshop, Sept. 27-29 1999 EPA Docket A-99-06 item U-G-141.
87. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission
Controls Association, June 1999 EPA Docket A-99-06 item II-G-139.
88. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission
Controls Association, June 1999 EPA Docket A-99-06 item II-G-139.
89. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
Incorporated, December 15, 1999 Air Docket A-99-06.
90. API Comments on the 2007 Heavy Duty Engine/Diesel Sulfur Proposed Rule, 14 August
2000. EPA Docket A-99-06 IV-D-343.
91. Testimony of Stephanie Williams - Director of Environmental Affairs, California
Trucking Association to EPA public hearing June 27, 2000 EPA Docket A-99-06 IV-F-
190.
m-135
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
92. "Sulfur-proof aftertreatment promised, European Fuels News, Volume 3, Issue 18,
September 22, 1999.
93. "The Impact of Sulfur in Diesel Fuel on Catalyst Emission Control Technology," report
by the Manufacturers of Emission Controls Association, March 15, 1999 Air Docket A-
99-06 item H-G-140.
94. "The Impact of Sulfur in Diesel Fuel on Catalyst Emission Control Technology," report
by the Manufacturers of Emission Controls Association, March 15, 1999, pp. 9 & 11 Air
Docket A-99-06 item U-G-140.
95. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emissions
Controls Association, June 1999 EPA Docket A-99-06 item II-G-139.
96. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emissions
Controls Association, June 1999 EPA Docket A-99-06 item II-G-139.
97. "Crankcase Emissions & Closed Crankcase Filtration", presented by Marty Barris,
Donaldson Corporation, Society of Automotive Engineers TOPTEC, September, 2000,
copy available in EPA Air Docket A-99-06.
98. "Advances In The Control of Crankcase Emissions From Diesel Engines", G. Dickson &
K. Edge, Diesel Progress, Nov. 1995, copy available in EPA Air Docket A-99-06.
99. Letter from Marty Barris, Donaldson Corporation, to Byron Bunker US EPA, March
2000. EPA Air Docket A-99-06.
100. "Crankcase Emissions & Closed Crankcase Filtration", presented by Marty Barris,
Donaldson Corporation, Society of Automotive Engineers TOPTEC, September, 2000,
copy available in EPA Air Docket A-99-06.
101. "Documentation of Closed Crankcase Systems for Model Year 1999 Daimler-Benz On-
highway Heavy-duty Diesel Engine Family", EPA Memorandum, Byron Bunker,
available in EPA Air Docket A-99-06.
102. Revolutionary Diesel Aftertreatment System Simultaneously Reduces Diesel Particulate
Matter and Nitrogen Oxides, Toyota Motor Corporation press release, July 25, 2000 EPA
Docket A-99-06 item IV-G-36.
103. Hawker, P. et al, Experience with a New Paniculate Trap Technology in Europe, SAE
970182. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
IH-136
-------
Chapter III: Emissions Standards Feasibility
104. Hawker, P. et al, Experience with a New Paniculate Trap Technology in Europe, SAE
970182. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
105. Phase I Interim Data Report No. 3: Diesel Fuel Sulfur Effects on Paniculate Matter
Emissions, Diesel Emission Control-Sulfur Effects Program, October 29, 1999 Air
Docket A-99-06.
106. Allansson, et al. SAE 2000-01-0480.
107. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, US EPA, Air Docket
A-99-06.
108. Telephone conversation between Dr. Barry Cooper, Johnson Matthey, and Todd
Sherwood, EPA, EPA Docket A-99-06.
109. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, US EPA, Air Docket
A-99-06.
110. Cooper, B., et al., "Role of NO in Diesel Paniculate Emission Control", SAE 890404.
Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
111. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase JJ Interim Data Report
No. 4, Diesel Paniculate Filters-Final Report, January 2000. Table Cl Air Docket A-99-
06.
112. Phase 1 Interim Data Report No. 2: NOx Adsorber Catalysts, Diesel Emission Control-
Sulfur Effects Program, October 1999 Air Docket A-99-06.
113. Dou, D., et al.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE 982594.
Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
114. Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.
115. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies," SAE 982595. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
116. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst", SAE 1999-01-3501. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
IH-137
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
117. Memo from Byron Bunker to Docket A-99-06, "Estimating Fuel Economy Impacts of
NOx Adsorber De-Sulfurization," December 10, 1999.
118. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst", SAE 1999-01-3501. Copy available in EPA Air Docket A-99-06,
docket item IV-G-131.
119. Whitacre, Shawn. "Catalyst Compatible" Diesel Engine Oils , Presentation at
DOE/NREL Workshop "Exploring Low Emission Diesel Engine Oils." January 31, 2000
Air Docket A-99-06 item U-G-148.
120. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 4, Diesel Paniculate Filters-Final Report, January 2000. Table Cl Air Docket A-99-
06.
121. See 65 FR 59896, October 6, 2000.
122. State of California, Air Resources Board Staff Report: Initial Statement of Reasons,
Proposed Amendments to California Exhaust and Evaporative Emission Standards and
Test Procedures for Passenger Cars, Light-duty Trucks and Medium-duty Vehicles "LEV
IF, September 18, 1998 Air Docket A-97-10 item U-G-356.
123. See 65 FR 6698, February 10, 2000.
124. See Chapter IV.A of the final Tier 2 RIA, contained in Air Docket A-97-10. See also,
McDonald, J., et al., "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty
Trucks," SAE 2000-01-1957. Copy available in EPA Air Docket A-99-06, docket item
IV-G-131.
125. McDonald, J., et al., "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty
Trucks," SAE 2000-01-1957. Copy available in EPA Air Docket A-99-06, docket item
IV-G-131.
126. State of California, Air Resources Board Staff Report: Initial Statement of Reasons,
Proposed Amendments to California Exhaust and Evaporative Emission Standards and
Test Procedures for Passenger Cars, Light-duty Trucks and Medium-duty Vehicles "LEV
IF, September 18, 1998 Air Docket A-97-10 item II-G-356.
127. See 65 FR 6698, February 10, 2000.
128. See 65 FR 59896, October 6, 2000.
IH-138
-------
Chapter III: Emissions Standards Feasibility
129. See 65 FR 6698, February 10, 2000. See also Chapter IV.A of the final Tier 2 RIA,
contained in Air Docket A-97-10. See also, McDonald, J., et al., "Demonstration of Tier
2 Emission Levels for Heavy Light-Duty Trucks," SAE 2000-01-1957. Copy available in
EPA Air Docket A-99-06, docket item IV-G-131.
130. See 65 FR 6698, February 10, 2000.
IH-139
-------
Chapter IV: Fuel Standard Feasibility
Chapter IV: Fuel Standard Feasibility
A. Feasibility of Removing Sulfur from Highway Diesel Fuel
1. Sources of Diesel Fuel Sulfur
The primary sources of sulfur in diesel fuel are the sulfur-containing compounds which
occur naturally in crude oil.1 Depending on the source, crude oil contains anywhere from
fractions of a percent of sulfur, such as less than 0.05 weight percent (500 ppm) to as much as
several percent.1 The average amount of sulfur in crude oil refined in the U.S. is about one
percent.2 Most of the sulfur in crude oil is in the heaviest boiling fractions. Since most of the
refinery blendstocks that are used to manufacture diesel fuel come from the heavier boiling
components of crude oil, they contain substantial amounts of sulfur.
The diesel fuel produced by a given refinery is composed of one or more blendstocks
from the crude oil fractionation and conversion units at the refinery. Refinery configuration and
equipment, and the range and relative volumes of products manufactured (the product slate) can
significantly affect the sulfur content of diesel fuel. The diagram on the following page
illustrates the configuration and equipment used at a typical complex refinery in the U.S.
1 Additives that contain sulfur are sometimes intentionally added to diesel fuel.For a discussion how the
addition of these additives will be affected under this program, see Section IV.D.5.
IV-1
-------
Natural
Gas
Vacuum Tower
Coker
*" Fuel Gas
LPG
Gasoline
Aromatics
Kerosine
Jet Fuel
On-Highway
Diesel
Off-roadDiesel
Fuel Oil
*- Resid
Coke
Figure IV.A-1. Diagram of a Typical Complex Refinery
IV-2
-------
Chapter IV: Fuel Standard Feasibility
Refineries differ from the model in the preceding diagram depending on the range of
crude oils used, and their product slate. For example:
- Refiners that process heavier crude oils are more likely to operate coker and/or
hydrocracker units.
- Refinery streams that can be used to manufacture diesel fuel can also be used in the
manufacture of heating oil, kerosene and jet fuel. Much of the distillate product from the
hydrocracker is often blended into jet fuel rather than diesel fuel.
On an aggregate basis, most of the highway diesel fuel volume manufactured in the U.S.
comes from the straight-run product from the crude fractionation tower (called straight run).
Most of the remainder comes from the fluid catalytic cracker (FCC) conversion unit (called light
cycle oil). The remaining small fraction of diesel fuel volume comes from a coker conversion
unit (called light coker gas oil), or from the hydrocracker conversion unit (called hydrocrackate).
To comply with the current federal regulatory requirement on the sulfur content of
highway diesel fuel (500 ppm cap), the blendstock streams from these process units are typically
further processed to reduce their sulfur content. Desulfurization of highway diesel blendstocks is
currently accomplished in fixed-bed hydrotreaters that operate at moderate pressures (500-700
psi), with a few exceptions at higher pressures such as the small portion of highway diesel which
comes from hydrocrackers. Most of the low-sulfur diesel blendstocks come from such
hydrotreaters. However, a small amount of low-sulfur diesel also comes from hydrocrackers.
The sulfur levels of the various highway diesel blendstocks and the fraction of the total volume
of highway diesel fuel that comes from each blendstock varies considerably from refinery to
refinery. A recent survey conducted by the American Petroleum Institute (API) and National
Petroleum Refiners Association (NPRA) in 1996 examined the typical blendstock properties for
the U.S. highway diesel pool as a whole.3 The results of this survey are contained in the
following tables (IV.A-1 and IV.A-2).
IV-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.A-1. Volume Fraction of U.S. Highway Diesel Pool
from each Blendstock Component4
Diesel Blendstock
Straight Run
Hydrotreated
Straight Run
Cracked Stock
Hydrotreated
Cracked Stock
Coker Gas Oil
Hydrotreated
Coker Gas Oil
Hydrocrackate
Percent of U.S. Highway Diesel Fuel Pool
per Blendstock Boiling Fraction
Naphtha
0.1
0.3
-
-
-
0.1
-
Light
Distillate
6.4
8.1
0.1
2.1
-
2.1
1.3
Heavy
Distillate
4.9
41.2
0.8
15.6
1.0
3.7
2.7
Light
Gas Oil
1.0
2.3
2.2
1.7
-
2.3
-
All Boiling Fractions
Combined
12.4
51.9
3.1
19.4
1.0
8.2
4.0
IV-4
-------
Chapter IV: Fuel Standard Feasibility
Table IV.A-2. Sulfur Levels of Highway Diesel Blendstocks (CA Excluded)5
Diesel Blendstock
Straight Run
Hydrotreated
Straight Run
Cracked Stock
Hydrotreated
Cracked Stock
Coker Gas Oil
Hydrotreated
Coker Gas Oil
Hydrocrackate
Sulfur Content (ppm) by Boiling Fraction of the Blendstock4
Naphtha
827
362
-
18
540
8C
-
Light
Distillate
1,770
119
2,219
37
1,800
25
12
Heavy
Distillate
2,269
394
2,892
939
3,419
310
120
Light
Gas Oil
4,980
548
6,347C
1,306C
-
400
-
All Boiling Fractions
Combined8
2,218
358
5,322
874
3,419
258
85
A The boiling ranges that define the four different boiling fractions of each diesel blendstock (naphtha, light
distillate, heavy distillate, and light gas oil) varied somewhat from refiner to refiner. There was also
definitional overlap in the boiling ranges provided by refiners.
B These values were derived by weighting the values for the four boiling fractions by the fraction they
represent of the highway diesel fuel blendstock (see Table IV-1).
C Indicates properties that were not reported in the refiner survey. These values were calculated using the
reported sulfur contents of like boiling fractions in other diesel blendstocks by assuming the same relative
sulfur levels between boiling fractions. This was necessary to allow the calculation of the sulfur content of
the blendstock as a whole.
As shown in Table IV. A-1, approximately 80 percent of all blendstocks used to
manufacture highway diesel fuel outside of California are hydrotreated to reduce their sulfur
content. Hydrocrackate is desulfurized to a substantial extent as a necessary element of the
hydrocracking process and is not further processed in a hydrotreater. The table also shows that
approximately 16 percent of highway diesel fuel comes from nonhydrotreated blendstocks.
The blendstocks used to manufacture highway diesel fuel used in California differ from
the rest of the nation due to the unique requirements of the California market and California's
specific regulatory requirements. As a result, California's highway diesel fuel averages 140 ppm
sulfur.6 Highway diesel fuel used in California is made primarily from hydrocrackate and
hydrotreated straight run in roughly equal proportions, with a small volume fraction of
IV-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
hydrotreated cracked stock and hydrotreated coker gas oil. No high-sulfur containing
blendstocks are used in the manufacture of California highway diesel fuel. California diesel fuel
requirements include a limit on aromatics content which limits the amount of light-cycle oil
(LCO) and light-coker gas oil (LCGO) that can be used in the manufacture of California highway
diesel fuel. LCO and LCGO have a high aromatics content which is not reduced by
desulfurization.
2. Current Levels of Sulfur in Highway Diesel Fuel
To determine the most cost-effective sulfur removal methods, it is important to evaluate
the amount of sulfur currently in highway diesel fuel. EPA set standards for highway diesel fuel
quality in 1990 (55 FR 34120, August 21, 1990). These standards have been effective since
1993. The standards limit the sulfur concentration in fuel to a maximum of 500, compared to a
pre-regulation average of 2500 ppm. They also protect against a rise in the fuel's aromatics
content from then-existing levels by setting a minimum cetane index of 40 (or, alternatively, a
maximum aromatics level of 35 volume percent).
California set more stringent standards in 1988 for motor vehicle diesel fuels used in the
South Coast air basin.7 These standards took effect statewide in 1993. They apply to both
highway and off-highway fuels (excluding marine and locomotive use), and limit sulfur levels to
500 ppm and aromatics levels to 10 volume percent, with some flexibility provisions to
accommodate small refiners and alternative formulations. Under the provisions for alternate
formulations, fuel manufacturers have certified highway diesel fuel for use in California with a
lower maximum sulfur content than 500 ppm ppm (California highway diesel fuel actually
averages 140 ppm) and a higher maximum aromatics content than 10 percent.
Alaska and certain U.S. territories currently have an exemption from federal highway
diesel fuel requirements. In these areas, the) an ASTM (the American Society for Testing and
Materials) specification on the maximum allowed sulfur content of diesel fuel (5,000 ppm)
applies.8 These regulatory and industry criteria set the upper bound on the sulfur content of
highway diesel fuel in the U.S.
To enable our cost analysis, we compiled the data by various regions called Petroleum
Administrative Districts for Defense (PADDs), as well as for California and the country as a
whole.9 The PADD regions are illustrated in the following figure (IV.A-2).
IV-6
-------
Chapter IV: Fuel Standard Feasibility
Petroleum Administration for Defense (PAD) Districts
11 ' -J:
*• ;£**'* r^«.
•TENN.X* ' "' "
J
i\.C. ?'
/
5.C. •*'
<5A. (^
HAWAII t_J
Figure IV.A-2. Map of U.S. Petroleum Administrative Districts for Defense
Our evaluation revealed relatively little difference in the sulfur content of highway diesel
fuel in PADDs 1, 2, 3, and 4. The sulfur content of highway diesel fuel in California is
considerably lower than that in the rest of the nation due to regulatory requirements specific to
California. The sulfur content of diesel fuel in PADD 5 outside of California and Alaska,
although higher than that within California, is lower than that in PADDs 1 through 4. This is due
to the fact that a large fraction of the highway diesel fuel used in PADD 5 outside of California
and Alaska is manufactured by refineries that are configured primarily to supply lower-sulfur
diesel fuel to the California market. Alaska currently has an exemption from federal highway
diesel sulfur requirements. Please refer to Chapter IX for a discussion of diesel fuel sulfur levels
in Alaska.
IV-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.A-3. Average Highway Diesel Fuel Sulfur Levels by Geographic Area
Sulfur Content
(PPM)
PADD1
340
PADD2
350
PADD3
360
PADD4
330
PADD5
O-CA&AK*
280
CA
140
U.S.
Avg. *
340
* Outside of California and Alaska.
3. Current Levels of Other Fuel Parameters in Highway Diesel Fuel
The refinery process options which could be used reduce the sulfur content of highway
diesel to under 15 ppm have the potential to affect other fuel parameters as well. Highway
diesel fuel is required to meet specifications on a range of fuel parameters .10 If process changes
made to comply with the proposed cap on sulfur content adversely affect other fuel parameters,
refiners may need to take additional steps to ensure that these other parameters meet
specifications. Thus, to determine the most cost-effective sulfur removal methods, it is also
important to evaluate current levels of the other fuel parameters which might be affected by
refinery process changes to meet the sulfur cap. Data on the current distillation characteristics,
API gravity, pour point, natural cetane level, and aromatics content of diesel fuel blendstocks are
contained in the following tables (TV.A-4, IV.A-5, and IV.A-6).
rv-8
-------
Table IV.A-4. Distillation Characteristics of Diesel Blendstocks (CA Excluded)11
Blendstock
Straight-Run
Hydrotreated
Straight Run
Cracked Stock
Hydrotreated
Cracked Stock
Coker Gas Oil
Hydrotreated
Coker Gas Oil
Hydrocrackate
Distillation Fraction
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Distillation ( °F)
T10
325
360
466
421
296
383
431
457
-
346
488
-
284
345
448
457
237
369
454
-
188
359
460
521
-
357
433
-
T30
349
394
510
456
375
412
492
528
-
357
-
508
-
360
501
524
-
382
-
-
210
375
-
564
-
393
501
-
T50
393
421
540
547
405
429
543
584
-
369
526
547
310
385
-
534
314
394
501
-
245
463
504
599
-
435
528
-
T70
422
443
567
575
432
454
576
-
-
384
541
599
-
440
565
-
-
410
531
-
275
494
534
-
-
459
556
-
T90
452
477
601
618
467
484
621
595
-
408
637
666
351
508
613
634
399
436
561
-
305
580
594
628
-
503
617
-
IV-9
-------
Table IV.A-5. Properties of Diesel Blendstocks (CA Excluded)12
Blendstock
Straight-Run
Hydrotreated
Straight Run
Cracked Stock
Hydrotreated
Cracked Stock
Coker Gas Oil
Hydrotreated
Coker Gas Oil
Hydrocrackate
Distillation
Fraction
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Naphtha
Light Distillate
Heavy Distillate
Light Gas Oil
Aromatics
(Vol%)
-
15.9
15.5
-
-
18.6
31.0
-
-
40.2
-
-
-
19.0
45.0
-
8.0
-
-
-
-
22.1
25.1
-
-
-
24.2
-
Cetane #
(Unadditized)
-
40.3
-
45.0
-
44.5
50.4
-
-
-
-
-
-
42.7*
44.1*
-
-
-
-
-
-
45.3
-
36.1
-
-
50.2
-
API
Gravity
50.0
42.2
35.2
30.3
47.1
42.9
34.4
29.9
-
33.1
26.8
22.3
52.6
45.0
30.7
-
51.7
42.4
32.2
-
-
43.1
34.8
29.9
-
41.8
32.9
-
Pour
Point (°F)
70
(additized)
-
-
11
-
-
-
3
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
IV-10
-------
Chapter IV: Fuel Standard Feasibility
While these values are quoted directly from the API/NPRA survey, they are high compared to values found
in other information sources. We understand the cetane number of light cycle oil is normally in the 20s.
IV-11
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.A-6. Average Highway Diesel Fuel Parameter Levels by Geographic Area13
Fuel Parameter
API Gravity
Cetane Number
Unadditized
Cetane Additive
(ppmv)
Cetane Number
[additized]
Pour Point (°F)
[additized]
Pour Point
Depressant
Additive (ppmw)
Distillation
(°F)
T10
T30
T50
T70
T90
Aromatics (Vol %)
PADD1
34.6
-
0
-
[10]
7
426
458
497
549
609
28.9
PADD2
34.2
42.9
83
-
[10]
47
427
470
505
549
600
25.8
PADD3
34.3
43.8
2
-
[2]
7
436
478
514
557
610
37.0
PADD4
36.2
-
12
-
0
11
405
435
495
519
598
27.1
PADD5
(OQ*
33.8
46.5
0
-
[2]
0
432
472
521
554
611
-
CA
33.6
42.6
183
-
6
0
447
-
525
-
612
28.8
U.S.
(OQ*
34.4
44.1
27
-
[5]
19
431
471
510
551
606
32.3
* Outside of California
4. Overview of Diesel Fuel Sulfur Control
As mentioned in Section A.I., the sulfur in diesel fuel comes from the crude oil processed
by the refinery. One way to reduce the amount of sulfur in diesel fuel, therefore, is to process a
crude oil that is lower in sulfur. Some refiners already do this. Others could switch to low or at
least lower sulfur crude oils. However, there is limited capability worldwide to produce low
sulfur crude oil. While new oil fields producing light, sweet crude oil are still being discovered,
most of the new crude oil production being brought on-line is heavier, more sour (i.e., higher
IV-12
-------
Chapter IV: Fuel Standard Feasibility
sulfur) crude oils. The incentive to use low sulfur crude oils has existed for some time and low
sulfur crude oils have traditionally commanded a premium price relative to higher sulfur crude
oils. While a few refiners with access to lower sulfur crude oil could potentially reduce their
diesel sulfur levels in this way, it is not feasible for most, let alone all U.S. refiners to switch to
low sulfur crude oils to meet a tighter diesel fuel sulfur standard. In addition, while helpful, a
simple change to a low sulfur crude oil would fall well short of compliance with the 15 ppm
sulfur cap. Thus, this analysis will not assume that this broad approach could be used to meet
the new highway diesel sulfur standard.
Another method to reduce diesel fuel sulfur is to chemically remove sulfur from the
hydrocarbon compounds which comprise diesel fuel. This is usually accomplished through
reaction with hydrogen at moderate to high temperature and pressure. A couple of specific
examples of this process are hydrotreating and hydrocracking. Another process was announced
recently which uses a moving bed catalyst to both remove and adsorb the sulfur using hydrogen
at moderate temperature and pressure. There are other low temperature and pressure processes
being developed, such as biodesulfurization, and chemical oxidation. Sulfur can be removed via
these processes up front in the refinery, such as from crude oil, before being processed in the
refinery into diesel fuel . Or, sulfur can be removed from those refinery streams which are to be
blended directly into diesel fuel. Finally, another method to reduce diesel fuel sulfur is to shift
sulfur-containing hydrocarbon compounds to other fuels produced by the refinery.
As discussed below, we expect that most of the sulfur reduction required by the sulfur
cap standard will be chemical removal via hydrotreating. Thus, this section will begin with a
relatively detailed discussion of the capabilities of this and similar processes. We also expect
refiners to use the other methods to obtain cost effective sulfur reductions which will
complement the primary sulfur reduction achieved via hydrotreating. These other methods, such
as FCC feed hydrotreating, adsorption, biodesulfurization, chemical oxidation, and undercutting
LCO, will be discussed following the primary discussion of distillate hydrotreating.
As mentioned above, this sulfur removal can occur either early or late in the refining
process. The most practical place to remove sulfur early in the process is prior to the FCC unit.
Hydrotreating feed to the FCC unit requires higher temperatures and pressures than hydrotreating
distillate streams used to produce diesel fuel because FCC feed contains much larger and heavier
molecules. Because of this, FCC feed hydrotreating is more expensive than distillate
hydrotreating. We expect that most refiners will enhance or expand their current distillate
hydrotreating capability to meet the sulfur cap, although the other benefits associated with FCC
feed hydrotreating could lead some refiners to add this technology. The remaining discussion of
hydrotreating will therefore begin with distillate hydrotreating, followed by a brief discussion of
FCC feed hydrotreating.
IV-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
5. Hydrotreating and Other Hydrogen-Based Processes Which
Remove Sulfur
Hydrotreating and similar processes generally combine hydrogen with a hydrocarbon
stream at high temperature and pressure in the presence of a catalyst. Refineries currently
employ a wide range of these processes for a number of purposes. For example, naphtha
(gasoline like material which itself does not meet gasoline specifications, such as octane level)
being fed to the refinery reformer is always hydrotreated to remove nearly all sulfur, nitrogen and
metal contaminants which would deactivate the noble metal catalyst used in the reforming
process. Similarly, feed to the FCC unit is often hydrotreated to remove most of the sulfur,
nitrogen and metal contaminants in order to improve the yield and quality of high value products,
such as gasoline and distillate (distillate refers to a range of similar products including kerosene,
diesel fuel, No. 2 heating oil and jet fuel), from the FCC unit. Nearly all refineries currently
hydrotreat the refinery streams used to produce highway diesel fuel in order to remove much of
the sulfur present and comply with the current 500 ppm sulfur cap. EPA expects that nearly all
refiners will hydrotreat the naphtha produced by the FCC unit to remove most of the sulfur
present to comply with the Tier 2 gasoline sulfur standards.14 These hydrotreating processes
remove sulfur, nitrogen, metal and other contaminants from the hydrocarbon streams. They also
can saturate some or most of the olefins present, depending on the process. However, they do not
generally otherwise affect the chemical nature of the hydrocarbon compounds.
If the temperature or pressure is increased sufficiently, or a carbon-oriented catalyst is
used, hydrotreating can more dramatically affect the chemical nature of the hydrocarbons, as well
as remove contaminants. For example, through a process called hydrocracking, smaller, lighter
molecules are created by splitting larger, heavier molecules. In the process, nearly all of the
contaminants are removed and olefins and aromatics are saturated into paraffins and naphthenes.
Outside the U.S., this process is commonly used to produce distillate from heavier, less
marketable refinery streams. The production of distillate via hydrocracking to produce gasoline
from poor quality distillate, such as LCO from the FCC unit.
A few refineries also currently hydrotreat their distillate more severely than is typical, but
not as severe as hydrocracking. Their intent is to remove the sulfur, nitrogen and metallic
contaminants and to also saturate most of the aromatics present. This is done primarily in
Europe to meet very stringent specifications for both sulfur and aromatics applicable to certain
diesel fuels and encouraged by reduced excise taxes. This severe hydrotreating process is also
used in the U.S. to "upgrade" petroleum streams which are too heavy or too low in quality to be
blended into the diesel pool, by cracking some of the material to lower molecular weight
compounds and saturating some of the aromatics to meet the distillation and cetane requirements.
A different catalyst which encourages aromatic saturation is used in lieu of one that simply
encourages contaminant removal.
IV-14
-------
Chapter IV: Fuel Standard Feasibility
To meet the 15 ppm diesel sulfur cap, EPA expects refiners to focus as much as possible
on sulfur removal. Other contaminants, such as nitrogen and metals, are already sufficiently
removed by existing refinery processes. While saturation of aromatics generally improves diesel
fuel quality, there is a significant additional cost involved, primarily the consumption of
additional hydrogen. Consequently, we anticipate refiners will choose desulfurization processes
that minimize the amount of aromatics saturation. Current diesel fuel already meets all
applicable specifications, and hydrotreating to remove sulfur should not degrade quality, except
possibly lubricity, as discussed in Section C. Thus, with this one exception, there should be no
need to improve diesel fuel quality as a direct result of this new diesel sulfur standard. Should a
refiner choose to do so, it would be to improve profitability,13 and not related to meeting the 15
ppm sulfur cap standard.
As mentioned above, this sulfur removal can occur either early or late in the refining
process. The most practical place to remove sulfur early in the process is prior to the FCC unit.
Hydrotreating feed to the FCC unit requires higher temperatures and pressures than hydrotreating
distillate streams used to produce diesel because FCC feed contains much larger and heavier
molecules. Because of this, FCC feed hydrotreating is more difficult and more expensive than
distillate hydrotreating. We expect that most refiners will enhance or expand their current
distillate hydrotreating capability to meet the 15 ppm sulfur diesel cap standard, although the
other benefits associated with FCC feed hydrotreating will therefore begin with distillate
hydrotreating, followed by a more brief discussion of FCC feed hydrotreating and other emerging
diesel desulfurization technologies.
a. Fundamentals of Distillate Hydrotreating
Essentially all distillate hydrotreater designs follow the same broad format. Liquid
distillate is heated and pumped to temperatures of 300-380°C and pressures of 500-700 psia
with hydrogen and reacted over a catalyst. Hydrogen reacts with sulfur and nitrogen atoms
contained in the hydrocarbon molecules, forming hydrogen sulfide and ammonia. The resulting
vapor is then separated from the desulfurized distillate. The desulfurized distillate is usually
simply mixed with other distillate streams in the refinery to produce diesel fuel and heating oil.
The vapor still contains a lot of valuable hydrogen, because the reaction requires the use
of a significant amount of excess hydrogen to operate efficiently and practically. However, the
vapor also contains a significant amount of hydrogen sulfide and ammonia, which inhibit the
desulfurization and denitrogenation reactions and must be removed from the system. Thus, the
b Refiners can choose to "upgrade" heavy refinery streams which do not meet the cetane and distillation
requirements for highway diesel fuel. The process for doing so is also called ring opening, since one or more of the
aromatic rings of heavy, aromatic molecules are opened up, improving the value of the stream. Upgrading the heavy
refinery streams to highway diesel fuel improves the stream's market price by 10-30 c/gal.
IV-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
hydrogen leaving the reactor is usually mixed with fresh hydrogen and recycled to the front of the
reactor for reaction with fresh distillate feed. However, by itself, this would cause a build up of
hydrogen sulfide and ammonia in the system, since it would have no way to leave the system. In
some cases, the hydrogen sulfide and ammonia are chemically scrubbed from the hydrogen
recycle stream. In other cases, a portion of the recycle stream is simply purged from the system
as a mixture of hydrogen, hydrogen sulfide and ammonia. The latter is less efficient since it
leads to higher levels of hydrogen sulfide and ammonia in the reactor, but it avoids the cost of
building and operating a scrubber.
Desufurization processes in use today in the U.S. generally use only one reactor, due to
the need to only desulfurize diesel fuel to 500 ppm or lower. However, a second reactor can be
used, particularly to meet lower sulfur levels. Instead of liquid distillate going to the diesel
fuel/heating oil pool after the first reactor, it would simply be mixed with fresh hydrogen and
sent to the second reactor.
Traditional reactors are cocurrent in nature. The hydrogen is mixed together with the
distillate at the entrance to the reactor and flow through the reactor together. Because the
reaction is exothermic, heat must be removed periodically. This is sometimes done through the
introduction of fresh hydrogen and distillate at one or two points further down the reactor. The
advantage of cocurrent design is practical, it eases the control of gas-liquid mixing and contact
with the catalyst. The disadvantage is that the concentration of hydrogen is the highest at the
front of the reactor and lowest at the outlet. The opposite is true for the concentration of
hydrogen sulfide. This increases the difficulty of achieving extremely low sulfur levels due to
the low hydrogen concentration and high hydrogen sulfide concentration at the end of the reactor.
The normal solution to this problem is to design a counter-current reactor, where the fresh
hydrogen is introduced at one end of the reactor and the liquid distillate at the other end. Here,
the hydrogen concentration is highest (and the hydrogen sulfide concentration is lowest) where
the reactor is trying to desulfurize the most difficult (sterically hindered) compounds. The
difficulty of counter-current designs in the case of distillate hydrotreating is vapor-liquid contact
and the prevention of hot spots within the reactor. The SynAlliance (consisting of ABB
Lummus, Criterion Catalyst Corp., and Shell Oil Co.) has patented a counter-current reactor
design called SynTechnology. With this technology, in a single reactor design, the initial portion
of the reactor will follow a co-current design, while the last portion of the reactor will be counter-
current. In a two reactor design, the first reactor will be co-current, while the second reactor will
be counter-current.
ABB Lummus estimates that the counter-current design can reduce the catalyst volume
needed to achieve 97 percent desulfurization by 16 percent relative to a co-current design.15 The
impact of the counter-current design is even more significant when aromatics control (or cetane
improvement) is desired in addition to sulfur control.
IV-16
-------
Chapter IV: Fuel Standard Feasibility
Sulfur containing compounds in distillate can be classified according to the ease with
which they are desulfurized. Sulfur contained in paraffins or aromatics with a single aromatic
ring are relatively easy to desulfurize. These molecules are sufficiently flexible so that the sulfur
atom is in a geometric position where it can make physical contact with the surface of the
catalyst. The more difficult compounds are contained in aromatics consisting of two aromatic
rings, particularly dibenzothiophenes. Dibenzothiophene contains two benzene rings which are
connected by a carbon-carbon bond and two carbon-sulfur bonds (both benzene rings are bonded
to the same sulfur atom). This compound is essentially flat in nature and the carbon atoms bound
to the sulfur atom hinder the approach of the sulfur atom to the catalyst surface. Despite this,
today's catalysts are very effective in desulfurizing dibenzothiophenes, as long as only hydrogen
is attached to the carbon atoms bound directly to the sulfur atom.
However, distillate can contain dibenzothiophenes which have methyl or ethyl groups
bound to the carbon atoms which are in turn bound to the sulfur atom. These extra methyl or
ethyl groups further hinder the approach of the sulfur atom to the catalyst surface.
Dibenzothiophenes with such methyl or ethyl groups are commonly referred to as being sterically
hindered. An example of a dibenzothiophene with a single methyl or ethyl group next to the
sulfur atom is 4-methyl dibenzothiophene. An example of a dibenzothiophene with two methyl
or ethyl groups next to the sulfur atom is 4,6-dimethyl dibenzothiophene. In 4,6-dimethyl
dibenzothiophene, and similar compounds, the presence of a methyl group on either side of the
sulfur atom makes it very difficult for the sulfur atom to react with the catalyst surface to assist
the hydrogenation of the sulfur atom.
Most straight run distillates (or straight run light gas oil (SRLGO)) contains relatively low
levels of these sterically hindered compounds. LCO contains the greatest concentration of
sterically hindered compounds, while other cracked distillate streams from the coker and the
visbreaker contain levels of sterically hindered compounds in concentrations between straight run
and LCO. Thus, LCO is generally more difficult to desulfurize than coker distillate which is
more difficult to treat than straight run distillate.16 In addition, cracked stocks, particularly LCO,
have a greater tendency to form coke on the catalyst, which deactivates the catalyst and requires
its replacement.
The greater presence of sterically hindered compounds in LCO is related to two
fundamental factors. First, LCO contains much higher concentrations of aromatics than typical
SRLGO.17 All sterically hindered compounds are aromatics. Second, the chemical equilibria
existing in cracking reactions favors the production of sterically hindered dibenzothiophenes over
unsubstituted dibenzothiophenes. For example, in LCO, methyl substituted aromatics are twice
as prevalent as unsubstituted aromatics. Di-methyl aromatics are twice as prevalent as methyl
aromatics, or four times more prevalent as unsubstituted aromatics. Generally, desulfurizing 4-
methyl dibenzothiophene using conventional desulfurization is 6 times slower than desulfurizing
similar non-sterically hindered molecules, while desulfurizing 4,6-dimethyl dibenzothiophene
IV-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
using conventional desulfurization is 30 times slower. Slower reactions mean that either the
volume of the reactor must be that much larger, or that the reaction must be somehow speeded
up. The latter implies either a more active catalyst, higher temperature, or higher pressure.
These alternatives will be discussed later below.
Because moderate sulfur reduction is often all that is required in distillate hydrotreating,
catalysts have been developed which focus almost exclusively on contaminant removal. The
most commonly used desulfurization catalyst consists of a mixture of cobalt and molybdenum
(Co/Mo). These catalysts interact primarily with the sulfur atom and encourage the reaction of
sulfur with hydrogen.
Other catalysts have been developed which encourage the saturation (hydrogenation) of
the aromatic rings. As mentioned above, this generally improves the quality of the diesel fuel
produced from this distillate. These catalysts also indirectly encourage the removal of sulfur
from sterically hindered compounds by eliminating one or both of the aromatic rings contained in
dibenzothiophene. Without one or both of the rings, the molecule is much more flexible and the
sulfur atom can approach the catalyst surface much more easily. Thus, the desulfurization rate of
sterically hindered compounds is greatly increased through the hydrogenation route. The most
commonly used hydrogenation/desulfurization catalyst consists of a mixture of nickel and
molybdenum (Ni/Mo).
There are a number important issues which should be highlighted about using the
hydrogenation pathway for desulfurization. As pointed out above, one or both of the aromatics
rings are being saturated which significantly increases the consumption of hydrogen. It is
important that one of the aromatic rings of a polyaromatic compound is saturated, as this is the
facilitating step which results in the desulfurization of a sterically hindered compound. If the
mono aromatics compounds are also saturated, there would likely be a further improvement in
the desulfurization reaction rate of the sterically hindered compounds, however at a large
hydrogen cost. In addition, certain diesel fuel qualities, such as cetane, would improve
significantly as more of the aromatic compounds are saturated. However, the vendors of diesel
desulfurization technology explained to us that if cetane improvement is not a goal, then the most
cost effective path to desulfurize the sterically hindered compounds is to saturate the
polyaromatic compounds to monoaromatic compounds, but not to saturate the monoaromatic
compounds. The vendors tell us that because the existence of the monoaromatic compounds is at
equilibrium conditions within the reactor, that the monoaromatic compounds are being both
saturated and unsaturated, which helps to enable the desulfurization of these compounds.
The vendors also point out a number of reasons why the cycle length of the catalysts
which catalyze hydrogenation reactions, which would likely occur in a second stage, is actually
longer than the first stage desulfurization catalyst. First, the temperature at which the
hydrogenation reactions occur to saturate the polyaromatic compounds to monoaromatic
IV-18
-------
Chapter IV: Fuel Standard Feasibility
compounds, but not to saturate the monaromatic compounds is significantly lower than the
higher temperatures of the first stage. The lower temperature avoids color changes problems and
reduces the amount of coke formation on the hydrogenation catalyst. Furthermore, since the first
stage has somewhat "cleaned" the diesel fuel of contaminants such as sulfur, nitrogen and metals,
the catalyst in this hydrogenation seconds stage is not degraded as quickly.
If refiners are "upgrading" their diesel fuel by converting heavy, high aromatic, low
cetane, stocks to highway diesel fuel under the 15 ppm highway diesel fuel sulfur standard, they
are intentionally reacting a lot of hydrogen with the diesel fuel. The hydrogen reactions with the
diesel fuel saturates many or most of the aromatics, increases cetane number and reduces sulfur.
The lower concentration of aromatics and improved cetane of the upgraded feedstock would then
allow the product to be sold as highway diesel fuel. The much higher sales price of the highway
diesel fuel compared to the lower value of the feedstock justifies the much larger consumption in
hydrogen and the cost of a larger reactor.
Up to a certain level of sulfur removal, the CoMo catalyst is generally preferred. It is
more active with respect to desulfurizing non-sterically hindered compounds, which comprise the
bulk of the sulfur in distillate, straight run or cracked. Below that level, the need to desulfurize
sterically hindered compounds leads to greater interest in NiMo catalysts. Acreon Catalysts had
indicated that NiMo are preferred for deep desulfurization around 15 ppm due to this catalyst's
ability to saturate aromatic rings and make the sulfur atom more accessible to the catalyst. On
the other hand, Haldor-Topsoe has performed studies which indicate that CoMo catalysts may
still have an advantage over NiMo catalysts, even at sulfur levels below 50 ppm.18
Two-stage processes may also be preferable to achieve ultra-low sulfur levels. Both
stages could emphasize desulfurization or desulfurization could be emphasized in the first stage
and hydrogenation/desulfurization emphasized in the second stage. In addition to this advantage,
the main advantage of two stages lies in the removal of hydrogen sulfide from the gas phase after
the first stage. Hydrogen sulfide strongly inhibits desulfurization reactions, as will be discussed
further in the next section. It can also recombine with non-sulfur containing hydrocarbon
compounds at the end of the reactor or even in subsequent piping, essentially adding sulfur to the
desulfurized distillate. Removing hydrogen sulfide after the first stage reduces the hydrogen
sulfide concentration at the end of the second stage by roughly two orders of magnitude,
dramatically reducing both inhibition and recombination.
In one study, Haldor-Topsoe analyzed a specific desulfurized 50/50 blend of SRGO and
LCO at 150 ppm sulfur and found that essentially all of the sulfur is contained in sterically
hindered compounds.19 This feed contains more LCO than would be processed in the typical
refinery. A refinery processing less LCO would presumably reach the point where the sulfur
compounds were dominated by sterically hindered compounds at a lower sulfur level. They also
compared the performance of CoMo and NiMo catalysts on a SRLGO feed at the same space
IV-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
velocity. The NiMo catalyst performed more poorly than the CoMo catalyst above 200 ppm
sulfur, and better below that level. This implies that much of the sulfur left at 200 ppm (and even
above this level) was sterically hindered. These two studies indicate that the amount of sterically
hindered compounds can exceed the 15 ppm sulfur cap by a substantial margin.
In addition to NiMo catalysts, precious metal catalysts are also very effective in
desulfurizing sterically hindered compounds. An example of a precious metal catalyst is the
ASAT catalyst developed by United Catalysts and Sud-Chemie AG, which uses both platinum
and palladium.20 They are most commonly used to more severely dearomatize distillate and
increase cetane by opening up the aromatic rings, a process called ring opening.
b. Meeting a 15 ppm Cap with Distillate Hydrotreating
Using distillate hydrotreating to meet a 15 ppm sulfur cap on highway diesel fuel has
been commercially demonstrated, as will be discussed below. Thus, meeting the 15 ppm cap is
quite feasible using current refining technology. Assessing the most reliable and economic
means of doing so is more complicated. Refiners already hydrotreat their highway diesel fuel to
meet a 500 ppm sulfur cap. These hydrotreaters use a variety of catalysts and have a range of
excess capacity. Thus, refiners are not all starting from the same place. Many refiners also
produce off-highway diesel fuel and heating oil, which have much less stringent sulfur
requirements and could, for example, provide a sink for sterically hindered sulfur containing
compounds. Finally, the amount of cracked stocks that a refiner processes into diesel fuel varies
widely. Those with a greater fraction of LCO will face a more difficult task of complying with a
15 ppm cap, than those processing primarily SRLGO.
To understand the types of modifications which can be made to distillate hydrotreaters in
order to improve their performance, it is useful to better understand the quantitative relationships
between the various physical and chemical parameters involved in hydrotreating. Haldor-Topsoe
has developed the following algebraic expression to describe the rate of desulfurization via both
direct desulfurization and hydrogenation/desulfurization.
Rate of = k * C.° * Pma + k * Cm * Pffib
Desulfurization (1 + Kms * Pms) (1 + KF * CF)
Per Catalyst
Surface Area
IV-20
-------
Chapter IV: Fuel Standard Feasibility
where: k, Kms and KF are various rate constants, which only vary with temperature
Cs is the concentration of sulfur in the distillate
PJC and Pms are the partial pressures of hydrogen and hydrogen sulfide in the
vapor phase
KF * CF is the total inhibition due to hydrogen sulfide, ammonia, and aromatics
n, m, a, and b are various constant exponents
The first term represents the rate of direct desulfurization, such as that catalyzed by
CoMo. This reaction rate increased by increasing the partial pressure of hydrogen. However, it
is inhibited by increasing concentrations of hydrogen sulfide, which competes with the distillate
for sites on the catalyst surface.
The second term represents the rate of desulfurization via hydrogenation of the aromatic
ring next to the sulfur atom. This rate of desulfurization also increases with higher hydrogen
partial pressure. However, this reaction is inhibited by hydrogen sulfide, ammonia, and
aromatics. This inhibition by aromatics leads to the presence of a thermodynamic equilibrium
condition which can prevent the complete saturation of aromatics. Also, this inhibition makes it
more difficult to desulfurize cracked stocks, which contain high concentrations of both sterically
hindered sulfur compounds and aromatics. While the literature generally expresses a preference
for NiMo catalysts for desulfurizing cracked stocks, Haldor-Topsoe has found situations where
this aromatics inhibition leads them to favor CoMo catalysts even for desulfurizing feeds with a
high concentration of sterically hindered compounds.
These relationships essentially identify the types of changes which could be made to
improve the performance of current distillate hydrotreaters. First, a more active catalyst can be
used. This increases the "k" terms in the above equations. Second, temperature can be
increased, which also increases the "k" terms in the above equations. Third, improvements can
often be made in vapor-liquid contact, which effectively increases the surface area of the catalyst.
Fourth, hydrogen purity can be increased. This increases the Pm term in the two numerator terms
of the equation. Fifth, the concentration of hydrogen sulfide in the recycle stream can be
removed by scrubbing. This decreases the Pms and CF terms in the two denominator terms of the
equation. Finally, more volume of catalyst can be used, which increases the surface area
proportionally.
Regarding catalysts, at least two firms have announced the development of improved
catalysts since the time that most distillate hydrotreaters were built in the U.S. to meet the 1993
500 ppm sulfur cap: Akzo Nobel / Nippon Ketjen Catalysts (Akzo Nobel) and Haldor-Topsoe.
Akzo Nobel currently markets four CoMo desulfurization catalysts: KF 752, KF 756 and KF 757
which have been available for several years, and KF 848, which was announced this year.21 KF
752 can be considered to be typical of an Akzo Nobel catalyst of the 1992-93 timeframe, while
KF 756 and 757 catalysts represent improvements. Akzo Nobel estimates that under typical
IV-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
conditions (e.g., 500 ppm sulfur), KF 756 is 25 percent more active than KF 752, while KF 757
is more than 50 percent more active than KF 752 and 30 percent more active than KF 756.22
However, under more severe conditions (e.g., <50 ppm sulfur), KF 757 is 35-75 percent more
active than KF 756. KF 848 is 15 - 50 percent more active than KF 757. Commercial experience
exists for both advanced catalysts. KF 756 is widely used in Europe (20 percent of all distillate
hydrotreaters operating on January 1, 1998), while KF 757 has been used in at least three
hydrotreaters commercially. KF 757 and KF 842 utilizes what Akzo Nobel calls STARS
technology, .Super Type U Active Reaction .Sites. Type U refers to a specific kind of catalyst site
which is particular good at removing sulfur from sterically hindered compounds.
In terms of sulfur removal, Akzo Nobel projects that a desulfurization unit producing 500
ppm sulfur with KF 752, would produce 405, 270 and 160ppm sulfur with KF 756, KF757, and
KF 842, respectively.
Haldor-Topsoe has also developed a more active catalyst. Its TK-554 catalyst is
analogous to Akzo Nobel's KF 756 catalyst, while its newer, more active catalyst is termed TK-
574. For example, in pilot plant studies, under conditions where TK-554 produces 400 ppm
sulfur in SRLGO, TK 574 will produce 280 ppm. Under more severe conditions, TK-554 will
produce 60 ppm, while TK 574 will produce 30 ppm. Similar benefits are found with a mixture
of straight run and cracked stocks.
UOP projects a similar reduction in sulfur due to improved catalyst. They estimate that a
hydrotreater producing 500 ppm sulfur distillate today (20% LCO, 10% light coker gas oil) could
produce 280 ppm sulfur distillate with a 50 percent more active catalyst.23
Over the last two years, Criterion Catalyst Company announced two new lines of
catalysts. One is called Century, and the other is called Centinel.24 These two lines of catalysts
are reported to be 45 - 70 percent and 80 percent more active, respectively, at desulfurizing
petroleum fuel than conventional catalysts used in the mid-90s. These improvements have come
about through better dispersion of the active metal on the catalyst substrate.
Thus, by itself, changing to a more active catalyst can reduce sulfur moderately. Based
on the history of the industry, improvements in catalyst performance can be anticipated over time
to result in roughly a 25 percent increase in catalyst activity every 4 years. Vendors have
informed EPA that the cost of these advanced catalysts is very modest relative to less active
catalysts. This will help to reduce the reactor size needed, but by itself would not appear to be
sufficient for most refiners to meet a 15 ppm
The second type of improvement is to reduce the concentration of hydrogen sulfide,
which reduces the inhibition of the desulfurization and hydrogenation reactions. Hydrogen
sulfide can be removed by chemical scrubbing. Haldor-Topsoe indicates that decreasing the
IV-22
-------
Chapter IV: Fuel Standard Feasibility
concentration of hydrogen sulfide at the inlet to a co-current reactor by three to six volume
percent can decrease the average temperature needed to achieve a specific sulfur reduction by 15-
20°C, or reduce final sulfur levels by more than two-thirds. UOP projects that scrubbing
hydrogen sulfide from recycle hydrogen can reduce sulfur levels from roughly 285 to 180 ppm in
an existing hydrotreater.
The third type of improvement which can be made to current distillate hydrotreaters is to
improve vapor-liquid contact. Akzo Nobel estimates that an improved vapor-liquid distributor
can reduce the temperature necessary to meet a 50 ppm sulfur level by 10 °C, which in turn
would increase catalyst life and allow an increase in cycle length from 10 to 18 months. Based
on the above data from Haldor-Topsoe, if temperature were maintained, the final sulfur level
could be reduced by 50 percent. Similarly, in testing of an improved vapor-liquid distributor in
commercial use, Haldor-Topsoe found that the new distributor allowed a 30 percent higher sulfur
feed to be processed at 25°C lower temperatures, while reducing the sulfur content of the product
from 500 to 350 ppm. Maintaining temperature should have allowed an additional reduction in
sulfur of more than two-thirds. Thus, ensuring adequate vapor-liquid contact can have a major
impact on final sulfur levels.
The fourth improvement possible is to increase hydrogen partial pressure and/or purity.
As discussed above, this increases the rate of both desulfurization and hydrogenation reactions.
Haldor-Topsoe indicates that increasing hydrogen purity is preferable to a simple increase in the
pressure of the hydrogen feed gas, since the latter will also increase the partial pressure of
hydrogen sulfide later in the process, which inhibits both beneficial reactions. Haldor-Topsoe
projects that an increase in hydrogen purity of 30 percent would lower the temperature needed to
achieve the same sulfur removal rate by eight to nine °C. Or temperature could be maintained
while increasing the amount of sulfur removed by roughly 40 percent. Hydrogen purity can be
increased through the use of a membrane separation system or a PSA unit. UOP project that
purifying hydrogen can reduce distillate sulfur from 180 to 140 ppm from an existing
hydrotreater.
The fifth type of improvement is to increase reactor temperature. Haldor-Topsoe has
shown that an increase of 14°C while processing a mix of SRLGO and LCO with its advanced
TK-574 CoMo catalyst will reduce sulfur from 120 ppm to 40 ppm.25 UOP projects that a 20 °F
increase in reactor temperature would decrease sulfur from 140 to 120 ppm. The downside of
increased temperature is reduced catalyst life (i.e., the need to change catalyst more frequently).
This increases the cost of catalyst, as well as affects highway diesel fuel production while the unit
is down for the catalyst change. Still, current catalyst life currently ranges from six to 60 months,
so some refiners could increase temperature and still remain well within the range of current
industry performance. The relationship between temperature and life of a catalyst is a primary
criterion affecting its marketability. Thus, catalyst suppliers generally do not publish these
figures.
IV-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Sixth, additional sulfur can be removed by increasing the amount of recycle gas sent to
the inlet of the reactor. However, the effect is relatively small. Haldor-Topsoe indicates that a
50 percent increase in the ratio of total gas/liquid ratio only decreases the necessary reactor
temperature by six to eight °C. Or, temperature can be maintained and the final sulfur level
reduced by 35-45 percent.
Seventh, reactor volume can be increased. UOP projects that doubling reactor volume
would reduce sulfur from 120 to 30 ppm.
These individual improvements described cannot be simply combined, either additively
or multiplicatively. As mentioned earlier, each existing distillate hydrotreater is unique in its
combination of design, catalyst, feedstock, and operating conditions. While the improvements
described above are probably indicative of improvements which can be made in many cases, it is
not likely that all of the improvements mentioned are applicable to any one unit; the degree of
improvement could either be greater than, or less than the benefits that are indicated.
Therefore, many refiners may have to implement one additional technical change listed
by UOP to be able to meet the 15 ppm cap standard. This last technical change is to install a
complete second stage to the existing, one-stage hydrotreater. This second stage would consist
of a second reactor, and a high pressure, hydrogen sulfide scrubber between the first and second
reactor. The compressor would also be upgraded to allow a higher pressure to be used in the new
second reactor. Assuming use of the most active catalysts available in both reactors, UOP
projects that converting from a one-stage to a two-stage hydrotreater could produce 5 ppm sulfur
relative to a current level of 500 ppm today.
In addition to these major technological options, refiners may have to debottleneck or
add other more minor units to support the new desulfurization unit. These units could include
hydrogen plants, sulfur recovery plants, amine plants and sour water scrubbing facilities. All of
these units are already operating in refineries but may have to be expanded or enlarged.
Overall, Akzo-Nobel projects that current hydrotreaters can be modified short of a
revamp to achieve 50 ppm sulfur. Acreon/IFP/Procatalyse is less optimistic, believing that more
than a catalyst change will be necessary to meet this sulfur level.26 BP-Amoco projects that a 70
percent improvement in catalyst activity could reduce sulfur from a current hydrotreater meeting
a 500 ppm sulfur specification to 30 ppm.27 While this improvement is somewhat greater than
the 50 percent improvement measured by Akzo Nobel at current desulfurization severity, it
indicates that it may be possible to improve current hydrotreaters to produce distillate sulfur
levels in the 50-100 ppm range. Thus, it appears that additional reductions needed to meet a 15
ppm cap would require additional measures. To assess the degree that these measures would be
needed, it is useful to examine the commercial and pilot plant performance of distillate
hydrotreaters to achieve very low sulfur levels.
IV-24
-------
Chapter IV: Fuel Standard Feasibility
After reviewing the technology for desulfurizing diesel fuel, and discussing the
advancements in catalyst technology, it is appropriate to turn to how refiners will invest to meet
the 15 ppm cap standard. Refiners have a choice of continuing to use their existing distillate
hydrotreater, or to not use that equipment and design an entirely new hydrotreater. As will be
shown below, numerous commercial examples exist where existing hydrotreaters have been
modified to improve their sulfur removal performance. The vendors of desulfurization
technology assert that refiners can meet the 15 ppm cap standard by revamping their existing
diesel hydrotreating units. However, several refiners we spoke to indicated that they foresee
replacing one or more of their existing diesel hydrotreaters with a brand new "grassroots" unit.
One refiner stated that they plan to use the idled units in other places in the refinery.
We gathered more information on whether refiners would revamp versus install a
grassroots unit during a session on diesel hydrotreating at this year's NPRA Q & A conference.
The refiners on the panel and in the audience were asked if they would scrap their existing diesel
hydrotreaters to install a new grassroots unit if they were faced with the proposed EPA highway
diesel standard. The response by one of the refiners was that refiners would not waste good
capital units in the refinery, suggesting that the refiners would revamp their existing diesel
hydrotreaters with additional capital. However the refiner went on say that some refiners may
choose to convert their existing diesel hydrotreaters to gasoline hydrotreaters, or to processing
nonroad diesel fuel, once any nonroad requirements are known, and then put in a grassroots unit
for diesel hydrotreating. That responder supposed that a refiner might choose to scrap a unit if it
"was very, very old," however, when considering the tone of the commenter's remarks, he
implied that few refiners would indeed scrap their existing highway diesel hydrotreaters.
Another refiner said that they currently are not producing as much highway diesel fuel as they
would like and that they might build a grassroots unit which would allow them to expand their
highway diesel production.
Charles River and Baker and O'Brien, in a study of the cost of desulfurizing diesel fuel
for API, also considered the issue of revamps versus grassroots units.28 The API contractors used
a set of assumptions to estimate how many of the desulfurization units that would be built to
meet a 15 ppm cap standard would be revamped units versus brand new grassroots units. An
important assumption of their analysis is that to meet a 15 ppm cap standard, both the first and
second stages of diesel desulfurization require moderate to high pressure (800 psi or higher) if
LCO is present in the feed to be treated. They also assume that all diesel desulfurization units
installed in 1993 to meet the 500 ppm highway diesel sulfur standard are capable of this pressure,
while the units which were converted over from another service are not. Finally, the study
assumes that a refinery with a hydrocracker is processing its LCO in the hydrocracker and not
processing it in the diesel hydrotreater. Based on these assumptions, the study assumed a refiner
would revamp a diesel desulfurization unit installed in 1993, and would revamp an older unit if
the refinery had a hydrocracker. By deduction, the study assumed that the refineries which had
IV-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
converted an existing hydrotreating unit to diesel desulfurization service in 1993 but did not have
a hydrocracker, would not revamp and put in a grassroots unit. According to API's contractors,
this set of assumptions would result in about 60 percent of the refineries revamping their existing
desulfurization units and 40 percent of the refineries putting in new grasssroots units. The
contractors did not include the information which served as the basis for their assumptions about
revamps versus grassroots in their final report, and when we asked API for the information, they
would not share it with us.
A cost study was conducted by the National Petroleum Council in response to a request
from the Department of Energy to study the cost of desulfurizing diesel fuel.29 This study
estimated the cost to desulfurize diesel fuel down to an average of 30 ppm. An assumption of the
cost study was that current diesel hydrotreaters producing 50 percent of the highway diesel fuel
would be retrofitted to meet a 30 ppm sulfur standard, while the other diesel hydrotreaters
producing 50 percent of highway diesel fuel would be replaced by new grassroots units. Despite
that this study analyzed a much less severe diesel sulfur standard, the study assumed that the
industry would have to resort to more grassroots units than the API study.
We project that some refiners will put in new grassroots units. We believe that those
refiners that put in grassroots units will do so because they can most economically meet the
combination of the Tier 2 gasoline sulfur standard and this highway diesel sulfur standard by
converting their existing diesel desulfurization unit to meet the gasoline sulfur requirement. Or,
in a few cases, refiners will put in grassroots units because the unit is too old to operate reliably
enough to produce diesel on a regular basis which meets the 15 ppm cap standard. However,
when we compare the assumptions made in the API and NPC studies to our discussions with
refiners and with the comments made by refiners at the NPRA Q & A meeting, we believe that
the API and NPC assumptions are too conservative. Regardless of the operating pressure of their
existing diesel desulfurization unit, refiners are able to desulfurize distillate down to under 500
ppm to meet the existing highway diesel sulfur standard, a sulfur reduction on the order of 95
percent. In meeting a 15 ppm cap standard, this existing sulfur reduction would provide an
important first stage reduction for meeting a 15 ppm sulfur cap standard. We also believe that
refiners would not have much to gain by replacing this first stage with a higher pressure first
stage. After considering the comments made by the refiners at the NPRA Q & A meeting, the
comments made by vendors, and considering that there are few compelling reasons for going
with a grassroots unit, we project that the percentage of refiners putting in grassroots units will be
between 10 to 30 percent. For our cost analysis we used the average of this range, which is 20
percent.
c. Low Sulfur Performance of Distillate Hydrotreating
Data from both pilot plant studies and commercial performance are available which
indicate the capability of various hydrotreating technologies to reduce distillate sulfur levels to
IV-26
-------
Chapter IV: Fuel Standard Feasibility
very low levels. While many studies are available which focus on reducing sulfur to 500 ppm or
below, studies of achieving lower sulfur levels (e.g., 10-50 ppm) also focus on reducing
aromatics content significantly. This combination is related to the fact that Swedish Class II
diesel fuel must meet a tight aromatics specification in 2005 along with a 10 ppm sulfur cap.
Reducing aromatic content along with sulfur content is generally desirable with respect to
diesel fuel quality, as aromatic reductions increase cetane levels and generally improve
combustion characteristics. However, reducing aromatics consumes hydrogen and increases the
cost of desulfurization relative to a case where only sulfur was being removed. A number of
catalyst and engineering firms have projected the technology necessary to just reduce sulfur
without a mandated aromatics reduction (and its cost) for EPA, the Engine Manufacturers
Association, the American Petroleum Institute, the National Petroleum Council and others.
These projections will be discussed in the next chapter on the economic impacts of this rule.
The discussion in this chapter will focus on the available pilot plant and commercial data
demonstrating the achievement of low sulfur levels. It is worth noting that until the recent
announcements by the German government to seek sulfur levels as low as 10 ppm, there had
been little effort by industry to develop technology capable of such a level across the diesel pool.
Recent advancements by catalyst manufacturers demonstrating the feasibility of producing diesel
fuel which meets these levels through pilot plant testing should be considered a first-generation
of technology, with new and continual advancements expected over time.
Starting with SynTechnology, as of August 2, 1999, there were 24 units either in
operation or in the process of being constructed. Their purposes range from desulfurization to
desulfurization plus dearomatization to mild hydrocracking. Of particular interest here is a
revamp of an existing two reactor distillate hydrotreater at the Lyondell / Citgo refinery in Texas.
The revamped unit was designed to process a low-cost feed very heavily weighted
towards cracked material (65-70 percent LCO and LCGO). One existing reactor was converted
to SynSat Technology, while the other was used simply as a flash drum. A new first-stage
reactor was added. Both reactors were designed to operate in a co-current fashion. Pilot plant
studies predicted average sulfur and aromatics levels of seven ppm and 31 volume percent,
respectively, based on feed sulfur and aromatics levels of 11,900 ppm and 53 volume percent,
respectively. The unit exceeded expectations in the case of sulfur, producing an average sulfur
level of less than five ppm from a feed sulfur level of 13,800 ppm. The actual aromatic level
achieved was above the target by four volume percent, but the feed aromatic level was five
volume percent higher than expected. Thus, the net reduction in aromatic content in terms of
volume percent was still higher than found in the pilot plant. ABB Lummus and Criterion
indicate that their catalyst technology is sufficiently flexible to focus on the deep desulfurization
with or without the significant aromatics reduction seen here. This is reflected in their projection
of the technology needed to meet a 15 ppm sulfur cap which is discussed in the next chapter.
IV-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
While this two-stage unit initially produced less than 5 ppm product, it does not do so
consistently. The primary purpose of the unit is to increase cetane so that the product can be
blended directly into diesel fuel. The primary sulfur reduction requirement is to protect the noble
metal catalyst in the second stage reactor. This generally requires that the product from the first
stage be less than 50 ppm. Thus, if the cetane specifications are being met at less severe
conditions, there is no incentive to reduce sulfur any further than necessary for catalyst
protection. In addition, the unit is seeing a heavier feedstock than designed, and the
desulfurization reactor is being operated at a lower temperature than designed to increase the
cycle lengths.
IFF, in conjunction with various catalyst manufacturers, offers its Prime D technology for
deep desulfurization, aromatics saturation and cetane improvement.30 Using a NiMo catalyst,
IFP's Prime D process can produce distillate sulfur levels of 10 ppm from SRLGO and of less
than 20 ppm from distillate containing 20-100 percent cracked material using a single stage
reactor. With a two-stage process, less than one ppm sulfur can be achieved.
United Catalysts and Sud-Chemie AG have published data on the performance of their
AS AT catalyst, which uses platinum and palladium.31 The focus of their study was to reduce
aromatics to less than 10 volume percent starting with a feed distillate containing up to 500 ppm
sulfur and at least 100 ppm nitrogen. Starting with a feed distillate containing 400 ppm sulfur
and 127 ppm nitrogen and 42.5 volume percent aromatics, the ASAT catalyst was able to reduce
sulfur to eight to nine ppm, essentially eliminate nitrogen and reduce aromatics to two to five
volume percent. Hydrogen consumption was 800-971 standard cubic feet per barrel (SCFB).
Akzo Nobel recently presented a summary of the commercial experience of about a years
worth of operations of their STARS catalyst for desulfurizing diesel fuel at the BP-Amoco
refinery in Grangemouth, UK.32 The original unit was designed to produce 35,000 barrels per day
of diesel fuel at 500 ppm treating mostly straight run material, but some LCO was treated as well.
Akzo Nobel's newest and best catalyst (KF 757 at that time) was dense-loaded0 into the reactor to
produce 45,000 barrels per day diesel fuel at 10 - 20 ppm (to meet the 50 ppm cap standard).
From the data, it was clear to see that as the space velocity changed, the sulfur level changed
inversely proportional to the change in space velocity. Usually when the space velocity dipped
below 1.0, the sulfur level dropped below 10 ppm. At that refinery, however, it was not
necessary to maintain the sulfur level below 10 ppm.
These studies indicate the commercial feasibility of producing diesel fuel with 10 ppm or
less sulfur. The primary issue remaining is to commercially demonstrate that the 15 ppm cap
standard can be met using the desulfurization/hydrogenation method without saturating much of
0 Dense loading is a process of packing a certain volume of catalyst into a smaller space than
conventional catalyst loading.
IV-28
-------
Chapter IV: Fuel Standard Feasibility
the aromatics in diesel fuel, especially with a feedstock blend which contains a substantial
amount of cracked material. The ease or difficulty of accomplishing this depends on the amount
of cracked stocks that the refiner blends into diesel fuel and the possibility of shifting some of the
sterically hindered compounds to fuels complying with less stringent sulfur standards, such as
off-highway diesel fuel and heating oil.
d. Undercutting Cracked Stocks
The primary stumbling block preventing the simple desulfurization of distillate to sulfur
levels meeting the 15 ppm cap is the presence of sterically hindered compounds, particularly
those with two methyl or ethyl groups blocking the sulfur atom. These compounds are aromatic
in nature, and are found in greatest concentration in LCO, which itself is highly aromatic. These
compounds can be desulfurized readily if saturated. However, due to the much higher hydrogen
cost of doing so, it is better economically if this can be avoided. Because these compounds are
inherently large in molecular weight due to their chemical structure, they distill near the high end
of the diesel range of distillation temperatures. Thus, it is possible to segregate these compounds
from the rest of the cracked stocks via distillation and avoid the need to desulfurize them.
Once separated, this LCO material could be mixed into the refinery streams currently
being used to produce off-highway diesel fuel and heating oil. These fuels would still have to
meet applicable quality specifications, such as cetane, density, sulfur and distillation. For
example, the industry specification for non-road diesel fuel is a minimum of 40 cetane number,
and a maximum sulfur concentration of 5,000 ppm.33 An analysis of off-highway diesel fuel
shows that off-highway diesel fuel averages 44.4 cetane number, 3,300 ppm sulfur, 34.5 API
gravity, T10 of 438 °F, T50 of 517 °F, and T90 of 600 °F.34 We anticipate that refiners would
need to use cetane additives to compensate for the addition of LCO to maintain off-highway fuel
cetane levels similar to those of current in-use fuels (see Section V). Additional cold-flow
additives might also be necessary for off-highway diesel fuel in the winter to maintain cold-flow
performance at current levels. We anticipate that refiners would allow other off-highway and
heating oil properties to change as a result of the addition of LCO, while continuing to ensure
that all specifications on these fuels are met.
Shifting LCO to off-highway diesel fuel and heating oil would prevent the need to
desulfurize a sizeable fraction of the sterically hindered compounds currently present in highway
diesel fuel. For example, Akzo Nobel studies indicate that a drop of 10 °C in the 95th percentile
distillation point (T95) of diesel fuel decreases sulfur from 50 - 60 ppm.35 Of course, such a
shift to non-highway diesel fuel markets would decrease the amount of highway diesel fuel
produced, about 3 percent for the typical refinery, if more easy to hydrotreat material was not
switched from non-highway diesel fuels to the highway diesel fuel pool. A decrease of T95 of
this magnitude effected by undercutting only LCO would decrease sulfur even more because the
sulfur levels in the heaviest portions of LCO are much greater than those in SRLGO and are the
most difficult to desulfurize. Shifting only heavy LCO would increase the sulfur reduction per
IV-29
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
volume of highway diesel fuel lost, but would still result in a net loss of highway diesel fuel
production if no other feedstocks replaced it.
While this heavy LCO material could be shifted to other markets, this does not
necessarily have to be the case. Under certain conditions, this material can be recycled to the
FCC unit.36 For this to be feasible, the refiner must hydrotreat the FCC feed at a pressure
sufficient to desulfurize the sterically hindered sulfur containing compounds and the feed
hydrotreater must have sufficient excess capacity to handle the additional material. This material
could also be sent to an existing hydrocracker, if sufficient capacity existed, and converted into
gasoline blendstock. Or, it could be hydrotreated separately under more severe conditions to
remove the sulfur, such as with SynAlliance's SynShift process. This would entail higher
hydrogen consumption per barrel of treated material because of some aromatic saturation.
However, the amount of material being processed would be small, so overall hydrogen
consumption would still be low.
A number of vendors of distillate desulfurization processes recently developed specific
projections of the technology needed to meet a range of highway diesel fuel sulfur levels. These
projections were developed to support refining cost studies conducted by the Engine
Manufacturers Association and the American Petroleum Institute, and the National Petroleum
Council.d These projections addressed compliance with three different average sulfur levels: 10,
30 and 100 ppm. Generally, these projections indicate that it will be possible for refiners to
meet the 10 ppm average sulfur level without resorting to catalysts and operating conditions
which reduce aromatic levels dramatically. Thus, it appears that the cost of providing sufficient
hydrogen to saturate these aromatics can be avoided. The specifics of these projections will be
addressed in more detail in the next chapter.
6. Other Desulfurization Technologies
a. Biodesulfurization
Biodesulfurization is essentially an alternative to distillate hydrotreating. This process is
being developed by Energy Biosystems. It involves the removal of sulfur-containing
hydrocarbon compounds from distillate or naphtha streams using bacteria. The distillate stream
is first mixed with an aqueous media containing the bacteria, caustic soda and nutrients for the
bacteria. Enzymes in the bacteria first oxidize the sulfur atoms and then cleaves some of the
sulfur-carbon bonds. The sulfur leaves the process in the form of hydroxyphenyl benzene
sulfonate, which can be used commercially as a feedstock to produce surfactants. Designs based
on pilot plant studies combine biodesulfurization with conventional hydrotreating to produce
diesel fuel containing 50 ppm sulfur.
d See Chapter V for additional discussion on these projections.
IV-30
-------
Chapter IV: Fuel Standard Feasibility
b. Chemical Oxidation and Extraction
Another desulfurization technology was announced by Petrostar this year which
desulfurizes diesel fuel using chemical oxidation.37 Desulfurization of diesel fuel is
accomplished by first forming a water emulsion with the diesel fuel. In the emulsion, the sulfur
atom is oxidized to a sulfone using catalyzed peroxyacetic acid. With an oxygen atom attached
to the sulfur atom, the sulfur-containing hydrocarbon molecules becomes polar and hydrophilic
and then move into the aqueous phase. Like biodesulfurization, some of the sulfones can be
converted to a surfactant which could be sold to the soap industry at an economically desirable
price. The earnings made from the sales of the surfactant would offset much of the cost of
oxidative desulfurization.
We are aware of another chemical oxidation process which currently is in the patenting
process. This process is similar to the Petrostar process, except instead of keeping the sulfone
intact, this process separates the oxidized sulfur atom from the hydrocarbon immediatly after the
oxidation reaction. The resulting sulfate is easily separable from the petroleum. While this
process does not create a valuable byproduct, it would likely be a less capital intensive means to
make the sulfur separation than the Petrostar process.
c. Sulfur Adsorption
A prospective diesel desulfurization process was recently announced by Phillips
Petroleum.38 This process is an extension of their S-Zorb process for gasoline. S-Zorb for diesel
contacts highway diesel fuel (typically with about 350 ppm sulfur) with a catalyst in a reactor at
relatively low pressures and temperature in the presence of hydrogen. The sulfur atom of the
sulfur-containing compounds adsorbs onto the catalyst. The catalyst next cleaves the sulfur atom
from the sulfur-containing hydrocarbon. To prevent the accumulation of sulfur on the catalyst,
the catalyst is continually removed from the reactor. In a separate regeneration vessel, the sulfur
is burned off of the catalyst and is sent to the sulfur plant. The regenerated catalyst is then
recycled back to the reactor for removing more sulfur. Because the catalyst is continuously being
regenerated, the catalyst should never force the unit to be shutdown, thus, Phillips estimates that
the unit will be able to operate 4-5 years between shutdowns. Because untreated distillate can
contain several percent sulfur, Phillips believes that its S-Zorb process for diesel could get
overwhelmed by the amount of sulfur which is adsorbing onto the catalyst. Thus, the S-Zorb
process may not be able to treat untreated distillate streams, but would likely be used to treat
distillate containing 500 ppm sulfur or less.
Phillips' diesel desulfurization process has only been demonstrated in the laboratory up to
this point. The laboratory testing has shown that diesel with LCO can be desulfurized down
below 5 ppm. However, Phillips is on the fast track to demonstrate this process in a pilot plant
and in a commercial unit. First, the company reports that its S-Zorb commercial demonstration
unit for gasoline is on schedule to startup the first quarter of 2001. Since the process has never
IV-31
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
been demonstrated commercially, this demonstration unit will go a long way toward proving that
the Phillips process works as designed. However, the sulfur compounds in diesel fuel are
different, usually more refractory, than those in gasoline. Phillips reports, though, that the
absorption catalyst more readily desulfurizes the sterically hindered sulfur compounds than the
thiophenes (single ring compounds which contain sulfur) which must be desulfurized in gasoline.
This suggests the possibility that S-Zorb for diesel may actually desulfurize current highway
diesel fuel more easily then S-Zorb for gasoline. Phillips projects that they will have an S-Zorb
diesel desulfurization pilot plant up and running by the third quarter of 2001, and a commercial
unit up and running during the fourth quarter of 2003. After hearing Phillips' timeline for
developing the S-Zorb diesel desulfurization process, and weighing the uncertainty associated
with S-Zorb, it seems that refiners may consider this process too risky for 2006. However, this
process could be far enough along in its development to be used by refiners in 2010.
d. FCC Feed Hydrotreating
At the beginning of Section 3., it was mentioned that sulfur could be removed from
distillate material early or late in the refining process. Early in the process, the most practical
place to remove sulfur early in the process is prior to the FCC unit. The FCC unit primarily
produces gasoline, but it also produces a significant quantity of distillate, called LCO as
described in Section A. 1 above. LCO is high in aromatics and sulfur and contains a relatively
high fraction of the sterically hindered sulfur compounds found in diesel fuel.
Many refineries already have an FCC feed hydrotreating unit. The LCO from these
refineries should contain a much lower concentration of sterically hindered compounds than
refineries not hydrotreating their FCC feed. Adding an FCC feed hydrotreating is much more
costly than distillate hydrotreating. Just on the basis of sulfur removal, FCC feed hydrotreating is
more costly than distillate hydrotreating, even considering the need to reduce gasoline sulfur
concentrations, as well. This is partly due to the fact that FCC feed hydrotreating by itself is
generally not capable of reducing the level of diesel fuel sulfur to those being considered in this
rule. However, FCC feed hydrotreating provides other environmental and economic benefits.
FCC feed hydrotreating decreases the sulfur content of gasoline significantly, as well as reducing
sulfur oxide emissions from the FCC unit. Economically, it increases the yield of relatively high
value gasoline and LPG from the FCC unit and reduces the formation of coke on the FCC
catalyst. For individual refiners, these additional benefits may offset enough of the cost of FCC
hydrotreating to make it a more economical than distillate hydrotreating. However, these
benefits are difficult to estimate in a nationwide study such as this. Thus, this study will rely on
distillate hydrotreating as the primary means with which refiners would meet the 15 ppm sulfur
cap. For those refiners who would choose FCC feed hydrotreating, their costs would be
presumably lower than distillate hydrotreating and the costs estimated in the next chapter can
then be considered to be somewhat conservative in this respect.
IV-32
-------
Chapter IV: Fuel Standard Feasibility
7. Will There Be Enough Supply of Highway Diesel Fuel?
First, in assessing the cost of desulfurizing highway diesel fuel, we began with the
assumption that sufficient desulfurization equipment would have to be constructed to adequately
supply diesel highway vehicles, as well as other users of highway diesel fuel. We examined
historic production and demand of highway diesel fuel, factored in growth per estimates by EIA,
and determined that amount of highway diesel fuel which would have to meet the 15 ppm cap
both during and after the initial period during which the temporary compliance option and
various hardship provisions are in effect.
The issue of future supply of highway diesel fuel was raised in the NPRM and received
considerable attention during the comment period. Numerous commenters to the proposed rule
indicated that they believed that the 15 ppm sulfur cap would cause shortages in highway diesel
fuel supply. A number of commenters also thought otherwise (i.e., that future supplies would be
adequate). These comments are summarized in Section 8.1.1 of the Response to Comments
document for this rule. The factors which affect refiners' decisions on how much 15 ppm diesel
fuel to produce fall into the following categories:
- Required investment per refinery
- Historic refining profit margins
- Variation in compliance costs between refiners
- Other markets for highway diesel fuel
- Uncertainty in requisite desulfurization control technology
Likely price and import response to new sulfur standard
- Impact of desulfurization processes on fuel volume, and
Impact of fuel transport on fuel supply
Each of these factors is addressed below. In addition, the findings of a study performed
by Charles River Associates and Baker and O'Brien for API concerning the potential supply
impacts of the new sulfur standard are discussed at the end of this section.
a. Required Investment per Refinery
The first issue is that the level of investment per refinery required to meet this diesel
sulfur standard is more than that required to meet the recent Tier 2 gasoline sulfur standard. This
is true. We projected that it would cost $44 million per refinery to meet the Tier 2 gasoline
sulfur standards, while we project that it will cost $50 million per refinery to meet the diesel fuel
sulfur cap. In addition, this $50 million figure represents the average of revamped units (which
will cost less) and new units (which will cost more). Revamping an existing diesel hydrotreater
(representing roughly 80 percent of all current units) will cost roughly $40 million, while a new
diesel hydrotreater will cost $80 million. Thus, roughly 25 refineries will face twice the
IV-33
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
investment cost to meet this diesel standard as they did to meet the Tier 2 gasoline sulfur
standards.
This difference in investment is to be expected. Nearly all of the sulfur in gasoline is
contained in the naphtha (material boiling in the gasoline boiling range) produced in the fluidized
catalytic cracker (FCC). Generally, this is the only material which needs to be hydrotreated in
order to meet the 30 ppm average sulfur standard. In contrast, all diesel blendstocks, except for
that from a hydrocracker (4 percent of all blendstocks) will need to be hydrotreated in order to
meet the 15 ppm cap. Refiners produce roughly the same volumes of FCC naphtha and highway
diesel fuel. However, diesel fuel desulfurization requires much higher temperatures and
pressures, and the material must be in contact with the catalyst for longer periods of time, so the
capital investment per unit volume of treated material is much higher. Finally, because refineries
make 3-4 times as much gasoline as highway diesel fuel, the investment per gallon of finished
fuel is 3-4 times higher for diesel fuel.
This higher investment per unit volume of product means that refiners will be putting
more investment at risk relative to potential profit in the case of diesel fuel than gasoline. As
will be discussed further below, the market sometimes allows refiners to recoup their full cost of
meeting environmental standards (operating plus capital costs) and sometimes only allows them
to recoup operating costs. The greater level of investment per unit volume of product means that
refiners would have to cover 3-4 times the investment cost per gallon of fuel if the market does
not reward them with a price increase which allows the recovery of capital plus a reasonable
return on this investment. Directionally, this means that refiners will look much more closely at
the market situation for diesel fuel before making the investment to meet the 15 ppm standard.
In particular, refiners are likely to carefully assess their competitors' actions to ensure that
significant overcapacity does not exist, which decreases refining margins.
The second issue is that, of late, relatively poor refining margins have not allowed
refiners to recoup the full cost of environmental standards. Two examples are the 500 ppm
sulfur diesel fuel standard and the RFG standards. In both cases, over-investment by the refining
industry led to over-supply of these fuels and low prices.
b. Historic Refining Profit Margins
Over the past year, refining margins have improved dramatically. Domestic refineries are
operating at full practical capacity and are expected to do so for the foreseeable future. Thus, the
market may have begun a long term period where refining margins will be strong and reward
refiners who invest in additional capacity. Refiners also know that even slight shortages in
highway diesel fuel supply would lead to significant price increases and substantial profits for
those in the market. Thus, as always there will be a tension between wanting to invest and reap
the rewards of a potentially short market and the concern over over-investment and the inability
to recover investment. The large investment per unit volume of product will make this a more
IV-34
-------
Chapter IV: Fuel Standard Feasibility
difficult situation to balance than in past regulatory requirements. However, the temporary
compliance option will help counter this difficulty, as it will allow many refineries currently
producing highway diesel fuel to delay their investment until 2010. Thus, refiners who are in a
better financial position to take the financial risk involved in such a significant investment can do
so, while those which are not in as good a position can wait until 2010, buying credits in the
meantime.
c. Variation in Compliance Costs Faced by Refiners
The third issue related to supply is the range of costs faced by refiners in complying with
the diesel sulfur standard. Our refinery by refinery analysis indicates that refineries face a wide
range of compliance costs. If each refinery currently producing highway diesel fuel invests to
just maintain their current production, costs range from under 3 cents per gallon to under 12 cents
per gallon. It is probably unreasonable for a refiner to expect the market to allow a long term
increase in the price of diesel fuel of 12 cents per gallon.6 However, our refinery model also
indicates that some refineries can produce 15 ppm diesel fuel from their current nonroad diesel
fuel blendstocks more cheaply than many refineries which currently produce highway diesel fuel.
In assessing where the additional volume of nonroad diesel fuel blendstocks would come
from, we considered both refineries which produce both highway and nonroad diesel fuel today,
as well as refineries which only produce nonroad diesel fuel. Based on the volume and
characteristics of each refinery's nonroad diesel blendstocks, we projected which refineries could
increase production of highway diesel fuel most economically. We found that many refineries
could increase production of highway diesel fuel from blendstocks currently used to produce
nonroad diesel fuel as or more economically than many refineries currently producing highway
diesel fuel. Thus, there appears to be a plentitude of distillate blendstocks available from which
to produce highway diesel fuel.
In addition to requiring that these nonroad diesel fuel blendstocks be processed to
compensate for the loss of highway diesel fuel during production and distribution, we also
allowed highway diesel fuel produced form these nonroad diesel fuel blendstocks to supplant
highway diesel fuel produced at refineries facing higher desulfurization costs. The following
table presents the maximum cost in each PADD before and after this adjustment. It should be
noted that shifts of fuel production across PADDs were very limited. No transfers were allowed
into or out of PADDs 4 and 5. PADD 3 was allowed to increase fuel shipments to PADD 1 and
to the southern portion of PADD 2. No shifts were allowed between PADDs 1 and 2.
e A few small refiners are among those facing higher compliance costs. Many of these refiners are
expected to take advantage of EPA's option to delay their compliance with the Tier 2 gasoline sulfur standards.
This delay in their gasoline related investment and operating costs will allow these refiners to recoup their diesel
fuel-related investment with a much lower price increase than would otherwise be the case.
IV-35
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.A-7. Maximum Cost of Meeting the 15 ppm Cap (cents per gallon) A
PADD1
PADD2
PADD3
PADD4
PADD 5 B
All Current Highway Diesel Fuel Producers Continue to Produce
2006-2010
20 10 and beyond
4.8
9.6
4.6
9.7
4.2
11.9
5.5
9.1
4.3
8.9
With Minor Shifts in Production Between Highway and Nonroad Diesel Fuel
2006-2010
20 10 and beyond
4.8
5.5
4.6
7.4
4.1
5.1
5.5
8.2
4.3
5.1
A Excludes small refiners which we project would chose to produce 100% 15 ppm diesel fuel in 2006 and in return
would be granted a three-year extension in meeting the Tier 2 gasoline sulfur standards/
B Excludes Hawaii and Alaska, where maximum costs are 4.8-5.3 cents per gallon.
The difference between the maximum costs during the initial years are nearly identical
regardless of whether shifts between highway and nonroad diesel fuel production occur.
However, greater differences appear when the entire highway diesel fuel pool must meet the new
standard. This occurs because it appears that a very small fraction of current highway diesel fuel
production faces very high desulfurization costs, primarily because of extremely low production
volumes (i.e., poor economies of scale). By shifting only 1.4 percent of current highway diesel
fuel volume to nonroad diesel, the reductions in the maximum costs shown in bottom line of
Table V.C-3 occur. Thus, only a very small percentage of current highway diesel fuel production
volume faces costs well above the average. Likewise, it appears that ample highway diesel fuel
can be produced from nonroad diesel fuel blendstocks at reasonable costs. The costs to produce
highway diesel fuel from nonroad diesel fuel blendstocks assume the use of two-stage,
conventional hydrotreating. Costs for those refineries not meeting the new standard until 2010
could be much lower if novel, lower cost technologies, such as Phillip's SZorb process perform
as well as expected.
Also discussed in Chapter 5 of this RIA, the temporary compliance option allows a large
number of refineries, roughly up to 58, to delay production of 15 ppm diesel fuel until 2010.
f These refiners were excluded because the three-year delay in meeting the Tier 2 gasoline sulfur standards
provides these refiners with economic benefits which can be used to compensate for the cost of meeting the 15 ppm
diesel fuel sulfur cap. Thus, the actual cost of meeting the diesel sulfur cap is lower than indicated by our refinery
model, which only considers the cost of diesel fuel desulfurization.
IV-36
-------
Chapter IV: Fuel Standard Feasibility
(This presumes that the roughly 63 refineries investing in desulfurization equipment in 2010 and
producing 100 percent 15 ppm diesel fuel make their credits available to these other refineries.)
First, these 58 refineries will be able to observe the performance of the various technologies
selected by the other 63 refineries for almost 2 years before making final decisions regarding the
technology they will utilize. Second, they will be able to observe the reaction of the market to
the new fuel, particularly demand for use in older vehicles and to some degree, prices. However,
prices during the initial years will likely differ from those once the program is fully implemented.
This brings us to the third benefit of the temporary compliance option, small refiner hardship,
and GPA provisions.
As indicated in Table IV.A-7, the refineries producing 15 ppm diesel fuel in 2006 should
face lower costs than those delaying production until 2010. This difference in compliance cost is
primarily due to lower capital-related costs. Since the biggest risk facing a refiner is the
possibility that the market price increase after the implementation of the new standard will not
allow the recovery of both operating and capital costs, the lower the capital cost, the lower the
risk that substantial amounts will not be recovered. Also, the 75 or so refineries will only be able
to delay investment until 2010 if they buy credits from those producing more than 70 percent of
their highway diesel fuel under the 15 ppm cap. Thus, these 75 refineries will be subsidizing the
cost of producing the 15 ppm fuel through the purchase of credits. The net cost of producing
both 15 and 500 ppm fuels should be the same. This is illustrated by the following example.
Assume refinery A produces 70,000 bbl/day of highway diesel fuel, all meeting the 15
ppm cap. Also assume that the cost of meeting the new standard is 4 cents per gallon. Refinery
B produces 30,000 bbl/day of highway diesel fuel meeting the 500 ppm cap. Its costs do not
change from today. Refinery A generates 21,000 bbl/day of credits, while refinery B needs
21,000 bbl/day of credits. The two refiners will negotiate a price for the credits, which will be a
function of how many other sources of credits are available. However, if we assume that
Refinery A is willing to sell its credits at cost, then Refinery A will sell 21,000 bbl/day of credits
at 4 cents per gallon. Since 21,000 bbl/day represents 30 percent of its production, selling these
credits reduces Refinery A's average cost to 2.8 cents per gallon. Refinery B, on the other hand,
paid 4 cents per gallon for 70 percent of its production. Thus, its average cost is 2.8 cents per
gallon; the same as Refinery A's average cost.
This example demonstrates that with credit trading, the refining costs of both 15 and 500
ppm fuels should be roughly the same. This should lead to the two fuels having similar prices at
retail where both fuels are sold. In fact, given that 15 ppm will be the dominant fuel being
produced and needs to be distributed throughout the U.S., it will likely be transported by
pipeline. 500 ppm fuel, on the other hand, need not be distributed everywhere, since all vehicles
can burn the 15 ppm fuel. Thus, distribution of 500 ppm fuel may be concentrated around
refining areas and along major pipeline corridors. The price of 500 ppm fuel is likely to be
slightly lower than that of 15 ppm fuel in these areas to encourage older vehicle owners to buy
IV-37
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
500 ppm fuel and minimize the areas to which the 500 ppm fuel must be distributed. Thus, the
risk of a large price differential encouraging misfueling should be low.
The temporary compliance option also makes this diesel fuel program more similar to the
RFG program wherein not every refiner need participate, at least for the initial years. By the time
that refiners need to make their final decisions on whether to construct new equipment in time to
meet the new standard by 2010, the desulfurization units built for 2006 will have been operating
for at least one year. This will give refiners evaluating compliance for 2010 considerable
confidence in both the cost and performance of the technologies which are available. These
refiners will also be able to observe the response of the market to the new fuel in terms of price.
While we project that the price of 15 ppm fuel will be very similar to the price of 500 ppm fuel at
refinery gates, this is due to the credit trading system. We expect that the price of both fuels will
increase with the implementation of this rule. Knowledge of the cost and performance of the
desulfurization technology and this market response to the new fuel should be particularly
helpful to refiners needing to borrow money to fund the new equipment.
d. Other Markets for Highway Diesel Fuel
Current highway diesel fuel easily meets the specifications for nonroad diesel fuel or
heating oil. However, the market for these other distillate fuels is not large enough, nor growing
fast enough to absorb much highway diesel fuel. Plus, the highway diesel fuel market is
currently in balance, so any decrease in domestic supply would have to made up by imports.
In order to assess the potential for refiners to sell their current highway diesel fuel or
some of the blendstocks used to produce highway diesel fuel into alternative markets, EPA
contracted with SwRI and Muse, Stancil to assess these other markets. Muse, Stancil found that
refiners would have very limited possibilities of disposing of highway diesel fuel or its
blendstocks domestically. Only PADD 1 imports significant quantities of nonroad diesel fuel or
heating oil. Refineries in PADD 1 could produce more of this fuel and back out imports.
However, refineries in other PADDs would have to export any fuel which they back out of the
highway diesel fuel market. Based on historical prices (i.e., highway diesel fuel priced under the
500 ppm sulfur standard), Muse, Stancil estimates that refiners outside of PADD 1 would lose 3-
6 cents per gallon in revenue if they shift even 5 percent of their highway diesel fuel to the
nonroad diesel fuel market. These losses increase to 4-20 cents per gallon if they shift over 5
percent of their current highway diesel fuel to these alternative markets. Refiners in PADDs 2
and 4 would be particularly hard pressed, as they would have to ship their product to the US Gulf
Coast prior to exportation. This adds significant transportation costs, as there are no pipelines
flowing from PADDs 2 or 4 to the Gulf.
Should refiners shift highway diesel fuel production to these other markets, it will not
only affect the price of the shifted product. The price of all nonroad diesel fuel and heating oil
will drop. Refiners trying to sell their highway fuel into these other markets will try to sell it
IV-38
-------
Chapter IV: Fuel Standard Feasibility
locally prior to exportation. These refiners will compete with those currently producing nonroad
diesel fuel and heating oil, depressing prices in the entire market. Despite lower prices, fuel
demand will not increase substantially, since the use of nonroad equipment is a very weak
function of fuel price. (For example, fuel costs are a small portion of the total cost of farming,
mining and construction, so one would not expect that the demand in these sectors of the
economy would be very responsive to fuel price.) Thus, refiners planning on shifting their
highway fuel to alternative markets will not only have to consider the decrease in market value of
the shifted product, but also the drop in value of their existing nonroad fuel and heating oil
production. This added cost of a drop in highway diesel fuel production would vary widely from
refinery to refinery since some refineries produce much more nonroad diesel fuel than highway
fuel and vice versa.
This loss in market price serves as a discouragement to shift highway diesel fuel to these
other markets. It basically provides refiners with a second reward for investing in desulfurization
equipment in order to stay in the highway diesel fuel market. First, investment allows them to
obtain the price rise of highway diesel fuel which should accompany the new sulfur standard to
be achieved. Second, investment allows the price drop associated with export to be avoided, as
well as reduces the potential for a drop in value in existing nonroad diesel fuel production. (This
last factor is a function of other refiners' decisions, as well, in this area.) Thus, a refiner should
desire to invest in meeting the new standard if he believes that the price increase in highway
diesel fuel will be at least the cost of meeting the standard minus the loss associated with export.
For example, if it costs up to 7 cents per gallon to meet the 15 ppm standard, then the required
price increase in highway diesel fuel price may only need to be 3 cents per gallon for refineries to
prefer meeting the 15 ppm standard versus taking a loss in the nonroad market of 5 cents per
gallon (ignoring any price drop for existing nonroad diesel fuel production). The lack of a ready
domestic alternative market for their product appears to be a strong discouragement to refiners
shifting their production away from highway diesel fuel.
e. Uncertainty in Requisite Desulfurization Technology
The next factor which could affect highway diesel fuel supply is uncertainty in what
technology will be required to meet the 15 ppm standard. As discussed in Section 8.1.2 below
and in the RIA, uncertainty does exist concerning the requisite desulfurization technology. Most
vendors project that two-stage conventional hydrotreating at low to moderate hydrogen pressure
will be sufficient to achieve the new standard, even with significant quantities of LCO. Most
refiners commenting on the rule, plus 1-2 vendors believe that moderate to high pressures will be
needed, accompanied by more aromatic saturation and hydrogen consumption. In addition,
Phillips Petroleum just announced that they have developed a new, low pressure process which
promises to consume no hydrogen. This process cannot yet be licensed, but Phillips hopes to
begin licensing next year. However, a commercial unit utilizing this technology will not start up
until 2004.
IV-39
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The uncertainty in current technology which will be necessary to achieve the new
standard could encourage some refiners to delay investment until the latest possible time while
still allowing time to build their equipment in time for the 2006 implementation date. The
promise of lower costs based on refiner's experience meeting the new standard in 2006 or with
the new Phillips technology could encourage refiners to delay the construction of new equipment
until beyond the 2006 implementation date. In some cases, particularly refiners located in
isolated areas where hydrogen costs are high, the promise of lower long term compliance costs in
one to two years could be preferable to the lower revenues obtained from selling highway diesel
fuel into the nonroad diesel fuel market in the short term.
Countering the benefit of more leadtime with respect to conventional hydrotreating
technology is the fact that vendors will have 2-3 years to generate both pilot plant and
commercial data to convince refiners of the efficacy of their processes. While no refiners are
currently required to meet a 15 ppm cap prior to 2006, numerous two-stage (and low space
velocity one-stage) hydrotreating units exist world-wide. Vendors and refiners are likely to
utilize these units to demonstrate their catalysts commercially. This is already being done with
some units in Europe. Thus, the largest detriment to investing to meet the new standard in 2006
is the potential cost savings associated with novel technologies such as Phillips' SZorb. These
technologies are sufficiently different from conventional hydrotreating that refiners are likely to
require full-size commercial operation for a year or two prior to betting tens of millions of dollars
on their effectiveness. With the temporary compliance option, refiners able to delay investment
until 2010 should be able to utilize the newer technologies, such as SZorb.
f. Likely Price and Import Response to the New Standard
Moving onto the likely price increase which will accompany the new standard, it is very
difficult to predict whether the future market price of highway diesel fuel will increase enough to
cover only operating costs or operating plus capital costs. No one can predict future prices, so
the real issue is what refiners project the price increase will be at the time they need to invest in
order to meet the new standard. As mentioned above, the 1990's were not good to refiners;
refining margins were poor. Refiners generally did not recover their capital investments which
were associated with environmental programs.
However, the demand for fuel continues to grow and domestic refinery capacity is
growing at only about half the rate of growth in demand. Imports of finished fuel, including
highway diesel fuel are increasing. Also, refining margins during the past year have been
excellent for most refiners. Integrated oil company profits have also been at record levels. The
net income of individual major oil companies over the 2nd and 3rd quarters of 2000 (e.g.,
ExxonMobil) was sufficient to fund all of the capital investment associated with this rule. If
these refining margins continue for any appreciable amount of time, the availability of capital
should not be an issue, even considering other environmental programs facing refiners. These
include the Tier 2 gasoline sulfur requirements and NESHAP standards for FCC units, reformers
IV-40
-------
Chapter IV: Fuel Standard Feasibility
and sulfur plants. We analyze the combined capital investments associated with the gasoline and
diesel fuel sulfur programs later in this chapter of the RIA. There, we found that the level of
capital investment per year will be lower than that occurring in the early 1990's, when most of the
programs associated with the Clean Air Act of 1990 were being implemented and when refining
margins were low. Thus, from an industry wide perspective, the availability of capital should not
be a problem. The temporary compliance option helps this situation substantially. A few
individual refiners could have difficulty raising sufficient capital to meet the new diesel sulfur
standard. We have included hardship provisions in this rule to accommodate at least some of
these situations. In addition, as discussed above, not every refiner currently producing highway
diesel fuel will need to continue to do so in order to meet future demand.
Overseas refiners may not be as able to produce diesel fuel under the new 15 ppm cap, as
they have been under the current 500 ppm cap. The three largest exporters of diesel fuel to the
U.S. are Canada, the Virgin Islands and Venezuela. The Canadian refineries which export to the
U.S. are located in the far eastern portion of Canada and send the vast majority of their
production to the U.S. The same is true of the largest Virgin Island refinery, which has U.S.
ownership. These refineries look to the U.S. as their main market. Thus, they are as likely to
invest to meet the new standard as any domestic refinery. Venezuelan refineries are in a
somewhat better position to send their diesel fuel elsewhere and could be less likely than
domestic refiners to invest in new desulfurization equipment. At the same time, Europe and
Japan are implementing 50 ppm diesel sulfur caps and Europe is already considering a 10 ppm
cap. Thus, export oriented refineries world-wide will have to invest to at least meet a 50 ppm
cap and will likely prepare for even lower standards. Even a refinery designed to produce 50
ppm sulfur diesel fuel is capable of producing some 15 ppm fuel. This may require reducing
volumetric throughput or cutting the endpoint of its most difficult to hydrotreat blendstocks.
However, such refineries should be able to send diesel fuel to the U.S. even if they do not design
to do so on a regular basis. Several overseas refiners are likely to closely observe the investment
patterns of U.S. refineries to assess the economics of exporting their diesel fuel under the new
standard.. Thus, overall, exporting fuel to the U.S. will be more difficult under the new standard,
but supplies should be available if necessary. Again, the temporary compliance option helps this
situation by allowing importers to import three gallons of 500 ppm fuel for every seven gallons
of 15 ppm brought into the country.
g. Impact of Desulfurization Processes on Fuel Volume
Conventional desulfurization processes both reduce the physical and energy density of
diesel fuel. Desulfurization actually increases the volume of diesel fuel produced, but each
gallon of diesel fuel contains less energy. Overall, the total amount of energy leaving the
hydrotreater in the form of diesel fuel decreases by roughly 1.5 percent. Vehicular fuel economy
is directly proportional to fuel energy density. Thus, in order in to provide the same number of
vehicle miles, refineries will need to increase the volume of blendstocks which they process by
IV-41
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1.5 percent. As discussed in Chapter 5, our cost projections consider this loss of diesel fuel
volume in assessing the hydrotreating capacity needed by refiners.
In terms of supply, the effect is much less. Most of the energy lost to diesel fuel is in the
form of naphtha or LPG. This increases the refinery's production of these products. This allows
the refinery to make other adjustments which increase diesel fuel production at the expense of
gasoline, bringing the net production of both products back into balance.
h. Impact of Fuel Transport on Supply
The final factor potentially affecting supply is the possibility that 15 ppm diesel fuel
produced at a refinery will be contaminated during shipment and becoming unsuitable for use in
2007 and later highway vehicles. As discussed in Chapter 5 of this RIA, we expect refiners to
produce diesel fuel with an average of 7 ppm under the new standard. However, some batches
are likely to be higher than this. Pipelines are likely to set their own limits below 15 ppm (e.g.,
10 ppm). This means that diesel fuel can only pick up 5 ppm sulfur during distribution, given the
testing allowance provided in the final rule. As also discussed in Chapter 5, we estimate that
current loss of highway diesel fuel to nonroad diesel market of 2.2 percent will double to 4.4
percent. This increases the production requirements for 15 ppm diesel fuel, but not for total
diesel fuel, since the volume lost during distribution can be used as nonroad diesel fuel or heating
oil.
As was done for the volume lost during hydrotreating, we considered that refineries
would have to process 2.2 percent more diesel fuel blendstocks to produce enough highway
diesel fuel to account for losses in the distribution system. This additional volume of
blendstocks came from blendstocks currently being used to produce nonroad diesel fuel and
heating oil.
i. Charles River Associates and Baker and O'Brien Study
The study by Charles River Associates (CRA) and Baker and O'Brien, which was
commissioned by API, assessed refiners ability to maintain an adequate supply of highway diesel
fuel under the 15 ppm cap. As part of this study, CRA polled refiners concerning their plans
under a 15 ppm sulfur cap. Using the results of this survey, as well as other information, CRA
projected refiners' costs of meeting the 15 ppm standard, as well as their likely production
volumes. CRA concluded that U.S. refiners would likely reduce their highway diesel fuel
production by an average of 12 percent, creating significant shortages and price spikes.
CRA's conclusions appear to have been strongly affected by their assumptions, as well as
the refiner survey they conducted. For example, CRA assumed that the new sulfur standard
would cause 10 percent more highway diesel fuel to be "lost" in the distribution system
compared to today (i.e., downgraded to off-highway diesel fuel). We believe based on the
IV-42
-------
Chapter IV: Fuel Standard Feasibility
analysis outlined in the RIA that 2.2 percent is a more accurate estimate, resulting in 9 percent
more 15 ppm fuel being available than CRA estimated. This difference alone accounts for 75
percent of the potential national supply shortfall projected by CRA.
CRA also concluded, with little explanation, that 20 refineries producing highway diesel
fuel today would not produce highway diesel fuel under the 15 ppm standard and that many more
would reduce production. Given the lack of information provided in the study, it was not
possible to evaluate CRA's criteria in selecting these 20 refineries, nor was it possible to
determine how much of the shortfall was attributable to this conclusion. While CRA evaluated
whether refiners currently producing highway diesel fuel would be likely to leave the market,
they did not assess whether any refineries currently not producing highway diesel fuel might
enter the market. EPA did conduct such an assessment. We found 2 refineries that produce
essentially no highway diesel fuel today which could meet the new standard for less than 5 cents
per gallon. Production from these refineries would increase highway diesel fuel production by 9
percent. We also found based on our assessment that 4 other refineries could produce highway
diesel fuel from their off-highway diesel fuel blendstocks for less than 5 cents per gallon.
Production from these 6 refineries would increase highway diesel fuel production by 7 percent.
Together with a more reasonable estimate of downgrades in the distribution system, this would
more than compensate for any potential lost production, even as estimated by CRA.
CRA also implicitly assumed that the material it projected could be removed from the
highway diesel market could be sold at a reasonable price. However, CRA did not analyze the
impact of this additional supply on the prices which could be obtained in these markets, or even
if these alternative markets could physically absorb all of this material. Much of this material is
not diesel fuel, but poor quality blendstock. It is not clear that such material could be blended
into non-highway diesel fuel and CRA did not analyze this likely problem. Our analyses,
supported by a study by Muse, Stancil and Co., indicate that any substantial quantities of
highway diesel fuel diverted to other markets will depress prices in those markets substantially.8
Hydrotreating diesel fuel to meet the 15 ppm standard avoids these depressed prices, reducing the
net cost of meeting the new standard. Since CRA only considered the cost to desulfurize
highway diesel fuel, and ignored the added cost of dumping this fuel into markets with depressed
prices, CRA's conclusions must be considered to be seriously flawed in this regard.
Furthermore, CRA ignored the fact that roughly 15 percent of today's highway diesel fuel
is consumed in engines and furnaces not requiring this fuel. Any shortage of highway diesel fuel
would lead many of these non-essential users to switch to nonroad diesel fuel or heating oil.
Only limitations in the fuel distribution system would cause these users to continue to burn
highway diesel fuel.
B "Alternate Markets for Highway Diesel Fuel Components," Muse, Stancil & Co., for Southwest
Research Institute, for U.S. EPA, September, 2000.
IV-43
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
These problems with CRA's analysis, plus the lack of detail available concerning the
specifics of the study, lead us to reject the study's conclusions that there will be significant
supply shortfalls under a 15 ppm sulfur standard.
Finally, if any potential for highway diesel fuel shortfalls exists by requiring all fuel to
meet 15 ppm sulfur in 2006, as CRA's analysis suggests, we believe that allowing some
continued supply of 500 ppm, as under the temporary compliance option and hardship provisions
contained in today's action, addresses this concern. By allowing some transition period before
the entire highway diesel pool is required to meet the 15 ppm sulfur standard, some refiners will
not need to change their current operations and will be able to continue producing 500 ppm fuel
during these years. Those refiners that delay production of low sulfur diesel fuel until the later
years of the program will tend to be the refiners with the highest cost to comply and, thus, the
greatest tendency not to invest and impact supply. Refiners that begin producing low sulfur
diesel fuel in the later years of the program will be able to take advantage of ongoing
improvements in desulfurization technology that will help avoid or reduce any potential losses in
highway diesel fuel production when the program requires full compliance with low sulfur diesel
fuel.
8. Conclusions
In order to meet the proposed 15 ppm sulfur cap, refiners are likely to further hydrotreat
their highway diesel fuel in much the same way as it is being done today to meet the 500 ppm
sulfur cap. Improvements to current hydrotreaters can be used to reduce diesel fuel sulfur beyond
that being done to meet the 500 ppm cap. However, these improvements alone do not appear to
be sufficient to provide compliance with the proposed 15 ppm cap. Based on past commercial
experience, it is very possible to incorporate current distillate hydrotreaters into designs which
provide compliance with the proposed 15 ppm cap. Thus, the equipment added to meet the 500
ppm standard in the early 1990's will continue to be very useful in meeting a more stringent
standard.
The primary changes to refiners' current distillate hydrotreating systems would be:
1) the use of a second reactor to increase residence time, possibly incorporating
counter-current flow characteristics, or the addition of a completely new second
stage hydrotreater,
2) the use of more active catalysts, including those specially designed to desulfurize
sterically hindered sulfur containing material,
3) greater hydrogen purity and less hydrogen sulfide in the recycle gas, and
4) possible use of higher pressure in the reactor.
IV-44
-------
Chapter IV: Fuel Standard Feasibility
Existing commercial hydrotreaters are already producing distillate with average sulfur
levels below 10 ppm, which should be more than sufficient to meet a 15 ppm cap. These
hydrotreaters are processing distillate with typical breakdowns of SRLGO, LCO and LCGO.
Therefore, the proposed 15 ppm cap appears to be quite feasible given today's distillate
processing technology. The only drawback of these commercial demonstrations is that they were
designed to reduce aromatics content, or improve cetane, as well as reduce sulfur. Therefore,
these units' hydrogen consumption and its associated cost are higher than that needed for simple
sulfur removal. This combination of sulfur and aromatics reduction has been encouraged by fuel
tax incentives in Europe. The incentive to reduce sulfur by itself to such low levels has not
existed, so refiners have generally had no incentive to produce such a product commercially.
Advances continue to be made in catalyst technology, with greater amounts of sulfur
being able to be removed at the same reactor size, temperature and pressure. Therefore, it is
reasonable to expect that distillate hydrotreaters put into service in the 2006 timeframe will
utilize even more active catalysts than those available today.
Other methods of reducing diesel fuel sulfur, such as FCC feed hydrotreating, removing
the heavy end of LCO, etc. help to reduce diesel fuel sulfur levels, but will generally not be
sufficient to provide compliance with a 15 ppm cap. However, we expect that a number of
refiners will utilize these techniques to reduce the severity of their distillate hydrotreaters and
reduce hydrogen consumption (particularly by avoiding aromatic saturation). Some of these
techniques would tend to increase the supply of highway diesel fuel (e.g., FCC feed
hydrotreating), while others would tend to decrease it (e.g., removing the heavy end of LCO).
Biodesulfurization technology holds promise to reduce distillate sulfur without the high
temperatures and pressures involved in hydrotreating. Efforts are underway to demonstrate that
this technology can achieve 50 ppm sulfur or less in the next few years. However, it is not clear
whether this technology would be sufficient to meet a 15 ppm cap.
9. Fuel Availability in 2006
a. Summary
We analyzed the refining and finished products distribution industries to determine the
minimum volume of 15 ppm diesel fuel that will assure it is widely available in all parts of the
country by September 1, 2006 and still provide for the supply of a modest amount of 500 ppm
fuel to mitigate concerns about supply shortfall. Small refiners, which contribute about 5 percent
to the national highway diesel fuel supply, have been given the opportunity to defer production of
15 ppm fuel for four years. We investigated how much production of 15 ppm fuel by the
remaining refiners would still assure adequate availability across the country. We determined
IV-45
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
that at least 80 percent of the highway production within each PADD by non-small refiners must
be converted to 15 ppm diesel to provide those assurances.
We feel it is important to understand, to the degree possible, how well balanced and operated
the refining, supply, and distribution industries in this country are. Everyday 110 million gallons
(2.6 million barrels) of diesel fuel (roughly 15 percent of total finished product volume) will be
produced by refineries and distributed via pipeline, truck, and other means to end-users by the
year 2006. These industries have developed and maintained a very efficient and safe, albeit
complex, system for converting crude oil into finished products and making them available in the
market at reasonable prices, especially compared with much of the rest of the world.
b. Diesel Fuel Refining Under the Temporary Compliance Option
There are currently 124 refineries in the country that produce highway diesel.
Historically, the Pacific and Gulf coasts have had the highest concentration of both large and
small refineries. For example, the refineries in Texas, Louisiana, and Mississippi, most of which
are located near the Gulf, produce roughly 43 percent of the highway diesel in the country. The
refineries in California, mostly located near the Pacific, produce about 12.5 percent of all
highway diesel. On the other hand, all the refineries located in PADD I (the Atlantic coast
region) produce about 6 percent of total highway diesel, and only 18 percent of expected PADD I
highway demand. The others are "scattered" across the country, although some are clustered in
certain regions or states, such as the Rocky Mountains, Kansas, Oklahoma, and Illinois.
There were two key considerations in our analysis of the refining industry. First, we
projected which refineries would make the investment to convert to 15 ppm diesel and which
would continue to produce 500 ppm fuel. Second, we evaluated where in each PADD each of
the sets of refineries are located with respect to each other, to pipelines, terminals and other
major fuel consuming markets.
We used the refinery cost model described in chapter V of the Regulatory Impact
Analysis to predict which refineries would most likely make the investment to produce 15 ppm
fuel. Early analysis showed a concentration of low cost refineries on the Gulf coast.
Consequently, restrictions on averaging and trading were necessary to prevent 15 ppm fuel from
being produced in limited areas of the country if wide spread availability was to be achieved. We
considered various regional restrictions, but concluded that the PADD regions provide a good
differentiation of the main fuel production and distribution regions of the country. Subsequent
analyses were conducted assuming averaging and trading would only take place among refineries
within each of the five PADD's. While the ABT program is PADD restricted, the transfer of fuel
between PADD's is important and refinery location with respect to other PADD's was also
considered. The small refiner hardship and GPA provisions of the rule were assumed in the base
case in these analyses. With one exception, refiners were assumed to produce either 100 percent
15 ppm fuel or 100 percent 500 ppm fuel, based on our analysis of cost and on discussions with
IV-46
-------
Chapter IV: Fuel Standard Feasibility
various refiners. The one exception was those refineries with hydrocrackers that can produce 15
ppm diesel very cheaply from their hydrocrackate but otherwise produced 500 ppm diesel. We
assumed that the "least-expensive-to-convert" refineries in each PADD would make the
investment to produce 100 percent of 15 ppm fuel and that the remaining refineries would
purchase available credits from them to continue producing 500 ppm diesel fuel. We want to
stress that each refinery was studied as thoroughly and equitably as the available data allowed.
With the exception of small changes from PADD to PADD due to varying numbers of small
refiners and their volume, the PADD specific ABT restrictions result in essentially the same
volume of 15 ppm produced within each PADD. Depending on the level of 500 ppm fuel
allowed to be produced, however within each PADD production of 15 ppm fuel may be limited
to certain areas. At an 80 percent level for 15 ppm fuel under the temporary compliance option,
production of 15 ppm fuel is projected to occur on a widespread basis across all the PADD's.
Table IV.A-8. Number of Refineries Producing 15 ppm Diesel by PADD
PADD I
PADDH
PADDHI
PADD IV
PADDV
3
13
24
8
11
/'. Pipelines, Terminals, and Bulk Plants
As important as it is for fuel availability for refineries to produce adequate volumes of 15
ppm fuel in all major regions of the country, it is equally important that pipelines handle that
fuel. At present, large volumes of both highway (< 500 ppm total sulfur) and off-highway diesel
(-3,000 ppm total sulfur) are transported long distances via pipeline to delivery or " break-out"
terminals for distribution by bulk and tank truck. There are approximately 127 pipeline
companies currently operating pipelines in the country. Trucking fuel over long distances is
prohibitively expensive and logistically, nearly impossible. A case in point, is the transfer of fuel
into PADD I from PADD ILL The Plantation pipeline, which runs from Louisiana to Indiana can
deliver 476,000 barrels per day. On the other hand, it would require 2,400 trucks, each carrying
200 barrels (8,400 gallons) to deliver that same volume, which includes running the trucks just
one-way. The distances involved would make the cost for distribution by truck prohibitively
high.
IV-47
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The experience and knowledge-base of the pipeline companies and shippers make it
possible to ship batches of millions of gallons of different products, such as gasoline, jet fuel,
kerosene, diesel, home heating oil down the same line, switching lines from time-to-time, cost-
effectively, and with a minimum of problems. For example, over the course of a year, the
Colonial Pipeline handles 38 different grades of gasoline, including reformulated gasoline and
multiple vapor pressures for each grade, seven grades of kerosene (including two for military), 16
grades of home heating oil and diesel fuel (including marine diesel fuel for the U.S. Navy and
light cycle oil) and one grade of transmix.. The Plantation Pipeline has a similar slate of
finished products. While these pipelines carry multiple grades of fuel, their ability to add another
fuel, especially in large volumes, is limited by tankage along the way. In that the northeast part
of the country uses most of the home heating oil in the country, most other pipelines don't
usually carry it on their slate. The TEPPCO Pipeline slate includes gasolines, distillates,
commercial jet fuel, military jet fuel, unfinished gasoline, and speciality blendstocks. The
Explorer Pipeline system transports primarily gasoline, fuel oil, and jet fuel. The Olympic
Pipeline carries gasoline, diesel, and jet fuel. Generally, from the mid-west through the west,
homes and businesses are heated with natural gas which precludes the need to ship heating oil.
Pipelines vary in lengths of from as little as a few thousand feet (the shorter sections are
usually referred to as stub lines) to thousands of miles, including connections. For instance the
Plantation system is 3,100 miles long; the Colonial system is about 4,300 miles long; the
Chevron Pipe Line Company's network of pipelines is about 5,000 miles long, which includes
their crude, chemical, and LPG capacity. The Olympic line is 400-miles, the Explorer 1,400-
miles, and the Kaneb 2,075-miles long.
Pipelines vary in diameter from 6 to 48-inches. For example the Plantation pipeline
varies from 6 to 30-inches; the Colonial pipeline has segments that vary from 8-inches to 48-
inches. The Explorer mainline pipe size is 28-inches from Port Arthur to Tulsa and 24-inches
from Tulsa to Hammond, Indiana. Most pipelines ship liquids at velocities of from 4 to 7-miles
per hour, or at an average of between 5 and 6-miles per hour. Capacities can be estimated using
the diameter of the pipe. Average velocities are probably most useful, because sections of many
pipelines vary in diameter. Velocities are also affected by the viscosities of the various products
in the line. For example, the Plantation pipeline delivers about 476,000 barrels per day through
their system. The Olympic ships about 306,000 barrels per day. Colonial pipeline's main line
batch sizes vary from 75,000 to 3.2 million barrels. The smallest main line batch is 75,000
barrels. Explorer's 28-inch section has a capacity of over 500,000 barrels per day; it's 24-inch
section can handle about 317,000 barrels per day.
Splitting the current single grade of highway diesel fuel into multiple grades raises
concerns about the size of the batches that move through the pipelines. All pipelines, regardless
of capacity, require minimum batch volumes to avoid the problems inherent with shipping small
batches. For instance, the Colonial Pipeline has a minimum batch size of 75,000 barrels, while
normal volumes range from 350,000 barrels or more per batch. The Plantation pipeline has a
IV-48
-------
Chapter IV: Fuel Standard Feasibility
minimum batch size of 25,000 barrels. The Chevron pipeline from Salt Lake City, Utah to
Boise, Idaho has a minimum batch size of 6,000 barrels.
One problem with small batches is the speed with which they pass through a section of
pipeline. In a 48-inch pipeline, 75,000 barrels would flow past a point in the line in just one and
one-quarter hours. To break-out a shipment under these conditions requires extra-ordinary care
and the possibility of contamination is high. Usually, the interface volume must be raised, which
results in an increase in the loss of the higher quality, more valuable fuel, either to reprocessing
or downgrading.
Most pipelines are common carriers and as such have fungibility requirements, usually
referred to as pipeline specifications. That is, a shipper can ship a given volume of product, but
in order to maximize the batch volume, a terminal will mix or co-mingle it with fuel from other
shippers, such as from other refiners, that meet the same fungibility specification. At the
destination terminal, the batch volume is "broken-out" into tanks from which bulk and tank
trucks make deliveries. While this fuel will have characteristics similar to the fuel the shipper
sent, it will not be "exactly the same fuel", since batches of similar fuels were mixed with it at
the origin of the shipment.
Another potential problem with small batches of 15ppm fuel is the need to "wrap" the
fuel more carefully and differently than other fuels. "Wrapping" refers to the choice of products
that precede and follow a particular batch of fuel in a pipeline. On the Explorer pipeline, a
typical sequence would be fuel oils, diesel fuels, jet fuels, and gasolines. A typical sequence on
the Colonial pipeline would be reformulated gasoline, low sulfur diesel, kerosene/jet fuel, high
sulfur diesel, conventional regular gasoline, all premium grades, and reformulated regular
gasoline. Each fuel is essentially "wrapped" by the fuel that precedes it and follows it. An
efficient way to "wrap" is to ship two products next to each other, one of which can be
"downgraded" to the other to avoid losing the interface to slop and reprocessing. For instance,
when Jet A is shipped either ahead of or following 500 ppm diesel. A small volume of the Jet A
can usually be downgraded to 500 ppm diesel, since it fits well within the highway diesel specs.
It's total sulfur content can be significantly higher, but the total volume of interface is usually
small relative to the much larger diesel volumes and the small amount of extra sulfur can be lost
to dilution. However, neither 500 ppm diesel nor 15 ppm diesel can be downgraded to Jet A
because the high endpoint would drive the Jet A out-of-spec. Jet A, on the other hand, cannot be
downgraded to 15 ppm diesel because it's sulfur content can range as high as 3000 ppm. As a
result, it is important to maximize the batch volumes of 15 ppm fuel, to the extent possible. The
volume of interface relative to the shipment volume makes the cost of shipping small batches of
15 ppm fuel prohibitively high. See Chapter IV, D.2.a of the RIA for a complete discussion of
pipeline interfaces.
We also considered whether we could expect the pipelines to handle 15 ppm diesel as a
"proprietary" or "specialty" fuel and thus perhaps ship small batches and still make the fuel
IV-49
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
widely and dependably available. Most pipelines will handle proprietary shipments (sometimes
referred to as "tenders") of certain products, but even then, the product usually has a specification
range that fits well with the specs of other products shipped on the line and can be efficiently
"wrapped"; interface losses are usually larger with tenders and extra tankage fees can add as
much as one-half to three-quarters of a cent per gallon. Another problem is that tenders must
usually fit into an established pipeline schedule. Consequently, it is difficult to rely on regular
deliveries of these products. While this is not impossible, as a practical matter it is difficult
because the pipelines are usually kept full on schedules up to a month or more in advance. We
expect this approach to be prohibitively expensive as well as making it difficult to havelS ppm
fuel widely available at all times.
The pipelines which currently ship only one grade of diesel fuel are expected to continue
doing so, until 15 ppm fuel becomes the predominant fuel. We expect them to switch to 15 ppm
diesel as their only grade of fuel. A few companies, for instance the Williams, the Cenex, the
Chevron, and the Pioneer pipelines in the Rocky Mountain area prefer to carry only one grade of
diesel, usually highway, and simply add the required dye when the fuel is loaded on to trucks at
the terminal and sell it as off-highway. Pipelines that carry highway and high-sulfur off-highway
diesels are also expected to switch to 15 ppm diesel rather than carry three fuels. However, we
also recognize that a few of the larger pipelines, such as the Colonial and Plantation, may choose
to ship three fuels. It is important to understand that not all terminals are owned by the pipelines
that deliver to or from them. Consequently, even though a pipeline may be willing to carry 15
ppm diesel in addition to 500 ppm diesel, there must be terminals in the appropriate locations
along the pipeline that are capable of taking delivery of the product.
As such, an important element of this analysis is to determine at what volumes the
pipelines and terminals will likely ship/handle sufficient volumes of 15 ppm diesel to make it
widely available, either as the only diesel fuel or in conjunction with other higher-sulfur, diesel
grade fuels. We anticipate that under the 80 percent temporary compliance option program that
the vast majority of the pipelines will just carry 15 ppm fuel. Some of the larger pipelines may
choose to carry both fuels, but will limit either the distance the 500 ppm fuel is carried or the
number of breakout locations. For this reason, in our analysis we assumed that 500 ppm fuel will
be sold in just 50 percent of the country.
Moving to terminals, an important distinction exists between the difficulties terminals
face and those that pipelines face when deciding to carry 15 ppm diesel if the volume of the 15
ppm fuel is much less or even nearly equal to the volume of 500 ppm fuel. As discussed above,
small batches, including minimum volume batches of 15 ppm fuel, are difficult to handle but in
fact most the difficulties take place in and around the terminals. It is in fact, at the terminals
where batches are sequenced into a line and broken out at delivery. Interfaces are also managed
at the terminal. Once the batch of 15 ppm fuel is in the line, it travels much the same way other
products do. Volume on the line does not necessarily change, since the 15 ppm fuel is displacing
a matching volume of 500 ppm fuel. However, at the terminal, the 15 ppm fuel must be broken
IV-50
-------
Chapter IV: Fuel Standard Feasibility
out into tankage separate from the 500 ppm, which means that unless the terminal has an extra
tank somewhere of sufficient volume to handle the product, a new one must be constructed.
Switching back and forth between the fuels is definitely problematic, since very small volumes of
500 ppm fuel could easily drive the 15 ppm fuel out of spec.
Today, thousands of pipeline terminals make diesel fuel accessible to thousands of bulk
and tanker trucks that can easily and economically deliver smaller volumes to bulk plants or
service stations, truck stops, fleets, and other end-users over distances of up to 150 to 200 miles.
For instance, the Plantation pipeline is connected to 130 shipper terminals in eight states. These
terminals are owned by petroleum refiners, marketers, military, and commercial fuel users.
Products are "tendered" to the system from nine refineries in Mississippi and Louisiana, from
other products pipeline systems, and via marine facilities on the Mississippi River. TEPPCO has
21 product delivery terminals and 31 storage facilities in 12 states. The Explorer has major
tankage and terminals at Port Arthur, Greenville and Grapevine, Texas; Glenpool, Oklahoma;
Wood River, Illinois, and Hammond, Indiana, and serves 70 major populations centers in 16
states. The Kaneb services parts of Wyoming, Colorado, North and South Dakota, Nebraska,
Iowa, and Kansas. The TEPPCO system includes 21 product delivery terminals and 31 storage
facilities in 12 states. The Olympic has 10 delivery stations between Anacortes, Washington and
Portland, Oregon. These represent just a few examples of the roughly 1,400 storage facilities and
terminals in the U.S.
At the production volumes and for the logistical reasons discussed above, the terminals
which currently handle only one grade of diesel fuel are expected to switch to 15 ppm diesel
rather than invest in the tankage and ancillary equipment necessary to carry two fuels.
Discussion with and comments with industry suggest that very few, if any, terminals have unused
tankage available to carry an extra fuel. A few of these companies actually supply some off-
highway diesel but because they prefer to carry only one grade, usually highway, they simply add
the required dye when the fuel is loaded on to trucks and sell it as off-highway. For example,
some of the refineries in PADD IV supply off-highway diesel to the mining industry.
Historically, most of the off-highway diesel in PADD IV has been relatively low sulfur and the
refiners and shippers have simply sold dyed highway diesel into this market. We estimate that
about 20 percent of the off-highway diesel in these areas is actually high sulfur fuel. Terminals
that handle highway and high-sulfur off-highway diesels are also expected to switch their
highway fuel to 15 ppm diesel rather than carry three fuels. However, we also recognize that the
terminals on a limited number of the larger pipelines, such as the Colonial, Plantation, and
perhaps the Explorer could choose to handle three fuels, and may need to build additional
tankage. It is through these terminals that the remaining 500 ppm fuel would be distributed.
ii. Bulk Plants
In addition to terminals, there are roughly 10,000 bulk plant across the country which
receive diesel fuel, usually by truck and then redistribute it in smaller quantities to retail outlets.
IV-51
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Many of the bulk plants are owned and operated by the owners of truck fleets and service stations
and are major source of supply of diesel fuel, particularly in the rural areas of the country. Bulk
plants typically have just one, roughly 20,000 gallon tank per product handled. As such, the
introduction of another 20,000 gallon grade of diesel fuel would require the to either add tankage
to carry both or specialize in supplying one or the other. At an 80 percent level for the temporary
compliance option, with the 15 ppm fuel being the only fuel available except near refineries
producing 500 ppm fuel or near those terminals that invest to distribute both fuels, we anticipate
that most bulk plants will not add tankage and will merely switch over to 15 ppm fuel. Where
available, 500 ppm fuel would likely be trucked directly from the terminal to the retail outlet or
centrally fueled fleet without going through the bulk plant.
Hi. Fleets & Card-locks
We expect fleet owners and card-lock companies will make the most economically
reasonable choices available. They will likely purchase 500 ppm fuel, if it is available and even
if it is only a penny cheaper, until they or their customers purchase a vehicle which requires 15
ppm fuel. At that time, we expect they will switch to 15 ppm fuel rather than build an extra tank,
since all of their diesel powered vehicles can use the 15 ppm fuel. There was some discussion as
to whether a fleet owner could arrange for a vehicle with the new emissions device to fill at a
facility, i.e., another fleet owner, carrying 15 ppm fuel rather than convert his fleet to!5 ppm
diesel. The proposition sounds simple, but in fact there are several costs involved. It is possible
that the nearest sister facility is several miles distant, at times in heavy traffic, and perhaps in the
direction opposite the one to be taken by the vehicle looking for the 15 ppm fuel. Distance and
time are both important factors. For example, some drivers are paid by the load and would likely
demand extra pay for time spent fueling a truck. Driving loaded trucks, at 4 to 5-miles per gallon
(perhaps even lower in slow or heavy stop-and-go traffic) can quickly add several cents to each
gallon of fuel purchased from a distant facility. We also considered whether a fleet owner could
fuel at a service station. Most service stations are designed for light vehicles only and are often
located in high traffic areas, such as at intersections. Unless the station owner installs special
accommodations for large trucks, fueling would be nearly impossible. Many stations also do not
have around-the-clock service. It was suggested that because some larger fleets have multiple
fueling depots, an owner could assign vehicles with the new emissions device to a particular
depot where 15 ppm fuel would be available. Flexibility is very often the key to success for a
trucking company. We do not expect owners would spend hundreds of thousands of dollars for a
new truck and then restrict it's use to a particular, confined region. Card-lock companies will
likely also switch with demand for the 15 ppm fuel, rather than add facilities to handle two fuels.
iv. Truckstops
Truckstops depend on never having to turn away customers. The 15 ppm fuel can be sold
to all customers while 500 ppm fuel can only be sold to the pre-2007 vehicles. Consequently, we
expect that most truckstops would choose to begin carrying 15 ppm fuel at the start of the
IV-52
-------
Chapter IV: Fuel Standard Feasibility
program, particularly at an 80 percent requirement under the temporary compliance option where
500 ppm fuel may be in short supply. It was suggested that truckstops could easily de-manifold
their current systems and incorporate 15 ppm fuel for sale with their 500 ppm fuel. Based on
discussions with truckstop owners we learned that many of them take delivery multiple times a
day to prevent their tanks from running dry. Because at some point, there would be equal
demand for both fuels, half the tankage would need to be available for each fuel. However, at the
beginning of the program the 15 ppm fuel would likely be in low demand, which would make it
nearly impossible to keep the inventory of 500 ppm fuel from running out. An additional
concern, is what it would cost in lost business to take the system down for the re-manifolding
period at the beginning. However, truckstops are likely to be the location most capable of and
have the greatest economic incentive to make both fuels available. Regardless, however, 15 ppm
fuel should be available.
v. Service Stations
We expect most retailers, such as service stations, will switch to 15 ppm fuel rather than
install extra tankage to carry both fuels, especially given the magnitude and brevity of the
optional compliance period. While a limited number of retailers choose to sell only 500 ppm
fuel, they risk turning away customers, since the actual price differential between 15 ppm and
500 ppm fuel in most markets under the 80 percent requirement is expected to be small. The risk
of losing customers would likely outweigh any economic advantage for continuing to sell 500
ppm fuel.
vi. Evaluation of Fuel Availability by PADD
Essentially, our line of thinking and analysis was that if a majority of the refineries
produce the 15 ppm fuel, and given that the majority of the major pipelines connected to these
refineries, the fuel will be made available in quantities sufficient to widely distribute it through
the bulk plants to retailers and other end users. We began the evaluation with PADD HI since it
supplies fuel into most of the other PADD's.
(1) PADD HI
The total volume of diesel produced in PADD in is about 1.2 million barrels per day.
Eighty percent of that or about 960,000 barrels per day of 15 ppm diesel will be produced by
twenty-four refineries, most of which are located along the Gulf coast, although a few in other
areas of the state will play an important role in assuring 15 ppm fuel is widely available, not only
within PADD IE, but also in PADD's I, n, and IV. Movement within the PADD is handled by
companies such as the Longhorn, Koch, Ultramar Diamond Shamrock, Citgo, Conoco, Shell,
Chevron, Mobil, Fina, Texaco, and Trust pipeline companies. At productions volumes less than
80 percent, we estimated that at least one region of Texas may have difficulty receiving 15 ppm
fuel, except by truck. At the lower production levels, the refineries in that area would likely
IV-53
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
continue producing 500 ppm fuel, most of which would be exported into PADD's I and II; small
volumes are also exported into PADD's IV and V. Since most, if not all the fuel from this region
is flowing by pipeline out of the area, there is limited expectation that 15 ppm fuel would flow
into the area by pipeline. At 80 percent production volume and for the reasons discussed above,
we expect the fuel of choice for most pipelines will be 15 ppm fuel and that most retailers will
carry the fuel. As noted above, this PADD is a significant source of diesel fuel for PADD's I and
II.
(2) PADD I
Three refineries are predicted to produce about 135,000 barrels per day of 15 ppm diesel,
which is 80 percent of the total highway diesel produced by refineries in PADD I. Although
these refineries will produce more than enough 15 ppm fuel to meet first and even second year
demand in the entire PADD, most of the fuel produced in PADD I comes from the area above
Virginia. As such, it is usually distributed northward, northwestwards and northeastwards within
the PADD, with some short-distance distribution southward, mostly by truck. To the extent that
500 ppm diesel is already shipped, mostly by truck, into the upper northeastern states today, it
should not be difficult or expensive to replace that volume with 15 ppm fuel. We expect that at
the production levels most retailers will carry 15 ppm diesel. Since highway diesel demand in
PADD I is approximately 820,000 barrels per day, about 82 percent of it must be imported from
PADD HI, via the Colonial and Plantation pipelines and through foreign imports. For instance, a
shipment on the Plantation pipeline takes about 20 days and costs approximately 2 cents per
gallon to travel from Baton Rouge, Louisiana to the Washington D.C. area. Shipments on the
Colonial are comparable both in time and price. In order to meet overall diesel demand, the
pipelines will likely carry bothlS ppm and 500 ppm fuel.
(3) PADD H
The total volume of diesel produced in PADD II is about 682,000 barrels per day.
Thirteen refineries, fairly strategically located in seven of the 13-PADD II states, are predicted to
produce about 80 percent or 546,000 barrels per day of 15 ppm diesel. At this volume we expect
the diesel of choice for the maj ority of pipelines will be the 15 ppm fuel and that most retailers
will carry the fuel. The Kaneb, Amoco, Marathon Ashland, Buckeye, Countrymark, Conoco,
Phillips, and Wolverine pipelines move much of the fuel around in this PADD. Most of the
pipelines are hooked into refinery terminals but most also take delivery from and supply into the
other pipelines. We found that at volumes less than 80 percent production, it was likely that two
or three refineries in at least two strategic locations may not choose to produce 15 ppm fuel. In
both cases these refineries were an important source of highway fuel for a fairly significant area.
If they continued to produce 500 ppm fuel, it would likely stop, or at the very least hinder, the
flow of sufficient 15 ppm fuel into those areas and prevent it from being widely available. At the
80 percent production level, sufficient fuel would be available in both areas. About 122,000
barrels per day of additional diesel must be imported, principally from PADD IE. The Explorer,
IV-54
-------
Chapter IV: Fuel Standard Feasibility
Williams, Citgo, TEPPCO, Phillips, and Conoco pipelines play an important role in this transfer.
For the reasons we discussed above, we expect the 15 ppm diesel will be the fuel of choice for
the pipelines from PADD HI. Considering the location of the refineries in this PADD and their
access to pipelines which are expected to carry 15 ppm diesel, we expect the fuel will be widely
available.
(4) PADD IV
The total volume of diesel produced in PADD IV is about 127,000 barrels per day.
Eighty percent of that volume or about 101,000 barrels per day of 15 ppm diesel is projected to
be produced by eight refineries. The Chevron, Pioneer, Conoco, Yellowstone, Cenex, and Kaneb
pipelines move most of the fuel in this PADD. This PADD makes up the majority of the GPA
and has several small refineries that may choose wait until 2010 to make the investment to
convert to 15 ppm fuel. We analyzed each pipeline with respect to the volume of 15 ppm fuel
produced by refineries serving it. We were very concerned that, given the potential small refiner
and GPA choices, there would be insufficient fuel to cause the pipelines and terminals to switch
to 15 ppm fuel. We felt that parts of eastern Montana and Wyoming, western parts of North and
South Dakota and an area in northeastern Montana may not receive enough 15 ppm fuel to make
it widely available. At the 80 percent production requirement, we expect the fuel of choice for
the pipelines will be 15 ppm diesel because it is the dominant fuel and that most retailers will
carry the fuel.
(5) PADD V
PADD V has a few characteristics that make it somewhat different from the other
PADD's. This is the only PADD that is really comprised of a number of separate and distinct
fuel distribution systems; California, Arizona, and Nevada; Washington and Oregon, and Hawaii
and Alaska. In California the SFPP, Shell, Mobil, and CalNev pipelines most of the diesel
within the state. Las Vegas, Nevada is serviced via the CalNev. The southern part of Utah
(PADD IV) is supplied from Las Vegas by truck. The SFPP services Reno and Fallen, Nevada
in the north.
Another somewhat unusual condition exits in this PADD, in that the western halves of
Washington and Oregon are somewhat isolated from service from either California or PADD IV.
If PADD trading was widely permitted under a production requirement of less than 80 percent, it
is possible that the refineries in the northwest could actually purchase credits and produce no 15
ppm fuel. Because the region is isolated from reasonable service out of either PADD IV or V,
there would be no 15 ppm fuel in this area. At the 80 percent level, we expect that at least two
refineries in the northwest will convert to produce 15 ppm fuel in volumes sufficient to meet
demand for at least the first year or two. The Olympic pipeline connects the refineries in
Washington with Portland, Oregon and the SFPP connects Portland to Eugene, Oregon. Due to
the unique situation described earlier for this PADD, Alaska and Hawaii were split off from
IV-55
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
PADD V and made their own trading area in order to ensure sufficient availability of 15 ppm
fuel. Consequently, we expect most retailers and truckstops will switch to 15 ppm fuel for the
reasons described above.
B. Interaction with Other Programs
In addition to the program proposed today, there are a number of other environmental
programs that may concurrently have an impact on the refining industry. The most significant of
these programs is the recently promulgated gasoline sulfur standards as part of the Tier 2
rulemaking. We have examined the impacts on engineering, construction, and capital
expenditures of gasoline sulfur control in conjunction with the diesel sulfur control program .
A particular concern has been raised to the Agency regarding the capability of the
engineering and construction (E&C) industries to be able to design and build diesel fuel
hydrotreaters while at the same time doing the same for gasoline, as well as accomplishing their
other objectives. Compliance with the 15 ppm sulfur cap for on-highway diesel fuel begins for
refiners on June 1, 2006. This is within the timeframe of the phase-in of the Tier 2 sulfur
standards applicable to gasoline. Thus, it is important to consider the requirements of complying
with the diesel fuel sulfur cap in the context of the requirements of the Tier 2 gasoline sulfur
standards. Two areas where it is important to consider the combined impact of two or more fuel
quality specifications are: 1) refiners' ability to procure design and construction services and 2)
refiners' ability to obtain the capital necessary for the construction of new equipment required to
meet the new quality specification.
1. Design and Construction Services
We evaluated the requirement for engineering design and construction personnel,
particularly three types of workers: front-end designers, detailed designers and construction
workers, needed to implement the Tier 2 gasoline sulfur program and this diesel fuel sulfur cap.
We developed estimates of the maximum number of each of these types of workers needed
throughout the design and construction process and compare those figures to the number of
personnel currently employed in these areas. It would also be useful to evaluate certain types of
construction workers which might be in especially high demand, such as pipe-fitters and welders.
However, good estimates of the number of people currently employed in these job categories are
not available. Thus, it is not possible to determine how implementing the diesel fuel sulfur cap
might stress these specific job categories.
IV-56
-------
Chapter IV: Fuel Standard Feasibility
The number of job-hours necessary to design and build individual pieces of equipment
and the number of pieces of equipment per project were taken from Moncrief and Ragsdale.h
Their paper summarizes analyses performed in support of the recent National Petroleum Council
study of gasoline and diesel fuel desulfurization, as well as other potential fuel quality changes.
These factors are summarized in Table IV.B-1.
Table IV.B-1. Design and Construction Factors for Desulfurization Equipment
Number of New Pieces of Equipment per Refinery
Number of Revamped Pieces of Equipment per Refinery
Gasoline
60
0
Diesel
15
30
Job hours per piece of new equipment *
Front End Design
Detailed Design
Direct and indirect construction
300
1200
9150
300
1200
9150
: Revamped equipment estimated to require half as many hours per piece of equipment.
The gasoline sulfur standards as promulgated last year phase in between 2004 and 2008,
with the potential for the generation of early sulfur reduction credits prior to 2004. However, a
number of small refiners and refiners selling gasoline in the Geographic Phase-in Area are
expected to take advantage of an option being afforded in this diesel fuel program. These
refiners will be able to delay their compliance with the 30 ppm average, 80 ppm cap standards for
gasoline for two years. Thus, the phase in of the Tier 2 gasoline sulfur program now extends
from 2004 to as late as 2010 for GPA refiners and 2011 for qualifying small refiners.
The sulfur standards phase in at equal 12 month intervals effective on January 1 of each
calendar year. Thus, it is convenient to break up the construction of gasoline desulfurization
units by the year in which they have to become operational. Table IV.B-2 shows our projection
of the number of gasoline desulfurization units which must be operational by January 1 of the
indicated year.
h Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low Sulfur
Timetable," NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.
IV-57
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.B-2. Number of Gasoline Desulfurization Units Becoming Operational on
January 1 of the Indicated Year '
Prior
to 2004
2004
2005
2006
2007
2008
2009
2010
Gasoline Units: After Promulgation of the Tier 2 Gasoline Sulfur Program
10
37
6
26
9
9
Gasoline Units: After Promulgation of the Diesel Fuel Sulfur Program
10
37
6
26
5
3
4
6
Diesel Units
63
58
The diesel fuel desulfurization units are projected to start either 2006 or 2010, according to the
temporary compliance and hardship provisions. With respect to the required number of job-
hours per unit, all of the gasoline units were assumed to be new, grassroots units. The diesel fuel
units were assumed to be 80 percent revamps and 20 percent new, grassroots units, consistent
with the analysis presented earlier in this chapter.
A worse case assumption would be that all of the units scheduled to start up on a
particular January 1 began and completed their design and construction at the exact same time.
However, this is not reasonable for a couple of reasons. One, an industry-wide analysis such as
this one assumes that all projects take the same amount of effort and time. This means that each
refinery is using every specific type of resource at exactly the same time as other refineries with
the same start-up date. However, refineries' projects will differ in complexity and scope. Even
if they all desired to complete their project on the same date, their projects would begin over a
range of months. Thus, two projects scheduled to start up at exactly the same time are not likely
to proceed through each step of the design and construction process at the same time. Second,
the design and construction industries will likely provide refiners with economic incentives to
avoid very temporary peaks in the demand for personnel. Thus, with respect to units starting up
in a given year, we assumed that the design and construction of these units would be spread out
throughout the year, with 25 percent of the units starting up per quarter. Given this assumption,
1 Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: The Tier 2 Motor
Vehicle Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA, December 1999, EPA 420-R-
99-023.
IV-58
-------
Chapter IV: Fuel Standard Feasibility
we developed the breakdowns of personnel requirements by month for a given project shown in
Table IV.B-3.
IV-59
-------
Table IV.B-3. Distribution of Personnel Requirements Throughout the Project
Duration per project j
Duration for projects starting
up in a given calendar year
Front-End Design
6
15
Detailed
Engineering
11
20
Construction
14
23
Fraction of total hours expended per month from start of that portion of the project
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
0.050
0.050
0.050
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.050
0.050
0.050
0.020
0.030
0.040
0.040
0.040
0.050
0.050
0.060
0.065
0.075
0.075
0.075
0.060
0.060
0.050
0.050
0.040
0.040
0.030
0.020
0.030
0.030
0.030
0.040
0.040
0.040
0.040
0.050
0.050
0.055
0.055
0.060
0.060
0.055
0.055
0.050
0.050
0.040
0.040
0.040
0.030
0.030
0.030
1 Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low Sulfur
Timetable," NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.
IV-60
-------
Chapter IV: Fuel Standard Feasibility
The fraction of total hours expended estimated in Table IV.B-3 was derived based on the
following. Front end design typically takes six months to complete. If 25 percent of the
refineries scheduled to start of in a given year start their projects every quarter, each subsequent
group of the refineries starts when the previous group is halfway through their front end design.
Overall, front end design for the four groups covers a period of 15 months, or 6 months for the
first group plus 3 months for each of the three subsequent groups. In spreading this work out
over the 15 months, we assumed that the total engineering effort would be roughly equal over the
middle 9 months. The effort during the first and last 3 month period would be roughly two-thirds
of that during the peak middle months. The same process was applied to the other two job
categories.
Finally, we assumed that personnel were able to actively work 1877 hours per year, or at
90 percent of capacity assuming a 40 hour workweek.
Applying the above factors, we projected the maximum number of personnel needed in
any given month for each type of job. The results are shown in Table IV.B-4, both assuming the
availability and unavailability of the temporary compliance option. In addition to total personnel
required, the percentage of the U.S. workforce currently employed in these areas is also shown.
IV-61
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.B-4. Maximum Monthly Demand for Personnel
Front-End Design
Detailed
Engineering
Construction
Tier 2 Gasoline Sulfur Program As Promulgated
Number of Workers
Percentage of Current
Workforce *
421
22%
1,277
13%
8,423
5%
Gasoline Plus Diesel Fuel Programs: No Temporary Compliance Option
Number of Workers
Percentage of Current
Workforce *
882
46%
2,570
27%
15,623
10%
Gasoline Plus Diesel Fuel Programs: With Temporary Compliance Option
Number of Workers
Percentage of Current
Workforce *
571
30%
1,669
17%
10,658
7%
Based on current employment in the U.S. Gulf Coast, assuming that half of all projects occur in the Gulf Coast.
As can be seen, the diesel fuel program without the temporary compliance option would
have had a large, impact on the required amount of E&C resources compared to only the Tier 2
gasoline program. Employment required in all three job categories would have essentially
doubled with the addition of the diesel fuel program. However, with the temporary compliance
option, the impact of the diesel fuel program is reduced dramatically, to the point where the
required resources for the two programs are only about 30 percent greater than those of the Tier 2
gasoline program alone.
With the temporary compliance option, the largest impact is on front end design, where
30 percent of available U.S. resources are required. Thus, we believe that the E&C industry is
capable of supplying the oil refining industry with the equipment necessary to comply with the
proposed diesel fuel sulfur cap on time. We believe that this is facilitated by the extended phase-
in we allowed regarding compliance with the Tier 2 gasoline sulfur standards and the diesel
sulfur cap.
The second aspect of the aggregate impact of the proposed diesel fuel sulfur cap and other
rules on refiners is their ability to procure adequate capital to fund the required investment in
IV-62
-------
Chapter IV: Fuel Standard Feasibility
new desulfurization equipment. Estimates of previous capital investments by the oil refining
industry for the purpose of environmental control are available from two sources: the Energy
Information Administration (EIA) and the American Petroleum Institute (API).
According to EIA, capital investment by the 24 largest oil refiners for environmental
purposes peaked at $2 billion per year during the early 1990's.k Total capital investment by
refiners for other purposes was in the $2-3 billion per year range during this timeframe. API
estimates somewhat higher capital investments for environmental purposes, with peaks of about
$3 billion in 1992-1993.'
In the Tier 2 gasoline sulfur control rule, we estimated the expenditure of capital for
gasoline desulfurization by year according to the phase in schedule described above."1 In that
analysis, we simply assumed that all of the capital investment occurred in the calendar year prior
to the requirement that the unit be on-stream. Here, we developed a somewhat more
sophisticated schedule for the expenditure of capital throughout a project. We projected that the
capital investment would be spread evenly over a 24 month period prior to the date on which the
unit must be on-stream. The results are shown in Table IV.B-5.
kRasmussen, Jon A., "The Impact of Environmental Compliance Costs on U.S. Refining profitability,"
EIA, October 29, 1997.
1 API Reported Refining and Marketing Capital Investment 1990-1998.
m Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: The Tier 2 Motor
Vehicle Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA, December 1999, EPA 420-R-
99-023.
IV-63
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table IV.B-5. Capital Expenditures for Gasoline and Diesel Fuel Desulfurization
Calendar Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
Gasoline
1.7*
1.11*
0.85
0.59
0.15
0.06
0.06
0.02
Diesel Fuel**
1.3
1.9
0.7
0.5
0.9
0.2
Total
1.7
1.11
2.15
2.49
0.85
0.06
0.56
0.92
0.2
* Includes capital related to the construction of desulfurization units built prior to 2004 for the generation of early
credits.
As can be seen, capital investment peaks in 2004 and 2005, at $2.15-2.49 billion. This is
about two-thirds the previous high levels of refining industry investment for meeting
environmental programs experienced during 1992-1994.39 Therefore, this level of investment
should be sustainable for a couple of years, particularly since the required level of investment
drops off dramatically after 2006, and inflation has degraded the value of money somewhat since
the early 90's. From 2002-2005, the required level of investment averages somewhat below $2.0
billion per year, or about one-half of the levels experienced during the early 1990's.
In addition to gasoline sulfur control there are other environmental programs that could
also concurrently have an impact on the refining industry. The phase-down of MTBE from
gasoline is currently under consideration. While the nature of the action on MTBE has not yet
been determined, if EPA acts to reduce or eliminate MTBE usage, we will consider cost impacts
on refiners and provide sufficient lead time to comply with such requirements.
C. The Need for Lubricity Additives
Note that much of the discussion in this section on lubricity was obtained from two
Society of Automobile Engineers (SAE) Technical Papers.40 They are referenced here once to
avoid numerous repetitive references in the text. Also, some studies are noted in the text without
references. These studies, unless otherwise noted, are also extracted from these two SAE papers.
IV-64
-------
Chapter IV: Fuel Standard Feasibility
1. What Impacts Will the Sulfur Change Have on Lubricity?
Diesel fuel lubricity is the characteristic of diesel fuel to provide sufficient lubrication to
protect each of the many contact types within fuel pumps and injection systems for reliable
performance. Unit injector systems and in-line pumps, commonly used in heavy-duty engines,
are actuated by cams lubricated with crankcase oil, and have minimal sensitivity to fuel lubricity.
However, rotary and distributor type pumps, commonly used in light and medium duty diesel
engines, are completely fuel lubricated, resulting in high sensitivity to fuel lubricity. Low fuel
lubricity has been associated with low-viscosity fuels, such as No. 1 diesel fuel or kerosenes,
which are typically used in cold climates. As a result, many rotary fuel injection systems
intended for use in cold climates contain components manufactured using improved metallurgy
specifically to tolerate the use of poorer lubricity fuels.
Experience has shown that it is very rare for a naturally high-sulfur fuel to have poor
lubricity, although most studies show relatively poor overall correlation between sulfur content
and lubricity. One study indicated a relationship between diesel fuel lubricity and the content
and composition of sulfur compounds. However, the artificial addition of sulfur compounds
seems to have no effect or even a slight detrimental effect at high concentrations. Another study
showed that fully-saturated hetrocyclic sulfur compounds are the most active naturally occurring
fuel lubricity agents. High molecular weight components, back-end volatility, napthalenes,
polyaromatics, nitrogen compounds, polar compounds (excluding sulfur and nitrogen
compounds) and oxygen compounds have been identified as potential lubricity agents. There is
some indicated correlation between total aromatics content and lubricity, as measured by
laboratory tests. The addition of aliphatic kerosene fractions to diesel fuel, which have inherently
lower lubricity, can also decrease the lubricity of the resulting blend.
Unfortunately, few consistent trends are visible in the literature, and some researchers
have shown that properties such as sulfur, aromatics, acidity and olefin content cannot be used
alone to predict fuel lubricity. At present, the most that can be said definitively regarding the
impact of fuel composition on lubricity is that a single fuel with low viscosity, low sulfur,
aromatics and acid content generally will tend to have poorer lubricity than those with higher
levels. Considerable research remains to be performed regarding the fuel components most
responsible for lubricity. Consequently, successful application of either a chemical test or
predictive model depends on a better understanding of the fuel and additive components
responsible for lubricity.
Hydrotreating, in addition to reducing sulfur content, can lead to a reduction in the
concentration of various compounds which may contribute to fuel lubricity such as aromatics and
high molecular weight hydrocarbons. As early as 1976, it was suggested that lowering the level
of aromatics, separation of sulfur compounds and polar substances, as well as separation of
IV-65
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
surface-active substances during the hydrotreating refining process, can result in a reduction of
the lubrication qualities of the fuel. One report41 suggested that reduction in lubricity is caused
by the removal of the sulfur that itself acts as a lubricant, and the removal of some compounds
such a furans, pyroles and thiophenes in the refining process. In one 1992-93 study, extremely
low aromatics content produced by hydrotreating caused catastrophic failure of rotary fuel
injection pumps. Unfortunately, few consistent trends are visible in the literature, and some
researchers have shown that properties such as sulfur, aromatics, acidity and olefin content
cannot be used alone to predict fuel lubricity.
Similarly, the aviation community has investigated the lubricity of jet fuel. The most
satisfactory explanation for wear on failed aviation components has been a simple corrosive
process, involving the repeated formation and removal of metal oxides during sliding. To a
lesser extent, wear has been explained by severe adhesive wear and scuffing as the component
dimensions were reduced beyond tolerable limits or if contact loads were unusually high. Severe
refinery treatment removes the naturally occurring corrosion inhibitors from the fuel, allowing
formation of a thick oxide layer. The removal of chemically active species to upgrade thermal
stability was found to be associated with poorer fuel lubricity. A number of studies by the
middle to late 1960s indicated that poor performance of high-purity jet fuel appeared to be
related to the availability of naturally occurring compounds, rather than gross physical or
chemical properties. Other studies indicated that fully saturated hetrocyclic compounds and
polynuclear aromatic hydrocarbons have a beneficial effect on lubricating characteristics. As
little as two percent aromatics greatly increased the load-carrying capacity of paraffins. Mixtures
of heavy aromatics and paraffins were much more effective than either compound used alone.
One study also found complex esters and, to a lesser extent, high molecular weight polymers to
be effective as anti-wear agents in turbine and diesel fuel. Another study found that the lubricity
of severely refined fuels could be improved by the addition of trace concentrations of surface-
active additives, such as corrosion inhibitors.
Some studies have indicated that the presence of water may have a significant effect on
lubricity values, although apparently only humidity values were monitored and controlled for
those studies. The U.S. Navy conducted a study to determine the effects of humidity and water
on distillate lubricity using the BOCLE, SLBOCLE, and HFRR tests42 (these tests are described
in the next subsection). The results of this study indicated that the test fuels were not noticeably
affected by any of the water introduction methods using any of the three test procedures.
Notwithstanding all these uncertainties, hydrotreating has been known to reduce the
lubricity of the treated fuel, depending on the severity of the treatment and characteristics of the
crude. If as anticipated, refiners increase the severity of their hydrotreating to comply with the 15
ppm sulfur standard, the lubricity of some batches of fuel may be reduced compared to today's
levels. To compensate for the potential impact on fuel lubricity, we have accounted for an
increased use of lubricity additives in highway diesel fuel in our cost calculations.
IV-66
-------
Chapter IV: Fuel Standard Feasibility
2. How Can One Determine Whether the Lubricity of a Fuel Is
Adequate?
Many researchers have demonstrated that the correlation between the different wear
mechanisms in fuel pumps is dependent on the fuel composition. This is particularly important
for dissimilar wear mechanisms, such as oxidative corrosion and adhesive scuffing. The most
successful wear tests appear to be those that reproduce the predominant (i.e., the most damaging)
wear mechanisms. However, there is considerable disagreement as to the relevant importance of
each mechanism and also to the appropriate lab oratory-scale test procedure to measure lubricity.
A number of studies have observed poor correlation between pump wear and the most widely
used laboratory test procedures, and no single wear test provides a complete description of
lubricity. In addition, these tests appear less effective when evaluating fuels that contain
additives, compared to the base fuels. Several studies have reported that the laboratory tests
predict negligible benefits from lubricity additives, but fuel trials indicate that lubricity additives
do provide acceptable lubricity.
Many laboratory fuel tests which are designed to operate under boundary lubricating
conditions are strongly correlated to viscosity. For many crude sources, a disproportionate
fraction of sulfur-containing compounds are contained in the higher molecular weight fuel
components, indicating an intrinsic relationship between chemical and physical fuel
characteristics. One researcher successfully developed a simple empirical relationship that
predicted fuel lubricity as measured using the SLBOCLE test (described below) using viscosity
and di-aromatic content. Unfortunately, such a model does not account for the effects of trace
constituents or lubricity additives. In 1993, the U.S. Army systematically defined the principal
wear mechanisms as oxidative corrosion, chemical corrosion, adhesion, and scuffing (severe
adhesion), with oxidative and scuffing predominating. In that study, the degree of pump wear
seemed to be highly sensitive to the availability of dissolved moisture, indicating the presence of
an oxidative mechanism.
The BOCLE (Ball-on-Cylinder Lubricity Evaluator) apparatus uses a ball-on-rotating
cylinder contact geometry. The primary wear mechanism produced by this test was found to be
oxidative corrosion and possibly the chemical corrosion mechanism found in high-sulfur fuels.
The U.S. Army sponsored development of a modified BOCLE - the SLBOCLE (Scuffing Load
Ball-on-Cylinder Lubricity Evaluator) - in 1994, to measure fuel load-carrying capacity. It
measures the applied load required to produce a transition from mild boundary lubricated wear to
adhesive scuffing. To minimize the effects of oxidative corrosion and abrasive wear, the
SLBOCLE uses a polished test ring in place of a ground specimen. A Society of Automobile
Engineers (SAE) paper concluded that the SLBOCLE test is a good tool to evaluate the lubricity
of base fuels, which contain no lubricity additive.43 However, this method can distinguish
additives only if large amounts are used, well above the concentrations required to protect the
equipment.
IV-67
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The HFRR (High Frequency Reciprocating Rig) was developed in Europe in 1986. The
apparatus consists of a AISIE52100 steel ball, which reciprocates against a polished plate of the
same material. The mean wear scar diameter formed on the ball is used as a measure of lubricity.
This test produces a very wide range of wear mechanisms, depending on the fuel being evaluated.
However, SAE paper 961948 concludes that the correlation between fuel injection pump rig and
HFRR results have not been satisfactory. Many fuels which were regarded good according to a
pump demerit wear rating were regarded poor according to the HFRR.
In 1991, the Society of Automotive Engineers (SAE) formed a committee to evaluate the
effects of reduced fuel lubricity and to identify an effective laboratory wear test procedure. In
Europe, the Coordinating European Council (CEC) was established for the same purpose. In
1992, both groups cooperated under the auspices of an International Standards Organization
(ISO) working group. Following a systematic evaluation of the available test procedures, the
group performed a round robin test program to compare the HFRR, two variants of the BOCLE,
and the Falex BOTS (Ball-on-Three Seats) test. This work was backed up by full-scale pump
tests performed by the original equipment manufacturers using a matrix of 12 fuels. A HFRR
result of 450 microns was recommended by the ISO working group, and 460 microns by the
CEC, as the maximum result for acceptable lubricity. No official minimum SLBOCLE result has
been defined, but the ISO working group data and most studies indicate that an SLBOCLE result
of approximately 3,000 grams delineates the transition between acceptable and poor lubricity.
The ISO is involved in a Phase 2 study to include additized fuels, which were largely ignored in
the original study. The objective is to evaluate the correlation between injection equipment rig
tests and the HFRR test for additized fuels. No conclusion was reached at the time of publication
of SAE 1999-01-1479 in May 1999.
The American Society for Testing and Materials (ASTM) lubricity task force evaluated
the information that had been generated by previous working groups, including ISO, and
recommended that the SLBOCLE and HFRR tests be adopted as ASTM test methods. However,
the ASTM group chose not to adopt ISO's 450 micron specification and has not included a
minimum lubricity requirement in ASTM D-975 specifications for diesel fuel quality. In
addition to the additive problem, the two ASTM test methods (SLBOCLE and HFRR) suffer
from poor precision and do not correlate well with each other. The ASTM group decided it
needed to conduct more work to improve the precision of the test methods, resolve the
discrepancy between the test results and the actual field experience, and modify the test methods
to apply to additized fuels. A fuel specification will be considered after the test issues are
resolved. The ASTM group is evaluating a recent BOTD (Ball on Three Disks) test, along with a
modification to the existing HFRR method.
Chevron conducted a limited number of tests with additized fuels. In all cases, the
HFRR test was indicated to be the least responsive to additive concentrations. This method does
not recognize the existence of any additive up to levels above 100 ppm, and full benefit is
indicated at levels between 200 and 500 ppm. The SLBOCLE test recognizes an additive effect
IV-68
-------
Chapter IV: Fuel Standard Feasibility
between 40 and 50 ppm, and detects the full potential of the additive between 50 and 200 ppm.
The BOTD test recognizes an additive effect at a level as low as 10 ppm, and detects the full
additive potential at around 20 to 40 ppm. Early reports by a recent effort at Southwest Research
Institute indicate that the HFRR test discrimination of additized fuels could be improved by
changes to the frequency and stroke.
3. What Experience Has There Been with Low-sulfur Fuels?
What has been the experience with aviation turbine engines'!
Aviation turbine kerosene (Jet A, Jet A-l) is the principle fuel used by commercial
airlines. A wide cut fuel spanning the gasoline and kerosene boiling range (JP-4, Jet B) has
historically been used by many national air forces. A higher flash point fuel (JP-5) has been used
for naval aircraft. Compared to both low and high sulfur diesel fuels, aviation kerosene fuels
tend to be poor lubricants.
In 1969, the British Ministry of Defense formed a Fuel Lubricity Panel to specify a
lubricity parameter for aviation turbine fuel. The Panel was unable to specify a lubricity test that
would accurately reflect the lubricity requirements of an aviation turbine fuel, but it did suggest
that a ball-on-cylinder machine showed promise. In the mid-1970's, the U.S. Navy began to
experience durability problems on equipment operated with fuels from refineries outside the
continental U.S. Failures were reported for afterburner hydrolic fuel pumps and also hang-up of
fuel controls. The U.S. Navy in cooperation with the Coordinating Research Council (CRC)
Aviation Fuel Lubricity Group began a detailed evaluation of the BOCLE apparatus. That study
recommended that the BOCLE apparatus continue to be used to evaluate lubricity of fuels from
the aircraft fleet, as well as the use of corrosion inhibitors in military aviation fuels. The study
also recommended that new aircraft and fuel system components be developed to operate
satisfactorily on low-lubricity fuel.
Since 1975, the approach of the commercial aviation community has been to maximize
equipment durability through improved materials and design, rather than through control or
measurement of fuel lubricating characteristics. However, the approach of the military has been
to add corrosion inhibitors to the fuel. Currently, many military aircraft use JP-8 fuel, which is
generally equivalent to Jet A-l treated with several additives, including a corrosion inhibitor to
improve lubricity. As a result of these efforts, the aviation community has reported only isolated
problems related to lubricity. Research is in progress relating to future advanced turbine engines
in which the incoming fuel will be exposed to temperatures reaching 163 °C and as high as 315
IV-69
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
What has been the experience with compression ignition engines?
Unit injector systems and in-line pumps are actuated by cams lubricated with crankcase
oil, and have minimal sensitivity to fuel lubricity. Rotary and distributor type pumps, commonly
used in light- and medium-duty diesel engines, are completely fuel lubricated, resulting in high
sensitivity to the effects of lubricity. As a result, the rotary fuel injection system has been the
primary focus of lubricity research. As noted as far back as 1970, blending diesel fuel with a
small concentration of good lubricity fuel has a disproportionately large effect on the wear
characteristics of a blend. Blending or mixing different fuels minimizes the effect of isolated
poor lubricity products. It has been observed that most equipment failures occur in fleets that are
supplied from a single fuel source.
What has been the experience of the U.S. Military?
Military vehicles are particularly susceptible to the effects of fuel lubricity, as a given
post or camp or station will use fuel from a single supplier for a minimum contract period of 12
months. As a result, little potential exists for blending of fuels from different sources. In
addition, due to harsher operating conditions, engines used in military vehicles (especially
tactical vehicles) are more vulnerable to lubricity problems than the equivalent engines operated
in commercial vehicles.44 In the 1970s, the Army approved JP-5 as an alternative to DF-2 (VV-
F-800) diesel fuel. In the 1980s, the Department of Defense (DOD) adopted a single fuel for the
battlefield and converted its tactical fleet of compression ignition powered vehicles from DF-2 to
aviation turbine fuel (MIL-T-83133). In March 1988, DOD specified JP-8 as the primary fuel
support for overseas ground forces, but considered it and Jet A-l equivalent fuels. Jet A-l does
not contain the corrosion inhibitor additives of the JP-8 fuels. During Operation Desert Storm,
the military experienced rotary diesel fuel pump failures on its vehicles when operated on Jet A-
1. While examinations of the failed fuel injection components indicated the majority of failures
were attributed to causes other than lubricity, the Jet A-l did appear to produce increased wear in
some areas of the pumps. Tests performed with rotary injection pumps on a motorized test stand
indicated very severe wear was produced with Jet A-l in as little as eight hours. Wear rate was
significantly reduced by the corrosion inhibitors specified for use in JP-8. Subsequently, the U.S.
military no longer considered those fuels to be equivalent.
The military noticed vehicle fuel system component wear when fuels with a SLBOCLE
value of less than 2,000 grams were used consistently. The wear became significant for fuels
with a SLBOCLE value of less than 1,600 grams. The DOD indicates in its comments to the
proposed rule that, since the introduction of 500 ppm sulfur diesel fuel in the United States in
1993, it has experienced lubricity problems particularly in the Midwest and Northwestern portion
of the United States, especially during the winter season. As a result, seven military bases
require lubricity additives in the diesel fuel they procure during the winter months.
IV-70
-------
Chapter IV: Fuel Standard Feasibility
What has been the experience of the U.S. commercial sector?
According to the literature, no widespread failures associated with poor fuel lubricity
have been reported in the United States, although on average, its diesel fuel has borderline
lubricity, based on the HFRR test. However, a few commenters indicated experience of lubricity
problems with existing diesel fuel, particularly in the United States. Fuel sulfur levels have
been restricted to 500 ppm nationwide since 1993, except for Alaska and certain territories. In
commercial vehicles, the beneficial effect of blending different fuels is likely to occur.
When lubricity has been a problem, failures that have been reported involved the use of
No. 1 type diesel fuels with viscosities below 2.0 cSt at 40°C. Very low ambient temperatures,
aside from the obvious effect on viscosity, greatly reduces the solubility of moisture in the fuel.
Dissolved moisture is necessary for the oxidative corrosion wear mechanism to occur. Many
rotary fuel injection systems intended for use in cold climates contain components manufactured
using improved metallurgy specifically for this reason. Many municipal bus fleets in the
continental United States operate year round using low viscosity diesel fuel, such as DF-1, to
minimize exhaust emissions. In practice, many operators procure aviation kerosene fuels,
particularly in more temperate southern areas where low viscosity diesel fuel is not readily
available. Anecdotal reports of injection system failures with these fuels are relatively common,
with replacements occurring as early as 15,000 miles in some instances.
What has been California's experience?
Low sulfur (500 ppm) diesel fuels have been marketed in Southern California since 1988.
Beginning Octoberl993, diesel fuels marketed in all of California had to meet the new Federal
sulfur standard of 500 ppm and a new state requirement of 10 percent aromatics by volume, or
equivalent emissions. On average, the sulfur content of California's diesel fuel is about 140 ppm.
In 1989, a few researchers, including fuel suppliers and engine and equipment manufacturers,
recognized that the regulations to reduce the aromatics content in 1993 would have the potential
to affect equipment if the fuel lubricity was reduced substantially. Of particular concern was the
protection of rotary distributor pumps used in passenger cars, light vans and trucks, and much of
California's agricultural equipment.45
A Governor's Task Force on diesel fuel was created to investigate concerns regarding
lubricity and other properties of the reformulated fuel. As a result, for three years from October
1993 through 1996, staff of the California Air Resources Board monitored fuel injection system
problems (and also price increases and reports of supply shortages), and recommended that fuel
suppliers monitor the lubricity of their fuel using the U.S. Army's SLBOCLE test or other
appropriate test and add lubricity enhancing additives to diesel fuel with a SLBOCLE test result
below 3,000 grams. The 3,000 gram level was a compromise between the 2,220 gram level
suggested by the American Petroleum Institute (API) and the 3,330 gram level requested by
Engine Manufacturers Association (EMA). Diesel fuels marketed in California are blended from
IV-71
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
various refinery products and contain lubricity enhancers and other additives. As previously
discussed, the SLBOCLE test lacks precision in evaluating additized fuels and underestimates the
benefit of lubricity additives. Thus, the test results were considered an indication of the lubricity
of the base fuel, and a worse case of actual fuel lubricity. Cetane-enhanced alternative
formulation fuels, with aromatic content near 20 percent, were not hydrogenated as severely, and
thus were not as dependent on additives for acceptable lubricity.46
During the three year monitoring period of 1993-1996, no lubricity-related fuel pump
damage was documented for diesel vehicles using California fuel. Also, analysis of the data
showed no strong correlation of lubricity with either sulfur or aromatic content. At first, only 30
percent of the fuels evaluated met or exceeded the 3,000 gram level. The average SLBOCLE
results for California fuels increased from 2700 grams in 1993 to 3,300 grams in 1996. This
improvement may be attributed to use of lubricity additives, combined with increasing
production of alternative formulations and blended products, which perform better on the
SLBOCLE test than do low aromatic products with lubricity additives.47
What has been Canada's experience?
Overall, Canadian fuels tend to have low density, low viscosity, and lighter distillation
characteristics than those used in the United States, and are among the worst lubricity fuels in the
world. Diesel fuel No. 1, as used for much of the year in Canada, is broadly similar to the
kerosene fuels that caused durability problems in military vehicles, municipal buses, and aviation
equipment. Even prior to the introduction of low-sulfur diesel fuel, Canada had reported
problems with reduced equipment life. These failures were typically associated with winter
grade diesel fuels, particularly when they were used in warmer conditions. Low-sulfur fuels have
been available in Canada since the 1980s, and a maximum sulfur content of 500 ppm was
mandated in 1994. Fleet testing repeatedly demonstrated catastrophic pump failure in less than
500 miles of operation on Canadian fuel. In 1997, Canada modified its low-sulfur diesel fuel
specification to address the lubricity of winter fuels - those having a viscosity below 1.9 cSt at
40°C or less and a cloud point of -30°C or lower. A fuel supplier can "qualify" its fuel using one
of several options, ranging from a field test to pump rig tests to the HFRR or BOCLE laboratory
scale test. A fuel supplier must use lubricity additives if the fuel fails the selected test.
What has been Sweden's experience?
Beginning in 1991, Sweden required very low concentrations of sulfur and aromatics in
its diesel fuels: maximum of 10 ppm sulfur and 5 percent by volume aromatics for Class I fuel,
and maximum of 50 ppm sulfur and 20 percent by volume aromatics for Class n fuel. Field trials
and research conducted by the fuel producers and equipment manufacturers indicated that these
fuels, without additives, would produce unacceptable wear of light-duty injection systems.
Failure of test rotary fuel pumps occurred between 5,000 and 19,000 miles for Class I fuel, and
between 8,000 and 48,000 miles Class n fuel. Heavy-duty in-line pumps were less susceptible to
IV-72
-------
Chapter IV: Fuel Standard Feasibility
low lubricity fuel. This experience drew more attention and interest to the possibility of a
widespread fuel lubricity problem in the rest of Europe and in North America at the time they
were introducing their lower sulfur and lower aromatic fuels. Since that time the use of lubricity
additives in Sweden's fuel has resulted in acceptable equipment durability.48 Beginning in 1995,
Sweden required nonroad equipment, excluding commercial boats, railroads, and stationary
engines, to use the very low sulfur diesel fuel (which, as noted above, includes the lubricity
additives). The use of the very low sulfur diesel fuel in nonroad applications in Sweden has not
resulted in any equipment durability problems.
What has been Great Britain's experience?
Since 1998, due to the use of tax incentives, nearly all highway diesel fuel in Great
Britain has met a 50 ppm sulfur level. A lubricity additive is added to the fuel. The use of the
low sulfur fuel in on-highway trucks in Great Britain has resulted in acceptable equipment
durability. Although nonroad diesel fuel in Great Britain is not low sulfur, a limited number of
applications do use the low sulfur diesel fuel. The use of low sulfur diesel fuel in nonroad
applications in Great Britain has not resulted in any equipment durability problems.
What has been the Experience in Asia and in South-Pacific Countries?
In the Far East, a number of countries have already or will soon implement a 500 ppm
sulfur maximum: Korea in 1996, Japan in 1997, Hong Kong in 1997, Taiwan in 1999, Thailand
in 1999, Philippines in 2000 . In addition, Australia in 2003, and New Zealand in 2005 will
implement a 500 ppm sulfur maximum. Research is being performed to determine the effects of
Asian low-sulfur fuel on injection system durability, and except for one study in Thailand, results
have not yet been published.
In Thailand, a field study was conducted to investigate the effects of low-sulfur diesel
fuel (500 ppm) without lubricity additives on rotary injector pumps operating in actual driving
conditions in Thailand.49 The study involved three vehicles each for two fuels for 30,000 km.
The first fuel was imported and made up of a blend of U.S. West Coast, Malaysian and locally
refined fuel having a HFRR test value of 358 jim. The second fuel was the first batch of locally
produced off highway fuel with a HFRR test value of 467 jim. Evidence of wear at the end of
the study were within the normal acceptable range of wear at 30,000 km for all six pumps,
although the pumps operated on the locally refined fuel showed nearly twice as much wear, on
average, as the pumps operated on the imported blend.
4. What Can Be Done About Poor Lubricity Fuels?
Blending poor lubricity diesel fuel with a small concentration of good lubricity fuel has a
disproportionately large effect on the wear characteristics of the blend. Thus, blending or mixing
IV-73
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
different batches of diesel fuel, such as that which occurs in the commercial market in the United
States, minimizes the effect of isolated poor lubricity fuels.
Also, blending small amounts of lubricity enhancing additives has increased the lubricity
of poor-lubricity fuels to acceptable levels. Laboratory testing, field experience and controlled
pump and vehicle testing indicate that additives can be effective in reducing pump wear. The
lubricity additives widely used in diesel fuels range from the corrosion inhibitors used in aviation
turbine fuel to fully synthetic non-acidic products. The additives' impact on pump wear appears
to be strongly influenced by fuel composition and distillation characteristics, with larger
improvements observed for poorer lubricity fuels. According to contacts in the industry and an
SAE report, refiners are likely blending additives to diesel fuel on a batch-to-batch basis when
poor lubricity fuel is expected. In one comment to the proposal, a producer of fuel systems
confirmed that much of the U.S. diesel fuel today contains lubricity enhancing additives,
including military fuels.
Sweden, Canada, and the U.S. military offer examples of experiences using additives to
improve the lubricity of diesel fuel. Since 1991, the use of lubricity additives in Sweden's fuel
has resulted in acceptable equipment durability.50 Since 1997, Canada has required that diesel
fuel not meeting a minimum lubricity be treated with lubricity additives. The U.S. military has
found that traditional corrosion inhibitor additives that it uses, such as di-linoleic acid, have been
highly effective in reducing fuel system component wear. Consequently, the U.S. Army now
blends 250 mg/L of MTL-I-25017E corrosion inhibitor additive to all fuels that show a
SLBOCLE result below 2,000 grams, and regularly for Jet A-l, JP-5 and JP-8 fuels. In addition,
seven military bases that procure commercial fuel from the Midwest and Northwestern portion of
the United States require lubricity additives, especially for that fuel they procure during the
winter months.
According to the literature, lubricity additives have not been found to significantly affect
exhaust emissions. However, adding too much can produce unwanted side effects, such as
deposits in in-line injection pumps, fuel filter plugging, injector tip deposits, plunger sticking,
and water haze problems. For these reasons, the selection and treat rate of an additive are
important. An SAE report indicated it is likely that the more recently developed non-acid based
additives provide improved performance with reduced problems from lubricant interaction. For
example, in early 1996, field problems occurred in Western Europe with vehicles fitted with in-
line diesel injection pumps. Fuel filters were being blocked with black sticky gel caused by the
use of lubricity additives based on a specific type of divalent acid.51
The literature indicates that treat rates typically vary from 20 to 200 mg/L. Higher
concentrations are occasionally used, although in general, benefits appear to decrease at
concentrations above 500 mg/L. Oxidative corrosion and associated sensitivity to moisture are
eliminated by trace quantities of corrosion inhibitor additives. However, these additives have
little or no effect on adhesion and scuffing wear mechanisms. While corrosion and rust inhibitor
IV-74
-------
Chapter IV: Fuel Standard Feasibility
additives are commonly blended to distillate fuels at 10-15 ppm by the petroleum producers to
protect transport pipelines, this low concentration provides little protection to consumers because
of leaching during transportation and handling.
Stanadyne, the National Biodiesel Board, and West Central Soy, in comments to the
proposal, indicate that blending biodiesel with low lubricity conventional diesel fuel can increase
the lubricity to acceptable levels. Biodiesel is a zero sulfur diesel fuel made from domestically
produced renewable fats and oils. Testing of biodiesel at Stanadyne indicated that the blending
of two percent biodiesel with any conventional diesel fuel will be sufficient to address the
lubricity concerns that we have with existing diesel fuels. However, more testing would be
required to determine the required level of biodiesel in fuels not yet being produced, such as the
15 ppm sulfur fuel required by today's action. Stanadyne indicates the inclusion of low blends of
biodiesel is desirable for two reasons. First, it would eliminate the inherent variability associated
with the use of other additives, and would also eliminate the question of whether sufficient
additive was used. Second, biodiesel is a fuel or a fuel component rather than an additive. It is
possible to burn pure biodiesel in conventional diesel engines. Thus, if more biodiesel is added
than required for adequate lubricity, there will not be any adverse consequences that might be
seen if other lubricity additives are used at too high a level.
Also, it is possible for equipment producers to design new injection system equipment to
tolerate lower lubricity fuels. Revised manufacturing practices to include improved materials
successfully allowed commercial aircraft to operate continuously with very poor lubricity
kerosene-based fuels. Studies sponsored by the U.S. Army also confirmed the possibility of
reducing or eliminating the effects of poor lubricity through the use of improved metallurgy.
This approach could be applied to fuel systems for commercial compression ignition engines.
Conversion kits are already available to allow many pump systems to operate on low-
lubricity/low viscosity fuels and are commonly used on engines in Arctic regions. For the
purposes of this rule, we will assume that such conversions will not be commonplace in the
commercial vehicle and vehicle engine market.
Recommendations by the commentors were largely split by industry and are briefly
summarized here. They are discussed in more detail in the Response to Comments document in
the public docket. The equipment manufacturers indicated that the "voluntary" approach is not
adequate today, and is not likely to be adequate with the 15 ppm sulfur standard. For example,
Stanadyne (and DOD) indicated that their experience with the current policy of treating fuel on
an as-needed basis has fallen far short of ensuring good fuel lubricity and that a voluntary
approach under the 15 ppm sulfur program will lead to wide scale lubricity problems. EMA,
Cummins, the Alliance of Automobile Manufacturers and Stanadyne recommended that we
specify lubricity measurement methods and set limits. DOD recommended that EPA stress to the
industry the importance of having the appropriate performance requirements in the ASTM
specification and to encourage the industry to develop standards by imposing a deadline for
industry-wide implementation.
IV-75
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
API, Marathon Ashland Petroleum, the American Trucking Association and Cenex
Harvest States Cooperatives also expressed concern about the impact of today's action on
lubricity and recommended that EPA fully explore the lubricity issue; but API and Marathon
Asland Petroleum also commented that regulating lubricity is not necessary. Stanadyne and the
National Biodiesel Board suggested that we could require biodiesel to be blended with the 15
ppm sulfur diesel fuel, thereby alleviating lubricity concerns otherwise apparent with the low
sulfur diesel fuel. By doing so, there would be no need for us to adopt a voluntary or minimum
lubricity standard.
5. Today's Action on Lubricity: A Voluntary Approach
We have decided not to establish a lubricity standard in today's action. We believe the
best approach is to allow the industry and the market to address the lubricity issue in the most
economical manner, avoiding an additional regulatory scheme. A voluntary approach should
provide adequate customer protection from engine failures due to low lubricity, while providing
the maximum flexibility for the industry. This approach will be a continuation of current
industry practices for diesel fuel produced to meet the current Federal and California 500 ppm
sulfur diesel fuel specifications, and benefits from the considerable experience gained since
1993. It recognizes the uncertainties of testing and measuring fuel for lubricity, and will also
include any new specifications and test procedures that we expect will be adopted by the
American Society for Testing and Materials (ASTM) regarding lubricity of highway diesel fuel
quality. We fully expect the refining industry, engine manufacturers and end users to work
together to resolve any issues as part of their normal process in dealing with customer and
supplier fuel quality issues.
We do not believe that an EPA regulation is appropriate for several reasons. First, the
expertise and mechanism for a lubricity standard already exist in the industry. According to the
comments, the industry has been working on a lubricity specification for ASTM D-975, and low
cost remedies for poor lubricity have already been proven and are already being used around the
world. Although some commenters expressed concerns that the ASTM process might move too
slowly to establish a lubricity specification by 2006, we fully expect that today's action will
increase the urgency of those working to establish an ASTM D-975 lubricity specification, and
we believe they will do so in time for the production and distribution of the low sulfur highway
diesel fuel. We will do our part to encourage the ASTM process be brought to a successful
conclusion.
Second, we have no firm basis to justify a lubricity specification in today's action. One
such basis might be adequate demonstration that a lubricity level below or above a certain
specification would either cause emissions to increase, or hinder the operation of emission
control equipment. However, we have no evidence that lubricity impacts emissions, or emission
control equipment. This issue is primarily a concern about equipment performance. Equipment
IV-76
-------
Chapter IV: Fuel Standard Feasibility
performance is more appropriately addressed by the industry rather than government regulation
by this Agency.
Third, even if we had a statutory basis to justify a lubricity standard, we are concerned
that establishing an EPA lubricity regulation would provoke the same disagreements that the
industry is now engaged in its efforts to establish an ASTM D-975 specification. We are in no
better position to judge those issues than the industry experts who are already involved. Further,
once a specification is put into the regulations and the industry subsequently determines that the
specification should be changed, based on new information or circumstances, the burden would
be on us to amend the mandated specification by rulemaking. This is a significant burden to put
on the Agency for an engine performance issue that can and should be resolved by the industry
without government intervention.
6. Are There Concerns Regarding the Impact of Diesel
Desulfurization on Other Fuel Properties?
EPA is not taking action today on any fuel properties other than sulfur. We have
examined the impact of fuel properties other than sulfur, such as aromatics, on the materials used
in engines and fuel supply systems. We do not believe there will be impacts on materials from
such other fuel properties.
While there were some problems with leaks from fuel pump O-ring seals made of a
certain material (Nitrile) after the introduction of 500 ppm sulfur diesel fuel in 1993, these issues
have since been addressed by equipment manufacturers who switched to materials that are
compatible with low aromatic fuels. The leakage from the Nitrile seals was determined to be due
to low aromatic levels in some 500 ppm fuel, not the low sulfur levels. In the process of
lowering the sulfur content of some fuel, some of the aromatics had been removed. Normally,
the aromatics in the fuel penetrate the Nitrile material and cause it to swell, thereby providing a
seal with the throttle shaft. When low aromatic fuel is used after conventional fuel has been
used, the aromatics already in the swelled O-ring will leach out into the low aromatics fuel.
Consequently, the Nitrile O-ring will shrink and pull away, thus causing leaks, or the stress on
the O-ring during the leaching process causes it to crack and leak. Not all off highway fuels will
cause this problem, because of the amount and type of aromatics will vary. Subsequently, one
engine manufacturer recommended replacement of the old O-ring seals in leaking fuel pumps
with a new part of the same material, reasoning that the new part is not worn or has not taken a
compression set. One fuel producer recommended switching all fuel injection pump applications
to a different material (Viton) . Fuel pumps using a Viton material for the seals did not
experience leakage.
In comments to the proposal, the EMA, American Trucking Association, API and
Marathon Ashland Petroleum expressed general concerns about potential impact of 15 ppm
IV-77
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
sulfur diesel fuel on elastomer compatibility. However, these comments did not suggest that the
15 ppm sulfur diesel fuel will likely cause elastomer compatibility problems, or that any
preemptive action by EPA or the industry is necessary. EMA clarified that the elastomer
compatibility difficulties in the early days of the 500 ppm sulfur fuel program were likely a result
of severe aromatic reductions in some of the fuels, not necessarily the removal of sulfur. We
have no reason to believe that additional problems will occur with a change of fuel from 500 to
15 ppm sulfur.
D. Feasibility of Maintaining Off highway Fuel in the
Distribution System
1. Overview
There are a number of potential links in the highway diesel fuel distribution system from
the source of the fuel (refinery or importer) to the fuel retailer or fleet operator (hereafter referred
to as the point-of-use).n Depending on the location of the point-of-use relative to the fuel source,
the path of any given batch of highway diesel fuel through the system can include various
combinations of these links. Some highway diesel fuel is transported directly from the refinery
rack to the point-of-use via tank truck. However, most fuel is transported via the pipeline system
to product terminals. From the storage tanks at such terminal facilities, fuel can either be trucked
to the point-of-use or transferred by tank truck to bulk plants for later transfer to the point-of-use.
In some circumstances, highway diesel fuel is also transported to a terminal (or to a
pipeline connection) by barge or marine tanker, such as along the eastern seaboard, from Texas to
Florida, and in the case of imports. In cases where pipeline service is limited, fuel is also shipped
to the terminal by rail car. Smaller tank trucks called tank wagons are used to deliver fuel to a
variety of users including smaller retailers, fleet operators, and heating oil customers. Tank
wagons normally have multiple tank compartments to accommodate the delivery of several
different fuel types in a single delivery circuit. Most tank wagons also have a separate delivery
system for each product. There is a trend towards the increased use of such separate delivery
systems.
The same facilities in the fuel distribution system that are used to handle highway diesel
fuel are sometimes used to handle other products, including those with a high sulfur content. As
a result, there is currently some mixing of high sulfur products into highway diesel fuel. Sulfur
contamination of highway diesel fuel from such mixing can occur at each link in the distribution
n For additional discussion of the make-up of the highway diesel fuel distribution system, please refer to the
National Petroleum Council's (NPC) report on U.S. Petroleum Refining (attachment #6 in docket item IV-D-343)
and to the Draft RIA for the proposed rule.
IV-78
-------
Chapter IV: Fuel Standard Feasibility
system and is cumulative. Therefore, fuel batches whose distribution paths contain more links
are subject to more instances where contamination can occur.
The pipeline system is the primary source of potential mixing of high sulfur products with
highway diesel fuel in the distribution system. Transport by pipeline can involve a number of
steps. For example, fuel is placed in storage tanks prior to injection into the pipeline and at
transfer points between a main pipeline and branch lines or another operators pipeline. Thus,
there are a number of instances when highway diesel fuel can come into contact with high sulfur
products during shipment by pipeline. Pipeline systems vary greatly in the factors that can
contribute to product mixing. These factors include in the pipeline diameter, length, flow rate,
and number of branches off the main pipeline line. The Association of Oil Pipelines (AOPL)
stated that as the complexity of the pipeline system increases, there are a greater number of
potential sources of sulfur contamination.52
The most significant opportunity for mixing of high sulfur products into highway diesel
fuel during shipment by pipeline is associated with the fact that different products are normally
shipped through the same line sequentially with no physical separation between the products.
The mixture between two products where they abut each other in the pipeline is referred to as
interface when it can be blended into another product, and transmix when it must be returned to
the refinery for reprocessing. Pipeline operators take care to sequence the different products they
carry in such a way as to minimize the amount of transmix generated and the cost of
downgrading interface volumes to a lower value product.
Smaller batches of highway diesel fuel are commonly drawn off from a larger batch at
various points as it travels along the length of a main pipeline. A batch of fuel can also be
injected into a pipeline at various points along its length. An additional volume of interface can
be generated during each of these transfers. To minimize the generation of additional interface
volumes, such transfers are accomplished within the "heart"0 the batch already in the pipeline
whenever possible. Additional interface volumes can also be generated when a batch of fuel is
passed between different pipeline systems. This is primarily due to the need for the fuel batch to
be temporally placed in a stationary storage tank to facilitate the transfer between pipeline
systems.
All of the product that must be downgraded to a lower value product because of mixing in
the pipeline is sometimes referred to as interface, although strictly speaking, interface is only
generated when two products abut each other in the pipeline. Relatively small volumes of mixed
products are commonly included in the statement of total interface volume, such as those
associated with purging products contained in the manifolds at tank farms and in preparing for
0 The heart of a batch of fuel in a pipeline is that portion far enough from either end of the batch to ensure
that no mixing occurs with other adjacent products.
IV-79
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
the injection of a batch of fuel into the pipeline. The various concerns related to potential sulfur
contamination during the distribution of highway diesel fuel by pipeline are discussed in the
following section on limiting contamination in the pipeline system.
There are also significant concerns specific to limiting sulfur contamination in the other
links in the distribution system (terminals / bulk plants, tank trucks / tank wagons, marine
transport). These concerns include the potential for contamination of 15 ppm highway diesel fuel
when it is put into a stationary storage tank, vehicle tank compartment, transfer line, or delivery
line that previously held high sulfur products. Sulfur contamination can also result from leaking
valves. Diesel fuel sulfur content can also be impacted by the use of additives that have a high
sulfur content. These concerns are discussed in separate subsections that follow on limiting
contamination at stationary storage facilities, during transport by surface vehicles, during marine
transport, and from diesel fuel additives. Issues related to limiting contamination at tank farms,
whether they are part of a pipeline operation or a terminal facility, are discussed in the section on
limiting contamination at stationary storage facilities.
The extent to which mixing of high sulfur products into highway diesel fuel can be
tolerated is dependent on the maximum allowable sulfur content for highway diesel fuel, the
sulfur level of highway diesel fuel as it leaves the refinery gate, and the sulfur content of the
product with the highest sulfur cap that shares the distribution system with highway diesel fuel.
The highest sulfur product that presents a concern with respect sulfur contamination of highway
diesel fuel from mixing in the distribution system is off highway diesel fuel, which has an
industry-standard maximum sulfur content of 5,000 ppm and often averages approximately 3,000
ppm sulfur. EPA's current cap on the sulfur content of highway diesel fuel is 500 ppm with
actual sulfur level at production averaging approximately 340 ppm.p Thus, currently there is a 1
to 10 ratio of the maximum allowable sulfur content of highway diesel fuel to the highest sulfur
content of other products in the distribution system. This ratio provides a reference regarding the
current experience of the distribution industry in limiting the sulfur contamination of highway
diesel fuel.
Another useful reference is provided by the ratio of the of the difference between actual
highway diesel fuel sulfur levels and the 500 ppm cap to a reasonably severe sulfur level in off
highway diesel fuel. The average sulfur level of current highway diesel fuel is 160 ppm below
the 500 ppm cap. This difference below the sulfur cap is hereafter referred to as the "headroom"
below the cap. Although the maximum sulfur level of of highway diesel fuel potentially is 5,000
ppm, fuel batches near this cap are likely to be very rare and to originate only from a very limited
number of refineries. In addition, batches of of highway diesel fuel that are near the 5000 sulfur
cap are likely to be diluted with batches of lower sulfur content before (or as) they are introduced
in the pipeline. Based on this, we believe that 4,000 ppm is an appropriately severe sulfur level
p See section IV.A.2. regarding current sulfur levels in highway diesel fuel.
IV-80
-------
Chapter IV: Fuel Standard Feasibility
to compare against the current headroom below the existing sulfur cap for highway diesel fuel.
Thus, currently there is a 1 to 25 ratio of the current headroom below the highway diesel fuel
sulfur standard to the highest sulfur level in a product that highway diesel fuel might reasonably
be expected to come into contact with in the distribution system.
When the 15 ppm sulfur cap for highway diesel fuel is implemented, of highway diesel
fuel could have a sulfur content of approximately 333 times the highway diesel fuel sulfur cap.
Under our sulfur program, we expect that highway diesel fuel designated as meeting the 15 ppm
cap on sulfur content will leave the refinery with an average sulfur concentration of
approximately 7 ppm. Consequently, for highway diesel fuel to comply with the 15 ppm sulfur
standard, sulfur contamination could contribute no more than 5-8 ppm to the final sulfur of the
fuel. This translates to a 1/500 - 1/800 ratio of the maximum allowable sulfur contamination in
highway diesel fuel to the highest sulfur level in a product that highway diesel fuel might
reasonably be expected to come into contact with in the distribution system.
Based on a comparison of the above ratios, batches of highway diesel fuel meeting a 15
ppm sulfur cap will be able to tolerate much less mixing with high sulfur products than can
current batches of 500 ppm highway diesel fuel. It follows that adequately limiting sulfur
contamination during the distribution of highway diesel fuel meeting a 15 ppm sulfur cap may be
significantly more challenging than under the current 500 ppm sulfur cap. A comparison with
instances where the distribution industry has managed other difficult contamination issues is
useful in evaluating the relative magnitude of the new challenge posed by the implementation of
a 15 ppm cap on the sulfur content of highway diesel fuel.
In Sweden, diesel fuel meeting a 10 ppm sulfur cap has been distributed for some time.
However, high sulfur fuel oils are typically distributed in a separate distribution system in
Sweden. Due to this separation, Swedish 10 ppm sulfur diesel fuel is mostly segregated from
high sulfur products. Therefore, it is difficult to draw inferences from the Swedish experience on
how well the U.S. distribution system will accommodate 15 ppm highway diesel fuel. ARCO
Petroleum currently markets highway diesel fuel meeting a 15 ppm cap in a limited fashion in
California. However, this effort has yet to expand to the extent that their product is shipped in
the common distribution system. Thus, current experience of marketing 15 ppm diesel fuel in
the U.S. does not provide a useful reference with respect to the conditions when our sulfur
program will be implemented.
The distribution system has experience in limiting contamination of other products it
handles that may provide techniques useful in adequately controlling the sulfur contamination of
15 ppm sulfur highway diesel fuel. For example, the presence of small quantities of gasoline in
IV-81
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
diesel fuel can cause the industry standard flash point for diesel fuel to be exceeded.q This is a
significant concern because diesel fuel with an inappropriately high flash point presents a
explosion hazard and its use can result in driveability problems/ In addition, it is very difficult
for a batch of diesel fuel that is out of compliance with the flash specification to be brought back
into compliance by blending in a quantity of compliant diesel fuel into the noncompliant batch.8
Consequently, extreme care is taken to prevent mixing gasoline into diesel fuel. One relevant
example is that in separating a batch of diesel fuel from a batch of gasoline that it abuts in the
pipeline, none of the interface is allowed into the batch of diesel fuel.
Although most of the fuels handled in the distribution system are fungible, some
segregated products are carried such as high cetane diesel fuel and Amoco's clear premium
gasoline. In the case of Amoco's clear gasoline, mixing with other products must be strictly
limited to maintain the clarity of the product. This product is routinely transported by pipeline.
Limiting contamination during the transportation of such specialty products involves unique
challenges that may provide techniques useful in limiting sulfur contamination of 15 ppm
highway diesel fuel.
The current experience with limiting dye contamination in highway diesel fuel provides
another useful point of reference regarding a contamination concern that is currently being
managed by the distribution system. EPA requires that highway diesel fuel must show no trace
of the red dye which is required to be present in of highway diesel fuel by the Internal Revenue
Service (IRS) to demonstrate its non-tax status. A very small quantity of dyed of highway diesel
fuel mixed into highway diesel fuel can cause in a visible trace, resulting in a violation of EPA
requirements. To satisfy IRS requirements at the terminal, red dye must be present in of highway
diesel fuel at a concentration of at least 3.9 pounds per 1000 barrels (approximately 13 ppm).53
Some pipeline operators commonly add a lesser amount of dye upstream in the pipeline. One
operator requires that 0.75 mg per liter or approximately 0.9 ppm is added to of highway diesel
fuel prior to its injection into the pipeline. This practice is followed to minimize the loss of red
dye during transportation of dyed fuel and to reduce the risk of contaminating highway diesel
fuel with red dye in the pipeline system. An additional quantity of dye is then added at the
terminal to meet IRS requirements. Approximately 0.1 mg per liter of red dye in diesel fuel is
sufficient to result in a visible trace. This translates to a ratio of 1 to 130 regarding the maximum
amount of dye contamination that can be tolerated in highway diesel fuel to the minimum
q The impact of mixing gasoline into diesel fuel on diesel flash point is a function of the high vapor
pressure of gasoline relative to diesel fuel. In a diesel storage tank, gasoline contained in the diesel fuel will
contribute a disproportionate fraction of the total fuel volume that is in the vapor phase. During the winter, this
contribution can cause the vapor phase in a storage tank to be combustible, resulting in an explosion hazard.
1 Driveability problems can result from premature ignition of the fuel-air charge in the cylinder.
s This procedure is the preferred method of handling noncompliant fuel batches.
IV-82
-------
Chapter IV: Fuel Standard Feasibility
concentration of red dye that must be in of highway diesel fuel to meet IRS requirements (3.9
pounds / 1000 barrels).
Since no red dye is intentionally added to highway diesel fuel, any dye that is present
must originate from contamination from of highway diesel fuel. This suggests that it may be
most appropriate to base a comparison of the experience in limiting dye contamination with the
difficulty in limiting sulfur contamination under our program on the 1/500 - 1/800 ratio of the
maximum amount of sulfur contamination that we expect could be tolerated in 15 ppm highway
diesel to the highest sulfur level in a product that highway diesel fuel might reasonably be
expected to come into contact with in the distribution system.
The fact that red dye is not added at the full concentration required by the IRS until after
off highway diesel fuel reaches the terminal, prevents a direct comparison of the experience in
limiting dye contamination with that of limiting sulfur contamination of highway diesel fuel
meeting a 15 ppm sulfur cap during transport by pipeline. However, the 1/130 ratio of the
concentration of dye allowed in highway diesel fuel to the minimum concentration required in of
highway diesel fuel does provide a useful reference regarding a the current ability of distributors
downstream of the terminal (such as tank truck and tank wagon operators) to limit
contamination.
The IRS can impose a more stringent chemical test to detect red dye in highway diesel
fuel at levels which do not cause a visible trace. A violation of IRS requirements can be
established based on the results of such a test. While we do not have information on the
concentration of red dye that could be detected by such a test, it is reasonable to assume that it
would be substantially lower than 0.1 mg per liter (which causes a visible trace). Therefore, the
ratio of the maximum concentration of dye allowed in highway diesel fuel to the minimum
concentration required in of highway diesel may be considerably closer to the 1/500 - 1/800 ratio
associated with limiting sulfur contamination of highway diesel fuel meeting a 15 ppm sulfur
cap. If this is the case, it would suggest that distributors downstream from the terminal are
currently coping with a level of difficulty in limiting contamination similar to that will be
encountered as a result of our sulfur program.
The experience in limiting lead contamination in unleaded gasoline from leaded gasoline
during the phase-out of lead in U.S. gasoline provides the most useful point of reference. The
lead in leaded gasoline was added fully at the refinery, so a comparison of this experience with
the difficulty of limiting sulfur contamination under our sulfur program should be valid
throughout the entire distribution system. The situations where unleaded gasoline came into
contact with leaded gasoline (or traces of leaded gasoline) also parallels the situations that will be
encountered regarding the transportation of highway and of highway diesel fuels under our sulfur
program. For example, batches of unleaded and leaded gasoline abutted each other during
shipment by pipeline, and the same storage tanks and delivery equipment would sometimes be
used to handle both types of gasoline. This further supports the applicability of comparing the
IV-83
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
experience in limiting lead contamination to that of limiting sulfur contamination under our
sulfur program.
The maximum lead concentration in unleaded gasoline has always been 0.05 grams per
gallon, with the additional requirement that no lead be intentionally added to unleaded gasoline.
The maximum lead concentration in leaded gasoline was reduced in steps. In 1980, EPA adopted
a "pool standard" of 0.5 grams lead per gallon. Compliance with this "pool standard" was based
on evaluating the lead added to leaded gasoline by a given refiner divided by all of the gasoline
that the refiner produced (unleaded and leaded). This standard resulted in typical lead levels in
leaded gasoline of approximately 1 gram per gallon. In 1982, EPA adopted a 1.1 gram per gallon
"leaded gallons standard". At this time, approximately 50 percent of the gasoline pool was
leaded gasoline. Compliance with this "leaded gallons standard" was based on evaluating the
lead added to leaded gasoline by a given refiner divided by the volume of leaded gasoline that the
refiner produced. The use of this "leaded-gallons standard" had little effect on the in-use lead
concentration in leaded gasoline, which remained at approximately 1 gram / gallon until the
standard was reduced to 0.5 grams of lead per gallon of leaded gasoline in 1985.
During the time when the lead content of leaded gasoline was typically 1 gram per gallon
(near the maximum allowed concentration), lead levels in unleaded gasoline were typically less
than 0.005 gram per gallon. This translates to approximately a 1 to 200 ratio of the typical
maximum concentration of lead in unleaded gasoline to the typical maximum lead concentration
in lead gasoline. Similar to the discussion above regarding dye contamination, the fact that the
lead in unleaded gasoline could only have originated from contamination from leaded gasoline
suggests that it is most appropriate to base our comparison with the leaded gasoline experience
on the 1/500 - 1/800 ratio of the maximum amount of sulfur contamination that we expect could
be tolerated in 15 ppm highway diesel to the highest sulfur level in a product that highway diesel
fuel might reasonably be expected to come into contact with in the distribution system.
IV-84
-------
Chapter IV: Fuel Standard Feasibility
The various ratios discussed above are summarized in the following table IV.D-1.
Table IV.D-1. Ratios Used in Comparing the Relative Difficulty in Limiting Contamination
During the Distribution of Various Fuels
500 ppm Highway Diesel Sulfur Cap
5,000 ppm Off highway Diesel Sulfur Cap
15 ppm Highway Diesel Sulfur Cap
5,000 ppm Off Highway Diesel Sulfur Cap
Current Headroom Under 500 ppm Cap
Severe Sulfur Level in Product that Contacts Highway Diesel
Expected Headroom Under 1 5 ppm Cap
Severe Sulfur Level in Product that Contacts Highway Diesel
Maximum Dve Concentration in Highway Diesel
Required Dye Concentration in Off Highway Diesel
Typical Maximum Lead Concentration in Unleaded Gasoline
Typical Maximum Lead Concentration in Leaded Gasoline
500
5,000
15
5,000
160
4,000
5 to 8
4,000
0.1 ppm
13 ppm
0.005 g/gal
1 g/gal
1
10
1
333
1
25
1 to 1
500 800
1
130
1
200
The Association of Oil Pipelines (AOPL) stated that their members believe the task of
preventing sulfur contamination in 15 ppm highway diesel fuel will be more difficult than the
transition from leaded to unleaded gasoline, the protection of the flash property of diesel fuel, or
the prevention of dye contamination.54 Comparing the ratios discussed above regarding limiting
dye contamination (1/130) and limiting lead contamination (1/200), with the ratio of the
anticipated headroom under the 15 ppm cap for highway diesel fuel to the highest sulfur
concentration in off highway diesel fuel (severe level referenced in table IV.D-1) that is likely to
contact highway diesel fuel in the distribution system (1/500 - 1/800), suggests that this is the
case. However, this comparison also suggests that the challenge of limiting sulfur contamination
in highway diesel fuel meeting a 15 ppm sulfur cap is not an order of magnitude different to the
challenge of limiting lead contamination in unleaded gasoline that was successfully managed 25
years ago. This suggests that meeting the new challenge can be accomplished by improving upon
existing techniques to limit contamination, rather than requiring a paradigm shift in the way
highway diesel fuel is distributed.
IV-85
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Not all pipelines, terminals, and other fuel distributors handle off highway diesel fuel. At
such facilities, the challenge of limiting sulfur contamination of highway diesel fuel is, and will
continue to be, somewhat less difficult than at facilities that handle off highway diesel fuel. This
is because other products that might share the same distribution system have a lower maximum
sulfur content than off highway diesel fuel. For example, jet fuel and kerosene is subject to an
industry-standard sulfur cap of 3,000 ppm. The ratio of the maximum amount of sulfur that will
be allowed in highway diesel fuel under our sulfur program to the maximum concentration
allowed in jet fuel or kerosene is 1 to 200. This ratio is identical to that associated with limiting
lead contamination during the lead phase-down program. This suggests that the difficulty of
limiting sulfur contamination during the distribution of 15 ppm highway diesel fuel at facilities
that do not handle off highway diesel fuel will be of a similar magnitude to that experienced in
limiting lead contamination during the lead phase-down program.
Although not within the scope of current EPA regulations, the difficulty of distributing
highway diesel fuel with a 15 ppm sulfur cap would be significantly reduced if the sulfur content
of nonoad diesel fuel were reduced by a future rulemaking. If this took place the product with
the highest sulfur content shipped by pipeline would be jet fuel or kerosene which have a
maximum sulfur content of 3,000 ppm
In the NPRM, we proposed that with relatively minor changes and associated costs, the
existing distribution system would be capable of adequately limiting sulfur contamination during
the distribution of highway diesel fuel with a 15 ppm sulfur cap. These projected changes
included an increase in the amount of highway diesel fuel that must be downgraded to a lower
value product in the pipeline system due to changes in interface handling practices, and increased
terminal testing for quality control purposes. We also recognized that distributors downstream of
the refinery such as tank truck operators would need to more carefully and consistently observe
current industry practices to limit contamination, but projected that this could be accomplished at
an insignificant cost. We proposed to hold diesel fuel additives to the same sulfur cap that would
apply to diesel fuel, and projected that this could be accomplished without causing a significant
burden to fuel marketers and additive manufacturers.
We received a number of comments on the proposed rule that substantial uncertainties
exist regarding the ability of the distribution system to adequately limit sulfur contamination of
highway diesel fuel meeting a 15 ppm sulfur cap. Some commenters stated that the only way to
adequately limit sulfur contamination in the distribution of diesel fuel meeting a 15 ppm sulfur
cap may be to create a completely segregated system.
Several commenters stated that EPA should conduct testing to further evaluate the ability
of the distribution system to limit contamination to the very low levels necessitated by the
implementation of a 15 ppm sulfur cap. The Department of Energy (DOE) called on EPA to
conduct a comprehensive technology review regarding EPA's sulfur control program in the 2003
time frame, including the feasibility of distributing diesel fuel with a 15 ppm sulfur cap.55 DOE
IV-86
-------
Chapter IV: Fuel Standard Feasibility
stated that such a review is warranted because the distribution industry has never faced a similar
challenge in limiting contamination and would need to take extraordinary actions to do so. DOE
suggested that EPA participate in an experimental shipment of highway diesel fuel meeting a 15
ppm sulfur cap to evaluate the difficulties faced in limiting contamination.
While we acknowledge that today's rule will pose a substantial new challenge to the
distribution system, we believe that the additional measures outlined in this section will
sufficiently address issues associated with limiting sulfur contamination during the distribution
of 15 ppm sulfur highway diesel fuel. We expect that the changes to distribution practices that
may be needed will be logical outgrowths and extensions of current practices. With modest
modifications, the existing distribution system will be capable of limiting contamination during
the distribution of 15 ppm highway diesel fuel. The changes needed in the distribution system as
a result of our sulfur program will be readily apparent once industry focuses on meeting the
challenge of limiting sulfur contamination during the distribution of 15 ppm highway diesel fuel.
Therefore, testing by EPA or a formal technology review of the ability of the system to limit
contamination during the distribution of 15 ppm highway diesel fuel is not necessary.
It is possible that sources of sulfur contamination which did not hitherto represent a
significant concern may need to be reevaluated to assess their potential impact on maintaining the
15 ppm cap on the sulfur content of highway diesel fuel. Although all of these potential minute
sources of sulfur contamination in the distribution system may not have been identified and
quantified, we believe that the total contamination from such sources, while made more
significant by the implementation of the 15 ppm sulfur cap, is not of a sufficient magnitude to
jeopardize the feasibility of distributing 15 ppm sulfur highway diesel fuel.
We anticipate that the distribution system will conduct an evaluation of the potential
sources of contamination to ensure that each segment in the system has a satisfactory margin of
compliance below the 15 ppm cap. As a result of this evaluation, we anticipate that industry
may take measures to help adequately limit sulfur contamination in addition to those specifically
identified at this time. However, we anticipate that these measures will be the exception rather
than the rule. We do not anticipate that such additional measures will result in an unacceptable
burden to the fuel distribution industry (see Section V.C.3.).
We anticipate that the distribution industry will resolve what minor issues that might
remain while gearing up for the implementation of our sulfur program. We also anticipate that
some refiners will begin producing 15 ppm diesel fuel well before the time they are required to
do so. The voluntary efforts currently under way to market 15 ppm diesel are also expected to
expand in the years before the implementation of our sulfur control program. This will facilitate
the evaluation by the distribution system of their ability to limit sulfur contamination, and help to
ensure that whatever additional changes that may be necessary are made before industry is faced
with a regulatory compliance requirement. Industry will also gain experience in limiting sulfur
contamination in complying with the recently finalized Tier 2 gasoline sulfur requirements.
IV-87
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Some commenters on the NPRM stated that tanks that handle highway diesel fuel
meeting a 15 ppm cap would need to be dedicated to that purpose, necessitating the construction
of a substantial number of storage tanks, tank trucks, tank wagons, and barges at unacceptably
high cost and with impacts on numerous small businesses. We do not believe that our sulfur
control program will cause a significant increase in the extent to which highway diesel fuel must
be segregated from high sulfur products in the distribution system beyond the segregation that
already exists in much of the system today. Many of the stationary storage tanks, tank trucks,
tank compartments and delivery systems on tank wagons, and tank compartments on barges
which are used to handle highway diesel fuel are already dedicated to this purpose. Further, we
understand that there is a trend to increase this level of dedication, at least among large
companies in the distribution industry. Although our program may encourage this trend, we
believe that situations where our program will require equipment to be dedicated to handling
highway diesel fuel will be the exception rather than the rule.
Fuel distributors commented that contamination during the distribution of fuel in tank
trucks, tank wagons, barges, and rail cars could not be successfully managed simply by careful
and consistent observation of current industry practices, as we asserted in the NPRM. As
discussed below, we continue to believe that in most cases current industry practices will be
sufficient to limit contamination if properly followed. The one exception is in the case of barges,
which may need additional flushing when switching from a high sulfur product to 15 ppm
highway diesel fuel (see Section IV.D.4.).
Several commenters on the NPRM stated that unavoidable contamination could cause
many batches of highway diesel fuel to be noncompliant with the 15 ppm cap, resulting in
shortages and high costs. These commenters also stated that the current practice of diluting
batches of highway diesel fuel that do not comply with sulfur requirements with batches of fuel
that have a sulfur content below the standard to bring the resultant mixture into compliance with
the sulfur specification would no longer be possible when a 15 ppm cap on the sulfur content of
highway diesel fuel was implemented. They related that batches of highway diesel fuel that were
found to be noncompliant with the 15 ppm sulfur cap would need to be shipped by truck back to
the refinery for reprocessing (treated as transmix), resulting in substantial disruption the market
and cost. The Association of Oil Pipelines stated that pipeline operators may need to change the
products they choose to place in the pipeline adjacent to batches of highway diesel fuel meeting a
15 ppm cap.56 If this were the case, additional volumes of transmix could be generated.
We believe that an insignificant additional volume of transmix will be generated as a
result of our sulfur program. The generation of such additional transmix volumes will be limited
to circumstances related to the transfer of products through the manifolds at stationary storage
facilities and in preparing for the injection of products into the pipeline (line fill). We expect that
no changes will be needed in the choice of products that abut highway diesel fuel in the pipeline.
Therefore, we believe that no significant additional volume of transmix associated with pipeline
interface will be generated as a result of our program.
IV-88
-------
Chapter IV: Fuel Standard Feasibility
We believe that there will not be a significant increase in the volume of highway diesel
fuel discovered to exceed the sulfur standard downstream of the refinery as a result of our sulfur
program. Distributors will quickly optimize the distribution system using the means described in
this section to avoid creating additional volumes of out of specification product. We anticipate
that the preferred method of coping with batches of highway diesel fuel that are discovered to
exceed the 15 ppm sulfur cap will continue to be to blend them back into compliance whenever
possible. We expect that only in the infrequent instances where other options do not exist, will
batches that exceed the 15 ppm cap need to be returned to the refinery for reprocessing (see
Section IV.D.6). We believe that such rare circumstances will not significantly increase the
difficulty (and cost) of handling out-of specification highway diesel batches under our sulfur
program.
We received comments that we had underestimated the amount of highway diesel fuel
that would need to be downgraded to a lower value product. Commenters stated that the amount
could be so large as to contribute to difficulties in supplying sufficient quantities of highway
diesel fuel.
In response to comments, we adjusted our estimate of the amount of highway diesel fuel
that would need to be downgraded to a lower value product. Our analysis indicates that the
magnitude of the additional volume that would need to be downgraded can be accommodated
without causing supply problems or other disruptions to the market (see Section IV.A.9.).
Additive manufacturers stated that holding additives to a 15 ppm cap would result in a
significant burden to additive manufacturers due to the need to reformulate their additive
packages. They also stated that for certain essential types of diesel additives, no low sulfur
alternatives exist.
Our analysis of the Fuel and Fuel Additive (F&FA) database indicates that additives with
a sulfur content below 15 ppm are available to meet every purpose in use. However, we agree
with commenters that the contribution of high sulfur additives can be adequately controlled
without holding such additives to a 15 ppm sulfur cap. Therefore, we included a provision to
allow the continued use of additives that exceed a sulfur content of 15 ppm provided that this
does not cause the 15 ppm cap on the sulfur content of highway diesel fuel to be exceeded. This
provision will prevent any significant impacts from our sulfur program related to the use of diesel
fuel additives. Although our sulfur program may encourage high sulfur additives to be retired
from the market, we have structured the program in a way that will not require this to happen.
A number of commenters stated that difficulties in complying with our sulfur program
would be eased substantially if EPA were to include a downstream tolerance on the 15 ppm
sulfur standard to reflect measurement variability. In response to comments, we incorporated a
downstream tolerance on the 15 ppm sulfur standard in the compliance provisions of our
program to accommodate measurement variability. As suggested in the comments, we believe
IV-89
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
this will substantially ameliorate concerns regarding the ability to comply with the 15 ppm sulfur
standard downstream of the refinery. We believe this allowance will not significantly impact the
average level of fuel sulfur in-use. Therefore, providing this measurement tolerance will not
significantly reduce the emissions benefits of our program.
We concluded that despite the heightened challenge to the distribution industry caused by
our sulfur program, it will be feasible to distribute 15 ppm highway diesel fuel with relatively
minor modifications to the existing system which can be accomplished at modest additional
costs. The potential sources of sulfur contamination and the additional measures that we
anticipate will need to be taken to limit such contamination are discussed in the following
sections. Areas where further changes may be found to be called for as a result of the anticipated
process of optimizing the distribution system to limit sulfur contamination are also discussed
below. Please refer to Section V.C. on the costs to the distribution system, and to the Response
to Comments (RTC) document for our reply to comments on the feasibility of distributing
highway diesel fuel under our sulfur program.
2. Feasibility of Limiting Sulfur Contamination in the Pipeline
System
The primary source of potential sulfur contamination in the pipeline system is associated
with the handling practices for interface volumes associated with shipments of highway diesel.
The Association of Oil Pipelines (AOPL) stated that other potential sources of sulfur
contamination include pipeline dead legs, line fill, tank heels, tank manifolds, and the fact that
some valves designed to facilitate batch changes take as long as 45 seconds or more to close.57
There may also be a heightened level of concern regarding leaking valves. AOPL also expressed
concern that their current physical methods' of evaluating when to make a cut between adjacent
batches in the pipeline may not be adequate for determining when a cut should be made between
a batch of 15 ppm diesel fuel and another product batch adjacent to it in the pipeline. The
Department of Energy (DOE) stated that sulfur contamination from internal surface accumulation
of high sulfur product along the sides of pipes and within tanks, which currently is considered
negligible, might become significant given the small amount of contamination that could be
tolerated in fuel that must meet a 15 ppm sulfur cap.58
Each potential source of contamination in the pipeline system is discussed in turn below.
Some of the concerns discussed in this section, such as those related to the interface handling
practices regarding pipeline shipments of highway diesel fuel that abut batches of jet fuel or
kerosene, line-fill, and leaking valves are also pertinent to limiting contamination in other parts
4 Pipeline operators often discern the interface between two products in the distribution system based on a
change in fuel density and/or a change in color. Tracking information from upstream in the pipeline is used to help
identify the approximate time when the interface between batches will arrive at a given point in the system.
IV-90
-------
Chapter IV: Fuel Standard Feasibility
of the distribution system such as terminals and bulk plants. These concerns are discussed here
because they are integral facets of the pipeline distribution system.
Pipeline owners operate storage tanks where product is fed into the pipeline, at points
along the line where product is exchanged, and at the juncture between two pipelines. These
storage tanks are necessary to facilitate the exchange of products in the various lines and to
ensure that the pipeline remains in steady operation. Interface and transmix can be generated
during this exchange. Concerns related to limiting sulfur contamination in storage tanks used to
facilitate pipeline operation are discussed in the section on limiting contamination in stationary
storage tanks.
Several commenters stated that our sulfur program would cause a substantial increase the
volume of transmix that is generated during the shipment of highway diesel fuel, resulting in an
unacceptable burden to industry. This concern is addressed within the sections that address the
various potential contamination sources.
a. Interface Handling Practices
/'. Current Downgrade Volume
Some pipeline operators currently cut as much as 25 percent of the interface volume
generated during the shipment of highway diesel fuel into the highway diesel batch. The other 75
percent of the interface volume is cut into the high sulfur product which abuts the batch of
highway diesel fuel in the pipeline. This practice is currently possible because of the large over-
compliance with the current 500 ppm standard by refiners." By allowing some high sulfur
product to mix with highway diesel fuel, the amount of highway diesel fuel that is downgraded to
a lower value product is reduced, thereby minimizing cost.
In addition to the amount of interface that is generated during the actual transport of fuel
through the pipeline, relatively minor volumes of interface are also generated during the transfer
of products into and out of storage tanks associated with pipeline (and terminal) operation, and in
preparing to inject a batch of fuel into the pipeline. Given the small diameter of the piping that
connects storage tanks and is used to "lay down" a batch of fuel prior to injection into the
pipeline relative to the diameter of the pipeline, and the short length of such lines compared to
11 Highway diesel fuel sulfur levels currently average 340 ppm in the United States outside of California.
California has its own requirements on highway diesel fuel sulfur content which result in an average sulfur content
of 140 ppm within the State of California. See section IV. A. 3.
IV-91
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
length of pipelines, the amount of interface generated during such product transfers is relatively
small compared to that generated in the pipeline."
The Association of Oil Pipelines (AOPL) related that the current downgrade accounting
system does not provide a ready means to estimate the current volume of highway diesel fuel
downgrade.59 AOPL stated that this raises concerns regarding the accuracy of the estimates of
current downgrade provided by their members and cautioned against their use in estimating the
economic impact of our sulfur program. They also stated that the diversity in the characteristics
of their members operations led to a wide range in the estimates of the current downgrade
volume (ranging from 0.2 percent to 10.2 percent of the total volume of low sulfur diesel fuel
shipped by pipeline). These estimates included all of the sources of downgraded highway diesel
fuel.
It is worth noting that some commenters on the proposed rule apparently used the upper
bound in this range of individual estimates and the assumption that downgrade volumes would
double under our program to estimate that 20 percent of 15 ppm highway diesel fuel supplied
would need to be downgraded to a lower value product due to mixing with high sulfur products
in the distribution system. This approach substantially overestimates the additional highway
diesel fuel that would need to be downgraded to a lower value product as a result of our sulfur
program because it assumes that the worst case condition with respect to the current downgrade
volume is applicable for the entire range of pipeline operators. This does not take into account
the diversity in the characteristics of pipeline owner's operations that AOPL related was linked to
the wide range in the estimates of the current downgrade volumes that they received from their
members.
We believe that the estimates provided by AOPL members provides an adequate
characterization of the range of current downgrade volumes across the diverse pipeline
distribution system. To derive an estimate of the average downgrade for the pipeline system as a
whole today, we used the range of downgrade estimates from AOPL and a characterization of the
pipeline distribution system in terms of pipeline diameter and length derived from the PennWell™
pipeline database. Due to the characteristics of fluids as they travel through a pipeline, the larger
the pipeline diameter and the longer a batch of product is pumped through a pipeline, the greater
degree of mixing with adjacent batches that will take place. Furthermore, larger diameter
pipelines tend to be relatively more complex than smaller diameter lines (i.e. have more tank
v The amount of interface generated is dependent on a number of factors, most prominent of which are the
diameter and length of the line through which the product flows. Issues related to "line fill" are specifically
discussed in a later section.
w A proprietary database of information on pipeline and terminal facilities in the United States produced by
PennWell MAPsearch Inc., P.O. Box 5237, Durango Colorado, mapsearch.com.
IV-92
-------
Chapter IV: Fuel Standard Feasibility
farms and connections to other lines) leading to a larger number of interface volumes being
generated for any given batch of fuel as it travels to its ultimate destination.
We assigned a specific estimate of percent downgrade from those provided by AOPL
members to each pipeline diameter included in the PennWell database, ranging from 10.2 percent
for the largest diameter pipeline to 0.2 percent for the smallest diameter line. In doing so, we
assumed that downgrade increases linearly with the cross sectional area of the pipeline. To
account for the impact of pipeline length on downgrade volume, we weighted the downgrade
estimate for each pipeline diameter by the fraction of total pipeline system length represented by
that diameter. By this method, we estimated that the average downgrade for the pipeline system
as a whole currently is approximately 2.5 percent of the highway diesel fuel shipped by pipeline.
Data from the Energy Information Administration (EIA) indicates that 85 percent of all
highway diesel fuel supplied in the U.S. is sold for resale. Therefore, we believe it is reasonable
to assume that only this 85 percent is shipped by pipeline, with the remaining 15 percent being
sold directly from the refiner rack or through other means that does not necessitate the use of the
common fuel distribution system. By multiplying 2.5 percent by 0.85 we arrived at an estimate
of the current amount of highway diesel fuel that is downgraded today to a lower value product
of 2.2 percent of the total volume of highway diesel fuel supplied.
/'/'. Downgrade Volume with 15 ppm Sulfur Standard for Highway Diesel Fuel
We are assuming that when the 15 ppm cap on highway diesel fuel sulfur content is
implemented, it will no longer be possible to cut any of the interface volume into highway diesel
fuel. This is referred to as a protective interface cut, and corresponds to a doubling of the volume
of highway diesel interface volume downgraded to a lower value product compared to the 25
percent / 75 percent cut described above. Some individual AOPL members stated that a
protective interface cut would be necessary to limit sulfur contamination during the shipment of
15 ppm highway diesel fuel. Some AOPL members also stated that the amount of highway
diesel fuel that would need to be downgraded to a lower value product would likely double as a
result of our sulfur program. However, they cautioned that actual losses may be higher
depending on the extent to which sulfur from preceding batch trails back into a batch of 15 ppm
fuel.
Some pipeline operators currently make a protective interface cut when separating a batch
of highway diesel fuel from other products which it abuts in the pipeline. This suggests that our
assumption that the amount of highway diesel fuel downgraded to a lower value product will
double as a result of the implementation of the 15 ppm cap on the sulfur content of highway
diesel fuel will yield a conservatively high estimate of our program's impact. However, given
the uncertainties regarding the various sources of highway diesel fuel that must be downgraded to
a lower value product, we believe that the use of this assumption provides an appropriate level of
IV-93
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
confidence that we are not underestimating the impact of our sulfur program. This estimate is
also in agreement with that provided by several commenters.
We believe that it is highly unlikely that any difference that might exist in the physical
properties of 15 ppm diesel fuel (e.g. density, viscosity) versus those of current highway diesel
fuel will cause a substantial change in the extent to which sulfur from preceding batches trails
back into batches of highway diesel fuel. Regardless, our estimate that downgrade volumes will
double will help to account for various unknowns that may cause downgrade volumes to
increase.
By applying the assumption that highway diesel fuel volumes will double as a result of
our sulfur program to the estimate of the current downgrade volume (2.2 percent of highway
diesel fuel supplied) we estimated that an additional 2.2 percent of the highway diesel supplied
will need to be downgraded to a lower value product to adequately limit sulfur contamination as
a result of the implementation of the 15 ppm sulfur standard under our program. In gaining
experience with the distribution of 15 ppm highway diesel fuel, we anticipate the pipeline
operators may cut a somewhat greater portion of highway diesel fuel batches into products that
they abut in the pipeline in order to ensure that no volume of interface is cut into the highway
diesel fuel batch. This may result in somewhat more highway diesel fuel being downgraded until
pipeline operators become more confident in their ability to make a protective interface cut.
However, we do not expect that the additional volumes will be significant and believe that
pipeline operators will quickly optimize their interface handling practices to limit the volume of
highway diesel fuel that must be downgraded. We anticipate that the expansion of voluntary
efforts to market 15 ppm diesel fuel will facilitate such fine tuning of the pipeline system to
handle 15 ppm highway diesel fuel prior to the implementation of our sulfur program. We
attributed costs for this optimization process in our cost analysis (see section V.C.3).
Hi. Changes to the Interface Handling Practices when Highway Diesel Fuel Abuts
Shipments of Jet Fuel or Kerosene
The industry specification for the end boiling point of kerosene and jet fuel is much
lower that the specification for the end boiling point of diesel fuel. Since the measured end
boiling point of a fuel is much more related to the presence of high boiling point fuel species
rather than their concentration, a small quantity of diesel fuel mixed into kerosene or jet fuel can
cause the end boiling point specification for these product to be exceeded. The current practice
when a batch of highway diesel fuel abuts a batch of jet fuel or kerosene in the pipeline is to cut
all of the interface generated into the batch of highway diesel fuel. Discussion at a recent
industry conference highlighted the fact that this practice will no longer be possible when all
highway diesel fuel is required to meet a 15 ppm sulfur cap because of the relatively high sulfur
content of jet fuel and kerosene (as high as 3000 ppm). It was stated that as a result the mixture
of highway diesel fuel meeting a 15 ppm sulfur cap and jet fuel or kerosene would need to be
IV-94
-------
Chapter IV: Fuel Standard Feasibility
returned from the terminal to the refinery for reprocessing, at high cost (i.e. would need to be
treated as transmix).
While we agree that handling procedures for this mixture will need to change, we believe
that it will not be necessary to treat it as transmix. We believe that there will be opportunity for
the mixture to be sold from the terminal into the off highway diesel pool or 500 ppm highway
diesel pool during the period when the temporary compliance option is available. We have
concluded that the increased volume of downgrade can be accommodated without disruption to
the fuel market. The increased cost associated with downgrading this mixture is included in our
analysis of distribution costs. The need for this additional downgrade results in an effective
reduction in the supply of highway diesel fuel. The increased cost of supplying an additional
volume of highway diesel fuel to compensate for this reduction is incorporated in our analysis of
refinery costs. Additional storage tanks will be needed to handle the mixture at those terminals
that currently do not handle off highway diesel fuel. The cost of these tanks has been fully
accounted for in the calculation of costs during the time period when the optional compliance
program is available.
iv. Conclusion
We conclude that the primary change needed to current distribution practices to limit
sulfur contamination of 15 ppm highway diesel fuel in the pipeline system (and for the
distribution system as a whole) will be the elimination of discretionary mixing of a fraction of the
interface volume associated with pipeline shipments of highway diesel fuel into the highway
diesel fuel pool. When the temporary compliance option expires, the additional volume of
highway diesel fuel that will need to be downgraded to a lower value product as a result of this
change will be sold into the off highway diesel fuel market. During the period when the
temporary compliance option is available, we estimated that a fraction of this volume would be
sold into the 500 ppm highway diesel fuel market. The relative volumes of downgrade that we
expect will be sold into the off highway vs the 500 ppm highway diesel market and the costs
associated this downgrade are discussed in section V.C. We concluded that the additional
downgrade can be accepted without significant disruption to either the off highway or highway
diesel fuel markets (see section IV.D.2.).
The need to produce an additional volume of 15 ppm highway diesel fuel to compensate
for the additional downgrade is accounted for in both our feasibility and cost analysis (see
Section IV.D.2. and V.C.). Given that in all cases there will be opportunity to downgrade the
volume of interface which currently is blended into highway diesel fuel to a lower value product,
we believe that the changes outlined above can be implemented without the generation of any
significant additional volumes of transmix from this source. The Association of Oil Pipelines
stated that pipeline operators may more frequently abut batches of highway diesel fuel with
batches of low sulfur gasoline in an attempt to limit sulfur contamination of highway diesel fuel
meeting a 15 ppm cap.60 If this were the case, additional volumes of transmix would be
IV-95
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
generated since mixtures of gasoline and diesel fuel must typically be returned to the refinery for
reprocessing. The changes to pipeline interface practices described in this section will be
sufficient to limit sulfur contamination from high sulfur products that abut batches of 15 ppm
highway diesel fuel in the pipeline. Pipeline operators will not be forced to increase the
frequency that batches of gasoline abut batches of highway diesel fuel in the pipeline. Since we
expect that no changes will be needed in the choice of products that abut highway diesel fuel in
the pipeline, we believe that no significant additional volume of transmix associated with
pipeline interface will be generated as a result of our program.
During the time period when the temporary compliance option is available, the ability to
abut some batches of 15 ppm highway diesel fuel with batches of 500 ppm highway diesel fuel in
pipelines that carry both fuels may ease the difficulty limiting sulfur contamination of 15 ppm
fuel. We believe that it will still be necessary to cut all of the interface between such products
into the 500 ppm fuel batch. Nevertheless, the lower sulfur content of 500 ppm highway diesel
fuel relative to off highway diesel fuel would mean that whatever mixing that does take place
from would have less impact on the sulfur content of 15 ppm highway diesel fuel.
b. Identifying the Location of the Interface Between Fuel Batches
The Association of Oil Pipelines (AOPL) expressed concern that their current physical
methods of evaluating when to make a cut between adjacent batches in the pipeline may not be
adequate for determining when a cut should be made between a batch of 15 ppm diesel fuel and a
batch of a different product adjacent to it in the pipeline.61 AOPL related that pipeline operators
currently do not use the measurement of fuel sulfur content to help determine when such a cut
should be made. They related that there would be no time to conduct a lab evaluation of fuel
sulfur content and that appropriate on-line sulfur measurement equipment is currently not
available. The recent National Petroleum Council (NPC) clean fuels report stated that they did
not expect that field test equipment such as that which might be used to measure the sulfur
content of fuel as it flows through the pipeline would be available by the time our sulfur control
program is implemented.62
We do not believe that pipeline operators will need to substantially change the methods
used to detect the interface between highway diesel fuel and adjacent products in the pipeline.
We believe that the difference between the physical properties of highway diesel fuel and other
products carried in the pipeline will continue to be as identifiable as they currently are.
Therefore, pipeline operators will have the same ability to discern the interface between two
products in the distribution system based on a change in fuel density and/or a change in color. In
addition, pipeline operators are already coping with instances where the physical properties of
fuels in the pipeline is very similar. When the normal methods of detecting an interface between
batches are thought to provide insufficient differentiation between batches, pipeline operators in
some cases today inject a small amount of dye or other fuel marker at the start of a new batch to
distinguish it from a previous batch. We expect this practice will continue in the future.
IV-96
-------
Chapter IV: Fuel Standard Feasibility
Tracking information from upstream in the pipeline is also used to help identify the approximate
time when the interface between batches will arrive at a given point in the system. This helps to
focus the attention of technicians who make the cuts between pipeline batches during the time
when the interface is expected to pass their observation post.
Making a protective interface cut will likely be critical to adequately limit sulfur
contamination during the distribution of 15 ppm diesel fuel. This may force additional measures
to ensure that there will be will be adequate time for the cut to be made. Such measures may
include: more rapid communication between the station at which the fuel is sampled in the line
and the control room where the valves are operated, providing means to control the valves at the
point where the product in the pipeline is measured, or moving the sampling point further
upstream in the pipeline. We believe that the need for such changes will be made within the
context of optimizing the distribution system to limit contamination. The costs of these changes
are accounted for in our cost estimate for this optimization process and should not cause a
significant disruption to pipeline operations.
c. Dead Legs
Dead legs are lengths of pipeline extending off from a main line (e.g. to serve a terminal
tank farm) that have a valve situated some distance from the junction of the two lines. There is
potential for some mixing of the fuel left in the dead leg (e.g. after an exchange of products
between a terminal and the pipeline) with other batches of fuel as it passes in the main pipeline.
If such mixing occurs quickly, the product left in the dead leg would tend to be included in the
interface between adjacent products in the pipeline. For short dead legs, we believe that the
turbulence at the junction of the two lines will ensure that this is the case. If a dead leg was long
enough, some product might remain trapped near the valve in the dead leg. During the operation
of a pipeline it is common for pressure fluctuations to occur. Such fluctuations could cause
product trapped in a long dead leg to be drawn out into the pipeline stream over time, resulting in
some contamination of a batch as it passes the dead leg. Commenters stated that the sulfur
contamination of highway diesel fuel from dead legs could be significant when the 15 ppm cap
on sulfur content is implemented.
We believe that existing concerns about limiting contamination has ensured that existence
of long dead legs is the exception rather than the rule. Such concerns will have already provided
a strong incentive to keep the volume of fuel contained in a dead leg to an minimum by careful
placement of the valve close to the junction of the lines. To the extent that there may still be
some long dead legs in the system, compliance with the Tier 2 gasoline sulfur requirements will
encourage their elimination well before the implementation of our diesel sulfur program. To the
extent that long dead legs exists when our diesel sulfur program is implemented, the problem can
be rectified by properly repositioning the valve. Given the limited extent that such instances are
likely to exist, this should not be a significant burden to the pipeline industry. The potential cost
of such valves is small enough to be accommodated in the costs we have attributed to the
IV-97
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
optimization of the distribution system to limit sulfur contamination (see section V.C.3.). We
believe that any instances where long dead legs do exist, will be identified and rectified before
the 15 ppm sulfur standard for highway diesel fuel is implemented. Given that most, if not all
dead legs, are already relatively short, and the fuel in such dead legs is exchanged in the interface
between fuel batches as they pass the leg in the pipeline, the highway diesel fuel downgrade
volume from such sources is already factored into our analysis of downgrade volumes (see
section IV.D.2.a.).
d. Line-Fill
Prior to the injection of fuel into a pipeline, the feed line from the terminal or refinery
tanks holding the batch of fuel to be injected must be "layed down" (filled) with the product to be
injected. There is a like situation at the tank farms where product is transferred from a main
pipeline to a branch line or to another operators pipeline. The term line-fill refers to the amount
of fuel in the feed line(s) from a tank farm to a pipeline. When product is received at a terminal
from the pipeline, the product which is resident in the feed line must be purged. To facilitate the
exchange of products in the feed line, most facilities have at least two lines from the pipeline to
the tank farm. When possible, the fuel that must be displaced from the feed line is directed to a
tank that contains the same product. However, inevitably some fuel will need to be directed into
a "slop tank" to be treated as transmix. Pipeline operators keep records of the products resident
in the various line segments to ensure proper routing and separation of product when the line is
being layed down.
Line-fill volume is fixed and will not increase as a result of our program. Pipeline and
terminal operators will need to exercise additional care to limit the mixing of high sulfur
products into highway diesel fuel when preparing for the injection of a batch of highway diesel
fuel into the pipeline. However, given the relatively small diameter and length of lines used in
this process, there should be little or no increase in the amount of interface or tranmix generated.
Thus, there should be no need for additional tanks to handle transmix and little or no impact on
the difficulty and costs associated with the line-fill process.
An analogous situation occurs when product is drawn off of the pipeline into a stationary
storage facility (terminal or pipeline brake out facility). The product contained in the receiving
line (which can also be used to inject products into the pipeline) must be properly directed when
receiving a batch of highway diesel fuel. For the reasons discussed above, the implementation of
our sulfur program should also not result in the a significant impact related to drawing fuel off of
the pipeline.
e. Leaking Valves
Contamination from leaking valves is a greater concern from single-seal valves. Existing
concerns about product contamination has encouraged the increased use of double-seal valves
IV-98
-------
Chapter IV: Fuel Standard Feasibility
throughout the distribution industry. As a result, much of the system already uses such valves
and there is an increasing trend towards their use. In addition, it is common practice to monitor
valves to ensure their proper operation. Therefore, there should be little potential for sulfur
contamination of 15 ppm highway diesel fuel from leaking valves. We anticipate that at those
locations where double-seal valves are not already utilized, distributors may be encouraged to
install such valves by the implementation of the Tier 2 gasoline sulfur requirements. To the
extent that single-seal valves continue to be in use when our diesel sulfur program is
implemented, our diesel program may further accelerate their replacement with double-seal
valves.x We expect that the locations where such replacement is advised will be identified
during the process of optimizing the distribution system to limit sulfur contamination in 15 ppm
highway diesel fuel, and that their installation where needed will not be burdensome.
f. Surface Accumulation of Sulfur-Containing Substances
The specter was raised in the comments on the NPRM that sulfur contamination from the
accumulation of substances on the walls of the pipeline and on the surfaces of stationary storage
tanks and the tank compartments in tank trucks, tank wagons, rail cars, and barges could
jeopardize the ability to comply with the 15 ppm cap on highway diesel fuel. No other
information was provided to substantiate this concern. We believe there is no reason to surmise
that contamination from surface accumulation in pipelines will represent a significant concern
under our sulfur program. To the extent that sulfur-containing molecules in a fuel batch adhere
to the wall of a pipeline, they would tend to be flushed back into the pipeline flow in the interface
between products. Whatever amount that might tend to tenaciously adhere to pipeline walls
would tend to remain in an aggregate formation rather than dissolving back into the stream. Such
accumulation would either be removed as part of normal pipeline cleaning processes, or if flaked
off into the pipeline flow, would be removed during the normal filtration process. To the extent
that products accumulated on pipeline walls might dissolve back into the pipeline flow, this
would be more likely to occur into lighter products which act as solvents such as gasoline. Based
on the above discussion, we believe that contamination from surface accumulation in pipelines
will not represent a significant concern. For these same reasons, we expect that surface
accumulation in storage tanks will not pose a significant contamination concern. In addition, to
the extent that contamination from surface accumulation may be a concern, it seems reasonable
to conclude that this issue would already be an issue since highway diesel fuel is very sensitive to
dye contamination from off highway diesel fuel.
To adequately limit sulfur contamination, it may become more important to allow
sufficient time for high sulfur fuel clinging to the walls of tanks to drain completely before
x Double valves were used to help prevent lead contamination from leaded to unleaded gasoline during the
phase-out of lead in U.S. gasoline. The lead phase-out presented perhaps the most difficult situation with respect to
limiting contamination up to this time.
IV-99
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
refilling the tank 15 ppm highway diesel fuel. However, this represents only more careful
observation of what is current industry practice and should not impose a significant burden.
Such concerns are discussed further in the section on limiting sulfur contamination at stationary
storage facilities and during transportation by surface vehicles and marine vessels.
3. Limiting Sulfur Contamination at Stationary Storage Facilities
This section addresses the following concerns related to limiting sulfur contamination at
stationary storage facilities such as terminals, bulk plants, and pipeline break-out tank farms:
Quality control testing
Switching products contained in storage tanks:
Contamination from high sulfur product left behind in a storage tank that will be
used to contain highway diesel
Tank manifolds:
Contamination from high sulfur product contained in lines that connect various
storage tanks to a common fuel transfer point such as a terminal rack
Because of their crosscutting nature, the following concerns regarding the limitation of
sulfur contamination at stationary storage facilities were addressed in the previous section on
limiting contamination in the pipeline system:
The need for additional storage tanks at certain terminals to accommodate a needed
change in the interface handling practices with respect to batches of highway diesel fuel
that abut batches of jet fuel or kerosene in the pipeline
Line fill
Leaking valves
Surface accumulation of high sulfur product
a. Quality Control Testing
We believe that a modest level of additional quality control testing will be needed at the
terminal level to ensure compliance with the 15 ppm sulfur cap. Further, we believe that such
additional testing can be conducted using existing equipment and will not represent a substantial
burden to industry. For additional discussion regarding the extent and costs of this testing please
see section V.C.3 in this RIA. For a discussion of the test procedures we expect will be used to
IV-100
-------
Chapter IV: Fuel Standard Feasibility
measure the sulfur content of highway diesel fuel please see section Vn in the preamble the final
rule for our diesel sulfur program.
b. Product Switching in Stationary Storage Tanks
In some instances, different products are held in the same tanks at different times. This
can occur at the tank facilities which are a part of the pipeline system, such as the facilities which
feed product into pipelines and at break-out tank facilities.y During the switching of products
held in a storage tank, sulfur contamination may result from high-sulfur product left behind in the
tank before being filled with highway diesel fuel. The foremost potential source of residual
product left behind during such switching operations is the tank heel. A tank heel is the amount
of product that remains in a tank after no additional product can be removed by normal means.
Pipeline operators have expressed concern that a tank heel volume of off highway diesel fuel that
today can be mixed with an incoming batch of highway diesel fuel without causing the current
500 ppm cap to be exceeded, would cause the 15 ppm cap to be exceeded.63 In such cases the
tank would need to be flushed before 15 ppm diesel fuel could be placed in it. This concern was
expanded upon by the Independent Fuel Terminal Operators Association (IFTOA) who stated
that storage tanks would need to be chemically cleaned before being used to store 15 ppm diesel
and be dedicated to holding only 15 ppm highway diesel fuel thereafter.64
The amount of tank heel varies depending on the type of tank floor. Some tanks have a
conical shaped floor typically constructed of concrete with a drain at the lowest point. The tank
heel for such a tank is nearly zero. New tanks are typically constructed with such a floor so that
they will drain completely. This greatly facilitates tank maintenance and facilitates the easy
removal of water that settles at the bottom of the tank.
Some (primarily older) tanks have a roughly level sand floor on which a liner rests. Such
tanks may or may not have a number of drains at various low points in the tank to facilitate the
elimination of water that settles to the bottom. The floors of such tanks can be quite irregular
and contain a number of low areas where fuel pools and can not readily be removed. This
volume is the tank heel. In addition, such tanks may have a side drain rather than a floor drain,
which can contribute further to the volume of the tank heel. The volume of the tank heel for such
tanks can be substantial. Therefore, high sulfur product in the tank heel could be a significant
source of sulfur contamination.
One fact which tends to limit the potential sulfur contamination from residual high sulfur
products in the flat bottom tanks is that water tends to settle into stagnant areas at the bottom of a
tank. This limits the volume of petroleum products that can reside in such stagnant areas since
y An example of a break-out tank facility is one that holds products the are stripped off of a main pipeline
before injection into another line.
IV-101
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
they tend to be filled with water. The water trapped in stagnant pools on the bottom of a tank is a
concern in and of itself. Biological organisms can grow at the interface between water and
petroleum products, especially diesel fuel. The residue from such organisms can cause
significant contamination. These concerns have contributed to the trend away from the use of
tanks with an irregular flat bottom. More importantly, the majority of storage tanks used to hold
highway diesel fuel are currently dedicated to this purpose and there is an increasing trend
towards this practice. It seems reasonable that due to the difficulty in switching the products
contained in storage tanks with irregular flat bottoms, such tanks would be the first tanks to be
dedicated to a single fuel. In the infrequent cases where this is not already the case, it seems
reasonable that where practicable the tank would be dedicated to a single fuel. In such cases,
only a one time cleaning would be required.
To the extent that some additional dedicated tankage will be necessary, we have more
than compensated for this in our analysis of the additional tanks that will be needed to facilitate
the presence of two grades of highway diesel fuels during the period covered by the temporary
compliance option. We have assumed that such tanks will be constructed at a number of
stationary storage facilities and anticipate that most storage facilities will have a dedicated tank to
handle 15 ppm diesel fuel. We also estimated that additional storage tanks may be needed at
some tank farms that service the pipeline system due to a change in the interface handling
practices regarding batches of highway diesel fuel and jet fuel or kerosene that abut each other in
the pipeline (see section IV.D.2.a. in this RIA). This also helps to compensate for whatever
additional tanks might be needed to address contamination concerns. The costs for these tanks is
incorporated in our estimation of tank costs to facilitate handling two grades of highway diesel
fuel during the period when
It seems likely that storage tanks would need to be flushed with highway diesel fuel prior
to being switched from containing off highway diesel fuel, jet fuel, or kerosene to 15 ppm
highway diesel fuel. We do not believe that there is any reason to suspect that the tank would
need to be chemically cleaned to remove residual high-sulfur products clinging to the interior
surfaces of the tank.2 It should be noted that due to concerns about dye contamination from off
highway diesel fuel and the impact of gasoline on the flash point of highway diesel fuel, properly
emptying a tank to hold highway diesel fuel is already a significant concern. Consequently, it is
not uncommon currently for a storage tank to be flushed with a quantity of highway diesel fuel
prior to being filled with highway diesel fuel if the previously held gasoline or off highway diesel
fuel. We believe that following such normal business practices when switching products
contained in a storage tank in most cases would provide sufficient protection against sulfur
contamination in 15 ppm highway diesel fuel. Some additional volume of highway diesel fuel
may need to be used in flushing tanks before switching a storage tank to highway diesel fuel
z See the earlier discussion on potential contamination from high sulfur products clinging to the walls of
pipelines (IV.D.l.e.).
IV-102
-------
Chapter IV: Fuel Standard Feasibility
service when our 15 ppm sulfur cap is implemented. However, because such switching occurs
infrequently, we believe that additional amount of downgrade caused by tank flushing will be
insignificant. In addition, our use of conservative assumptions in estimating the additional
downgrade volume from the changes needed in interface handling practices will more than offset
any additional downgrade volume that may result from tank flushing.
c. Tank Manifolds
The array of tanks at tank farms is connected by a network of pipes that resemble the
network of rail lines at a railroad yard. At the juncture between the feed lines from the pipeline
and the tank manifold system, a branching takes place such that products directed toward two
different tanks may flow down a single connecting line for a time. Similar to the line fill
situation, product downgrade and transmix can result from different products sharing the same
line. To the extent that the tank that contains highway diesel fuel is located at the end of the pipe
network, there may be more opportunity for mixing of high sulfur products into a batch of
highway diesel fuel as it moves through the manifold to and from the storage tank.
When off highway diesel fuel abuts highway diesel fuel in a tank manifold, it is common
practice to make a protective interface cut between the two batches (i.e.: all of the mixed product
is cut into the off highway diesel fuel). This practice is followed because the interface volumes
in manifold pipes are small and there is little incentive or ability to fine tune the amount of
interface which is cut into the different fuel batches. We expect that this procedure will continue
in the future and will be sufficient to limit the sulfur contamination of 15 ppm highway diesel
fuel in tank manifold systems. Therefore, we believe that the handling of 15 ppm highway diesel
fuel in tank manifold systems will not result in the generation of significant volumes of
additional product that must be downgraded to a lower value product or treated as transmix.
Another factor which tends to minimize concerns related to tank manifolds is that only a small
volume of product resident in the pipe networks must be displaced when moving a batch of
highway diesel fuel, even in those cases where the storage tank that holds the highway diesel fuel
is at the end of the manifold system.
As discussed in the previous section, we estimated that many facilities will construct an
additional tank dedicated to 15 ppm diesel fuel. To the extent that contamination concerns exist
regarding the placement of highway diesel fuel storage tanks in the manifold system, we
anticipate that new tanks will be located in a way that minimizes these concerns.
4. Limiting Sulfur Contamination During Transport by Surface
Vehicles
Highway diesel fuel is transported by the following types of surface vehicles: tank truck,
tank wagon, and rail car. Tank trucks are the largest capacity road haul vehicles that carry
IV-103
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
petroleum products. They deliver product to truck stops, service stations, and large fleet
operators, as well engaging in other road movement of petroleum products as needed. Tank
wagons usually serve smaller customers, such as centrally-fueled fleets, smaller service stations,
and in certain circumstances heating oil customers such as homeowners. Tank wagons normally
have multiple tank compartments to accommodate the delivery of several different fuel types in a
single delivery circuit. Tank wagons have a smaller total capacity than tank trucks. In cases
where pipeline service is limited, fuel is sometimes shipped to the terminal by rail car.
In the proposed rule, we stated that concerns related to limiting sulfur contamination
during the transport of 15 ppm diesel fuel by tank truck, tank wagon, and rail car could be
adequately addressed by careful and consistent observation of current industry practices used to
limit contamination. Based on this assessment, we concluded that our program would result in a
significant additional burden regarding the transportation of highway diesel fuel by such vehicles.
The Independent Fuel Terminal Operators Association (IFTOA) stated that all storage
tanks, including those on surface transport vehicles would need to be chemically cleaned before
being used to store 15 ppm diesel and would need to be dedicated to this purpose thereafter.65
The American Petroleum Institute (API) stated that it is unlikely that "consistent and careful"
observation of current practices will be sufficient to limit sulfur contamination during transport
of 15 ppm diesel fuel as EPA asserted in the proposed rule.66 The American Trucking
Association (ATA) stated that our assertion that enhanced observation of current industry
practices by truckers that distribute highway diesel fuel was incorrect.67 ATA argued that an
additional burden results whenever a trucker must alter current handling practices and that this
additional burden would impact truckers who are small businesses the most. ATA offered no
additional detail on the nature of the potential burdens. We did not receive information to
substantiate the concerns raised in these comments.
In their comments on the Advance Notice of Proposed Rulemaking (ANPRM), the
Petroleum Marketers Association of America (PMAA) stated that contamination concerns would
cause a large number of tank wagon operators to purchase new trucks with dedicated tank
compartments for 15 ppm highway diesel fuel and dedicated delivery systems.68 PMAA stated
that this would cause much disruption to the fuel distribution industry and an unacceptable
burden to its members. We responded to these comments from PMAA in the Draft Regulatory
Impact Analysis (DRIA) for the proposed rule.69
We continue to believe that sulfur contamination during the transport of 15 ppm diesel
fuel by tank truck, tank wagon, and rail car can be adequately controlled by the careful and
consistent observation of current industry practices used to limit contamination. These practices
include making sure that the tank truck, tank wagon, or rail car is properly leveled and allowing
sufficient time for the tank compartment to drain completely prior to filling with 15 ppm
highway diesel fuel. The tank compartments in such vehicles are designed to drain completely.
As discussed earlier (see section IV.D.2.e.), we do not believe that the accumulation of high
IV-104
-------
Chapter IV: Fuel Standard Feasibility
sulfur products on the walls of storage tanks will be significant source of sulfur contamination.
Therefore, assuring that such compartments drain completely will be sufficient to limit sulfur
contamination. There are no unique concerns related to limiting contamination during the
transport of 15 ppm highway diesel fuel by tank trucks and rail car.
PMAA stated that it would not be possible to adequately limit sulfur contamination when
delivering 15 ppm highway by a tank wagon that has a common delivery system. In such cases,
the delivery system would need to be purged of high-sulfur product prior to its use to delivery 15
ppm highway diesel fuel. Current industry practice is to switch the product flow just prior to the
end of the delivery so that the delivery system is charged with the product intended for delivery at
the next stop. PMAA stated that this is not always feasible because the receiving tank may fill
more quickly than was expected, or the next customer may not need the product that has been put
into the hose. We believe such occurrences will be rare and can be further limited by more
careful filling and delivery scheduling practices. Tank wagon operators currently schedule
deliveries of highway diesel fuel as a first stop in the delivery circuit whenever possible to help
minimize contamination concerns.
In transit!oning from the delivery of off highway diesel fuel to highway diesel fuel,
PMAA related that after switching to highway diesel fuel at the end of the delivery of off
highway diesel, tank wagon operators typically observe the color of the product being delivered
and do not turn off the flow until the stream is clear. Since off highway diesel (including heating
oil) typically contains a red dye, a clear stream indicates that the delivery system is completely
flushed of off highway diesel fuel. This practice will continue to be sufficient to ensure that the
delivery system is charged with highway diesel fuel since a minute quantity of dye causes a
visible trace in highway diesel fuel (see section IV.D.l).
Since the practices described above are currently in common use due to existing product
contamination concerns, we continue to believe that there should be not be a significant
additional burden associated with ensuring their consistent and careful observance. Some
marketers may need to reeducate their employees regarding the importance of these practices. To
the extent that such employee education is needed at all, we anticipate that it might be
accomplished in regular employee meetings or employee bulletins at negligible cost.
In any event, the concerns discussed above should represent the exception rather than
rule. Most highway diesel fuel is distributed to retail facilities and centrally fueled fleets where
off highway diesel fuel is not used. Thus, the circumstances where the same tank compartment is
used to alternately handle off highway and highway diesel fuels are limited. This also means that
cases where a tank wagon's delivery circuit includes off highway diesel fuel would be limited.
Such cases would primarily be limited to areas where diesel fuel oil is used for home heating,
such as in the North-East during the home heating season.
IV-105
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
More importantly, the tank compartments in tank trucks and tank wagons are for the most
part dedicated to carry a single fuel. In addition, most tank wagons have a separate delivery
system for each product. Further, there is an increasing industry trend towards dedicating such
equipment to handle a single fuel. In cases where such dedication exists, sulfur contamination
will not be a concern. Thus, the circumstances under which the concerns discussed above arise
are expected to be uncommon and to decrease over time.
5. Limiting Sulfur Contamination During Transport by Marine
Vessels
The Independent Fuel Terminal Operators Association (TFTOA) stated that the current
practice of flushing tanks on a barge with fuel when a supplier switches from a fuel with a higher
sulfur content to highway diesel fuel would no longer be possible when the 15 ppm cap on the
sulfur content of highway diesel fuel is implemented.70 IFTOA stated that it would be necessary
to clean the tank compartments with high powered water jets which is a difficult and expensive
process ($30,000 to $50,000 per barge).
During the three month transition period between the time when refiners are require to
produce 15 ppm highway diesel fuel and when it is required downstream, we anticipate that
distributors stationary storage tanks will gradually be blended down so that any residual product
is removed. Thus, for dedicated stationary storage tanks we expect that contamination from
residual high sulfur fuel will not be a significant concern. Similar to stationary storage tanks, we
expect that barges will experience sufficient turn overs of the fuel contained in their tank
compartments to ensure that sulfur contamination from residual high sulfur product is not a
significant concern.
It may be reasonable to presume that barges are equipped with sumps from which the
residual product can be completely removed. If this were the case, one might conclude a barge
cold be made ready to carry 15 ppm highway diesel fuel by allowing sufficient time for fuel to
drain into these sumps to be removed. If this were not sufficient, flushing the barge with 15 ppm
diesel fuel might provide the necessary cleaning action. If this were so, the situation would be
similar to that discussed above for stationary storage tanks, for which we concluded that sulfur
contamination from residual product should not be a significant concern.
Due to existing contamination concerns, most tank compartments in marine vessels used
to transport highway diesel fuel are already dedicated to this purpose and there is an increasing
trend toward such dedication. Some barges plying the eastern seaboard may on occasion switch
seasonally between highway diesel and heating oil. However, this is the exception rather than the
rule. Consequently, we expect that there would be few instances when this concern would arise
which would decrease in time. To the extent that such instances might occur, we believe that the
associated tank cleaning costs would not substantially add to the cost of our program. In
IV-106
-------
Chapter IV: Fuel Standard Feasibility
addition, the volume of heating oil shipped under such circumstances is small fraction of the total
volume shipped by barge. Hence, any impact would be insignificant in the context of our entire
program.
We received no comments to suggest that there are unique concerns for other marine
vessels such as coastal tankers.
6. Limiting Sulfur Contamination from Diesel Fuel Additives
Diesel fuel additives include corrosion inhibitors, cold-operability improvers, and static
dissipaters. Use of such additives is distinguished from the use of kerosene by the low
concentrations at which they are used and their relatively more complex chemistry.aa We
proposed that diesel fuel additives used in highway diesel fuel meet the same cap on sulfur
content required for the fuel itself. Additive manufacturers commented that there was no need to
impose a 15 ppm sulfur cap on such additives in order to effectively limit the sulfur content of
finished diesel fuel.71 They asserted that imposing such a cap would result in unjustified costs
and disruptions to the producers and users of diesel additives. Additive manufacturers also stated
that for certain additives, such as static dissipaters needed to prevent explosion hazards at
terminal facilities, there are currently no effective alternatives that comply with a 15 ppm cap on
sulfur content.
In response to these comments, we are allowing the use of diesel fuel additives with a
sulfur content greater than 15 ppm provided their use does not result in an exceedence of the 15
ppm cap on the sulfur content of highway diesel fuel.
Our review of data submitted by additive and fuel manufacturers to comply with EPA's
Fuel and Fuel Additive Registration (F&FAR) requirements (40 CFR Part 79), which is
summarized below, indicates that additives to meet every purpose (including static dissipation)
are currently in common use which meet a 15 ppm cap on sulfur content. The ability of industry
to provide additives for use in 15 ppm highway diesel fuel is further supported by the fact that
diesel fuel meeting a 10 ppm cap on sulfur content has been marketed in Sweden for some time,
and ARCO Petroleum recently began marketing fuel meeting a 15 ppm sulfur cap in California.
Even if low sulfur additives were not yet available for certain purposes, we believe that it is
reasonable to assume that they would become available before our sulfur program is
implemented in 2006. The summary of the data in the F&FAR database also indicates that the
aa Diesel fuel additives are used at concentrations commonly expressed in parts per million. Diesel fuel
additives can include specially-formulated polymers and other complex chemical components. Kerosene is used at
much higher concentrations, expressed in volume percent. Unlike diesel fuel additives, kerosene is a narrow
distillation fraction of the range of hydrocarbons normally contained in diesel fuel. See Section VII.C.4. above
regarding the requirements associated with the addition of kerosene to diesel fuel.
IV-107
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
industry could adapt to use only additives that contain less than 15 ppm sulfur. However, we
agree that it is not necessary to force the additives that contain greater than 15 ppm sulfur to be
retired. By allowing their continued use, we avoid any significant impacts from our sulfur
program related to diesel fuel additives.
Summary of Information Contained in the F&FAR Database on Diesel Fuel Additives
• Most sulfur containing additives registered with the EPA currently meet the 15 ppm cap.
There are approximately 3500 diesel additives registered with the EPA.
Of the diesel additives registered with EPA, 463 additives manufactured by 104
companies contain sulfur.
Of the sulfur-containing additives, 176 additives (38 of such additives)
manufactured by 51 companies (49 of companies that manufacture sulfur-
containing additives) have a sulfur content greater than 15 ppm
There are 226 sulfur-containing additives manufactured by 65 companies that
have a sulfur content less than 5 ppm.
• In 1999, 5.5 percent of the total volume the additives used in diesel fuel contained sulfur.
• In 1999, 47 percent of the diesel fuels registered by fuel manufacturers had sulfur
containing additives listed (of all purposes in-use). These fuel formulations represent 65
percent of the total diesel fuel volume.
• Several dozen different additives registered with EPA have anti-static (static dissipater)
listed as a purpose in-use (PIU). EPA data shows that there are 40 additives that list anti-
static as a PIU.
64 percent of these additives have an elemental sulfur level greater than 15 ppm.
Nearly a dozen different anti-static additives registered with the EPA have zero
amount of sulfur in their formulations.
Since such off highway additives are currently in widespread use side-by-side with high-
sulfur additives, it is reasonable to conclude that there is not a significant difference in their cost.
The unusually high sulfur content of a few additives may discourage their use in diesel
fuel that meets a 15 ppm sulfur cap. However, it will generally continue to be possible for
additive manufacturers to market additives that contain greater than 15 ppm sulfur for use in
highway diesel fuel. Such additives can also continue to be used in off highway diesel fuel.
Additive manufacturers that market such additives and blenders that use them in highway diesel
fuel will have additional requirements to ensure that the 15 ppm sulfur cap on highway diesel
fuel is not exceeded. Although our sulfur program may encourage the gradual retirement of
additives that do not meet a 15 ppm sulfur cap for use in highway diesel fuel, we do not
anticipate that this will result in disruption to additive users and producers or a significant
IV-108
-------
Chapter IV: Fuel Standard Feasibility
increase in cost. Additive manufactures commonly reformulate their additives on a periodic
basis as a result of competitive pressures. We anticipate that any reformulation that might need
to occur to meet a 15 ppm sulfur cap will be substantially accommodated within this normal
cycle.
In some cases, blenders may not find it feasible to conduct testing, or otherwise obtain
information on the sulfur content of the fuel either before or after additive blending, without
incurring substantial cost. Without such information, a blender would not have documentation
with which to evaluate what impact the use of an additive which exceeds 15 ppm would have on
the fuel's final sulfur content.bb We anticipate that blenders will manage the risk associated with
the use of additives above 15 ppm in sulfur content under such circumstances with actions such
as the following:
selecting an additive with minimal sulfur content above 15 ppm that is used at a low
concentration, and
working with their upstream suppliers to provide fuel of sufficiently low sulfur content to
accommodate the small increase in sulfur content which results from the use of the
additive.
This is similar to the way distributors will manage contamination from their distribution
hardware (tank trucks, etc.). Distributors will not necessarily test for fuel sulfur content after
each opportunity for contamination, but rather will rely on mechanisms set up to minimize the
contamination, and to obtain fuel sufficiently below the standard to accommodate the increase in
sulfur content from the contamination.
7. Handling Batches of Highway Diesel Found to Exceed the Sulfur
Standard Downstream of the Refinery
We believe that there will not be a significant increase in the volume of highway diesel
fuel discovered to exceed the sulfur standard downstream of the refinery as a result of today's
rule. We believe this will be the case both during the transition of the program and after the
sulfur requirements are fully implemented. We anticipate that distributors will quickly optimize
their practices to avoid sulfur contamination. We also anticipate that distributors will gain some
bb The transfer of an additive with a sulfur content greater than 15 ppm will be required to be accompanied
by a product transfer document which provides information in the sulfur content of the additive and the extent to
which its use at the maximum recommended concentration would increase the sulfur content of the finished fuel.
This information will allow the blender to assess the potential impact of the additive's use on their compliance with
the requirement that the use of additives not cause the 15 ppm cap on the sulfur content of highway diesel fuel to be
exceeded.
IV-109
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
experience in reducing sulfur contamination in the distribution system through complying with
the recently finalized Tier 2 low sulfur gasoline requirements (65 FR 6698, February 10, 2000).
While outside the scope of this final rule, it is worth pointing out that potential difficulties in
distributing 15 ppm diesel fuel would be lessened if the sulfur content of off highway diesel fuel
is reduced by a future rulemaking. We anticipate that the batches of highway diesel fuel that are
discovered to exceed the 15 ppm sulfur cap will be coped with as follows:
When possible, by blending highway diesel fuel that is below the 15 ppm cap with the out
of specification batch to bring the resulting mixture into compliance. This practice will
be more difficult than it is currently because the amount of fuel needed to blend the out of
specification batch into compliance may increase. However, we expect it to continue to
be the method of choice for handling out of specification highway diesel whenever
possible.
By downgrading the batch either to off highway diesel fuel or to 500 ppm highway diesel
during the initial years of our program when the temporary compliance option is
available.
By reprocessing the batch to meet the 15 ppm cap, but only in those infrequent instances
where the previous options do not exist.
We do not believe that the cost of handling out-of specification highway diesel batches
will increase significantly as a result of our sulfur program.
E. Misfueling
1. Introduction
As noted in the feasibility discussion of Chapter in, we believe that, in order to comply
with the 2007 and later model year heavy-duty diesel engine emission standards, low sulfur
diesel fuel is needed. For this reason, refiners will be required to begin producing 15 ppm sulfur
highway diesel fuel starting in mid-2006. Once 2007 and later model year heavy-duty vehicles
are sold and operated in the fleet, it will be very important that these vehicles are refueled with
low sulfur fuel to ensure proper operation of the emission control systems. Misfueling a 2007
and later model year heavy-duty vehicle with a fuel that has a sulfur level above 15 ppm could
poison the emission control system and eliminate any emissions benefit from the 2007 standards.
There is the potential for misfueling a 2007 and later heavy-duty vehicle because there are
a number of situations where vehicle owners could have access to diesel fuels with sulfur levels
significantly above 15 ppm. First, hardship provisions allow small refiners to continue producing
and selling as highway fuel, current highway diesel fuel (which can have a sulfur level of up to
IV-110
-------
Chapter IV: Fuel Standard Feasibility
500 ppm) until mid-2010. Second, we are adopting a temporary compliance option that allows
refiners to continue producing current highway diesel during the transition of the program.
Third, even without the temporary compliance option for highway diesel fuel, off highway diesel
fuel (which can have a sulfur level of up to 5,000 ppm) will continue to be available in the
market. Last, United States vehicles operated in Canada or Mexico may have access to fuels that
do not meet the 15 ppm sulfur limit being adopted for the United States.
Misfueling can happen for a number of reasons. A vehicle owner may choose to misfuel
deliberately if they perceive there would be an economic benefit to doing so. For example,
misfueling was a significant problem during the transition from leaded gasoline to unleaded
gasoline in the mid-1970s and 1980s when unleaded gasoline was required to be used in 1975
and later model year cars. On average, leaded gasoline was significantly cheaper than unleaded
gasoline at the retail level and provided a strong incentive for some owners to misfuel with the
wrong fuel. A vehicle owner may also misfuel accidentally, and not even realize they are using
the wrong fuel. This situation could happen currently at those retail outlets which carry both
highway and off highway diesel fuel, and could happen during the transition period to low sulfur
diesel fuel when both 15 ppm and 500 ppm sulfur will be available.
Depending on the level of concern over misfueling, there are a range of options that could
be taken to limit the occurrence of misfueling. Options include simple, low cost programs that
require labels on the fuel pump and labels on the vehicle that specify what fuel should be used in
a vehicle, or color-coding nozzles to alert operators to what fuel is being pumped. More
complicated and higher cost options include setting fuel nozzle size limits and fuel inlet
restrictors, or requiring computer chips on fuel pumps and vehicles that allow the vehicle and
fuel pump to "talk to each other" and ensure that the vehicle is getting the appropriate fuel.
The following section presents the steps being taken in this rule to ensure that 2007 and
later model year heavy-duty diesel vehicles will be fueled with 15 ppm sulfur fuel. We then
present our analysis of whether the steps being taken are sufficient to address concerns over
misfueling. The misfueling discussion is divided into two sections addressing deliberate
misfueling and accidental misfueling separately.
2. What Provision Are We Adopting to Ensure 2007 and Later
Heavy-Duty Diesel Vehicles Use 15 ppm Sulfur Fuel?
As noted above, there are a number of situations where vehicle owners could have access
to diesel fuels with sulfur levels significantly above 15 ppm. In order to ensure that operators of
2007 and later model year heavy-duty diesel vehicles are able to identify the proper fuel needed
in their vehicle when they refuel, we are adopting the following provisions. First, model year
2007 and later heavy-duty diesel vehicles must be equipped with labels on the dashboard and
near the refueling inlet that say: "Use Low Sulfur Diesel Fuel Only." Second, heavy-duty vehicle
IV-111
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
manufacturers must notify each purchaser of a model year 2007 or later diesel-fueled vehicle that
the vehicle must be fueled only with low sulfur diesel fuel. We believe this requirement is
necessary to alert vehicle owners to avoid high sulfur fuel that will be available in this country
and outside the United States as well. Third, all highway diesel fuel pumps and co-located
nonroad diesel fuel pumps must be prominently labeled to identify what type of fuel is dispensed
from the pump.
3. Are Additional Requirements Necessary to Address Deliberate
Misfueling?
A vehicle operator who deliberately misfuels will do so because they expect to realize
some type of benefit from using the wrong fuel in the vehicle. The benefit the operator expects
to realize might be economic (if the required fuel is more expensive than other fuels available) or
it might be a performance benefit (if the operator believes the required fuel is inferior in some
property to the other fuels available). For many heavy-duty diesel vehicles, particularly line-haul
trucks, fuel costs can be as much as 20 percent of annual operating costs, so operators have a
strong incentive to save on fuel costs. Therefore, one factor that would stongly encourage
deliberate misfueling would be if there was a price differential between the 15 ppm sulfur diesel
fuel (required for 2007 and later model year heavy-duty vehicles) and 500 ppm sulfur highway
diesel fuel.
As described in Chapter V, the cost of producing 15 ppm sulfur fuel will be more
expensive than current highway diesel fuel by approximately 4 cents per gallon. However, given
the requirements adopted today, we believe there should not be a large price differential between
the 15 ppm sulfur fuel and the 500 ppm sulfur fuel at retail outlets. Under the credit trading
program, to produce 500 ppm fuel, most refiners will have to purchase credits from other refiners
producing 15 ppm fuel, increasing the cost of the 500 ppm fuel, while decreasing the cost of the
15 ppm fuel. At the refinery gate, the cost of both fuels should be approximately the same. In
addition, given the amount of 15 ppm fuel required under the temporary compliance option, 15
ppm fuel will be distributed through essentially the entire pipeline system. The distribution of
500 ppm fuel, on the other hand, will be more limited, due to its much lower volume. We expect
that the 500 ppm fuel will be distributed by truck in the areas nearby refineries producing this
fuel and through a few major pipelines to a limited number of major fuel consuming areas.
Overall, the better economies of scale of transporting 15 ppm fuel should compensate for any
additional handling cost due to the need to more carefully avoid contamination with high sulfur
fuels. For these reasons, we expect the price to consumers of 500 ppm sulfur fuel to be generally
close to that of 15 ppm sulfur fuel and, therefore, there should not be a significant economic
incentive to misfuel with 500 ppm sulfur fuel. Nevertheless, any price differential could cause
some operators to consider misfueling. Therefore, it is important to examine how price
differential has affected misfueling in past fuel programs.
IV-112
-------
Chapter IV: Fuel Standard Feasibility
The main experience with any significant level of misfueling in the past was during the
unleaded gasoline fuel phase-in that began in the mid-1970s. Throughout the early years of the
unleaded gasoline phase-in, the retail price of unleaded fuel was typically around 7 to 8 cents per
gallon more than leaded gasoline. This price differential represented ten percent or more of the
average retail price of gasoline at the time.cc Primarily because of this significant price
difference, deliberate misfueling of cars with leaded fuel was a significant problem, resulting in
poisoned catalytic converters and a loss of emission benefits. Based on the current retail price
for highway diesel fuel from the Energy Information Agency of approximately $1.60 per gallon,
the small price difference expected under today's program between 15 ppm sulfur and 500 ppm
sulfur fuels is expected to be significantly less (i.e., a price difference of around one to two
percent) than the difference that existed between leaded and unleaded gasoline. With such a
small difference in price between the 15 ppm and 500 ppm sulfur fuels, we do not believe there
will be any significant short-term economic benefit for operators to misfuel 2007 and later heavy-
duty vehicles.
Beyond the lack of an economic incentive, we believe there are several factors that will
likely serve as deterrents to deliberate misfueling. First, the potential risk associated with
voiding any manufacturer emission warranty or damaging the engine and exhaust system on an
expensive vehicle might cause owners and operators of heavy-duty trucks to be more careful in
ensuring that their vehicles are fueled properly. Second, as discussed in Section ni.F. of this
RIA, misfueled vehicles could experience a loss in performance, such as poor acceleration or
engine stalling. Third, under our fuels regulations it is unlawful for any person to fuel a 2007 and
later model year heavy-duty diesel vehicle with any fuel other than low sulfur highway diesel
fuel.
With respect to the likelihood that operators would deliberately misfuel with nonroad
diesel fuel, we do not believe the new fuel requirements will increase this likelihood. Nonroad
diesel fuel is taxed significantly less than highway diesel fuel (approximately 24 cents per gallon
less), so there is already a large price difference between the two fuels. Under the requirements
of the new highway diesel fuel program, the price differential between highway and nonroad
diesel fuels will stay the same or get slightly larger. However, any increase should be relatively
minor and shouldn't result in any large increase in the likelihood of people deliberately
misfueling with nonroad fuel.
The likelihood of deliberate misfueling in Canada is minimal and lessened by the
prospects for eventual harmonization of standards. Canada has recently expressed its intent to
DOE Comments on the NPRM, docket item IV-G-28, Enclosure 1.
IV-113
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
harmonize its fuel regulations with our fuels standards.dd This would offer vehicle owners the
option of refueling with low sulfur fuel there. Even if Canada were to lag behind the U.S. in
mandating low sulfur fuels, there is less potential for U.S. commercial vehicles needing low
sulfur fuel to refuel in Canada because Canadian diesel fuel is currently much more costly than
U.S. fuel due to higher fuel taxes. As a result, most vehicle owners will prefer to purchase fuel
in the U.S., prior to entering Canada, whenever possible. This is facilitated by large tractor-
trailer trucks that have long driving ranges of up to 2,000 miles per tankful and the fact that most
of the Canadian population lives within 100 miles of the United States/Canada border.
Consequently, most U.S. diesel vehicles would not have a critical need to refuel in Canada, and
for those that do, low sulfur fuel would likely become available along major through routes to
serve the needs of U.S. commercial traffic that have the need to purchase it.
With regard to Mexico, the entrance of U.S. trucks beyond the border commercial zone
has been prohibited since before the conclusion of the North American Free Trade Agreement
(NAFTA) in 1994. This prohibition applies in the U.S. as well, as entrance of Mexican trucks
into the U.S. beyond the border commerce zone is also not allowed. Because these prohibitions
are contrary to the intent of the Free Trade Agreement, a timetable was established to eliminate
them.ee However, these prohibitions remain in force at this time. As a result, there is little
opportunity or need for misfueling in Mexico.
The NAFTA negotiations included creation of a "corridor" where commercial truck travel
occurs, and where Mexico is obligated to provide "low sulfur" fuel. At the time of the NAFTA
negotiations, "low sulfur" fuel was considered 500 ppm, which was the level needed to address
the needs of engines meeting the 1994 emission standards. The travel prohibition currently in
place may be lifted at some point. At that time, the issue of assuring, for U.S. vehicles, the
availability of 15 ppm sulfur fuel needed by the 2007 and later heavy-duty vehicles may need to
be addressed.
In summary, for the reasons described above, we do not believe there is cause for concern
over any significant level of deliberate misfueling of 2007 and later heavy-duty vehicles.
Although there is likely to be a limited amount of deliberate misfueling, we believe that people
who are intent on deliberately misfueling will quickly find ways around any requirements
dd "Process Begins to Develop Long Term Agenda to Reduce Air Pollution from Vehicles and Fuels",
Environment Canada press release, May 26, 2000.
ee See NAFTA, Volume II, Annex I, Reservations for Existing Measures and Liberalization Commitments,
Pages I-M-69 and 70, and Pages I-U-19 and 20.
IV-114
-------
Chapter IV: Fuel Standard Feasibility
designed to limit misfueling. For example, based on our experience with unleaded gasolineff,
many car owners physically removed the fuel inlet restrictor from their vehicles (which were
meant to prevent an owner from refueling with the larger sized leaded fuel nozzles) so that they
could refuel with cheaper leaded gasoline. We believe the best approach for minimizing the
level of deliberate misfueling is making sure operators of 2007 and later model year heavy-duty
diesel vehicles are educated about the negative effects on vehicle performance from using diesel
fuel with a sulfur level above 15 ppm.
4. Are Additional Requirements Necessary to Address Accidental
Misfueling?
There is also the possibility that a truck operator may misfuel accidentally, and not realize
they are refueling with the wrong fuel. As noted above, there are a number of reasons a truck
operator may find fuel other than 15 ppm sulfur highway diesel fuel when they pull into a retail
outlet to refuel. First, the temporary compliance option and hardship provisions will allow
refiners to produce two highway diesel fuels during the transition compliance period. Second,
there are a number of retail outlets that carry both highway diesel fuel and nonroad diesel fuel at
the same location.
With regard to the potential for accidental misfueling of 2007 and later heavy-duty
vehicles with 500 ppm sulfur highway diesel fuel during the transition to low sulfur fuel, we
believe the labeling requirements described earlier will lower the potential for accidental
misfueling. The labels should help vehicle operators identify which fuel is required for their
vehicle and help the operator identify the appropriate fuel when they refuel. Although the
possibility exists that an operator would not see the fuel pump label and accidentally misfuel with
500 ppm sulfur fuel, we do not believe this will be a common occurrence. Most retail outlets
(except truck stops) will likely only carry one grade of highway fuel, and because 15 ppm sulfur
fuel is the predominate fuel required even at the start of the program, it will likely be 15 ppm
sulfur fuel. Furthermore, the small refiner option lasts for only four years when the number of
vehicles needing 15 ppm fuel is relatively small but the majority of fuel out there will be 15 ppm
fuel. Last of all, as discussed in Chapter HI, Section A.7., a one time misfueling event with 500
ppm fuel will not necessarily irreversibly destroy the emissions control equipment. For these
reasons, we believe that a labeling program for both vehicles and fuel pumps will satisfactorily
address any concerns over accidental misfueling.
ff" An Analysis of the Factors Leading to the Use of Leaded Gasoline in Automobiles Requiring Unleaded
Gasoline," September 29, 1978, Sobotka & Company, Inc., "Motor Vehicle Tampering Survey - 1983," July 1984,
U.S. EPA, Office of Air and Radiation, and "Anti-Tampering and Anti-Misfueling Programs to Reduce In-Use
Emissions From Motor Vehicles," May 25, 1983 (EPA/AA/83-3). All contained in Docket A-99-06.
IV-115
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
With regard to the potential for accidental misfueling with nonroad diesel fuel, few retail
outlets currently carry both nonroad and highway diesel fuel. Those retail outlets that do also
carry nonroad diesel fuel, the nonroad fuel pump is often located away from the highway fuel
pump. Thus, it would be difficult to accidentally misfuel with nonroad diesel fuel. Therefore,
we do not believe there will be any significant amount of accidental misfueling of 2007 and later
model year heavy-duty diesel vehicles with nonroad diesel fuel. However, as noted earlier, we
are requiring that all nonroad fuel pumps at retail outlets carrying both nonroad diesel fuel and
highway diesel fuel be labeled. We believe the label requirements are sufficient to address
concerns over the potential for misfueling with nonroad diesel fuel.
In summary, for the reasons noted above, we believe that the simple labeling
requirements being adopted will help vehicle owners identify and use the correct fuel and will be
sufficient to address the level of concern regarding accidental misfueling.
IV-116
-------
Chapter IV: Fuel Standard Feasibility
Chapter IV. References
1. Baseline Submissions for the Reformulated Gasoline Program.
2. Swain, Edward J., Gravity, Sulfur Content of U.S. Crude Slate Holding Steady, Oil and
Gas Journal, January 13, 1997.
3. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
4. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
5. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
6. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
7. Title 13 of the California Code of Regulations, § 2281, "Sulfur Content of Diesel Fuel."
8. American Society for Testing and Materials (ASTM ), "Standard Specification for Diesel
Fuel Oils", ASTMD 975.
9. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
10. American Society for Testing and Materials (ASTM), "Standard Specification for Diesel
Fuel Oils", ASTM D 975. Some pipeline companies that transport diesel fuel have limits
for density and pour point, which are properties that ASTM D 975 does not provide
specifications on.
11. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
12. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
13. Final Report, 1996 American Petroleum Institute / National Petroleum Refiners
Association, Survey of Refining Operations and Product Quality, July 1997.
IV-117
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
14. Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles, Tier 2
Motor Vehicle Emission Standards and Gasoline Sulfur Control Requirements,
Environmental Protection Agency, December 1999.
15. Hamilton, Gary L., ABB Lummus, Letter to Lester Wyborny, U.S. EPA, August 2, 1999.
16. Mayo, S.W., "Mid-Distillate Hydrotreating: The Perils and Pitfalls of Processing LCO,"
17. Peries, J-P., Jeanlouis, P-E, Schmidt, M, and Vance, P.W., "Combining NiMo and CoMo
Catalysts for Diesel Hydrotreaters," NPRA 1999 Annual Meeting, Paper 99-51, March
21-23, 1999.
18. Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.
19. Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.
20. Tungate, F.L., Hopkins, D., Huang, D.C., Fletcher, J.C.Q., and E. Kohler, "Advanced
distillate Hydroprocessing, ASAT, A Trifunctional HDAr/HDS/HDN Catalyst," NPRA
1999 Annual Meeting, Paper AM-99-38., March 21-23, 1999.
21. Gerritsen, L. A., Production of Green Diesel in the BP Amoco Refineries, Presentation by
Akzo Nobel at the WEFA conference in Berlin, Germany, June 2000.
22. Gerritsen, L.A., Sonnemans, J.W M, Lee, S.L., and Kimbara, M., "Options to Met Future
European Diesel Demand and Specifications,"
23. UOP, Paper to be presented at the 98th NPRA Annual Meeting, March 27-28, 2000, San
Antonio, Texas.
24. Centinel Hydroprocessing Catalysts: A New Generation of Catalysts for High-Quality
Fuels, Criterion Catalysts and Technologies Company, October 2000.
25. Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.
26. Peries, J-P., Jeanlouis, P-E, Schmidt, M, and Vance, P.W., "Combining NiMo and
CoMo Catalysts for Diesel Hydrotreaters," NPRA 1999 Annual Meeting, March 21-23,
1999.
27. Wilson, R., "Cost Curves for Conventional HDS to Very Low Levels," February 2, 1999.
IV-118
-------
Chapter IV: Fuel Standard Feasibility
28. Moncrieff, T. Ian, Montgomery, W. David, Ross, Martin T., Charles River Associates
Inc., Ory, Raymond E., Carney, Jack T., Baker and O'Brien Inc., An Assessment of the
Potential Impacts of Proposed Environmental Regulations on U.S. Refinery Supply of
Diesel Fuel, A study prepared by Charles River and Associates Inc. and Baker and
O'Brien Inc. for the American Petroleum Association, August 2000.
29. U. S. Petroleum Refining, Assuring the Adequacy and Affordability of Cleaner Fuels, A
Report by the National Petroleum Council, June 2000.
30. "Processes for Sulfur Management," IFF.
31. Tungate, F.L., Hopkins, D., Huang, D.C., Fletcher, J.C.Q., and E. Kohler, "Advanced
distillate Hydroprocessing, ASAT, A Trifunctional HDAr/HDS/HDN Catalyst," NPRA
1999 Annual Meeting, Paper AM-99-38., March 21-23, 1999.
32. Gerritsen, L. A., Production of Green Diesel in the BP Amoco Refineries, Presentation by
Akzo Nobel at the WEFA conference in Berlin, Germany, June 2000.
33. American Society for Testing and Materials (ASTM ), "Standard Specification for Diesel
Fuel Oils", ASTMD 975.
34. "Diesel Fuel Oils, 1996", October, 1996, Cheryl Dickson, and Gene Sturm, Jr., National
Institute for Petroleum and Energy Research, Bartlesville, Oklahoma, NIPER-197 PPS,
96/5.
35. Gerritsen, L.A., Sonnemans, J.W M, Lee, S.L., and Kimbara, M., "Options to Met Future
European Diesel Deman and Specifications,"
36. Mayo, S.W., "Mid-Distiallate Hydrotreating: The perils and Pitfalls of Processing LCO,"
Akzo Nobel Catalysts.
37. Chapados, Doug, Desulfurization by Selective Oxidation and Extraction of Sulfur-
Containing Compounds to Economically Achieve Ultra-Low Proposed Diesel Fuel Sulfur
Requirements, Paper presented at the 2000 NPRA Annual Meeting.
38. Kidd, Dennis, S-Zorb - Advances in Applications of Phillips S-Zorb Technology,
Presented at the NPRA Q & A meeting, October 2000.
39. U. S. Petroleum Refining, Assuring the Adequacy and Affordability of Cleaner Fuels, A
Report by the National Petroleum Council, June 2000.
40. Much of the discussion in Section IV.C. on lubricity was obtained from Society of
Automobile Engineers (SAE) Technical Paper No. 982567: Fuel Lubricity Reviewed, P.
IV-119
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
I. Lacey and S. A. Howell; and SAE 1999-01-1479: Diesel Fuel Lubricity: On the Path to
Specifications, Manuch Nikanjam.
41. From Internet site of Highspeed Energy Care: highspeed.co.uk.
42. Society of Automobile Engineers (SAE) Technical Paper No. 982568: Effects of Water
on Distillate Fuel Lubricity.
43. Society of Automobile Engineers (SAE) Technical Paper No. 942014: Diesel Fuel
Lubricity Additive Study, M. Nikanjam.
44. Department of Defense comments, docket no. IV-D-298.
45. Society of Automobile Engineers (SAE) Technical Paper No. 981362: Lubricity of
California Diesel Fuel, F. Cameron.
46. Society of Automobile Engineers (SAE) Technical Paper No. 981362: Lubricity of
California Diesel Fuel, F. Cameron.
47. Society of Automobile Engineers (SAE) Technical Paper No. 981362: Lubricity of
California Diesel Fuel, F. Cameron.
48. The conclusion that Sweden's fuel has resulted in acceptable durability is based on the
two primary references: SAE Technical Paper No. 982567: Fuel Lubricity Reviewed, P. I.
Lacey and S. A. Howell; and SAE 1999-01-1479: Diesel Fuel Lubricity: On the Path to
Specifications, Manuch Nikanjam; and also from a letter from Lennart Erlandsson, MTC
to Michael P. Walsh dated 10/16/00, docket no. IV-G-42.
49. Society of Automobile Engineers (SAE) Technical Paper No. 982570: Low Sulfur Diesel
Field Test Study in Thailand, P. Siangsanorh and P. Boonchanta.
50. Letter from MTC to Michael P. Walsh, dated 10/16/00, docket no. IV-G-42.
51. Society of Automobile Engineers (SAE) Technical Paper No. 982571: The No-Harm
Performance of Lubricity Additives for Low Sulfur Diesel Fuels, E. Mozdzen, S. Wall,
and W. Byfleet.
52. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
53. Internal Revenue Service Publication 510, Fuel Excise Taxes for 1999, Internet page
http://www.irs. ustreas.gov/prod/forms_pubs/pubs/p51005. htm.
IV-120
-------
Chapter IV: Fuel Standard Feasibility
54. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
5 5. Comments of the Department of Energy (DOE) on the NPRM, docket item IV-G-28.
56. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
57. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
5 8. Comments of the Department of Energy (DOE) on the NPRM, docket item IV-G-28.
59. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
60. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
61. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
62. National Petroleum Council's (NPC) report on U.S. Petroleum Refining, attachment #6 to
the comments of the American Petroleum Institute (API) on the NPRM, docket item IV-
D-343.
63. Comments of the Association of Oil Pipelines (AOPL) on the NPRM, docket item IV-D-
325.
64. Comments of the Independent Fuel Terminal Operators Association (IFTOA) on the
NPRM, docket item IV-D-217.
65. Comments of the Independent Fuel Terminal Operators Association (IFTOA) on the
NPRM, docket item IV-D-217.
66. Comments of the American Petroleum Institute (API) on the NPRM, docket item IV-D-
343.
67. Comments of the American Trucking Association (ATA) on the NPRM, docket item
IV-D-269.
68. Comments of the Petroleum Marketers Association of America (PMAA) on the Advance
Notice of Proposed Rulemaking (ANPRM), docket item II-D-73.
IV-121
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
69. Draft Regulatory Impact Analysis (DRIA), docket item ffl-B-01.
70. Comments of the Independent Fuel Terminal Operators Association (IFTOA) on the
NPRM, docket item IV-D-217.
71. Comments of the American Chemistry Council on the NPRM, docket item IV-D-183.
IV-122
-------
Chapter V: Economic Impact
Chapter V: Economic Impact
A. Economic Impact of the 2007 Model Year Heavy-Duty
Diesel Standards
This section contains an analysis of the economic impacts of the emission standards for
heavy-duty diesel vehicles. First, a brief outline of the methodology used to estimate the
economic impacts is presented, followed by a summary of the technology packages that are
expected to be used to meet the standards. Next, the projected costs of the individual
technologies are presented, along with a discussion of fixed costs such as research and
development (R&D), tooling and certification. Following the discussion of the individual cost
components is a summary of the projected per-vehicle cost of the regulations. Finally, an
analysis of the aggregate cost for the new engine technologies is presented. Unless noted
otherwise all costs presented here are in 1999 dollars.
1. Methodology for Estimating Costs
While the following analysis is based on a relatively uniform emission control strategy for
designing the different categories of engines, this is not intended to suggest that a single
combination of technologies will actually be used by all manufacturers. In fact, depending on
basic engine emission characteristics, EPA expects that emission control technology packages
will gradually be fine-tuned to each application. Furthermore, EPA expects manufacturers to use
averaging, banking, and trading programs as a means to deploy varying degrees of emission
control technologies on different engines. EPA nevertheless believes that the projections
presented here provide a cost estimate representative of the different approaches manufacturers
are likely to take.
Because many of the technologies which we believe will be used by the industry in order
to meet the standards are being applied on a large scale for the first time, we have sought input
from a large section of the regulated community, seeking their estimation of the future costs to
apply these technologies. Under contract from EPA, ICF Consulting provided surveys to nine
engine manufacturers seeking their input on expectations for cost savings which might be
enabled through the use of low sulfur diesel fuel and seeking their estimations of the cost and
types of emission control technologies which might be applied with low sulfur diesel fuel. Based
on responses to these surveys, EPA estimated cost savings to the current and future fleets. The
survey responses were also used as the first step in estimating the costs for advanced emission
control technologies which may be applied in order to meet the 2007 heavy-duty vehicle
V-l
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
standards.1 These costs were then further refined by EPA based upon input from members of the
Manufacturers of Emission Control Association.
Projected heavy-duty vehicle sale estimates are used in several portions of this analysis.
Based on data submitted by engine manufacturers, we estimated 1995 engine sales to be 280,000
for light heavy-duty engines, 140,000 for medium heavy-duty engines, and 220,000 for heavy
heavy-duty engines (including those sold into urban bus applications). These numbers are
projected to grow at an annual rate of two percent of the base year without compounding through
2035 in this analysis and are included in table V.A-20.2
Costs of control include variable costs (for incremental hardware costs, assembly costs,
and associated markups) and fixed costs (for tooling, R&D, and certification). For technologies
sold by a supplier to the engine manufacturers, costs are either estimated based upon a direct cost
to manufacture the system components plus a 29 percent markup to account for the supplier's
overhead and profit, or when available, based upon estimates from suppliers on expected total
costs to the manufacturers (inclusive of markups).3 Estimated variable costs for new
technologies include a markup to account for increased warranty costs. Variable costs are
additionally marked up to account for both manufacturer and dealer overhead and carrying costs.
The manufacturer's carrying cost was estimated to be four percent of the direct costs accounting
for the capital cost of the extra inventory, and the incremental costs of insurance, handling, and
storage. The dealer's carrying cost was marked up three percent reflecting the cost of capital tied
up in inventory. This approach to individually estimating manufacturer and dealer markups, to
better reflect the value added at each stage of the cycle, was adopted by EPA based upon industry
input.4
EPA has also identified various factors that will cause cost impacts to decrease over time,
making it appropriate to distinguish between near-term and long term costs. Research in the
costs of manufacturing has consistently shown that as manufacturers gain experience in
production, they are able to apply innovations to simplify machining and assembly operations,
use lower cost materials, and reduce the number or complexity of component parts.5 The
analysis incorporates the effects of this learning curve as described in section A.6 of this chapter.
Finally, manufacturers are expected to apply ongoing research to make emission controls more
effective and to have lower operating costs over time.
Fixed costs for R&D are assumed to be incurred over the five-year period preceding
introduction of the engine, tooling and certification costs are assumed to be incurred one year
ahead of initial production. Fixed costs are increased by seven percent for every year before the
start of production to reflect the time value of money, and are then recovered with a five-year
amortization at the same rate. The analysis also includes consideration of lifetime operating
costs where applicable. Projected costs were derived for the four service classes of heavy-duty
diesel vehicles listed in Table V.A-1. The cost for each technology applied to urban buses is the
V-2
-------
Chapter V: Economic Impact
same as the cost of that technology when applied to heavy heavy-duty vehicles, unless specified
otherwise.
Table V.A-1. Service Classes of Heavy-Duty Vehicles
Service Class
Light
Medium
Heavy
Urban Bus
Vehicle Class
2B-5
6-7
8
—
GVWR (Ibs.)
8,500- 19,500
19,501 -33,000
33,001 +
—
2. Heavy-Duty Diesel Technologies for Compliance with the Standards
Several new technologies are projected for complying with the 2007 model year emission
standards. We are projecting that NOx adsorbers and catalyzed diesel particulate filters will be
the most likely technologies applied by the industry in order to meet our emissions standards.
We also anticipate the introduction of closed crankcase filtration systems for turbocharged
heavy-duty diesel engines due to the elimination of the current exception granted to these
engines. The fact that manufacturers have several years before implementation of the new
standards ensures that the technologies used to comply with the standards will develop
significantly before reaching production. This ongoing development will lead to reduced costs in
three ways. First, research will lead to enhanced effectiveness for individual technologies,
allowing manufacturers to use simpler packages of emission control technologies than we would
predict given the current state of development. Similarly, the continuing effort to improve the
emission control technologies will include innovations that allow lower-cost production. Finally,
manufacturers will focus research efforts on any drawbacks, such as fuel economy impacts or
maintenance costs, in an effort to minimize or overcome any potential negative effects.
We anticipate a combination of primary technology upgrades for the 2007 model year.
Achieving very low NOx emissions will require basic research on NOx emission control
technologies and improvements in engine management to take advantage of the aftertreatment
system capabilities. The manufacturers are expected to take a systems approach to the problem
optimizing the engine and aftertreatment system to realize the best overall performance possible.
Since most research to date with aftertreatment technologies has focused on retrofit programs
there remains room for significant improvements by taking such a systems approach. We have
estimated that the catalyst companies will spend approximately $220 million to further develop
the NOx and PM/HC control technologies described here. Further we have estimated that the
V-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
engine manufacturers will spend approximately $385 million dollars on R&D to develop the
control systems needed to take advantage of the advanced emission control technologies
described here. The NOx adsorber technology in particular is expected to benefit from re-
optimization of the engine management system to better match the NOx adsorber performance
characteristics. The majority of the $385 million dollars we estimated for engine research is
expected to be spent on developing this synergy between the engine and NOx aftertreatment
systems. PM/HC control technologies are expected to be less sensitive to engine operating
conditions as they have already shown good robustness in retrofit applications with low-sulfur
diesel fuel. Nevertheless the manufacturers are expected to take a global systems approach that
will optimize operation with consideration to both NOx and PM/HC emission control
subsystems.
EPA contracted with ICF Consulting to 1) Estimate the variable cost for advanced
emission control technologies which would be enabled by low sulfur diesel fuel, and 2) Estimate
the impacts of low sulfur diesel fuel for engine durability and maintenance costs. Task 1 was
completed by Engine, Fuel and Emissions Engineering and is referenced here as "Economic
Analysis of Diesel Aftertreatment System Changes Made Possible By Reduction of Diesel Fuel
Sulfur Content, Task 1," or as the EF&EE cost report. Task 2 was completed by ICF Consulting
and is referenced here as "Economic Analysis of Vehicle and Engine Changes Made Possible by
the Reduction of Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced
Maintenance," or as the ICF low sulfur benefits report.
The results of our cost analysis are considered in the following paragraphs and
summarized in Table V.A-2. Technology costs are described in section 3, fixed costs are
described in section 4, and maintenance cost savings are described in section 5.
V-4
-------
Chapter V: Economic Impact
Table V.A-2. Summary of Near and Long Term Cost Estimates
(net present value in year of sale)
Near Term (2007) Light Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber System
Catalyzed Diesel Participate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$87
$41
$0
$0
$0
$0
$128
Variable Cost
$925
$690
$206
$37
$0
$0
$1,858
Operating Cost
$0
$55
$0
$31
$576
($153)
$509
Long Term (2012+) Light Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Particulate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$0
$0
$0
$0
$0
$0
$0
Variable Cost
$592
$425
$132
$23
$0
$0
$1,172
Operating Cost
$0
$55
$0
$26
$609
($153)
$537
V-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Near Term (2007) Medium Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Participate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$231
$98
$0
$0
$0
$0
$329
Variable Cost
$1,080
$852
$261
$42
$0
$0
$2,235
Operating Cost
$0
$56
$0
$59
$1,077
($249)
$943
Long Term (2012+) Medium Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Particulate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$0
$0
$0
$0
$0
$0
$0
Variable Cost
$691
$527
$167
$27
$0
$0
$1,412
Operating Cost
$0
$56
$0
$48
$1,141
($249)
$996
V-6
-------
Chapter V: Economic Impact
Near Term (2007) Heavy Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Participate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$191
$89
$0
$0
$0
$0
$280
Variable Cost
$1,456
$1,103
$338
$49
$0
$0
$2,946
Operating Cost
$0
$208
$0
$218
$3,969
($610)
$3,785
Long Term (2012+) Heavy Heavy-Duty Diesel Vehicles
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Particulate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Maintenance Savings
Total
Fixed
Cost
$0
$0
$0
$0
$0
$0
$0
Variable Cost
$932
$686
$216
$32
$0
$0
$1,866
Operating Cost
$0
$208
$0
$172
$4,209
($610)
$3,979
V-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Near Term (2007) Urban Buses
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Participate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Current Oxidation Catalyst Removed
Maintenance Savings
Total
Fixed
Cost
$191
$89
$0
$0
$0
$0
$0
$280
Variable Cost
$1,456
$1,103
$338
$49
$0
($338)
$0
$2,608
Operating Cost
$0
$98
$0
$107
$4,772
$0
($352)
$4,625
Long Term (2012+) Urban Buses
(1999 Dollars per Engine)
Item
NOx Adsorber Catalyst
Catalyzed Diesel Particulate Filter
HC and H2S Clean Up Catalyst
Closed Crankcase System
Low Sulfur Diesel Fuel
Current Oxidation Catalyst Removed
Maintenance Savings
Total
Fixed
Cost
$0
$0
$0
$0
$0
$0
$0
$0
Variable Cost
$932
$686
$216
$32
$0
($216)
$0
$1,650
Operating Cost
$0
$98
$0
$92
$4,959
$0
($352)
$4,797
V-8
-------
Chapter V: Economic Impact
3. Technology/Hardware Costs for Diesel Vehicles and Engines
The following discussion presents the projected costs of the primary technological
improvements expected for complying with the emission standards detailing the variable costs of
the individual technologies. EPA believes that a small set of technologies integrated into a single
emission control system will represent the primary changes manufacturers must make to meet the
2007 model year standards. This integrated system is expected to include elements which could
be individually identified as a NOx adsorber catalyst, a catalyzed diesel particulate filter, a diesel
oxidation catalyst, and 15 ppm sulfur diesel fuel to enable the aforementioned emission control
technologies. In order to comply with the requirement to eliminate crankcase emissions from all
heavy-duty diesel engines, we are projecting the introduction of closed crankcase filtration
systems. Lean NOx catalysts and compact SCR systems were not considered in this analysis, not
because the control they offer is an incidental benefit, but because it appears unlikely that they
will be part of 2007 model year technology packages.
a. NOx Adsorber Catalyst Costs
NOx adsorber catalysts have been developed and are being applied today for stationary
power NOx emission control and for lean burn gasoline engine control. The application of this
catalyst technology to diesel engines is relatively new. Therefore we have projected that there
will be significant enhancements of the technology in order to better match the characteristics of
diesel engines. Nevertheless the basic components of the NOx adsorber catalyst are well known
and include, 1) an oxidation catalyst, typically platinum, 2) an alkaline earth metal to store NOx,
typically barium, 3) a NOx reduction catalyst, typically rhodium, and 4) a substrate and can to
hold and support the catalyst washcoat. Cost estimates for the NOx adsorber catalysts in 2007 are
presented in Table V. A-3 below.
The material costs listed in Table V.A-3 represent costs to the engine manufacturers
inclusive of supplier markups. The total direct cost to the manufacturer includes an estimate of
warranty costs for the NOx adsorber system. Hardware costs are additionally marked up to
account for both manufacturer and dealer overhead and carrying costs. The manufacturer's
carrying cost was estimated to be four percent of the direct costs accounting for the capital cost of
the extra inventory, and the incremental costs of insurance, handling, and storage. The dealer's
carrying cost was marked up three percent reflecting the cost of capital tied up in inventory. This
approach to individually estimating manufacturer and dealer markups, to better reflect the value
added at each stage of the cycle, was adopted by EPA based upon industry input.6
We have estimated the cost of this system based on information from the following
reports:
V-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1. Estimated Economic Impact of New Emission Standards for Heavy-Duty On-
Highway Engines, March 1997, EPA 420-R-97-009.
2. Cost Estimates for Heavy-Duty Gasoline Vehicles, September 1998, EPA Air
Docket A-99-06 Item No. II-A-13.
3. Economic Analysis of Diesel Aftertreatment System Changes Made Possible By
Reduction of Diesel Fuel Sulfur Content, December 1999, Air Docket A-99-06.
The individual assumptions used to estimate the cost for the system are documented in the
following subsections.
Catalyst Volume
The Engine Manufacturers Association was asked as part of a contractor work assignment
to gather input from their members on likely technology solutions including the NOx adsorber
catalyst.7 The respondents indicated that the catalyst volume for a NOx adsorber catalyst could
range from 1.5 times the engine displacement to as much as 2.5 times the engine displacement
based on today's washcoating technology. Based on current lean burn gasoline catalyst designs
and engineering judgement we have estimated that the NOx adsorber catalyst will be sized on
average 1.5 times the engine displacement.
Substrate Cost
The ceramic flow through substrates used for the NOx adsorber catalyst are estimated to
cost approximately $5 per liter. This cost estimate is based upon the relationship developed for
current heavy-duty gasoline catalyst substrates as documented in Cost Estimates for Heavy-Duty
Gasoline Vehicles of
C= $4.67 xV+ $1.50
where:
C = cost to the vehicle manufacturer from the substrate supplier
V = substrate volume in liters.
Washcoating and Canning
The report entitled, "Economic Analysis of Diesel Aftertreatment System Changes Made
Possible By Reduction of Diesel Fuel Sulfur Content," estimates a "value-added" engineering
and material product, called washcoating and canning, based on feedback from members of the
Manufacturers of Emission Control Association (MECA). By using a value added component
that accounts for fixed costs (including R&D), overhead, marketing and profits from likely
suppliers of the technology, we can estimate this fraction of the cost for the technology apart
V-10
-------
Chapter V: Economic Impact
from the other components which are typically more widely available as commodities (e.g,
precious metals and catalyst substrates). Here, we have taken the washcoating and canning costs
estimated in the above mentioned report and have split out 11 percent of that cost for R&D, with
the remaining 89 percent continuing to be called washcoat and canning. The R&D fraction is
then used to estimate a total R&D expenditure for the industry due to the 2007 HD rule of $133
million recovered over the first five years of the program. We arrived at a value of 11 percent for
R&D by looking at R&D costs as a fraction of gross profits from the annual report of one of the
larger catalyst manufacturers.
Precious Metals
The total precious metal content for the NOx adsorber is estimated to be 50 g/ft3 with
platinum representing 90% of that total and Rhodium 10%. The costs for rhodium and platinum
are the same as estimated in the Tier 2 RIA (EPA420-99-023) and are $868/troy oz. for rhodium
and $412 / troy oz. for platinum.
Barium
The cost for barium carbonate (the primary NOx storage material) is assumed to be less
than $1 per catalyst as estimated in "Economic Analysis of Diesel Aftertreatment System
Changes Made Possible By Reduction of Diesel Fuel Sulfur Content."
Can Housing
The material cost for the can housing is estimated based on the catalyst volume plus 20%
for transition cones, plus 20% for scrappage (material purchased but unused in the final product)
and a price of $.98/lb for 16 gauge stainless steel as estimated in contractor report "Economic
Analysis of Diesel Aftertreatment System Changes Made Possible By Reduction of Diesel Fuel
Sulfur Content." The resulting material costs are summarized in the table below.
NOx Regeneration System
The NOx regeneration system is likely to include a NOx/O2 sensor, a means for exhaust
air to fuel ratio control (one or more exhaust fuel injectors or in-cylinder means), a temperature
sensor and possibly a means to control mass flow through a portion of the catalyst system (a
"dual-bed" system). The cost for such a system is $300 for light and medium heavy-duty
vehicles and $350 for heavy heavy-duty vehicles as estimated in contractor report "Economic
Analysis of Diesel Aftertreatment System Changes Made Possible By Reduction of Diesel Fuel
Sulfur Content."
V-ll
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Direct Labor Costs
The direct labor costs for the catalyst are estimated based upon an estimate of the number
of hours required for assembly and established labor rates. Additional overhead for labor was
estimated as 40 percent of the labor rate.8
Warranty Costs
We have estimated the warranty costs based upon a 1% claim rate, and an estimate of
parts and labor costs per incident. The labor rate is assumed to be $50 per hour, and a parts cost
are estimated as 2.5 times the OEM component cost. These costs are summarized in the NOx
absorber summary table below.
Manufacturer and Dealer Carrying Costs
The manufacturer's carrying cost was estimated at 4% of the direct costs. This reflects
primarily the costs of capital tied up in extra inventory, and secondarily the incremental costs of
insurance, handling and storage. The dealer's carrying cost was estimated at 3% of the
incremental cost, again reflecting primarily the cost of capital tied up in extra inventory.
V-12
-------
Chapter V: Economic Impact
Summary - Total System Estimate
Table V.A-3. 2007 NOx Adsorber Cost Estimate
NOx Adsorber Catalyst
Catalyst Volume
Material Cost
Substrate
Washcoat (value added engineering)
Platinum
Rhodium
Alkaline Earth Oxide
Can Housing
NOx Regeneration System
Direct Labor Costs
Total Direct Cost to Mfr.
Warranty Costs (1% Claim Rate)
Mfr. Carrying Cost
Total Cost to Dealer
Dealer Carrying Cost
Total Cost to Customer
Vehicle Class
LHDD
9
MHDD
12
HHDD
20
$47
$223
$189
$44
$1
$9
$300
$37
$851
$22
$26
$899
$27
$925
$63
$267
$253
$59
$1
$13
$300
$37
$992
$26
$30
$1,048
$31
$1,080
$103
$312
$411
$96
$1
$17
$350
$49
$1,339
$34
$40
$1,413
$42
$1,456
b. Catalyzed Diesel Particulate Filter Costs
Catalyzed diesel particulate filters are already in limited production for retrofits in
markets were low sulfur diesel fuel is available. The final design configurations and catalyst
compositions that these technologies are likely to have in 2007 can be estimated with some
accuracy. Based on current systems and input from industry, costs for catalyzed diesel
particulate filters in 2007 were estimated and are presented in Table V.A-4 below. These cost
are reduced here by $45 for light heavy-duty vehicles, $50 for medium heavy-duty vehicles and
$55 for heavy heavy-duty vehicles to reflect the fact that diesel particulate filters also serve the
function of a muffler, eliminating the need for that device.
Material costs for the catalyzed diesel particulate filter given here are inclusive of
supplier markups as they reflect the expected cost to the engine manufacturer to purchase the
V-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
hardware from a supplier. The total direct cost to the manufacturer includes an estimate of
warranty costs for the catalyzed diesel particulate filter. Hardware costs are additionally marked
up to account for both manufacturer and dealer overhead and carrying costs. The manufacturer's
carrying cost was estimated to be four percent of the direct costs accounting for the capital cost of
the extra inventory, and the incremental costs of insurance, handling, and storage. The dealer's
carrying cost gives a three percent markup reflecting the cost of capital tied up in inventory. This
approach to individually estimating manufacturer and dealer markups, to better reflect the value
added at each stage of the cycle, was adopted by EPA based upon industry input.9
Diesel Particulate Filter Volume
The Engine Manufacturers Association was asked as part of a contractor work assignment
to gather input from their members on catalyzed diesel particulate filters for heavy-duty
applications.10 The respondents indicated that the particulate filter volume could range from 1.5
times the engine displacement to as much as 2.5 times the engine displacement based on today's
experiences with cordierite filter technologies. The size of the diesel particulate filter is selected
largely based upon the maximum allowable flow restriction for the engine. Generically the filter
size is inversely proportional to its resistance to flow (a larger filter is less restrictive than an
similar smaller filter). We have estimated that the diesel particulate filter will be sized to be 1.5
times the engine displacement in 2007 based on these responses and on-going research aimed at
improving filter porosity control to give a better trade-off between flow restrictions and filtering
efficiency.
Diesel Particulate Filter Costs
Cost estimates for cordierite diesel particulate filters (the most common type used today)
were provided by several members of the Manufacturers of Emission Control Association
(MECA) for each vehicle class. The cost estimates showed a non-linear relationship with
particulate filter size with larger filters being somewhat less expensive per liter of filter volume.
Here we have used an average of the MECA provided cost estimates for each of the classes to
arrive at our cost estimate.a
a MECA member companies provided estimates of future cordierite filter costs to EPA's contractor
EF&EE. EF&EE estimated the cost of future filters with a linear fit to the estimates provided. In this analysis, we
have estimated the future cost of the cordierite filters by averaging the MECA member estimates for each vehicle
class, rather than using the contractor's linear fit estimate. We used this alternate approach for estimating the cost
of the cordierite filter due to the non-linear nature of the cost estimates provided by MECA. This change from the
contractor's estimate increases the cost for light heavy-duty vehicles while decreasing the cost for heavy heavy-duty
vehicles due to the non-linear nature of the cost estimates. The MECA estimates were identified as Confidential
Business Information when provided to EF&EE and are therefore not provided in the docket associated with this
RIA.
V-14
-------
Chapter V: Economic Impact
Washcoating and Canning
Washcoating and canning costs are estimated and accrued in the same manner as for the
NOx adsorber technology discussed above. The resulting variable costs for washcoating and
canning are $134 for light heavy-duty DPFs, $178 for medium heavy-duty DPFs, and $223 for
heavy heavy-duty DPFs. Per filter R&D costs were estimated in the same manner as described
above for the NOx adsorber catalyst and are estimated to be $16, $22, and $27 for diesel
particulate filters applied to light, medium and heavy heavy-duty vehicles respectively.
Aggregating these R&D costs over the projected engine volumes during the first five years of the
program allows us to estimate the total R&D expense for catalyzed diesel particulate filters as
$87 million.
Precious Metals
The total precious metal content for catalyzed diesel particulate filters is estimated to be
30 g/ft3 with platinum as the only precious metal used in the filter. The cost for platinum is the
same as estimated in the Tier 2 RIA (EPA420-99-023) and is $412/troy ounce.
Can Housing
The material cost for the can housing is estimated based on the filter volume plus 20% for
transition cones, plus 20% for scrappage and a price of $.98/lb for 16 gauge stainless steel as
estimated in contractor report "Economic Analysis of Diesel Aftertreatment System Changes
Made Possible By Reduction of Diesel Fuel Sulfur Content." The resulting material costs are
summarized in the table below.
Differential Pressure Sensor
We have assumed that the catalyzed diesel particulate filter system will require the use of
a differential pressure sensor to provide a diagnostic monitoring function of the filter. A cost of
$45 per sensor has been assumed as estimated in contractor report "Economic Analysis of Diesel
Aftertreatment System Changes Made Possible By Reduction of Diesel Fuel Sulfur Content."
Direct Labor Costs
The direct labor costs for the catalyzed diesel particulate filter are estimated in contractor
report "Economic Analysis of Diesel Aftertreatment System Changes Made Possible By
Reduction of Diesel Fuel Sulfur Content" based upon an estimate of the number of hours
required for assembly and established labor rates.
V-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Warranty Costs
We have estimated the warranty costs based upon a 1% claim rate, and an estimate of
parts and labor costs per incident. The labor rate is assumed to be $50 per hour, and a parts cost
are estimated as 2.5 times the OEM component cost. These costs are summarized in the
catalyzed diesel particulate filter summary table below.
Manufacturer and Dealer Carrying Costs
The manufacturer's carrying cost was estimated at 4% of the direct costs. This reflects
primarily the costs of capital tied up in extra inventory, and secondarily the incremental costs of
insurance, handling and storage. The dealer's carrying cost was estimated at 3% of the
incremental cost, again reflecting primarily the cost of capital tied up in extra inventory.
Muffler Costs
The diesel particulate filter costs are reduced here by $45 for light heavy-duty vehicles,
$50 for medium heavy-duty vehicles and $55 for heavy heavy-duty vehicles to reflect the fact
that diesel particulate filters also serve the function of a muffler, eliminating the need for that
device.
V-16
-------
Chapter V: Economic Impact
Summary - Total System Estimate
Table V.A-4. 2007 Catalyzed Diesel Particulate Filter Cost Estimate
Catalyzed Diesel Particulate Filter
Trap Volume (liters)
Material Cost
Filter Trap
Washcoat (value added engineering)
Platinum
Can Housing
Differential Pressure Sensor
Direct Labor Costs
Total Direct Cost to Mfir.
Warranty Costs (1% Claim Rate)
Mfir. Carrying Cost
Total Cost to Dealer
Dealer Carrying Cost
Savings by removing muffler
Total Cost to Customer
Vehicle Class
LHDD
9
MHDD
12
HHDD
20
$300
$134
$126
$7
$45
$49
$670
$16
$27
$713
$21
($45)
$690
$360
$178
$168
$10
$45
$49
$822
$20
$33
$875
$26
($50)
$851
$420
$223
$274
$14
$45
$62
$1,056
$25
$42
$1,124
$34
($55)
$1,103
c. Diesel Oxidation Catalyst (HC and H2S "Clean-Up" Catalyst)
The NOx adsorber regeneration and desulfation functions may produce undesirable by-
products in the form of momentary increases in HC emissions or in odorous hydrogen sulfide
(H2S) emissions. In order to control these potential products we have assumed that
manufacturers may choose to apply a diesel oxidation catalyst (DOC) downstream of the NOx
adsorber technology. The DOC would serve a "clean-up" function to oxidize any HC and H2S
emissions to more desirable products as outlined in Chapter 3.
We have estimated the cost of diesel oxidation catalysts below in Table V. A-5 as $206
for a light heavy-duty diesel vehicle, $261 for a medium heavy-duty diesel vehicle and $338 for a
heavy heavy-duty diesel vehicle. The individual component costs for the DOC were estimated in
the same manner as for the NOx adsorber and CDPF above.
V-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.A-5. 2007 Diesel Oxidation Catalyst Cost Estimate
Catalyzed Diesel Paniculate Filter
Catalyst Volume (liters)
Material Cost
Substrate
Washcoat (value added engineering)
Platinum (5 g/ft3)
Can Housing
Direct Labor Costs
Total Direct Cost to Mfr.
Warranty Costs (1% Claim Rate)
Mfr. Carrying Cost
Total Cost to Dealer
Dealer Carrying Cost
Total Cost to Customer
Vehicle Class
LHDD
6
MHDD
8
HHDD
13
$32
$125
$14
$4
$13
$187
$5
$7
$200
$6
$206
$42
$150
$19
$6
$13
$237
$6
$9
$253
$8
$261
$69
$175
$30
$9
$13
$308
$8
$12
$328
$10
$338
d. Closed Crankcase Filtration Systems
New engines introduced in Europe in the 2000 model year must have closed crankcases
as part of the EURO in emission standards. The most common technology solution to this
requirement is a closed crankcase filtration system which separates oil and other contaminants
from the blow-by gases and then routes the blow-by gases into the engines intake system
downstream of the air filter. An analysis of this type of control system was made as part of the
2004 heavy-duty rulemaking and system costs were estimated.11 We have estimated the new
vehicle cost of this type of closed crankcase system in Table V.A-6.
V-18
-------
Chapter V: Economic Impact
Table V.A-6. 2007 Closed Crankcase Filtration System Cost Estimate
12
Closed Crankcase Filtration
Hardware Costs
Filter Housing
Service Filter (30,000 mile interval)
PCV Valve
Tubing (plumbing)
Assembly
Total Variable Cost to Manufacturer
Markup (@ 29%)
Total CCV RPE
Vehicle Class
LHDD
MHDD
HHDD
$10
$10
$5
$2
$1
$28
$8
$37
$12
$12
$5
$2
$1
$32
$9
$42
$15
$15
$5
$2
$1
$38
$11
$49
Additionally there is a recurring cost for this type of system associated with the
replacement of a service filter on a 30,000 mile interval. The cost for the service filter is
estimated to be $10, $12, and $15 for light, medium, and heavy heavy-duty vehicles respectively.
These operating costs are summarized in section 5 below along with other diesel vehicle
operating costs.
4. Fixed Costs
Fixed costs are costs to the manufacturer which are non-recurring and include costs for
research and development, tooling and new engine certification. The fixed costs for the diesel
control portion of this rulemaking are given below. Expected expenditures are reported in the
year incurred as non-annualized costs for PM/HC and NOx control separately. In general fixed
costs are incurred prior to the introduction of the new vehicles and are assumed to be recovered
over a five year period beginning with the first year of vehicle sale. Fixed costs are increased by
seven percent for every year before the start of production to reflect the time value of money.
The assumed recovery values for fixed costs associated with NOx and PM/HC control are given
in the tables as annualized values.
a. Research and Development
The advanced emission control technologies which are likely to be applied in 2007 are
already relatively well developed and are seeing application in retrofit markets where low sulfur
diesel fuel is available or in other fields, such as power generation. Further development of these
V-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
catalyst technologies to better adapt them to diesel applications is still needed however. We have
estimated, based on current industry practices, that expenditures to further develop these
advanced emission control technologies by the catalyst suppliers will be approximately $87
million for the CDPF technology and $133 million for the NOx adsorber technology (see
description of these estimates section V.A.S.a and V.A.S.b above for each of these technologies).
Developing the integrated electronic engine control systems required to take advantage of
these new emission reduction technologies for diesel engines will be a significant challenge for
the diesel engine manufacturers. This is a large task which will entail complete re-optimization
of diesel engine operation away from minimizing engine out emissions to minimizing total
system emissions. In addition the manufacturers will need to develop a full understanding of the
long term durability of the total emission control system in order to ensure compliance over the
useful life of the vehicle and in order to develop deterioration factors (DFs) for the systems. We
have therefore estimated that each of the 11 major diesel engine manufacturers will invest
approximately $7 million per year on research and development over a period of five years to
adapt their engine technology to the advanced emission control technologies described here.
Seven million dollars represents the approximate cost for a team of more than 21 engineers and
28 technicians to carry out advanced engine research, including the cost for engine test cell time
and prototype system fabrication. In total we have estimated that the engine manufacturers will
spend approximately $385 million on R&D. Although we believe the manufacturers will take a
total system approach optimizing the engine control system for PM/HC control and for NOx
control concurrently, we have apportioned these research dollars separately for NOx and PM/HC
due to the more complicated changes required to enable the NOx adsorber technology. We have
apportioned 25 percent of the $385 million estimated for engine R&D to PM/HC control and the
remaining 75 percent for development of the systems required for NOx control. These R&D
costs are further apportioned between each vehicle classes based on the ratio of the number of
engine families in a vehicle weight class to the total number of heavy duty diesel engine families.
The R&D costs for the advanced PM/HC emission control technologies are assumed to
be incurred over the five year period from 2002 through 2006 and then recovered over the five
year period starting in 2007. Research and development costs for the NOx adsorber system are
assumed to be incurred in ratio to the NOx standard phase-in timetable and as such are spread
over an eight year period beginning in 2002. For the vehicles introduced as part of the 50 percent
NOx phase-in in 2007 these costs are assumed to be accrued in the five years preceding 2007 and
to be fully recovered by 2011.
Tables V.A-7, V.A-8, and V.A-9 provide a year by year breakdown of the annualized and
non-annualized costs for research and development for the light, medium and heavy heavy-duty
vehicle categories. Fixed costs for urban buses are included in the cost estimates for heavy
heavy-duty vehicles.
V-20
-------
Chapter V: Economic Impact
Table V.A-7. Annualized and Non-Annualized R&D Costs for Light Heavy-Duty Diesel
Engines
Calendar
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
0
0
0
341,000
346,600
352,200
357,800
363,400
369,000
374,600
380,200
385,800
meeting
NOxStd
0
0
0
0
0
170,500
173,300
176,100
357,800
363,400
369,000
374,600
380,200
385,800
PM/HC Control
non-
annualized
$9,420,675
$9,420,675
$9,420,675
$9,420,675
$9,420,675
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$13,212,984
$13,212,984
$13,212,984
$13,212,984
$13,212,984
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$39
$38
$38
$37
$36
$0
$0
$0
$0
NOx Control
non-
annualized
$10,300,813
$10,300,813
$10,300,813
$20,601,625
$20,601,625
$10,300,813
$10,300,813
$10,300,813
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$14,447,422
$14,447,422
$14,447,422
$28,894,845
$28,894,845
$14,447,422
$14,447,422
$14,447,422
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$85
$83
$82
$81
$80
$78
$77
$76
$0
V-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.A-8. Annualized and Non-Annualized R&D Costs for Medium Heavy-Duty
Diesel Engines
Calendar
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
0
0
0
173,600
176,400
179,200
182,000
184,800
187,600
190,400
193,200
196,000
meeting
NOxStd
0
0
0
0
0
86,800
88,200
89,600
182,000
184,800
187,600
190,400
193,200
196,000
PM/HC Control
non-
annualized
$11,161,150
$11,161,150
$11,161,150
$11,161,150
$11,161,150
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$15,654,090
$15,654,090
$15,654,090
$15,654,090
$15,654,090
$0
$0
$0
$0
ann.
per
vehicle
$0
$0
$0
$0
$0
$90
$89
$87
$86
$85
$0
$0
$0
$0
NOx Control
non-
annualized
$13,811,325
$13,811,325
$13,811,325
$27,622,650
$27,622,650
$13,811,325
$13,811,325
$13,811,325
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$19,371,098
$19,371,098
$19,371,098
$38,742,196
$38,742,196
$19,371,098
$19,371,098
$13,371,098
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$223
$220
$216
$213
$210
$207
$203
$201
$0
V-22
-------
Chapter V: Economic Impact
Table V.A-9. Annualized and Non-Annualized R&D Costs for Heavy Heavy-Duty Diesel
Engines and Urban Buses
Calendar
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
0
0
0
272,800
277,200
281,600
286,000
290,400
294,800
299,200
303,600
308,000
meeting
NOxStd
0
0
0
0
0
136,400
138,600
140,800
286,000
290,400
294,800
299,200
303,600
308,000
PM/HC Control
non-
annualized
$16,165,875
$16,165,875
$16,165,875
$16,165,875
$16,165,875
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$22,673,476
$22,673,476
$22,673,476
$22,673,476
$22,673,476
$0
$0
$0
$0
ann.
per
vehicle
$0
$0
$0
$0
$0
$83
$82
$81
$79
$78
$0
$0
$0
$0
NOx Control
non-
annualized
$18,102,013
$18,102,013
$18,102,013
$36,204,025
$36,024,025
$18,102,013
$18,102,013
$18,102,013
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$25,389,009
$25,389,009
$25,389,009
$50,778,018
$50,778,018
$25,389,009
$25,389,009
$25,389,009
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$186
$183
$180
$178
$175
$172
$170
$167
$0
b. Tooling Costs
Capital costs for new, or changes to existing machine tooling, required to produce new
engines to meet the standard are a fixed cost and are assumed to be incurred one year prior to the
introduction of a new vehicle meeting the emission standard. The cost for the advanced
aftertreatment systems, the NOx adsorber and catalyzed diesel particulate filter, discussed in
section V. A.3 have been estimated based on cost to the engine manufacturer and are therefore
inclusive of tooling cost to manufacture those items. Changes to the electronic control system
and to the fuel and air management systems on the diesel engine may lead to some changes in
tooling cost which are accounted for here. These systems are themselves expected to use the
same hardware components developed to meet the 2004 heavy duty engine emission standards.
Some changes may be necessary however, to accommodate the advanced aftertreatment systems
described here. These changes are not expected to change the cost of the hardware itself in an
appreciable way, but some tooling changes may be required. Since these tooling costs are
intended to account for engine changes to the electronic control system and to the fuel and air
management systems of the engine similar to those required for the Phase 1 standards, we have
used the same tooling estimate for the Phase 2 engines here. These possible tooling costs have
V-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
been estimated to be approximately $6 million for light heavy-duty engines, $9 million for
medium heavy-duty engines, and $10 million for heavy heavy-duty engines and urban buses.13
The tooling costs have been apportioned evenly between NOx and PM/HC control
technologies as these system changes are likely to be made based on optimizations for both types
of aftertreatment system. The tooling charges apportioned for the NOx control technologies are
assumed to occur in two equal steps sequenced with the phase-in period of the NOx standard.
The tooling costs for each vehicle weight class are given in Tables V.A-10, V.A-11, and V.A-12.
Table V.A-10. Annualized and Non-Annualized Tooling Costs for Light Heavy-Duty
Diesel Engines
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HC Std
0
0
341,000
346,600
352,200
357,800
363,400
369,000
374,600
380,200
385,800
meeting
NOx Std
0
0
170,500
173,300
176,100
357,800
363,400
369,000
374,600
380,200
385,800
PM/HC Control
non-
annualized
$0
$2,775,000
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$724,172
$724,172
$724,172
$724,172
$724,172
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$2
$2
$2
$2
$2
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$1,387,500
$0
$0
$1,387,500
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$362,086
$362,086
$362,086
$724,172
$724,172
$362,086
$362,086
$362,086
$0
ann. per
vehicle
$0
$0
$2
$2
$2
$2
$2
$2
$2
$2
$0
V-24
-------
Chapter V: Economic Impact
Table V.A-11. Annualized and Non-Annualized Tooling Costs for Medium Heavy-Duty
Diesel Engines
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
173,600
176,400
179,200
182,000
184,800
187,600
190,400
193,200
196,000
meeting
NOxStd
0
0
86,800
88,200
89,600
182,000
184,800
187,600
190,400
193,200
196,000
PM/HC Control
non-
annualized
$0
$4,443,000
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$1,159,459
$1,159,459
$1,159,459
$1,159,459
$1,159,459
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$7
$7
$6
$6
$6
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$2,443,650
$0
$0
$2,443,650
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$637,702
$637,702
$637,702
$1,275,405
$1,275,405
$637,702
$637,702
$637,702
$0
ann. per
vehicle
$0
$0
$7
$7
$7
$7
$7
$7
$7
$7
$0
Table V.A-12. Annualized and Non-Annualized Tooling Costs for Heavy Heavy-Duty
Diesel Engines and Urban Buses
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
272,800
277,200
281,600
286,000
290,400
294,800
299,200
303,600
308,000
meeting
NOxStd
0
0
136,400
138,600
140,800
286,000
290,400
294,800
299,200
303,600
308,000
PM/HC Control
non-
annualized
$0
$5,132,750
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$1,339,458
$1,339,458
$1,339,458
$1,339,458
$1,339,458
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$5
$5
$5
$5
$5
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$2,566,375
$0
$0
$2,566,375
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$669,729
$669,729
$669,729
$1,339,458
$1,339,458
$669,729
$669,729
$669,729
$0
ann. per
vehicle
$0
$0
$5
$5
$5
$5
$5
$5
$4
$4
$0
V-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
c.
Certification Costs
Manufacturers will also incur costs to certify the range of engine families to the emission
standards. EPA previously developed a methodology for calculating certification costs which
results in an estimated certification cost of $30,000 per engine family.14 Here we have assumed
that all engine families will require certification in 2007 with the introduction of the new PM and
HC standards. Additionally as engine families are phased-in to meet the new NOx standards they
will again require certification. We have assumed that in the first year of the NOx phase-in
period 100 percent of the engine families will require certification and that in the fourth year of
the phase (when 100 percent are phased in) that 50 percent of the engine families will require
certification.
The total cost for certifying engines under this program can be rounded up to $5 million.
Distributing those costs across the different engine categories, amortizing the costs over five
years, and dividing by the number of projected sales for each category results in per-engine costs
between $1 and $3 for each category of heavy-duty diesel vehicles. These costs are detailed in
Tables V.A-13, V.A-14, and V.A-15 for each of the heavy-duty vehicle weight classes.
Table V.A-13. Annualized and Non-Annualized Certification Costs for Light Heavy-
Duty Diesel Engines
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Vehicle Sales
meeting
PM/HCStd
0
0
341,000
346,600
352,200
357,800
363,400
369,000
374,600
380,200
385,800
meeting
NOxStd
0
0
170,500
173,300
176,100
357,800
363,400
369,000
374,600
380,200
385,800
PM/HC Control
non-
annualized
$0
$480,000
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$125,262
$125,262
$125,262
$125,262
$125,262
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$0.4
$0.4
$0.4
$0.4
$0.3
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$0
$0
$0
$240,000
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$62,631
$62,631
$62,631
$62,631
$62,631
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$0.3
$0.3
$0.3
$0.3
$0.3
$0
V-26
-------
Chapter V: Economic Impact
Table V.A-14. Annualized and Non-Annualized Certification Costs for Medium Heavy-
Duty Diesel Engines
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
173,600
176,400
179,200
182,000
184,800
187,600
190,400
193,200
196,000
meeting
NOxStd
0
0
86,800
88,200
89,600
182,000
184,800
187,600
190,400
193,200
196,000
PM/HC Control
non-
annualized
$0
$1,020,000
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$266,182
$266,182
$266,182
$266,182
$266,182
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$1.5
$1.5
$1.5
$1.5
$1.4
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$0
$0
$0
$510,000
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$133,091
$133,091
$133,091
$133,091
$133,091
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$1.5
$1.4
$1.4
$1.4
$1.4
$0
Table V.A-15. Annualized and Non-Annualized Certification Costs for Heavy Heavy-
Duty Diesel Engines and Urban Buses
Calendar
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
0
0
272,800
277,200
281,600
286,000
290,400
294,800
299,200
303,600
308,000
meeting
NOxStd
0
0
136,400
138,600
140,800
286,000
290,400
294,800
299,200
303,600
308,000
PM/HC Control
non-
annualized
$0
$1,200,000
$0
$0
$0
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$313,156
$313,156
$313,156
$313,156
$313,156
$0
$0
$0
$0
ann. per
vehicle
$0
$0
$1.2
$1.1
$1.1
$1.1
$1.1
$0
$0
$0
$0
NOx Control
non-
annualized
$0
$0
$0
$0
$600,000
$0
$0
$0
$0
$0
$0
annualized
$0
$0
$0
$0
$0
$156,578
$156,578
$156,578
$156,578
$156,578
$0
ann. per
vehicle
$0
$0
$0
$0
$0
$1.1
$1.1
$1.1
$1.1
$1.0
$0
V-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
d. Summary of Fixed Costs
The total annualized fixed costs are summarized here for light, medium and heavy heavy-
duty vehicles. Fixed costs for urban buses are included in the estimates for heavy heavy-duty
diesel vehicles. Research and Development costs account for over 90 percent of the total fixed
costs per engine in our analysis. Tables V.A-16, V.A-17 and V.A-18 below summarize fixed
costs in each year of the program.
Table V.A-16. Annualized Fixed Costs for Light Heavy-Duty Diesel Engines
Calendar
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HCStd
341,000
346,600
352,200
357,800
363,400
369,000
374,600
380,200
385,800
meeting
NOxStd
170,500
173,300
176,100
357,800
363,400
369,000
374,600
380,200
385,800
PM/HC Control
annualized
$14,062,419
$14,062,419
$14,062,419
$14,062,419
$14,062,419
$0
$0
$0
$0
annualized
per vehicle
$41
$41
$40
$39
$39
$0
$0
$0
$0
NOx Control
annualized
$14,809,509
$14,809,509
$14,809,509
$29,681,648
$29,681,648
$14,872,140
$14,872,140
$14,872,140
$0
annualized
per vehicle
$87
$85
$84
$83
$82
$81
$79
$78
$0
Total
annualized
$28,871,92
$28,871,92
$28,871,92
$43,744,06
$43,744,06
$14,872,14
$14,872,14
$14,872,14
$0
annualized
per vehicle
$128
$126
$124
$122
$121
$81
$79
$78
$0
Table V.A-17. Annualized Fixed Costs for Medium Heavy-Duty Diesel Engines
Calendar
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HC Std
173,600
176,400
179,200
182,000
184,800
187,600
190,400
193,200
196,000
meeting
NOx Std
86,800
88,200
89,600
182,000
184,800
187,600
190,400
193,200
196,000
PM/HC Control
annualized
$17,079,731
$17,079,731
$17,079,731
$17,079,731
$17,079,731
$0
$0
$0
$0
annualized
per vehicle
$98
$97
$95
$94
$92
$0
$0
$0
$0
NOx Control
annualized
$20,008,800
$20,008,800
$20,008,800
$40,150,691
$40,150,691
$20,141,891
$20,141,891
$20,141,891
$0
annualized
per vehicle
$231
$227
$223
$221
$217
$215
$212
$209
$0
Total
annualized
$37,088,531
$37,088,531
$37,088,531
$57,230,422
$57,230,422
$20,141,891
$20,141,891
$20,141,891
$0
annualized
per vehicle
$329
$324
$318
$315
$309
$215
$212
$209
$0
V-28
-------
Chapter V: Economic Impact
Table V.A-18. Annualized Fixed Costs for Heavy Heavy-Duty Diesel Engines and Urban
Buses
Calendar
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
Projected Vehicle Sales
meeting
PM/HC Std
272,800
277,200
281,600
286,000
290,400
294,800
299,200
303,600
308,000
meeting
NOxStd
136,400
138,600
140,800
286,000
290,400
294,800
299,200
303,600
308,000
PM/HC Control
annualized
$24,326,090
$24,326,090
$24,326,090
$24,326,090
$24,326,090
$0
$0
$0
$0
annualized
per vehicle
$89
$88
$86
$85
$84
$0
$0
$0
$0
NOx Control
annualized
$26,058,738
$26,058,738
$26,058,738
$52,274,054
$52,274,054
$26,215,316
$26,215,316
$26,215,316
$0
annualized
per vehicle
$191
$188
$185
$183
$180
$178
$175
$173
$0
Total
annualized
$50,384,828
$50,384,828
$50,384,828
$76,600,144
$76,600,144
$26,215,316
$26,215,316
$26,215,316
$0
annualized
per vehicle
$280
$276
$271
$268
$264
$178
$175
$173
$0
5. Operating Costs
Operating costs include the cost for vehicle and engine maintenance, and the cost for
vehicle consumables such as fuel, oil, filters and tires. The new standards and technologies
introduced beginning in 2007 are expected to change vehicle operating costs. Costs for the
refining and distribution of diesel fuel are expected to change due to the 15 ppm sulfur
requirement. These costs are examined in detail later in this chapter (section V.D), but are also
summarized here on a per vehicle basis. The closed crankcase systems we have described here
include a paper filter element which is changed on a fixed service interval. The cost of this filter
is included here as an ongoing operating cost. In addition the reduction of the sulfur content in
diesel fuel is expected to lead to reduced maintenance costs or other cost savings in the design of
future diesel engines. These cost savings are discussed in detail for both new and existing
engines in section V.C and are summarized here on a per vehicle basis. The advanced emission
control technologies expected to be applied in order to meet the NOx and PM/HC standards
involve wholly new system components integrated into engine designs and calibrations, and as
such may be expected to change the fuel consumption characteristics of the overall engine
design. A discussion of the potential impacts of these technologies on vehicle fuel economy, and
an explanation of why we do not expect vehicle fuel economy levels to change from today's
levels are given here. All of these operating cost impacts are described here and are used to
present a total per vehicle cost for control in tables V.A-2 and V.A-19.
V-29
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
a. Low Sulfur Diesel Fuel
Low sulfur diesel fuel is a primary enabling technology without which the other
previously mentioned emission control technologies could not be applied. As an essential part of
the technology package which enables the standards its cost are summarized here and in table
V. A-2 on a per-vehicle cost basis (NPV).
The low-sulfur diesel fuel required to enable these technologies is expected to have a long
term incremental cost of approximately $0.05/gallon as discussed in more detail later in this
chapter. This per gallon cost can be accounted for on a per vehicle basis by considering the
mileage typically driven by a class of vehicle at each year of its life and the average fuel
economy. Using that approach and bringing the total cost back to a net present value in the year
of sale gives a long term per vehicle low sulfur fuel cost of $609 for a light heavy-duty vehicle,
$1,141 for a medium heavy-duty vehicle, $4,209 for a heavy heavy-duty vehicle and $4,959 for
an urban bus. For a more detailed discussion of the cost associated with low sulfur diesel fuel
please refer to section V.D in this RIA.
b. Maintenance Costs for Closed Crankcase Ventilation Systems (CCV)
We have eliminated the exception that allows turbo-charged heavy-duty diesel engines to
vent crankcase gases directly to the environment without accounting for these emissions,
sometimes called open crankcase systems, and are projecting that manufacturers will rely on
engineered closed crankcase ventilation systems which filter oil from the blow-by gases in order
to satisfy the emission standard. An integral part of the system described in Chapter HI of this
RIA is a paper filter designed to capture oil mist in the blow-by gases, coalesce this oil and return
this filtered oil to the oil sump. These filters are expected to require replacement on a fixed
interval of 30,000 miles.
The cost of these filters in 2007 has been estimated to be $10, $12, and $15 for light,
medium, and heavy heavy-duty vehicles respectively. The variable cost for these replacement
filters are reduced in future years due to the learning curve effect as described in section 6 below.
The long term total life cycle operating cost for the filter replacements expressed as a net present
value in the year of sale is $26, $48, and $172 for light, medium, and heavy heavy-duty vehicles,
respectively. Urban bus life cycle operating costs are estimated to be $92. To account for the
aggregate cost of filter replacement the filter costs are estimated on a per mile basis for each class
of vehicle (for example for heavy heavy-duty this is $15/30,000 miles) and then are estimated in
total using typical mileage accumulation rates given in each year of a vehicles life from our
inventory emissions model. The results of this calculation along with the maintenance costs for
CDPFs are reported in table V.A-21.
V-30
-------
Chapter V: Economic Impact
c. Maintenance Costs for Catalyzed Diesel Particulate Filters
The particulate matter (PM) emitted from diesel engines consists primarily of elemental
carbon formed during the combustion process from diesel fuel. This elemental carbon is
captured in the CDPF and then oxidized to CO2 and emitted from the engine. A very small
fraction of the PM consists of inorganic metals which are also captured by the CDPF but are not
emitted later from the CDPF. Instead this inorganic "ash" accumulates in the PM filter over time
slowing filling the filtering passages of the CDPF. Current engine oil formulations are the
primary source of this inorganic ash due to metal additives used in the oil.
The inorganic ash captured in the CDPF can be cleaned from the CDPF by removing it
from the vehicle and reverse flushing the ash out of the CDPF with compressed air or water.
Current industry guidelines suggest a maintenance interval for retrofit applications of
approximately 60,000 miles for CDPF cleaning. This guideline reflects a fairly short
maintenance interval because
• PM rates in retrofit applications are high (many retrofits are EURO 0,1, & II engines)15
• Oil consumption rates on older retrofit engines can be very high
• Current engine oils are highly additized to maintain Total Base Number (TEN).
We have estimated that for CDPF equipped vehicles in 2007 and beyond that the
maintenance interval will increase to 100,000 miles for light heavy-duty vehicles and 150,000
miles for medium and heavy heavy-duty vehicles. We expect that this interval will be planned to
coincide with other engine maintenance events and can be extended to these higher intervals
because
• PM rates are lower for modern diesel engines
• Modern diesel engines have low oil consumption rates (to meet the PM standard)
• Low sulfur diesel fuel will allow the use of "low ash" engine oils.
We have estimated the cost of this service based upon the assumption that the service is
scheduled to coincide with other service intervals and that the dominant cost for the service is the
cost labor cost to remove and clean the filter. We have assumed that this removal and
reinstallation will take approximately one hour. We have used a labor rate for this service event
of $65 / hour. These costs are aggregated on a fleet wide basis in each year of the program and
reported in table V.A-21 along with the maintenance costs for the closed crankcase ventilation
(CCV) system. The CDPF maintenance costs can also be expressed as a net present value in the
year of sale for an individual vehicle as $55 for a light heavy-duty vehicle, $56 for a medium
heavy-duty vehicle, $208 for a heavy heavy-duty vehicle and $107 for an urban bus.
V-31
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
d. Maintenance Savings due to Low Sulfur Diesel Fuel
In addition to its role as a technology enabler, low sulfur diesel fuel gives benefits
in the form of reduced sulfur induced corrosion and slower acidification of engine lubricating oil,
leading to longer maintenance intervals and lower maintenance costs. These benefits are
described in detail in section V.C and result in an estimated savings of $153 for light heavy-duty
vehicles, $249 for medium heavy-duty vehicles, and $610 for heavy heavy-duty vehicles and
urban buses.
e. Fuel Economy Impacts
Diesel particulate filters are anticipated to provide a step-wise decrease in diesel
particulate (PM) emissions by trapping PM and by oxidizing the diesel PM and hydrocarbon
(HC) emissions. The trapping of the very fine diesel PM is accomplished by forcing the exhaust
through a porous filtering media with extremely small opening and long path lengths.13 This
approach results in filtering efficiencies for diesel PM greater than 90 percent but requires
additional pumping work to force the exhaust through these small openings. The additional
pumping work is anticipated to negatively impact fuel economy by approximately one percent.16
However as detailed in the following discussion this fuel economy penalty is more than offset
through optimization of the engine-PM trap-NOx adsorber system, as discussed below.
NOx adsorbers are expected to be the primary NOx control technology introduced in
order to provide the reduction in NOx emissions necessary to meet the NOx standard. NOx
adsorbers work by storing NOx emissions under fuel lean operating conditions (normal diesel
engine operating conditions) and then by releasing and reducing the stored NOx emissions over a
brief period of fuel rich engine operation. This brief periodic NOx release and reduction step is
directly analogous to the catalytic reduction of NOx over a gasoline three-way-catalyst. In order
for this catalyst function to occur the engine exhaust constituents and conditions must be similar
to normal gasoline exhaust constituents. That is, the exhaust must be fuel rich (devoid of excess
oxygen) and hot (over 250°C). Although it is anticipated that diesel engines can be made to
operate in this way, it is assumed that the fuel economy of the diesel engine operating under these
conditions will be worse than normal. This increase in fuel consumption can be minimized by
carefully controlling engine air-to-fuel (A/F) ratios using the EGR systems introduced in order to
meet the 2004 heavy duty engine emission standards. The lower the engine A/F ratio, the lower
the amount of fuel which must be added in order to give rich conditions. In the ideal case where
the engine A/F ratio is at stoichiometry, and additional fuel is required only as a NOx reductant
the fuel economy penalty is virtually zero. We are projecting, that practical limitations on engine
A/F control will mean that the NOx adsorber release and reduction cycles will lead to a one
b Typically the filtering media is a porous ceramic monolith or a metallic fiber mesh.
V-32
-------
Chapter V: Economic Impact
percent decrease in the engine fuel economy. Again, we believe this fuel economy impact can be
regained through optimization of the engine-PM trap-NOx adsorber system.
In addition to the NOx release and regeneration event, another step in NOx adsorber
operation may affect fuel economy. NOx adsorbers are poisoned by sulfur in the fuel even at the
low sulfur levels we have set today. Chapter in of this RIA describes how the sulfur poisoning
of the NOx adsorber can be reversed through a periodic "desulfation" event. The desulfation of
the NOx adsorber is accomplished in a manner similar to the NOx release and regeneration cycle
described above. However it is anticipated that the desulfation event will require extended
operation of the diesel engine at rich conditions.17 This rich operation will, like the NOx
regeneration event, will lead to an increase in the fuel consumption rate and will cause an
associated decrease in fuel economy. With a 15 ppm fuel sulfur cap, we are projecting this fuel
economy penalty to be one percent or less as described in more detail in chapter in of this RIA.
Again, we believe this fuel economy impact can be regained through optimization of the engine-
PM trap-NOx adsorber system.
While NOx adsorbers require non-power producing consumption of diesel fuel in order to
function properly and, therefore, have an impact on fuel economy, they are not unique among
NOx control technologies in this way. In fact NOx adsorbers are likely to have a very favorable
NOx to fuel economy trade-off when compared to other popular NOx control technologies like
cooled EGR and injection timing retard. EGR requires the delivery of exhaust gas from the
exhaust manifold to the intake manifold of the engine and causes a decrease in fuel economy for
two reasons. The first of these reasons is that a certain amount of work is required to pump the
EGR from the exhaust manifold to the intake manifold; this necessitates the use of intake
throttling or some other means to accomplish this pumping. The second of these reasons is that
heat in the exhaust, which is normally partially recovered as work across the turbine of the
turbocharger, is instead lost to the engine coolant through the cooled EGR heat exchanger. In the
end, cooled EGR is only some 50 percent effective at reducing NOx below the current 4 g/bhp-hr
NOx emission standard. Injection timing retard is another strategy that can be employed to
control NOx emissions. By retarding the introduction of fuel into the engine, and thus delaying
the start of combustion, both the peak temperature and pressure of the combustion event are
decreased; this lowers NOx formation rates and, ultimately, NOx emissions. Unfortunately, this
also significantly decreases the thermal efficiency of the engine (lowers fuel economy) while also
increasing PM emissions. As an example, retarding injection timing eight degrees can decrease
NOx emissions by 45 percent, but this occurs at a fuel economy penalty of more than seven
1 8
percent.
Today, most diesel engines rely on injection timing control (retarding injection timing) in
order to meet the 4.0 g/bhp-hr NOx emission standard. For 2002/2004 model year compliance,
we expect that engine manufacturers will use a combination of cooled EGR and injection timing
control to meet the 2.0 g/bhp-hr NOx standard. Because of the more favorable fuel economy
V-33
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
trade-off for NOx control with EGR when compared to timing control, we have forecast that less
reliance on timing control will be needed in 2002/2004. Therefore, fuel economy will not be
changed even at this lower NOx level. NOx adsorbers have a significantly more favorable NOx
to fuel economy trade-off when compared to cooled EGR or timing retard.19 We expect NOx
adsorbers to be able to accomplish a greater than 90 percent reduction in NOx emissions, while
themselves consuming significantly less fuel than that lost through alternative NOx control
strategies such as retarded injection timing.0 Therefore, we expect manufacturers to take full
advantage of the NOx control capabilities of the NOx adsorber and project that they will decrease
reliance on the more expensive (from a fuel economy standpoint) technologies, especially
injection timing retard. We would, therefore, predict that the fuel economy impact currently
associated with NOx control from timing retard will be decreased by at least three percent. In
other words, through the application of these advanced NOx emission control technologies, we
expect the NOx trade-off with fuel economy to continue to improve significantly when compared
to today's technologies. This will result in much lower NOx emissions and potentially overall
improvements in fuel economy, improvements that could easily offset the one percent fuel
economy loss projected to result from the application of PM filters. For our analysis of economic
impacts, no penalty or benefit for changes to fuel economy is assumed.
In order to illustrate the sensitivity of cost to fuel economy, we have calculated the
benefit (or cost) of a one percent change in vehicle fuel economy as a sensitivity analysis to these
possible changes. For a light heavy-duty engine a one percent change in vehicle fuel economy
expressed as a net present value in the year of sale is approximately $100, for a medium heavy-
duty engine it is approximately $200, for a heavy heavy-duty engine it is approximately $800.
The amount of the benefit (or cost) of a one percent change in fuel economy expressed in terms
of its annual impact on the entire fleet of engines meeting the 2007 NOx standards can be
estimated as $155 million in 2010 and $459 million in 2030. These potential benefits (or costs)
represent approximately 4 percent of the total program cost in 2010 and less than 11 percent in
2030.
6. Summary of Near and Long Term Costs
We have estimated in section V.A.3 the cost of a technology package which is
representative of the technologies we expect industry to apply to meet our standards. These cost
estimates represent an expected incremental cost of engines in the 2007 model year. EPA has
also identified various factors that would cause cost impacts to decrease over time, making it
0 EPA has estimated the fuel consumption rate for NOx regeneration and desulfation of the NOx adsorber
as approximately 2 percent of total engine fuel consumption. This differs from the contractor report by EF&EE
which estimates the total consumption as approximately 2.5% of total fuel consumption. Additionally the
contractor's estimate of NOx adsorber efficiency ranges from 80-90 percent, while EPA believes over 90 percent
control is possible as discussed fully in Chapter III of this RIA.
V-34
-------
Chapter V: Economic Impact
appropriate to distinguish between near-term and long term costs. These factors are described
below and the resulting near and long term per vehicle costs are presented here.
First, initial fixed costs for tooling, R&D, and certification are recovered over a five-year
period phased with the NOx standard phase-in period. Fixed costs are therefore accrued in four
periods corresponding to each of the phase-in years of the NOx standard. The accrued costs are
then recovered over a five year period.
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.20
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
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 services.21 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-35
-------
15
10
CD
3
D"
CD
0
Distribution of Progress Ratios
55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 101 103 105 107
Progress Ratio
From 22 field studies (n = 108).
Figure V.A-1. Distribution of Progress Ratios
(Button and Thomas, 1984)
V-36
-------
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).
EPA believes the use of the learning curve is appropriate to consider in assessing the cost
impact of heavy-duty engine emission controls. The learning curve applies to new technology,
new manufacturing operations, new parts, and new assembly operations. Heavy-duty diesel
engines currently do not use any form of NOx aftertreatment and have used diesel particulate
filters in only limited application. These are therefore new technologies for heavy-duty diesel
engines and will involve new manufacturing operations, new parts, and new assembly operations.
Since this will be a new and unique product, EPA believes this is an appropriate situation for the
learning curve concept to apply. Opportunities to reduce unit labor and material costs and
increase productivity (as discussed above) will be great. EPA believes a similar opportunity
exists for the new control systems which will integrate the function of the engine and the
emission control technologies. While all diesel engines beginning in 2004 are expected to have
the basic components of this system, advanced engine control modules (computers), advanced
engine air management systems (cooled EGR, and variable geometry turbocharging) and
advanced fuel systems including common rail systems, they will now be applied in new ways.
Additionally some new components will be applied for the first time. These new parts and new
assemblies will involve new manufacturing operations. As manufacturers gain experience with
these new systems, comparable learning is expected to occur with respect to unit labor and
material costs. These changes require manufacturers to start new production procedures, which,
over time, will improve with experience.
We have applied a p value of 80 percent beginning in 2007 in this analysis. That is,
variable costs were reduced by 20 percent for each doubling of cumulative production. With one
year as the base unit of production, the first learning curve is applied at the start of 2009. The
second doubling of production occurs at the end of the 2010 model year, therefore variable costs
are reduced a second time by 20 percent beginning in the 2011 model year. In Tier 2, and in the
heavy-duty gasoline cost analysis presented in section B of this chapter, the learning curve
reduction was applied only once because we anticipated that for the most part the standards will
be met through improvements to existing technologies rather than through the use of new
technologies. With existing technologies, there will be less opportunity for lowering production
costs.
Fixed costs for this program have been allocated for two separate groups of vehicle
representing vehicles introduced in the first and fourth years of NOx phase in period. In this way
fixed costs on a per vehicle basis are appropriately weighted for the number of vehicles
introduced in that model year. The manufacturers are expected to accrue fixed cost in proportion
to the number of vehicles being introduced in a model year as we have done here. This means
that fixed costs are assumed to begin accruing in 2002 for vehicles intended for introduction in
V-37
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
2007 and to continue to be accrued through 2009 for vehicles intended for introduction in 2010.
Fixed costs are therefore assumed to be recovered beginning in 2007 (for vehicles introduced in
2007) and continuing through 2014 for vehicles introduced in 2010, the final year of the NOx
phase-in. For all per vehicle costs, the fixed costs are reported for vehicles first introduced in
2007 and are therefore fully recovered by 2012. For a more complete description of fixed costs
see section V.A.4 of this RIA.
The resulting hardware and life cycle operating costs for new vehicles developed to meet
the new 2007 heavy-duty vehicle standards are summarized in table V.A-19 below.
Table V.A-19. Projected Incremental Diesel Engine/Vehicle Costs
(net present value at point of sale in 1999 dollars)
Vehicle Class
Light heavy-duty
Medium heavy-duty
Heavy heavy-duty
Urban Bus
Model
Year
2007
2009
2012
2007
2009
2012
2007
2009
2012
2007
2009
2012
Change
—
20 percent learning curve applied to
variable costs
Fixed costs expire; 20 percent learning
curve has been applied to variable costs
—
20 percent learning curve applied to
variable costs
Fixed costs expire; 20 percent learning
curve has been applied to variable costs
—
20 percent learning curve applied to
variable costs
Fixed costs expire; 20 percent learning
curve has been applied to variable costs
—
20 percent learning curve applied to
variable costs
Fixed costs expire; 20 percent learning
curve has been applied to variable costs
Hardware
Cost
$1,986
$1,601
$1,173
$2,564
$2,096
$1,412
$3,227
$2,618
$1,866
$2,889
$2,347
$1,650
Life-cycle
Operating
Cost (NPV)
$509
$509
$537
$943
$943
$996
$3,785
$3,785
$3,979
$4,625
$4,625
$4,797
V-38
-------
Chapter V: Economic Impact
It is appropriate to compare the impact of these incremental costs to the total cost to
purchase and operate these vehicles. The analysis for the 2004 heavy duty engine standards
included work to document the cost to purchase and operate heavy duty vehicles. That analysis
is carried forward here and is given in Table V.A-20 after being adjusted to 1999 dollars. From
the table we can see that in the near term and long term vehicle operating costs can be expected
to increase by less four percent for all vehicle weight classes. Near term vehicle costs on average
will be expected to increase by approximately five percent. In the long term vehicle costs will be
increased by less than five percent for light heavy-duty vehicles, by less than three percent for
medium heavy-duty vehicles, and by less than two percent heavy heavy-duty vehicles and urban
buses.
Table V.A-20 Baseline Costs for Heavy-Duty Engines and Vehicles 22
Vehicle Class
Light heavy-duty
Medium heavy-duty
Heavy heavy-duty
Urban Bus
Engine Cost
$8,527
$13,555
$23,722
$24,050
Vehicle Cost
$24,600
$50,430
$105,481
$244,871
Operating
Costs
$13,610
$34,153
$118,093
$477,885
7. Total Incremental Nationwide Costs for 2007 Heavy-Duty Diesel
Engines
The above analysis develops per-vehicle cost estimates for each vehicle class. With
current data for the size and characteristics of the heavy-duty vehicle fleet and projections for the
future, these costs can be translated into a total cost to the nation for the emission standards in
any year. The result of this analysis are presented in the following tables which summarize the
total incremental cost for new vehicles introduced into the fleet for each model year.
Fixed costs have been previously developed for each class of heavy duty vehicle and are
presented in section V.A.4 of this RIA. Those costs have been totaled here to present the total
annualized and non-annualized fixed costs for the engine control under this program. Variable
costs are computed as a product of one full year of heavy-duty vehicle sales and the cost increase
for the new hardware on a per vehicle basis as developed previously. The operating cost for the
closed crankcase filtration systems and for cleaning CDPF catalysts are included here as well.
The operating cost associated with low sulfur diesel fuel and the savings associated with low
sulfur diesel fuel are summarized on an aggregate basis later in this chapter.
V-39
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The total annualized cost for the hardware changes are given in table V.A-21 below.
Non-annualized costs are also given below in table V. A-22.
V-40
-------
Chapter V: Economic Impact
Table V.A-21. Estimated Annualized Nationwide Costs
Associated with the 2007 Emission
(1999 dollars)
for Heavy-Duty Diesel Engines
Standard
Calendar
Year
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
2035
Projected
Vehicle Sales
787,400
800,200
813,000
825,800
838,600
851,400
864,200
877,000
889,800
902,600
915,400
928,200
941,000
953,800
966,600
979,400
992,200
1,005,000
1,017,800
1,030,600
1,043,400
1,056,200
1,069,000
1,081,800
1,094,600
1,107,400
1,120,200
1,133,000
1.145.800
Fixed Costs
$116,345,286
$116,345,286
$116,345,286
$177,574,633
$177,574,633
$61,229,347
$61,229,347
$61,229,347
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Variable Costs
$1,373,511,459
$1,395,802,627
$1,126,415,636
$1,521,698,170
$1,227,885,771
$1,246,599,207
$1,265,312,642
$1,284,026,077
$1,302,739,513
$1,321,452,948
$1,340,166,383
$1,358,879,819
$1,377,593,254
$1,396,306,689
$1,415,020,124
$1,433,733,560
$1,452,446,995
$1,471,160,430
$1,489,873,866
$1,508,587,301
$1,527,300,736
$1,546,014,172
$1,564,727,607
$1,583,441,042
$1,602,154,478
$1,620,867,913
$1,639,581,348
$1,658,294,784
$1.677.008.219
CCVandCDPF
Maintenance
Costs
$22,066,902
$58,732,341
$82,987,152
$110,217,085
$123,307,106
$143,989,773
$162,942,107
$180,369,083
$196,453,205
$211,356,424
$225,221,746
$238,175,421
$250,327,361
$261,771,775
$272,586,752
$282,835,130
$292,570,241
$301,857,197
$310,779,520
$319,383,848
$327,711,027
$335,796,605
$343,671,733
$351,363,512
$358,926,832
$366,345,968
$373,585,640
$380,771,649
$387.852.302
Total Costs
$1,511,923,648
$1,570,880,255
$1,325,748,074
$1,809,489,888
$1,528,767,511
$1,451,818,326
$1,489,484,096
$1,525,624,507
$1,499,192,718
$1,532,809,372
$1,565,388,129
$1,597,055,240
$1,627,920,615
$1,658,078,464
$1,687,606,876
$1,716,568,690
$1,745,017,236
$1,773,017,627
$1,800,653,386
$1,827,971,149
$1,855,011,763
$1,881,810,777
$1,908,399,340
$1,934,804,554
$1,961,081,310
$1,987,213,881
$2,013,166,988
$2,039,066,433
$2.064.860.521
V-41
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.A-22. Estimated Non-Annualized Nationwide Costs for Heavy-Duty
Diesel Engines Associated with the 2007 Emission Standard
(1999 dollars)
Calendar
Year
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
2035
Fixed Costs
$78,961,850
$78,961,850
$78,961,850
$121,176,000
$142,624,275
$42,214,150
$42,214,150
$49,961,675
$0
$0
$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 Costs
$0
$0
$0
$0
$0
$1,373,511,459
$1,395,802,627
$1,126,415,636
$1,521,698,170
$1,227,885,771
$1,246,599,207
$1,265,312,642
$1,284,026,077
$1,302,739,513
$1,321,452,948
$1,340,166,383
$1,358,879,819
$1,377,593,254
$1,396,306,689
$1,415,020,124
$1,433,733,560
$1,452,446,995
$1,471,160,430
$1,489,873,866
$1,508,587,301
$1,527,300,736
$1,546,014,172
$1,564,727,607
$1,583,441,042
$1,602,154,478
$1,620,867,913
$1,639,581,348
$1,658,294,784
$1.677.008.219
CCVandCDPF
Maintenance Costs
$0
$0
$0
$0
$0
$22,066,902
$58,732,341
$82,987,152
$110,217,085
$123,307,106
$143,989,773
$162,942,107
$180,369,083
$196,453,205
$211,356,424
$225,221,746
$238,175,421
$250,327,361
$261,771,775
$272,586,752
$282,835,130
$292,570,241
$301,857,197
$310,779,520
$319,383,848
$327,711,027
$335,796,605
$343,671,733
$351,363,512
$358,926,832
$366,345,968
$373,585,640
$380,771,649
$387.852.302
Total Costs
$78,961,850
$78,961,850
$78,961,850
$121,176,000
$142,624,275
$1,437,792,511
$1,496,749,118
$1,259,364,463
$1,631,915,255
$1,351,192,877
$1,390,588,980
$1,428,254,749
$1,464,395,160
$1,499,192,718
$1,532,809,372
$1,565,388,129
$1,597,055,240
$1,627,920,615
$1,658,078,464
$1,687,606,876
$1,716,568,690
$1,745,017,236
$1,773,017,627
$1,800,653,386
$1,827,971,149
$1,855,011,763
$1,881,810,777
$1,908,399,340
$1,934,804,554
$1,961,081,310
$1,987,213,881
$2,013,166,988
$2,039,066,433
$2.064.860.521
V-42
-------
Chapter V: Economic Impact
B. Economic Impact of the 2008 Model Year Heavy-Duty
Gasoline Standards
This chapter contains an analysis of the economic impacts of the emission standards for
2008 model year heavy-duty gasoline vehicles and engines. First, a brief outline of the
methodology used to estimate the economic impacts is presented, followed by a summary of the
technology packages that are expected to be used to meet the standards. Next, the projected costs
of the individual technologies is presented, along with a discussion of fixed costs such as
research and development (R&D), tooling and certification costs. Following the discussion of
the individual cost components is a summary of the projected per-vehicle cost. Finally, an
analysis of the aggregate cost to society of the new standards is presented. The costs presented
here are in 1999 dollars.
1. Methodology for Estimating Heavy-Duty Gasoline Costs
This analysis uses the emission control technology packages assumed for the final Phase
1 gasoline standards as a baseline from which changes will be made to comply with the new
Phase 2 standards. The Phase 1 standards go into effect for the 2004 or 2005 model year.23 That
is, we have identified the changes we expect to be made to the assumed 2005 baseline vehicles in
complying with the new 2008 standards. The 2005 baseline technology packages are consistent
with those being implemented to meet California's Low Emission Vehicle (LEV I) standards.
The technology packages assumed for the 2008 model year are consistent with those expected to
meet the California LEV-II medium-duty vehicle standards and our light-duty Tier 2 standards/'24
The catalyst system costs of these technologies are taken from the Phase 1 RIA, which are based
on a report done for EPA by Arcadis Geraghty & Miller.25 Other system costs are taken from the
final Tier 2 RIA, which are based in part on California's LEV-II analysis and the same Arcadis
Geraghty & Miller report.
The costs of meeting the 2008 emission standards include both variable costs
(incremental hardware costs, assembly costs, and associated markups) and fixed costs (tooling,
R&D, and certification costs). Supplier markups, those markups occurring between the part or
emission control system supplier to the vehicle or engine manufacturer, are applied to catalyst
costs in this analysis because the cost we estimated for each element comprising the catalyst are
the supplier cost rather than the vehicle or engine manufacturer cost. This contrasts with the
diesel cost analysis discussed in Section V.A where the cost of each element comprising a PM
trap or a NOx adsorber are costs to the vehicle manufacturer (i.e., they already contain a supplier
d While the Tier 2 standards are light-duty standards, and do not apply to the vehicles and engines covered
by this analysis, we expect that the technologies employed to meet the Tier 2 standards will transfer in large part
into the heavy-duty gasoline fleet; therefore, the types of technology packages are expected to be very similar.
V-43
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
markup). An exception to applying the supplier markup has been made for precious metals.
Vehicle manufacturers typically provide catalyst suppliers with precious metals for use in the
catalysts their suppliers manufacture. Thus, the 29 percent supplier markup is not applied to the
cost of precious metals. The supplier markup is already reflected in the non-catalyst system costs
(e.g., EGR system, secondary air injection system, etc.) presented in this section.
The variable costs to the manufacturer have then been marked up twice.26 The first
markup, at a four percent rate, covers manufacturer carrying costs reflecting primarily the costs
of capital tied up in extra inventory, and secondarily the incremental cost of insurance, handling,
and storage. The second markup, at a three percent rate, covers dealer carrying costs reflecting
the cost of capital tied up in extra inventory. These markups were discussed in more detail in
section A of this chapter. Fixed costs were amortized at a seven percent rate and recovered over
a five year period.
2. Technology Packages for Compliance with the 2008 Model Year
Heavy-Duty Gasoline Standards
The various technologies that could be used to comply with the proposed regulations
were discussed in Chapter 3. We expect that the technology mixes used to meet the California
LEV-n standards, and our Tier 2 standards, fairly accurately represent those that will be used to
comply with the 2008 heavy-duty gasoline standards. Thus, in developing costs for the
technology packages we expect to be used, we started with the technology packages assumed to
be implemented on HD gasoline vehicles and engines to meet the 2005 standards. Table 5.B-1
shows both the expected 2005 technology packages, the baseline for this analysis, and the
expected 2008 technology packages for both complete and incomplete gasoline vehicles. The
expected technologies for 2008 are consistent between vehicles and engines; we make this
assumption based on our belief that the standards for vehicles and engines are equivalently
stringent.
This table only shows the technologies which are expected to change in some way or to
be applied in different percentages to meet the 2008 standards. A technology like sequential
multi-port fuel injection, while important to meeting the new standards, is expected on 100
percent of the 2005 vehicles and engines, and its design is not expected to fundamentally change
for 2008. As a result, we expect no incremental changes or costs associated with that technology,
and it is not included in the table. However, the table does contain technologies we believe will
be more widely implemented, but which have no associated costs for their implementation. One
such example, spark retard on engine start up, is expected to be more widely implemented for the
2008 standards, but there are no costs associated with implementing that technology. Such
technologies are included in these tables for completeness, but do not appear in later tables
showing the incremental costs associated with the 2008 standards.
V-44
-------
Chapter V: Economic Impact
Table V.B-1. 2005 (Phase 1) and Expected 2008 (Phase 2) Technology Packages for
Heavy-Duty Gasoline Vehicles excluding Medium-Duty Passenger Vehicles
Technology
CatalystsA
Oxygen sensors8
EGR
Adaptive learning
Heat managed
exhaust0
Secondary air
injection with
closed-loop
control
Spark retard at
start-up
2005
Complete Vehicles
13% single underfloor
50% dual underfloor
37% dual close-
coupled with
dual underfloor
13% dual heated
87% four heated
85% - All electronic
80%
40%
30%
0%
2005
Incomplete Vehicles
(Engine-Based)
13% single underfloor
87% dual underfloor
13% triple heated
87% four heated
85% - All electronic
80%
0%
50%
0%
2008
Expected for Complete
and Incomplete
Vehicles
50% dual underfloor
50% dual close-
coupled with
dual underfloor
100% four heated
with two being
fast light-off
100%-- All electronic
100%
80%D
50%
100%
In addition to the change in catalyst configurations shown, we expect that catalyst washcoat and precious
metal compositions and loadings will change.
The estimated breakdown for 2005 reflects OBD requirements for all HDGEs. However, OBD is only
required on HDGEs under 14,000 Ibs GVWR (approximately 60 percent of HDGEs).
c May include air gaps, thin walls, low thermal capacity manifold, insulation, etc.
D 100 percent of those having dual underfloor catalysts, and 60 percent of those having dual close-coupled w/
dual underfloor catalysts.
V-45
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
3. Technology/Hardware Costs for Gasoline Vehicles and Engines
The following sections present the costs of the technologies we expect will be used to
comply with the 2008 standards. Because most heavy-duty gasoline manufacturers offer more
than one engine for their heavy-duty gasoline product line, cost estimates have been developed
for a standard engine size and a larger engine size.
a. Improved Catalysts and Catalyst Systems
Improvements in catalyst systems fall into two broad categories: changes in catalyst
system configuration and changes in the catalyst precious metal and washcoat compositions and
loadings. In addition to estimating costs for these improvements, we have estimated the
increased costs of substrates and packaging (cans) for the improved catalysts.
/'. Changes in Catalyst Configurations
For heavy-duty gasoline vehicles and engines, we expect there to be generally three
catalyst configurations for meeting the 2005 and 2008 standards — the single underfloor, the dual
underfloor, and the dual close-coupled combined with the dual underfloor. With the single
underfloor catalyst system, the exhaust streams from both banks of engine cylinders " Y" into a
single catalyst. With the dual underfloor catalyst system, each bank of engine cylinders exhausts
into its own catalyst. With a dual close-coupled catalyst system, each bank of engine cylinders
exhausts directly into a small, often called "pipe," catalyst, and then into a dual underfloor main
catalyst system.
For 2005, we estimate that: 13 percent of vehicles will employ a single underfloor
catalyst; 50 percent of vehicles will employ dual underfloor catalysts; and, 37 percent of vehicles
will employ dual close-coupled with dual underfloor catalysts. For 2008, we expect that 50
percent of vehicles will employ dual underfloor with the remaining 50 percent employing dual
close-coupled catalysts with a dual underfloor. For engine based systems in 2005, we estimate
that: 13 percent of engines will employ a single underfloor catalyst; and, 87 percent will employ
dual underfloor catalysts. For 2008, we expect that engines will employ the same configurations
as outlined above for vehicles. We believe these vehicle and engine catalyst configuration
estimates to be reasonable given the estimated catalyst configuration employment in our Tier 2
analysis for MDPVs (80 percent with dual close-coupled and either single or dual underfloor
configurations), and some previously done Arcadis estimates for standards similar to our 2008
standards.27
V-46
-------
Chapter V: Economic Impact
/'/'. Changes in Catalyst Volumes and Precious Metal Loadings
The catalyst configuration changes and associated costs discussed above do not include
changes in the precious metal and washcoat compositions and loadings. Gasoline vehicle
catalysts have typically used some combination of platinum (Pt), palladium (Pd) and rhodium
(Rh). These precious metals, or platinum group metals (PGM), account for a significant portion
of the catalyst cost. Historically, a Pt/Rh combination has been used, but Pd has been seeing
increased use in recent years. Pd is more thermally stable than Pt and Rh, which makes it a good
choice for close-coupled catalysts, which are typically 100 percent Pd, where much higher
temperatures are experienced. For 2005, we estimate a Pt/Pd/Rh ratio of 0/10/1 applied at a
PGM loading of 4 grams/liter (g/L) for vehicles and 4.5 g/L for engines. For 2008, we estimate
that the ratio will change to 1/14/1, consistent with Tier 2, at a loading of 5 g/L.28
We have also estimated that catalyst volumes will increase. For 2005, we assume catalyst
volumes will be 4.8 liters for the standard engines and 5.8 liters for the larger engines. Because
the 2008 standards are more stringent, we expect that catalyst volumes will need to increase to
5.2 liters and 6.4 liters, respectively. In our Tier 2 analysis, we assumed that catalyst volumes
would increase to equal engine displacement volume; however, we assumed no increase in
precious metal loading.6 While the catalyst volumes we are assuming for 2008 may be low for
some applications and high for others (2000 model year certified engine displacements ranged
from 4.2 L to 8.0 L), we believe that we have chosen the appropriate middle ground of likely
catalyst volumes.
The estimated costs associated with increased use of precious metals are summarized in
Table V.B-2.
e We assume a higher precious metal loading than our recent Tier 2 analysis because heavy-duty vehicles,
by definition, undergo more rigorous operation during normal use. Therefore, more precious metals would probably
be required to maintain acceptable emissions durability characteristics.
V-47
-------
Table V.B-2. Costs Associated with the Increased Use of Precious Metals
Vehicles
Standard
Engine
Larger
Engine
Projected
2005 Projected
Catalyst 2008
Volume Catalyst
(L) Volume (L)
4.8 5.2
5.8 6.4
2005 2008
Catalyst Catslyst
Loading Loading 2005 2008
(g/L) (g/L) Pt/Pd/Rh Pt/Pd/Rh
4 5 0/10/1 1/14/1
4 5 0/10/1 1/14/1
2005 R 2005 Pd 2005 Rh 2008 R 2008 Pd 2008 Rh Increased Increased Increased
(g) (g) (g) (g) (g) (g) Pt(g) Pd(g) Rh(g)
0.000 17.455 1.745 1.625 22.750 1.625 1.625 5.295 -0.120
0.000 21.091 2.109 2.000 28.000 2.000 2.000 6.909 -0.109
2008
2005 PGM
PGM Cost
Cost ($) ($)
267.60 352.17
323.35 433.44
Engines
Standard
Engine
Larger
Engine
Projected
2005 Projected
Catalyst 2008
Volume Catalyst
(L) Volume (L)
4.8 5.2
5.8 6.4
2005 2008
Catalyst Catslyst
Loading Loading 2005 2008
(g/L) (g/L) Pt/Pd/Rh Pt/Pd/Rh
4.5 5 0/10/1 1/14/1
4.5 5 0/10/1 1/14/1
2005 R 2005 Pd 2005 Rh 2008 R 2008 Pd 2008 Rh Increased Increased Increased
(g) (g) (g) (g) (g) (g) Pt(g) Pd(g) Rh(g)
0.000 19.636 1.964 1.625 22.750 1.625 1.625 3.114 -0.339
0.000 23.727 2.373 2.000 28.000 2.000 2.000 4.273 -0.373
2008
2005 PGM
PGM Cost
Cost ($) ($)
301.05 352.17
363.77 433.44
Precious Metal Costs (9/29/99)
$/Troy Oz $/gram
Platinum 412 13.25
Paladium 390 12.54
Rhodium 868 27.91
V-48
-------
Chapter V: Economic Impact
/'/'/'. Changes in Catalyst Washcoat
In addition to the changes to precious metals just discussed, we expect that the 2008
standards will also result in changes to the catalyst washcoat compositions and loadings. Current
washcoats are typically a combination of a cerium oxide blend (ceria) and aluminum oxide
(alumina). Current ratios of these two components range from 75 percent ceria/25 percent
alumina to 100 percent alumina. Of the two common washcoat components, ceria is more
thermally stable and, thus, is expected in higher concentrations in close-coupled catalysts. We
assume that a 75/25 ratio of ceria to alumina will be used to comply with the 2005 vehicle-based
standards and that an even higher 80/20 ratio of ceria to alumina will be used to comply with the
engine-based standards. For 2008, we are assuming that all washcoats will use an 80/20 ratio of
ceria to alumina.
Current washcoat loadings range from 160 to 220 g/L of catalyst substrate volume. For
2005, we assume an average loading of 190 g/L for vehicle-based systems, and 220 g/L for
engine-based systems. For 2008, we are assuming a loading of 220 g/L for all substrates. In
addition, we expect that a new technique of layering the washcoat and precious metals will be
employed. Currently, the precious metals and washcoat are applied to the catalyst substrate in a
single slurry. Under the layering approach, there is a separate slurry for each precious metal,
with the second slurry being applied after the first dries. This process allows for more reaction
surface area, resulting in a more efficient catalyst.
iv. Catalyst Substrates
The substrate that the precious metals and washcoat are affixed to are typically ceramic
substrates of 400 cells per inch. Increasing efforts are going into developing metallic substrates,
which offer better temperature and vibration stability, as well as requiring less precious metal
loading to achieve the same emission benefits. Since the increased costs of the metal substrates
will tend to cancel out any savings in precious metal costs, we assumed that the current ceramic
substrate would continue to be used to comply with the 2005 standards. We are assuming the
same for the 2008 standards. The following linear relationship has been shown to be accurate for
ceramic substrates sized from 0.5 L to 4 L:29
C = $4.67V+$1.50
where:
C = cost to the vehicle manufacturer from the substrate supplier
V = substrate volume in liters
We are including an increased substrate cost due to the larger expected catalyst volumes; larger
catalysts will need larger substrates. Generally, catalyst substrates for heavy-duty gasoline
V-49
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
vehicles and engines are manufactured in bricks no larger than 2.5 L, with a catalyst of greater
than 2.5 L being comprised of more than one brick.
v. Catalyst Packaging
The final cost component of the catalyst system is the catalyst can. The catalyst substrate
is typically packaged in a can made of 409 stainless steel and around 0.12 centimeters thick (18
gauge). The increased catalyst volumes expected for 2008 model year catalysts will result in
more stainless steel and, therefore, more cost. The cost of the can is a very small portion of the
overall catalyst cost.
vi. Summary of Catalyst Costs
Table V.B-3 shows our estimates of the total catalyst system cost for each of the three
configurations previously discussed for the 2005 and 2008 standards. This table includes catalyst
costs for standard size and larger size engines for applications certified to the vehicle or the
engine standards. The Pt/Pd/Rh costs are taken from Table V.B-2 and do not have a supplier
markup applied because we have been informed that the vehicle manufacturer purchases the
precious metals and provides them to their catalyst supplier. Included in the table are
incremental costs for ease of comparison. No costs are shown for a single underfloor catalyst
system for 2008 because we do not expect any such applications in 2008.
V-50
-------
Chapter V: Economic Impact
Table V.B-3. Costs Associated with Various Catalyst Configurations
Single Underfloor Catalyst System
Catalyst Volume (liters)
Substrate*
Washcoat**
Pt/Pd/Rh
Can (18 gauge 409 SS)**
Total Material Cost
Labor
Labor Overhead @ 40%
Supplier Markup @ 29% ***
Manufacturer Cost
Manufacturer Carrying Cost @ 4%
Total Cost to Dealer
Incremental Cost
Complete \
2005 Vehicle
Standard Larger
4.8 5.8
$25 $31
$18 $22
$268 $323
$5 $5
$321 $387
$4 $4
$2 $2
$8 $9
$335 $402
$13 $16
$348 $418
/ehicles
2008 Vehicle
Standard Larger
n/a n/a
n/a n/a
Incomplete^
2005 Engine
Standard Larger
4.8 5.8
$25 $31
$22 $26
$301 $364
$5 $5
$358 $431
$6 $6
$2 $2
$10 $11
$377 $451
$15 $18
$392 $469
i/ehicles
2008 Engine
Standard Larger
n/a n/a
n/a n/a
Dual Underfloor Catalyst System
Catalyst Volume (liters)
Substrate*
Washcoat**
Pt/Pd/Rh
Can (18 gauge 409 SS)**
Total Material Cost
Labor
Labor Overhead @ 40%
Supplier Markup @ 29% ***
Manufacturer Cost
Manufacturer Carrying Cost @ 4%
Total Cost to Dealer
Incremental Cost
Complete \
2005 Vehicle
Standard Larger
4.8 5.8
$25 $31
$18 $22
$268 $323
$5 $6
$321 $388
$7 $8
$3 $3
$10 $11
$340 $410
$14 $16
$354 $427
/ehicles
2008 Vehicle
Standard Larger
5.2 6.4
$27 $34
$24 $29
$352 $433
$6 $7
$415 $510
$11 $13
$4 $5
$13 $16
$443 $543
$18 $22
$461 $565
$107 $139
Incomplete^
2005 Engine
Standard Larger
4.8 5.8
$25 $31
$22 $26
$301 $364
$5 $6
$358 $432
$11 $12
$4 $5
$12 $14
$386 $463
$15 $19
$401 $482
i/ehicles
2008 Engine
Standard Larger
5.2 6.4
$27 $34
$24 $29
$352 $433
$6 $7
$415 $510
$11 $13
$4 $5
$13 $16
$443 $543
$18 $22
$461 $565
$60 $84
Dual Close-coupled with Dual Unde
Catalyst Volume (liters)
Substrate****
Washcoat**
Pt/Pd/Rh
Can (18 gauge 409 SS)**
Total Material Cost
Labor
Labor Overhead @ 40%
Supplier Markup @ 29% ***
Manufacturer Cost
Manufacturer Carrying Cost @ 4%
Total Cost to Dealer
Incremental Cost
>rfloor Catalyst Syste
Complete \
2005 Vehicle
Standard Larger
4.8 5.8
$28 $33
$19 $23
$268 $323
$6 $7
$325 $392
$14 $15
$6 $6
$13 $15
$358 $428
$14 $17
$372 $445
m
/ehicles
2008 Vehicle
Standard Larger
5.2 6.4
$30 $36
$24 $29
$352 $433
$7 $8
$418 $513
$18 $20
$7 $8
$16 $19
$460 $560
$18 $22
$478 $582
$106 $137
Incomplete^
2005 Engine
Standard Larger
4.8 5.8
$28 $33
$19 $23
$301 $364
$7 $8
$360 $434
$18 $20
$7 $8
$15 $17
$400 $479
$16 $19
$416 $498
i/ehicles
2008 Engine
Standard Larger
5.2 6.4
$30 $36
$24 $29
$352 $433
$7 $8
$418 $513
$18 $20
$7 $8
$16 $19
$460 $560
$18 $22
$478 $582
$62 $84
*2.5L bricks; use C=$4.67V+$1.50 (Arcadis, 9/30/99) with the $1.50 applied per2.5L brick (Note: C is cost to mfr, thus not marked up in tables).
"Baseline from 2005 FRM RIA; 2008 from Arcadis 9/30/98.
***Not applied to precious metals or Substrate (substrate costs already include supplier markup).
****From 2005 FRM RIA and Arcadis, 9/30/98.
V-51
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
b. Oxygen Sensors
Largely because we expect catalyst configurations to change, we expect oxygen sensor
usage to change. Oxygen sensors are used both for fuel control and for OBD catalyst monitoring.
Therefore, different catalyst configurations would likely result in different oxygen sensor usage.
For 2005, we assume that 13 percent of heavy-duty gasoline vehicles and engines will employ
dual heated oxygen sensors, and 87 percent will employ four heated oxygen sensors. For 2008,
we assume that all vehicles and engines will use four heated oxygen sensors, with two of those
being fast light-off sensors for better cold start performance. We have estimated the cost of a
heated oxygen sensor at $20 per sensor, and a fast light-off sensor at $28 per sensor.
c. Exhaust Gas Recirculation (EGR)
Electronically controlled EGR is currently used on about 85 percent of non-California
gasoline heavy-duty vehicles. The percentage of the fleet with EGR is not expected to change as
a result of the 2005 standards. For 2008, we assume that 100 percent of vehicles and engines
will use electronically controlled EGR. In addition, some minor changes in control algorithms
may be necessary to improve upon EGR performance. These changes are expected to cost from
$5 to $12 per vehicle. For this analysis, we have used a cost of $10 per vehicle, applied only to
those 15 percent adding EGR for 2008.
d. Secondary Air Injection with Closed Loop Control
The hardware cost for vehicles which use secondary air injection to reduce HC and CO
emissions is estimated to be about $65 per vehicle. For 2005, we estimate a secondary air
injection usage rate of 30 percent on vehicles and 50 percent on engines. For 2008, we estimate
that 50 percent of vehicles will use secondary air injection, while the percentage of engines using
it will remain at 50 percent.
e. Exhaust Systems
We expect that heat managed exhaust systems will be used on some applications to
improve catalyst light-off time. Heat managed exhaust systems can include any combination of
thin walled components or otherwise low thermal-capacity components, air gapped components,
insulation, etc. We estimate that such systems will cost $40 per vehicle when they are used. For
2005, we estimate that they will be used on 40 percent of the vehicles, and none of the engines.
For 2008, we estimate that they will be used on 60 percent of vehicles and engines having a dual
close-coupled with a dual underfloor catalyst system, and 100 percent of vehicles and engines
having only a dual underfloor catalyst system.
V-52
-------
Chapter V: Economic Impact
f. Evaporative Emission Control Systems
There are two approaches to reducing evaporative emissions for a given fuel. One is to
minimize the potential for permeation and leakage by reducing the number of hoses, fittings and
connections. The second is to use less permeable hoses and lower loss fittings and connections.
Manufacturers are already employing both approaches. The 2008 evaporative emission 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.
In our proposal, we estimated the cost associated with our new evaporative standards at
$4 per vehicle. However, we received comments that our $4 per vehicle cost was not appropriate
for heavy-duty gasoline vehicles. Those comments suggested the cost would be as high as $32 to
$45 per vehicle, with claims that a new canister array, a returnless fuel system, an upgrade of fuel
system materials, and possible air inlet control measures would be needed.
The $4 estimate used in our proposal was developed for light-duty applications under our
Tier 2 cost analysis.30 Given that the Tier 2 estimate was for light-duty applications, it may
represent an under estimate of the cost for heavy-duty applications. Despite the fact that most
heavy-duty gasoline vehicles currently can meet the emission levels being finalized, we believe
that manufacturers will improve upon their designs so as to improve upon compliance margins.
We also believe that the $32 to $45 cost estimate supplied via comment represents a
worst case estimate rather than an average cost that can be applied across the HD gasoline fleet.
Therefore, we have increased our estimated cost from $4 to $21 to represent a conservative
estimate of the typical cost. The $21 estimate is a middle ground estimate appropriate for
application to the entire heavy-duty gasoline fleet. This seems reasonable considering the $4 cost
at the lighter end of the range where vehicles are similar to the Tier 2 MDPVs, and the $32 to
$45 cost for vehicles at the heavier end of the range where larger fuel tanks and longer fuel lines
present more challenge. This $21 cost is applied to all heavy-duty gasoline vehicles and engines
for the purpose of estimating the overall cost of the new standards regardless of their current
emission levels.
g. Summary of Technology/Hardware Costs
The costs associated with technology, or hardware, are summarized in Table V.B-4.
V-53
-------
Table V.B-4. Summary of Hardware Costs for the Proposed 2007 Heavy-Duty Gasoline Standards
Catalyst Costs
Oxygen Sensors
EGR
Heat Managed Exhaust*
Secondary Air Injection with
Closed Loop Control
Evap System Improvements
Total Dealer Cost
Dealer Carrying Cost @ 3%
Total Cost to the Consumer
Increased Cost to the
Consumer
2005 Vehicle
Standard Larger
System System
$360 $432
$75 $75
$9 $9
$16 $16
$20 $20
$0 $0
$479 $551
$14 $17
$493 $568
Complete Vehicles
2008 Vehicle
Standard Larger
System System
$470 $574
$96 $96
$10 $10
$32 $32
$33 $33
$21 $21
$661 $765
$20 $23
$681 $788
Increment
Standard Larger
System System
$110 $141
$21 $21
$2 $2
$16 $16
$13 $13
$21 $21
$183 $214
$188 $220
2005 Engine
Standard Larger
System System
$400 $480
$77 $77
$9 $9
$0 $0
$33 $33
$0 $0
$518 $598
$16 $18
$534 $616
Incomplete Vehicles
2008 Engine
Standard Larger
System System
$470 $574
$96 $96
$10 $10
$32 $32
$33 $33
$21 $21
$661 $765
$20 $23
$681 $788
Increment
Standard Larger
System System
$70 $94
$19 $19
$2 $2
$32 $32
$0 $0
$21 $21
$143 $167
$147 $172
*May include air gaps, thin walls, low thermal capacity manifold, insulation, etc.
Note: Some values may not add up precisely due to rounding.
V-54
-------
Chapter V: Economic Impact
As Table V.B-4 shows, the incremental technology costs for heavy-duty gasoline vehicles
and engines associated with the 2008 standards are $188 and $220 for standard and large sized
engines in vehicle-based applications, respectively, and $147 and $172 for standard and large
sized engines in engine-based applications, respectively.
Weighting these costs assuming a standard/large split of 75/25 percent, gives incremental
costs of $196 for complete vehicles and $153 for incomplete vehicles. For the long-term, there
are factors we believe are likely to reduce the costs to manufacturers. As noted below, 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 consistently
shown that as manufacturers gain experience in production, they are able to apply innovations to
simplify machining and assembly operations, use lower cost materials, and reduce the number or
complexity of component parts. These effects are often described as the manufacturing learning
curve as described in Chapter V.A.6 of this Regulatory Impact Analysis.
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. This
differs from the heavy-duty diesel cost analysis where we did apply the learning curve beyond the
third year. We applied the learning curve reduction only once for gasoline because we anticipate
that, for the most part, the 2008 heavy-duty 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 through factors consistent with the application of 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. Also, we have not applied the learning curve to evaporative emission control
system costs.
Therefore, as a result of the learning curve, the variable costs per vehicle, minus the
precious metal costs, would decrease by 20 percent beginning in the 2010 model year.
Thereafter, the incremental technology costs would fall to $179 and $138 for vehicles and
engines, respectively.
V-55
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
4. Heavy-Duty Gasoline Fixed Costs
The fixed costs are broken into four main components: research and development,
tooling, certification, and in-use testing. These costs are discussed individually in the following
sections.
a. R&D and Tooling Costs
The 2008 vehicle-based standards will essentially require the application of California
LEV-II and Tier 2 technology to heavy-duty gasoline vehicles nationally. Since this technology
is being developed in response to those rules, we are assuming that considerable carry-across will
occur from those R&D efforts to the heavy-duty gasoline systems. 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 2008 standards. It also includes engine
modifications where necessary and air/fuel ratio calibration work. Manufacturers will take
differing approaches in their research programs. In our Tier 2 analysis, we assumed an R&D cost
of $5 million per vehicle line estimating that this would cover about 25 engineering staff person
years and about 20 development vehicles/ We estimated such a large R&D effort because
calibration and system optimization was expected to be a critical part of the effort to meet the
Tier 2 standards. However, we believe those R&D costs are likely overstated for purposes here
because the projection ignores the carryover of knowledge from the first vehicle lines designed to
meet the standard to others phased-in later. For this heavy-duty gasoline analysis, we assume an
R&D cost of $2.5 million per line due to the carryover from Tier 2 and LEV II R&D efforts.
According to 2000 model year certification data, there is one engine family certified as an
incomplete vehicle federally with no corresponding engine certified for sale in California. We
have assumed that engine will require R&D efforts to comply with today's proposed standards.
We have also assumed that four other engines (those having six liters or more displacement
typically used in larger applications) currently being certified to engine standards will continue to
be so certified and will require R&D efforts to comply with today's engine standards. That gives
four more engines requiring R&D efforts, for a total of five engines to which we have applied the
$2.5 million R&D cost.
In our Tier 2 analysis and our proposal, we estimated tooling costs at $2 million per line.
Tooling costs include facilities modifications necessary to produce and assemble components and
vehicles meeting the new standards. We believe that this is a reasonable estimate based on
engineering judgement and review of previous estimates of tooling costs for emissions control
components.31 We have applied tooling costs only to those engines requiring R&D efforts.
f This estimate is based on staff cost of $60 per hour and development vehicle cost of $100,000 per vehicle line.
V-56
-------
Chapter V: Economic Impact
R&D costs are spread out evenly over the three year period prior to the first year of
implementation and grown at a seven percent rate. Tooling costs are assumed to occur one year
prior to implementation and are grown for one year at a seven percent rate. These costs are then
amortized over a five year period following implementation, again at a seven percent rate. This
results in R&D and tooling costs of just over $9 per complete vehicle and $23 per incomplete
vehicle. The costs are higher for the incomplete vehicles because of the lower sales over which
to spread the same total costs as estimated for complete vehicles. These costs become zero five
years after implementation because we assume the costs will have been recovered.
b. Certification Costs
Manufacturers incur an annual cost as part of certification and compliance and would
incur those costs without any change to the standards. However, we 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 will have to be generated for the
new 2008 model year vehicles rather than being carried-over from previous model years.
Therefore, we believe it is appropriate to include the cost of generating new emissions test and
durability data. We have estimated certification costs at $30,000 per engine family.32 This
estimate does not account for the ability of manufacturers, in most cases, to carry-over
certification data from California certified systems. Such a practice would lower certification
costs.
We have applied the certification cost to the 17 complete and 26 incomplete engine
families, the number certified for the 2000 model year. Certification costs would be incurred, on
average, one year before the start of production. Thus, this cost is increased at a rate of seven
percent for one year and applied to the appropriate vehicle certifications and engine
certifications. The costs are then amortized over five years and divided by the appropriate
complete and incomplete sales projections. This results in projected per-vehicle certification
costs of $0.42 for complete vehicle configurations and $1.59 for incomplete vehicle
configurations during the first five years of the program. After five years, the certification costs
become zero as manufacturers fall into their normal practice of carrying-over data from one year
to the next.
5. Summary of Heavy-Duty Gasoline Costs
Table V.B-5 contains a summary of per-vehicle costs associated with the 2008 standards
for heavy-duty gasoline vehicles and engines. The hardware cost components include a part or
emission control system supplier markup of 29 percent, and both manufacturer and dealer
carrying costs of four percent and three percent, respectively. The costs are presented as
incremental cost increases from the 2005 system costs.
V-57
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.B-5. Summary of Incremental Costs to Meet the 2008 Heavy-Duty
Gasoline Emission Standards
Near
Term
Long
Term
Technol ogy/Hardware
Fixed Costs
Incremental Cost
Technol ogy/Hardware
Fixed Costs
Incremental Cost
Complete
Vehicles
$196
$10
$206
$179
$0
$179
Incomplete
Vehicles
$153
$25
$178
$138
$0
$138
HDGVs
$184
$14
$198
$167
$0
$167
6. Total Nationwide Costs for 2008 Heavy-Duty Gasoline Vehicles
The above analyses developed incremental per vehicle manufacturer and consumer cost
estimates for heavy-duty gasoline vehicles designed to the new Phase 2 gasoline standards. 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 annualized costs to the nation for the
new Phase 2 gasoline standards. Table V.B-6 presents the results of this analysis.
To prepare these estimates, we projected sales for heavy-duty gasoline vehicles. We
estimated current vehicle sales based on 1996 sales data submitted by vehicle manufacturers as
part of certification. These sales correlated reasonably well with other available sales
information. We assumed a mix of 71 percent complete vehicles and 29 percent incomplete
vehicles based on these sales data, excluding an estimated 70,000 units counted in the Tier 2
analysis as medium-duty passenger vehicles. California sales were excluded from this analysis
because California emissions standards apply to those vehicles. We have projected vehicle sales
to grow two percent from 1996 through 2007, then at a constant number of vehicles (two percent
of 1996 sales) for each year thereafter. Table V.B-6 contains those sales projections.
V-58
-------
Table V.B-6. Estimated Annualized Nationwide Vehicle Costs Associated with the 2008
Heavy-Duty Gasoline Emission Standards
Fraction of
Fleet
Year Projected Sales Fixed Costs Complying Variable Costs Operating Costs Total Cost
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
2035
424,560
431,520
438,480
445,440
452,400
459,360
466,320
473,280
480,240
487,200
494,160
501,120
508,080
515,040
522,000
528,960
535,920
542,880
549,840
556,800
563,760
570,720
577,680
584,640
591,600
598,560
605,520
612,480
619,440
$0
$6,213,290
$6,213,290
$6,213,290
$6,213,290
$6,213,290
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0%
50%
100%
1 00%
100%
1 00%
100%
100%
1 00%
100%
1 00%
100%
1 00%
100%
1 00%
100%
1 00%
100%
1 00%
1 00%
100%
1 00%
100%
1 00%
100%
1 00%
100%
1 00%
1 00%
$0
$39,635,728
$73,362,727
$74,527,215
$75,691,703
$76,856,190
$78,020,678
$79,185,166
$80,349,654
$81,514,141
$82,678,629
$83,843,117
$85,007,604
$86,172,092
$87,336,580
$88,501,068
$89,665,555
$90,830,043
$91,994,531
$93,159,019
$94,323,506
$95,487,994
$96,652,482
$97,816,970
$98,981,457
$100,145,945
$101,310,433
$102,474,920
$103,639,408
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$45,849,018
$79,576,017
$80,740,505
$81,904,993
$83,069,481
$78,020,678
$79,185,166
$80,349,654
$81,514,141
$82,678,629
$83,843,117
$85,007,604
$86,172,092
$87,336,580
$88,501,068
$89,665,555
$90,830,043
$91,994,531
$93,159,019
$94,323,506
$95,487,994
$96,652,482
$97,816,970
$98,981,457
$100,145,945
$101,310,433
$102,474,920
$103,639,408
Per
Vehicle
Cost
$0
$213
$181
$181
$181
$181
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
$167
V-59
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
As shown in Table V.B-6, we have projected a total cost starting at $46 million in 2008
and peaking at $83 million in 2012. In 2013, the costs decrease due to the elimination of fixed
costs. Thereafter, costs gradually increase with projected sales. Operating costs are $0 because
the technologies expected should have no impact on fuel economy or maintenance costs. 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. These
costs include exhaust and improved evaporative control systems. These estimates do not include
costs due to improved fuel quality, which were presented in the Tier 2 Regulatory Impact
Analysis for gasoline.33
Table V.B-7 shows the non-annualized costs.
V-60
-------
Chapter V: Economic Impact
Table V.B-7. Estimated Non-Annualized Nationwide Vehicle Costs Associated with the
2008 Heavy-Duty Gasoline Emission Standards
Fraction of
Fleet
Year Projected Sales Fixed Costs Complying Variable Costs Operating Costs Total Cost
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
2035
403,680
410,640
417,600
424,560
431,520
438,480
445,440
452,400
459,360
466,320
473,280
480,240
487,200
494,160
501,120
508,080
515,040
522,000
528,960
535,920
542,880
549,840
556,800
563,760
570,720
577,680
584,640
591,600
598,560
605,520
612,480
619,440
$0
$4,166,667
$4,166,667
$14,946,667
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
0%
0%
0%
0%
50%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
$0
$0
$0
$0
$39,635,728
$73,362,727
$74,527,215
$75,691,703
$76,856,190
$78,020,678
$79,185,166
$80,349,654
$81,514,141
$82,678,629
$83,843,117
$85,007,604
$86,172,092
$87,336,580
$88,501,068
$89,665,555
$90,830,043
$91,994,531
$93,159,019
$94,323,506
$95,487,994
$96,652,482
$97,816,970
$98,981,457
$100,145,945
$101,310,433
$102,474,920
$103,639,408
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$4,166,667
$4,166,667
$14,946,667
$39,635,728
$73,362,727
$74,527,215
$75,691,703
$76,856,190
$78,020,678
$79,185,166
$80,349,654
$81,514,141
$82,678,629
$83,843,117
$85,007,604
$86,172,092
$87,336,580
$88,501,068
$89,665,555
$90,830,043
$91,994,531
$93,159,019
$94,323,506
$95,487,994
$96,652,482
$97,816,970
$98,981,457
$100,145,945
$101,310,433
$102,474,920
$103,639,408
V-61
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
C. Diesel Fuel Costs
In this section, we first lay out the methodology for our analysis of the cost of
desulfurizing highway diesel fuel. Then we present the estimated cost of desulfurizing highway
diesel fuel.
1. Methodology
a. Overview
For the proposed rule, we estimated the cost of desulfurizing highway diesel fuel to meet
a 15 ppm cap sulfur standard based on a characteristic refinery, which was sized to represent the
average cost for all U.S. refineries. Although we felt confident in the cost estimates made with
this model, the analysis did not allow us to adequately address certain issues, particularly the
comments which we received concerning the future supply of highway diesel fuel. For this final
rule, we expanded upon our analysis to allow us to better understand the range of situations faced
by refiners to supply highway diesel fuel. This section presents an overview of our expanded
cost analysis.
Our cost estimate for desulfurizing diesel fuel is based on hydrotreating process
operations and capital cost information received from two licensors of conventional distillate8
desulfurization technology. In addition, information obtained from two other vendors of diesel
desulfurization technology further corroborated the information provided by the first two
vendors. The costs for desulfurizing diesel fuel were estimated for each refinery in the country
which was producing highway diesel fuel during 1998 and 1999. Each refinery's production
volumes were projected to 2006 using a ratio of the projected consumption of highway diesel
fuel in 2006 by EIA versus the production in 1998 and 1999. We presume that each refinery
producing highway diesel fuel starts with a highway diesel fuel sulfur level of about 340 ppm and
reduces it to between 5 to 10 ppm, or 7 ppm on average. We believe that refiners would have to
desulfurize their diesel fuel to about 7 ppm to reliably and continually meet the proposed 15 ppm
cap standard. Construction and operating cost factors and utility costs for each refinery are based
on values calculated for each PADD and are applied to all the refineries operating in that PADD.
For each refinery we estimated the fraction of straight run distillate, light cycle oil (LCO), and
other cracked stocks (coker, visbreaker, thermal cracked) in the highway diesel fuel, and the cost
to desulfurize each of those stocks. The average desulfurization cost for each refinery was based
on the volume-weighted average of desulfurizing each of those blendstocks. We based our cost
estimate on the premise that the refining industry will be able to revamp 80 percent of the
g Distillate refers to a broad category of fuels falling into a specific boiling range. Distillate fuels have a
heavier molecular weight and therefore boil at higher temperatures than gasoline. Distillate includes diesel fuel, jet
fuels, kerosene and home heating oil.
V-62
-------
Chapter V: Economic Impact
existing diesel hydrotreater capacity, while the other 20 percent will have to install brand new
"grassroots" units. Since we do not know which refineries would install revamps units and
which would install grassroots units, we calculated the revamp and grassroots cost for each
refinery, and based 80 percent of the cost on the revamped cost, and 20 percent on the grassroots
cost. For determining the grassroots cost of a refinery currently producing highway diesel fuel,
we used the operating cost of a revamped unit and the capital cost of a grassroots unit. Using the
operating cost of a revamped unit is appropriate because that refinery is incurring operating cost
now for meeting the current 500 ppm sulfur highway diesel fuel standard.
The final rule provides the refining industry a temporary compliance option which
refiners them to continue selling up to about 20 percent of the highway diesel pool at this higher
sulfur level until 2010, at which point all highway diesel fuel must meet the 15 ppm cap sulfur
standard. We estimated the cost of refiners using this option based on the assumption that the
refineries which can meet the 15ppm cap sulfur standard at the lowest cost will meet the
requirements in 2006. The balance of refineries are presumed hold off making their investments
to meet the 15 ppm sulfur standard until 2010.
We received a number of comments from the refining industry which suggested that some
refiners may choose to partially or completely leave the highway diesel fuel market which could
result in a shortfall in highway diesel fuel supply. Arguably, the refiners which are most likely to
exit the highway diesel market would be those which are facing the highest cost to desulfurize
their highway diesel fuel. Those most likely to maintain highway production, or even expand
production to fill market demand would be the lowest cost producers. In some cases a portion of
the market demand for 15 ppm sulfur highway diesel fuel could be met by today's predominant
or exclusive producers of nonhighway diesel fuel. To understand ths possibility, we assessed the
cost to offhighway diesel fuel producers of desulfurizing their offhighway diesel fuel to make up
a potential supply shortfall in highway diesel fuel. In fact, current highway diesel fuel producers
which decide they must install a grassroots unit to meet the 15 ppm cap standard would have no
advantage over current nonhighway producers producing a similar volume of fuel and processing
a similar type of crude oil. The cost analysis allowing for such production shifts between diesel
fuel markets by refineries is presented as a sensitivity analysis further below.
Finally, the cost of desulfurizing diesel fuel to meet the 15 ppm cap standard was
estimated by several other entities. Mathpro provided estimates for the Engine Manufacturers
Association. The National Petroleum Council used the Mathpro estimates and adjusted them
based on some concerns which they had on costs. The American Petroleum Institute funded a
study by Charles River and Baker and O'Brien to study this issue. Finally, the Department of
Energy hired Ensys to estimate the cost of meeting the 15 ppm cap standard. These various cost
studies are summarized at the end of this section and the cost estimates are compared to our costs
if an appropriate comparison can be made.
V-63
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The analyses and discussion associated with these issues is contained in the following
sections.
b. Derivation of the Fraction of LCO and other Cracked Blendstocks in
Highway Diesel Fuel for Each Refinery
In Chapter IV, we established that an important challenge for refiners in meeting the
proposed 15 ppm sulfur cap was the LCO fraction of their highway diesel fuel pool. Thus, the
first step in segregating refineries according to the difficulty of desulfurization is to estimate each
refinery's LCO fraction of their highway diesel fuel pool. This data is generally not publically
available, so we estimated these fractions from other sources of information.
First, estimates of the volumes of high and low sulfur distillate produced in the last half
of 1998 and the first half of 1999 by each U.S. refinery were obtained from the Energy
Information Administration (EIA). According to EIA, U.S. refiners produce a total of 49 billion
gallons of distillate per year, with 32 billion gallons (about 65 percent) of that being low sulfur
diesel fuel. We determined that highway diesel fuel is produced by 121h different refineries
throughout the U.S.
Second, we estimated the volume of LCO produced by each refinery using information
from the Oil and Gas Journal (OGJ).34 The OGJ publishes information on the capacity of major
processing units for each refinery in the country, including the FCC unit. We assumed that FCC
units operate at 90 percent of capacity, which is consistent with the API/NPRA survey of
Refining Operations and Product Quality.35 We first assumed that 17 percent of the feedstock
volume to the FCC unit is converted into LCO based on confidential information shared with
EPA by a vendor of fluidized cat cracker units. Next we assumed that refineries with distillate
hydrocrackers send their LCO to the distillate hydrocracker and convert it to gasoline.
Furthermore, FCC feed hydrotreaters can affect the sulfur level and the treatability of
light cycle oil. FCC feed hydrotreaters hydrotreat the gasoil fed to the FCC unit, usually at a
pressure much higher than distillate hydrotreaters. The resulting cracked blendstock from the
FCC unit is much lower in sulfur, and, most important, some of the sterically hindered
compounds are desulfurized. However, only high pressure feed hydrotreaters (i.e., 1500 psi
units) can convert a significant portion of these sterically hindered compounds.36 We don't have
h There are four refineries in Alaska producing diesel fuel which is exempted from the current 500 ppm
sulfur cap standard for highway diesel fuel. Consequently, the diesel fuel they produce is used for both highway
and offhighway purposes without regard to the end use. Since only an estimated 5 percent of diesel fuel in Alaska
is consumed in highway applications, for our cost analysis we assumed only one would, in the future, produce
highway diesel fuel to supply demand. Thus, we also included that one refinery in this analysis of blendstock
quality.
V-64
-------
Chapter V: Economic Impact
any specific information on what fraction of these hydrotreaters are high pressure, however,
industry experts estimated that about 20 percent of the FCC feed hydrotreaters are high pressure,
with most or all of these being in California. Since we don't know which feed hydrotreaters are
high pressure, we conservatively presume that only the California feed hydrotreaters are high
pressure. Since most California refineries already have distillate hydrocrackers, the fact that they
have high pressure feed FCC hydrotreaters is a moot point and does not affect the fraction of
LCO of these refineries. Consequently, we have not made any adjustments in our cost
methodology to account for the presence of FCC feed hydrotreaters.
Based on these assumptions, we calculated the fraction of LCO to total distillate
production to be about 15 percent. To independently check this estimate, we compared our
estimate of the LCO fraction of total distillate production with that reported in the API/NPRA
survey. The API/NPRA survey shows that, on average for the U.S. refining industry as a whole,
light cycle oil comprises about 21 percent of number two distillate. For highway diesel fuel, the
API/NPRA Survey shows the percentage of LCO to the total pool of highway diesel fuel to be 22
percent, and both of these percentages are much higher than our initial estimate. In our distillate
production model, if we increase the fraction of FCC feedstock converted to LCO from 17
percent to 25 percent, our model matches the fraction of LCO to distillate shown by the
API/NPRA survey for the highway diesel pool. Thus, we used 25 percent for the ratio of LCO
product to FCC feed in our refinery model.
Applying these assumptions using the EIA and OGJ information, we calculated the
fraction of LCO relative to the total distillate production for each refinery. We then categorized
the refineries based on the fraction of their distillate pool which is LCO at 5 or 10 percent
intervals from 0 to 60 percent. The distribution of refineries by fraction of LCO is summarized
in Table V.C-1.
V-65
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-1. Presence of Light Cycle Oil in the Distillate of U.S. Refineries Producing
Highway Diesel Fuel
Number of
Refineries
Cumulative
Percentage of
US Highway
Diesel Volume
Percentage ofLCO in the Distillate Pool
0%
49
27
<10%
51
29
<15%
54
32
<20%
59
36
<25%
71
47
<30%
93
77
<40%
113
95
<50%
116
98
<60%
118
99
In Table V.C-1, our analysis shows that distillate contains anywhere from no LCO to 60
percent LCO. Our analysis also shows that 49 U.S. refineries which produce about 27 percent of
the distillate in the U.S. blend no LCO into this distillate, while the distillate from the remaining
72 refineries averages about about 28 percent LCO by volume. This is important because of the
large difference in fractions of LCO in the highway diesel pool for the U.S refining industry.
Refineries which blend no LCO into their distillate pool do so because they either do not have an
FCC unit, or because they have a distillate hydrocracker which is used to "upgrade" their LCO to
gasoline. Refineries with LCO in their distillate have an FCC unit, and they likely do not have a
hydrocracker. The refineries in both groups have distillate hydrotreaters for producing
onhighway diesel fuel for meeting the current 500 ppm cap standard.
We also estimated the fraction of other cracked stocks, which includes coker, thermally
cracked and visbreaker distillate, in each refinery's distillate fuel. We first estimated the volume
of these other cracked stocks produced by each refinery using information from the Oil and Gas
Journal (OGJ). Similar to how we calculated the fraction of LCO, we assumed that delayed and
fluid cokers, visbreakers, and thermal crackers all operate at 90 percent of capacity. Based on a
conversation with a refining industry consultant, we assumed that 30 percent of delayed coker
and 15 percent of the other units' product is distillate blended into the distillate pool. Unlike
LCO, we do not assume that refineries with hydrocrackers send their other cracked stocks to the
hydrocracker for conversion to gasoline. While most refineries probably do not send their other
cracked stocks to their hydrocracker, it is also likely that some do for at least some of their other
cracked stocks, so our assumption is probably somewhat conservative. After analyzing each
refinery's other cracked stock distillate production and averaging that production over the entire
industry, we estimate that about 8 percent of the entire highway diesel fuel volume is comprised
of these other cracked stocks. This value agrees well with the API/NPRA survey.37
V-66
-------
Chapter V: Economic Impact
Table V.C-2. Presence of Other Cracked Blendstocks in the Distillate of U.S. Refineries
Producing Highway Diesel Fuel
Number of
Refineries
Cumulative
Percentage of US
Highway Diesel
Volume
Percentage of Other Cracked Stocks in the Distillate Pool
0%
89
55
<10%
95
67
<15%
103
77
<20%
111
88
<25%
112
89
<30%
118
95
<40%
120
100
<50%
121
100
As depicted in Table V.C-2, our analysis shows that over half of distillate fuel in the U.S,
which is produced by 89 refineries, does not contain other cracked stocks from cokers,
visbreakers and thermal crackers. Of the refineries which are projected to blend other cracked
stocks into their distillate pool, we estimate that, on average, the distillate from these refineries
contains approximately 18 percent of other cracked stocks.
Next we set out to determine the cost of desulfurizing highway diesel fuel. We met with
Criterion Catalyst/ABB Lummus, UOP, Akzo Nobel and Haldor Topsoe and a number of
refiners. One of these vendors provided diesel desulfurization unit operation and capital cost
information for different levels of LCO in diesel fuel, which included none, 15 percent, 23
percent and 30 percent, and varying amounts of coker distillate. Another vendor provided
significant cost information for 25 percent LCO in diesel fuel, and 10 percent coker distillate. In
addition, information from the other two vendors helped to corroborate the operating and cost
information obtained from the first two vendors. This information provided by these vendors
allowed us to estimate the cost of desulfurizing the different diesel fuel blendstocks.
The information provided by the vendors is based on typical diesel fuels, however, in
reality diesel fuel (especially LCO, and to a lesser degree other cracked stocks) varies in
desulfurization difficulty based on the amount of sterically hindered compounds present in the
fuel, which is determined by the endpoint of diesel fuel, and also by the type of crude oil being
refined. The vendors provided cost information based on diesel fuels with T-90 distillation
points which varied from 605 °F to 630 °F, which would roughly correspond to distillation
endpoints of 655 °F to 680 °F. These endpoints can be interpreted to mean that the diesel fuel
would, as explained in Chapter IV above, contain sterically hindered compounds. However, a
summertime diesel fuel survey for 1997 shows that the endpoint of highway diesel fuel varies
from 600 °F to 700 °F, thus the lighter diesel fuels would contain no sterically hindered
V-67
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
compounds, and the heavier diesel fuels would contain more.38 Our analysis attempts to capture
the cost for each refinery to produce highway diesel fuel which meets the 15ppm cap sulfur
standard, however, we do not have specific information for how the highway diesel endpoints
vary from refinery to refinery, or from season to season.
Similarly, we do not have information on what type of crude oil is being processed by
each refinery as the quality of crude oil being processed by a refinery affects the desulfurization
difficulty of the various diesel fuel blendstocks. For example, North Slope crude oil from Alaska
contains a higher fraction of aromatic compounds than most other crude oils.39 If the highway
diesel fuel produced from Alaskan crude oil has a high endpoint, the highway diesel fuel would
be expected to contain more sterically hindered compounds compared to another diesel fuel
produced from a lighter crude oil, such as Western Texas Intermediate, with the same endpoint
and the same mix of cracked stocks.
As discussed in Chapter IV, refiners which are producing their highway diesel fuel with a
higher endpoint and refining heavier, more aromatic crude oils, they are doing so with an
economic incentive. The economic incentive is that those heavier, more sour crude oils are 1 to
2 dollars per barrel less expensive than lighter, sweater crude oils. Also, if the heaviest fraction
of highway diesel fuel containing the sterically hindered compounds earns at least 10 dollars per
barrel (about 25 c/gal) more when it has been upgraded and blended into highway diesel fuel
instead of the most likely alternative, which is to be sold in the resid market.40 In sum, diesel fuel
processed by a particular refiner can either be easier or more difficult to treat than what we
estimate depending on how their diesel fuel endpoint compares to the average endpoint of the
industry, and depending on the crude oil used. For a nationwide analysis, it is appropriate to base
our cost analysis for each refinery on what we estimate would be typical or average qualities for
each diesel fuel blendstock. Some estimates of individual refinery costs will be high, others will
be low, but be representative on average.
c. Technology and Cost Inputs from Vendors
The most significant cost involved in meeting a more stringent diesel sulfur standard
would be the cost of constructing and operating the distillate desulfurization unit. For estimating
the cost of building and operating these units, we obtained detailed information on the raw
material and utility needs, the capital costs and the desulfurization capabilities from licensors of
two different desulfurization technologies.4142 43 Each vendor provided most of the information
needed to allow us to cost out a retrofit to an existing desulfurization unit, and also cost out the
building of a new desulfurization unit from grass roots. We also met with two other vendors of
desulfurization technology, though they did not provide enough information to develop an
independent cost estimate.
V-68
-------
Chapter V: Economic Impact
In addition to the information which we obtained directly from the vendors, we reviewed
the vendor submissions made to the National Petroleum Council (NPC) by Akzo Nobel,
Criterion, Haldor Topsoe, UOP and IFF.44 Of the five vendors which provided information to
the NPC; we met with all of them except IFF. These vendors provided information for
retrofiting existing diesel hydrotreaters and many of them also provided information on the
combined operations of the existing hydrotreater and the revamp together. The full set of
submissions made to the NPC allowed us to compare all these vendor's information to each other
on the same basis. With one exception, these submissions corroborated the costs we had
developed earlier. In one case, though, the vendor's information suggested that a significant
amount of hydrogen would be consumed to remove the sulfur, which would also cause a
significant increase in API gravity (the diesel fuel would be made less dense). However, the
other vendors' information indicated that the sulfur can be removed from diesel fuel without
dramatic differences in diesel fuel quality, and with only a modest amount of hydrogen
consumption. Thus, we based our estimate of hydrogen consumption on the estimates of
hydrogen consumption, as reflected by the majority of the vendors. Conversely, API has
indicated that they believe that very high hydrotreating pressures (e.g., 1200 psi or more) will be
necessary to meet a 15 ppm cap standard, although their contractor for their cost study indicated
that pressures under 1000 psi would be adequate. None of the vendors projected that pressures
more than 900 psi would be necessary and most of the vendors projected that 650 psi would be
sufficient. Likewise, a number of refiners have indicated that pressures well below 1000 psi
would be sufficient. Thus, we based our estimate of capital cost on two different vendor
submissions which were based on units operating at 650 and 900 psi pressure.
Since refineries already have a distillate hydrotreater in place to desulfurize highway
diesel fuel down to under 500 ppm, the vendors concluded that it would only be necessary to
retrofit an existing diesel hydrotreating unit with a number of different vessels, such as a reactor,
a hydrogen compressor, a recycle scrubber an interstage stripper and other associated process
hardware. Despite the fact that each vendor is basing their cost information on retrofits, the two
vendors who provided us information on our cost analysis, still differed in individual cost
elements due to differences in the capital equipment used, although the overall cost ended up
being roughly the same.
The differences in the estimated capital and operating costs between the two vendors is
largely due to the differences in technical approaches assumed by each vendor for meeting the
proposed diesel sulfur standards. One vendor, which we will call Vendor A,1 chose to estimate
1 Vendor A wished to keep its name confidential. For consistency in our tables we are labeling the second
vendor, UOP, as Vendor B.
V-69
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
operating and capital costs for a two-stage revamp, which is operated at a higher pressure.j Thus,
this vendor would recommend the use of a two stage unit right away instead of opting for other
subunits at the higher diesel fuel sulfur levels. The other vendor, which we will call Vendor B,
chose to estimate the operating and capital costs for a single stage revamp for moderate levels of
desulfurization, which included a larger reactor, hydrogen purification, a recycle gas scrubber,
and a color reactor to address the implications of increased reactor temperature. Then, to
desulfurize diesel fuel to under 10 ppm, Vendor B would recommend a two stage unit, but
without hydrogen purification and at lower temperature which negates the need to install a color
reactor. While there are substantial hardware differences between the two vendors for
desulfurizing diesel down to levels above 10 ppm, the differences between the vendors
diminishes with deeper desulfurization as both vendors use a two stage approach. We believe
that there are merits of using either approach and that both approaches would be used by different
refiners. Thus, we based our rule on the cost of both vendors representing both approaches and
we averaged them together. The technical approach generally used by each vendor to achieve
reduced diesel fuel sulfur levels is summarized in the following table. The vendors assumed that
the existing desulfurization unit in place would provide a number of hydrotreater subunits which
would save on both capital and operating costs for a one or two stage revamp compared to whole
new grassroots unit. These subunits include heat exchangers, a heater, a reactor filled with
catalyst, two or more vessels used for separating hydrogen and any light ends produced by
cracking during the desulfurization process, a compressor, and sometimes a scrubber. The
desulfurization subunits listed here are discussed in detail in the feasibility section contained in
Chapter IV.
1 Vendor A provided cost inputs for both low pressure and intermediate pressure units to NPC. The diesel
desulfurization costs were similar for each, which suggests that one approach does not have a predictable advantage
over the other, however, refinery configuration may provide an advantage of one approach over the other for each
individual refiner.
V-70
-------
Chapter V: Economic Impact
Table V.C-3. Technology Projected to be Used to Achieve Various Diesel Fuel Sulfur
Levels
Average Diesel
Fuel Sulfur Level
Vendor A
VendorB
30 ppm
Change to a more active catalyst
Install recycle gas scrubber
Modify compressor
Install a second reactor, high
pressure (900 psi)
Use existing hot oil separator for
interstage stripper
Change to a more active catalyst
Install a recycle gas scrubber
Purify make-up hydrogen
Install a second reactor (650 psi)
Increase temperature in the second
reactor and install a color reactor
10 ppm
Same as above
Use more catalyst
Increase the size of the second
reactor
Same as above
Use more catalyst
Increase the size of the second
reactor
<10ppm
Same as above
Increase catalyst volume further
Use an even larger second reactor
Raise temperature in the second
reactor
Same as above,
Install an interstage stripper, which
negates the need to purify
hydrogen and increase the reactor
bed temperature
Increase size of the second reactor
Increase catalyst volume
Prior to presenting the vendor inputs which allowed us to estimate the cost of meeting the
15 ppm cap standard, we will first qualify the information in terms of its perceived accuracy of
the actual cost of desulfurizing diesel fuel. We received several comments from refiners which
assert that the vendor costs are optimistic and need to be adjusted higher to better assess the
costs. While the vendors costs may be optimistic, we believe that there are a multitude of
reasons why the cost estimates should be optimistic.
First, capital costs can be lower than what the vendors project. Many refiners have used
reactors, compressors, and other vessels which can be employed in a new or revamped diesel
hydrotreating unit. We do not know to what extent that additional hydrotreating capacity can be
met by using used vessels, however, we believe that at least a portion of the capital costs can be
offset by used equipment.
V-71
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
There are also operational changes which refiners can make to reduce the difficulty and
the cost of desulfurizing highway diesel fuel. Based on the information which we received from
vendors and as made apparent in our cost analysis which follows, refiners with LCO in their
diesel fuel would need to hydrotreat their highway diesel pool more severely resulting in a higher
cost to meet the cap standard. We believe that these refiners could potentially avoid some or
much of this higher cost by pursuing two specific options. The first option which we believe
these refiners would consider would be to shift LCO to distillate fuels which do not face such
stringent sulfur control, such as off-highway diesel fuel and heating oil. When we analyze the
refineries which blend LCO into their diesel fuel, we find that a number of them also produce a
significant quantity of high sulfur distillate. The lenient sulfur limits which regulate heating oil
and off-highway diesel provide ample room for blending in substantial amounts of LCO.
Because of the low cetane value inherent with LCO, refiners cannot simply dump a large amount
into off-highway diesel since off-highway diesel must meet an ASTM cetane specification.
Thus, we believe that refiners could distill its LCO into a light and heavy fraction and only shift
the heavy fraction to off-highway diesel fuels. Essentially all of the sterically hindered
compounds distill above 630 °F, so if refiners undercut their LCO to omit these compounds, they
would cut out about 30 percent of their LCO. We expect that refiners could shift the same
volume of non-LCO distillate from the highway distillate pool to the highway pool to maintain
current production volumes of all fuels. In addition to the cetane limit which limits blending of
LCO into off-highway diesel, the T-90 maximum established by ASTM limits would limit the
amount of LCO, and especially heavy LCO, which can be moved from highway diesel fuel into
the high sulfur distillate streams. For those refineries which could trade the heavy portion of
LCO with other blendstocks in the high sulfur pool from own refinery or other refineries, we
presume that those refiners could make that separations cheaply by using a splitting column for
separating the undercut LCO from the uncracked heavy gasoil in the FCC bottoms.
Another option for refineries which are faced with treating LCO in its highway diesel fuel
would be to sell off or trade their heavy LCO to refineries with a distillate hydrocracker. This is
a viable option only for those refineries which are located close to another refinery with a
distillate hydrocracker. The refinery with the distillate hydrocracker would upgrade the
purchased LCO into gasoline or high quality diesel fuel. To allow this option, there must be a
way to transfer the heavy LCO from the refinery with the unwanted LCO to the refinery with the
hydrocracker, such as a pipeline or some form of water transport. We asked a refinery consultant
to review this option. The refinery consultant corroborated the idea, but commented that trading
the of blendstocks between refineries is a complicated business matter which is not practiced
much outside the Gulf Coast, and that the refineries with hydrocrackers that would buy up and
process this low quality LCO may have to modify their distillate hydrocrackers.45 The
modification which may be needed would be due to the more exothermic reaction temperature of
treating LCO which could require refiners to install additional quenching in those hydrocrackers.
Additionally, LCO can demand 60 to 80 percent more hydrogen for processing than straight run
material. The refineries which can take advantage of selling or trading their LCO to these other
V-72
-------
Chapter V: Economic Impact
refineries are mostly located in the Gulf Coast where a significant number of refineries have
hydrocrackers and such trading of blendstocks is commonplace. However, we also identified
other refineries outside the Gulf Coast which could take advantage of their very close location to
another refinery with a distillate hydrocracker. Through a quick analysis, we identified that these
refineries which could sell off or trade their heavy LCO to other refineries with hydrocrackers
produce about 25 percent of the highway diesel fuel in this country.
As we summarized in Chapter IV, catalysts are improving and expected to continue to
improve. Our costs are based on vendor submissions and incorporate the most recent catalysts
which they have to offer, however, as catalysts continue to improve, the cost of desulfurizing
diesel fuel will continue to decrease.
Emerging technologies provide another opportunity for the cost of desulfurizing diesel
fuel to be much lower than what we have estimated. Enchira Biosystems Corp., which was
Energy BioSystems Corp., created and has been developing a process which uses genetically
enhanced bacteria for oxidizing the sulfur molecules in diesel fuel, and then extracts the oxidized
sulfur-containing petroleum molecules to sell as a surfactant on the chemicals market.46 Another
similar process has been created by Petrostar. The Petrostar process also oxidizes the sulfur
molecules in diesel fuel, but uses an oxidation compound to do so.47 Finally, Phillips has adapted
their gasoline desulfurization process, which relies on adsorption, to diesel fuel. These various
processes are still being developed, though, and may not be ready in time for making the
implementation date of this final rule.
In summary, if the vendor cost estimates are optimistically low, there are a number of
reasons why the cost of desulfurizing highway diesel fuel to meet the 15 ppm cap standard are
likely to be low. Vendors are expected to continue to improve their desulfurization technology
such as the activity of their catalysts. Also, refiners have several cost cutting options at their
disposal such as using existing spare equipment to lower their capital costs. Also, refiners may
be able to resort to either of two operational options to reduce the amount of LCO in their
highway diesel fuel. Furthermore, refiners could choose to use emerging technology which could
offer significant reductions in the cost of desulfurizing diesel fuel.
We next present diesel fuel desulfurization information provided by the vendors for
typical diesel fuel blends containing 8 percent and 10 percent coker, 23 percent and 25 percent
LCO and the balance straight run, and another containing only straight run. This information is
summarized below in Tables V.C-4 & 5. This information was provided either for a revamp or
for a grassroots unit, which is indicated.
V-73
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-4. Process Projections to Desulfurize a Typical Diesel FuelA
(Information Provided for a Retrofit Unless Indicated)
Capacity
(bbl/stream day)
Capital Cost
(ISBL) (MM$)
LHSV (Liquid
Hour Space
Velocity (Hf ')
Chemical
Hydrogen
Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/bbl)
Yield Loss
(wt%) Diesel
Naphtha
LPG
Fuel Gas
Vendor A
50 ppm
900 psi
Hydrotreat.
25,000
15-18
2.5
1.25*
100
325B
0.30
_
-2.2C
0.06
1.42B
-0.89B
-0.05B
-0.09B
Vendor A
10 ppm
900 psi
Hydrotreat.
25,000
15-18
1.5
1.0*
160
375B
0.36
_
-2.9
0.08
1.51B
-1.06B
-0.06B
-0.10B
Vendor A
7 ppm
900 psi
Hydrotreat.
25,000
1 more than
at 10 ppm
0.8B
20 more
than at 10
ppm
NP
_
NP
NP
NP
NP
NP
NP
Vendor B
30 ppm
650 psi
Hydrotreat.
31,200
5.5
1.5
70
330B
0.5
_
100
0.14
NP
NP
NP
NP
Vendor B
10 ppm
650 psi
Hydrotreat.
31,200
7
0.9
115
375B
0.6
_
100
0.41
NP
NP
NP
NP
Vendor B
7 ppm
650 psi
Hydrotreat.
31,200
15
NP
NP
NP
_
NP
NP
NP
NP
NP
NP
This diesel fuel contains 23% LCD, 8% coker, and 69% straight run for Vendor A, and 25% LCD, 10% coker
and 65% straight run for Vendor B.
Sulfur levels in the table are averages.
NP = not provided.
B Information provided for a grassroots unit.
c Information provided for achieving 30 ppm; negative values indicate exothermic reactions.
V-74
-------
Chapter V: Economic Impact
Table V.C-5. Process Projections to Desulfurize 100% Straight Run Diesel Fuel
(Information is for a Grassroots Unit)
Capacity BPSD
(bbl/day)
Capital Cost (ISBL)
(MM$)
LHSV (Liquid Hour Space Velocity
(Hr1)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/BPSD)
Yield Loss (wt%) Diesel
Naphtha
LPG
Fuel Gas
Vendor A
50 ppm
800 psi Hydrotreating
25,000
NP
1.6
210
NP
-
NP
34
NP
Vendor A
10 ppm
800 psi Hydrotreating
25,000
NP
1.25
225
NP
-
NP
45
NP
NP = not provided.
Sulfur levels in the table are averages.
We are aware that there are potentially other capital and operating costs in the refinery
which would contribute the projected cost of desulfurizing diesel fuel beyond that provided to us
by the vendors. For example, refiners may need to expand their amine plant or their sulfur plant
to enable the processing of the sulfur compounds removed from diesel fuel. Then the small
amount of additional sulfur compounds treated would incur additional operating costs. Thus, we
adjusted the projected capital and operating costs upward to account for these other potential
costs which we have not accounted for directly. Our contingency factors, described further
below, are 1.18 for capital and 1.12 for operating costs.
V-75
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
d. Development of Diesel Desulfurization Cost Projections
After obtaining the information from Vendors A and B, and corroborating their
submissions based on some other information which we obtained from other vendors, we needed
to apply this information to estimate the cost of meeting the 15 ppm highway diesel fuel cost
standard. However, in many cases the information provided by the vendors was not sufficient for
inserting directly into our cost model. Vendors A and B provided most of the information
needed to cost out both a revamp and a grassroots unit. However, for some of the cost inputs for
our refinery model, the information provided by the vendors is for a grassroots unit and it must
be adjusted to reflect the impact or cost of a revamp, and vice versa. In other cases, no
information was presented at all so we developed a method for estimating the necessary cost
inputs.
In the case where we only received information for a grassroots unit for a specific cost,
we typically estimated the cost of a revamp using ratios of the liquid hour space velocity (LHSV)
provided by the vendor for a revamp. Using LHSV seems reasonable considering that the value
is inversely proportional to the catalyst and reactor volume projected to be necessary to
accomplish the required desulfurization. Thus, applying the inverse of LHSV for meeting
differing sulfur levels should be a good surrogate for the ratio of costs. We did not receive
information from Vendor B for desulfurizing 100% straight run diesel fuel, but instead of relying
only on the information from Vendor A, we projected Vendor B's costs using the percentage
difference in costs estimated by Vendor A for treating a 100% straight run feed compared to a
typical feed. Using information from both vendors for estimating the cost for the sensitivity
analysis results in a better comparison with the case which assumed a typical mix of diesel
blendstocks. For meeting the 15 ppm cap standard, which we estimate to mean achieving 7 ppm
on average, the vendors did not provide specific cost information for many of the individual cost
elements, thus we extrapolated the costs. While hydrogen consumption and space velocity
information was provided by Vendor A specifically, the other cost elements, such as catalyst
cost, yield loss and utility costs were projected using the ratio of the LHSV or by extrapolating
the costs from the higher sulfur levels. These extrapolations are described in detail below Tables
V.C-6 and V.C-7.
Cost Projections for a Typical Feed
The adjusted vendor capital and operating cost information is summarized in Tables V.C-
6. and V.C-7. below.
V-76
-------
Table V.C-6. Process Projections for Revamping an Existing Highway Diesel Hydrotreater for Further Desulfurizing a
Typical Diesel Fuel4
Average Sulfur Level
Capacity (bbl/stream
day)
Capital Cost (ISBL)
(MM$)
LHSV (Liquid Hour
Space Velocity (Hf ')
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/bbl)
Yield Loss (%)
Diesel
Naphtha
LPG
Fuel Gas
900 psi (Based on Vendor A)
50 ppm
25,000
16
2.5
125
0.24
-1.5
0.06
0.8
-0.5
-0.03
-0.05
10 ppm
25,000
18
1.5
185
0.36
-2.9
0.08
1.0
-0.71
-0.04
-0.07
7 ppm
25,000
19
1.2
205
0.37
-3.0
0.1
1.3
-0.88
-0.05
-0.08
650 psi (Based on Vendor B)
30 ppm
31,200
5.5
1.5
95
0.5
100
0.14
0.9
-0.54
-0.03
-0.05
10 ppm
31,200
7
0.9
154
0.6
100
0.41
1.0
-0.71
-0.04
-0.07
7 ppm
31,200
15
0.7
160
0.6
100
0.51
1.3
-0.88
-0.05
-0.08
This typical diesel fuel contains 23% LCO, 8% coker, and 69% straight run for Vendor A, and 25% LCO, 10% coker and 65% straight run for Vendor B.
V-77
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
When available, the information contained in Table V.D-6. reflects exactly the
information provided by the two vendors. However, the vendors did not provide projections for
some of the relevant factors. These factors were estimated from the information provided by the
other vendor or otherwise, as described below.
As stated above under Table V.C-4., Vendor A provided a range of $15 - $18 million for
the capital costs of desulfurizing diesel fuel from the base to 50 ppm and from the base down to
10 ppm. Consistent with the methodology laid out above, we assigned the capital cost of
desulfurizing diesel fuel with 23 percent LCO down to 50 ppm as $16 million, and the cost of
producing 10 ppm diesel as $18 million. For achieving a sulfur level of 5 ppm, Vendor A
estimated the additional capital cost to be $1 million more, which we used for our estimated 7
ppm case. For Vendor B, we have two sources of information for the capital costs which seem to
vary at the 10 ppm level. We based the cost analysis on the explicit cost provided by Vendor B.
However, interpolating the capital cost from Vendor B's second information source suggests that
the capital cost for desulfurizing diesel fuel to the 10 ppm level may be fifty percent higher.
We are aware that small leaks in the heat exchangers of existing highway diesel
hydrotreating unit can lead to contamination of the product stream. Even a small leak of tenths
of a percent in volume of high sulfur feed into the very low sulfur product could ruin batches of
the product. For this reason, many refiners who chose to revamp their existing diesel
hydrotreaters are expected to take preventative measures against contamination by welding the
heat exchanger tubes to the plates, or by replacing their heat exchangers altogether.48 To account
for this added cost we assumed that each refinery would invest a million dollars to revamp or, in
some cases, completely replace their highway diesel heat exchangers to ensure that they could
meet a 15 ppm diesel fuel sulfur standard.49
Since neither Vendor A nor Vendor B provided estimates of the LHSV for a retrofit unit
down to 5 ppm, we calculated Vendor A's ratio of the LHSV for achieving 5 ppm to the LHSV
for achieving 10 ppm for a grassroot unit, and applied the ratio to the LHSV values for retrofits
for both Vendor A and Vendor B for 10 ppm.
Vendor A estimated hydrogen consumption for achieving 5 ppm as 25 SCF/bbl higher
than that for achieving 10 ppm. To desulfurize down to 7 ppm from 10 ppm, we assume that an
additional 20 scf/bbl would be necessary. Since Vendor B did not provide a estimate for
achieving 7 ppm, we applied Vendor A's increased hydrogen consumption to Vendor B. At all
levels of desulfurization, we assume that each characteristic refinery would lose 25 standard
cubic feet per barrel (SCF/bbl) hydrogen due to solution and purge losses for the revamp.50 51
Solution losses of hydrogen is the hydrogen which becomes entrained in the highway diesel fuel
and thus is no longer available to recycle back to the diesel hydrotreater. Purge losses is the
intentional bleeding off of the hydrogen stream and sending that stream to plant gas to prevent a
V-78
-------
Chapter V: Economic Impact
high concentration of nonreactive gases, such as methane, from being recycled back to the
reactors.
The electricity necessary for achieving 7 ppm sulfur is extrapolated from the 10 ppm and
50 ppm cases for both Vendor A and Vendor B.
The catalyst cost for achieving 7 ppm for a revamp for Vendor A and Vendor B is
estimated by multiplying Vendor A's ratio of the LHSV for 10 ppm divided by the LHSV for 7
ppm for a grassroots unit times the LHSV for 10 ppm for a revamp.
The yield loss and resulting by products produced which was provided by Vendor A for a
grassroots unit was adjusted to project the yield loss for a revamped unit using the ratio of the
LHSV of a grassroots unit to the LHSV of a retrofitted unit. Since Vendor B did not provide
yield loss information, Vendor A's yield loss and by-product information was applied to Vendor
B. This seems reasonable because the LHSV (which indicates the contact time which diesel has
with the catalyst) for both vendors is similar and yield loss would likely be proportional to the
contact time of diesel fuel with the catalyst.
V-79
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Cost Projections for a Straight Run Feed
Table V.C-7. Process Projections for Revamping an Existing Highway Diesel
Hydrotreater for Desulfurizing 100% Straight Run Diesel Fuel
Capacity (bbl/stream day)
Capital Cost (ISBL)
(MM$)
LHSV
Liquid Hour Space Velocity
(Hr1)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/bbl)
Yield Loss (wt%) Diesel
Naphtha
LPG
Fuel Gas
800 psi (Based on Vendor A)
50 ppm
25,000
15
2.8
95
0.28
-1.5
0.03
0.6
-0.4
-0.02
-0.04
10 ppm
25,000
17
1.9
100
0.35
-2.9
0.05
0.8
-0.6
-0.03
-0.05
7 ppm
25,000
18
1.5
107
0.35
-3.0
0.07
1.0
-0.7
-0.04
-0.07
650 psi (Based on Vendor B but
adjusted using Vendor A 's
information )
30 ppm
31,200
5.5
1.7
80
0.5
100
0.11
0.7
-0.4
-0.03
-0.04
10 ppm
31,200
6.2
1.1
84
0.6
100
0.33
0.8
-0.6
-0.03
-0.05
7 ppm
31,200
11
0.9
90
0.6
100
0.41
1.0
-0.7
-0.04
-0.07
When available, the information contained in Table V.C-7. reflects exactly the
information provided by the two vendors. However, the vendors did not provide projections for
some of the relevant factors. These factors were estimated from the information provided by the
other vendor or otherwise, as described below.
Vendor A did not provide a specific capital cost for a 100 percent straight run diesel case.
Instead, the vendor estimated a capital cost of $15-18 million for a refinery processing different
amounts of LCO to meet a range of final sulfur levels of 10-50 ppm. Based on discussions with
V-80
-------
Chapter V: Economic Impact
the vendors, we surmised that increased amounts of LCO provides a similar extent of difficulty
for desulfurization as decreasing the sulfur level in this range of desulfurization. Thus, we
estimated the capital cost for the 100 percent straight run case for 50 ppm sulfur to be at the
lowest end of the range ($15 million) and to be $16 million for 10 ppm, since diesel fuel without
LCO is easier to desulfurize than diesel containing LCO. Also, the increment of $1 million was
the cost estimated by this vendor of reducing sulfur from 10 ppm to 5-10 ppm for LCO
containing material, so we used the same increment for this case as well. In Table V.C-6. above,
the capital cost for treating a typical diesel fuel falls within the upper part of Vendor A's capital
cost range.
Vendor B also did not provide capital costs for a no LCO case. Since we had no
information from Vendor B for how it would allocate its capital costs for varying levels of LCO,
we assumed that the capital costs for the no LCO cases producing sulfur at 10 ppm or higher
would be the same as those for the 23 percent LCO case. While this assumption may be
conservative, we felt comfortable with this assumption because of the low capital costs projected
by Vendor B. However, below 10 ppm, instead of the large increase in capital cost projected for
the 23 percent LCO case, we projected that the capital cost would be halfway between the
increase for the 23 percent case, which would be $11 million. This assumption seemed
reasonable since straight run contains some sterically hindered compounds which requires more
reactor volume to treat, although still much less than that of the 23 percent LCO case.
The hydrogen consumption for this retrofit case was calculated using the ratios of the
retrofit case for the case with 23 percent LCO. Vendor B's hydrogen consumption for a
grassroots case with no LCO was estimated first assuming the same hydrogen consumption as
Vendor A, however, the retrofit hydrogen consumption for Vendor B is a smaller ratio than that
of Vendor A.
The LHSV for both vendors' retrofit technology for the no LCO case was estimated from
the information which they provided for the grassroots units. The ratio of the LHSV for the
grassroots units treating no LCO to the LHSV for the grassroots unit treating 23 percent LCO
was applied to the LHSV for the retrofit unit treating 23 percent LCO to project the LHSV for the
retrofit unit treating no LCO.
Electricity consumption for the no LCO cases was assumed to be 97 percent of that for
the 23 percent LCO cases based on the ratio of specific gravities for the two different feeds, since
the density of the fuel governs the pumping energy consumed for moving the fuel. Fuel gas
consumption for treating the non-LCO feed was assumed to be the same as that for the 23 percent
LCO case. The catalyst cost for the non-LCO feed was assumed to be proportional to the ratio of
the LHSV of the no LCO and 23 percent LCO cases. The yield loss of the no LCO case was
adjusted downward from the 23 percent LCO case using ratios of the LHSV; since Vender B did
V-81
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
not provide yield loss information, Vendor A's information was applied to Vendor B's
technology as well.
e. Development of Desulfurization Cost Factors for Individual Diesel
Blendstocks
Once we established the inputs for estimating the cost of desulfurizing a typical diesel
fuel containing both straight run and cracked stocks, we set out to estimate the inputs for each
individual blendstock. Configuring our cost analysis to estimate costs based on the estimated
highway diesel blend of each refinery gave us more confidence in our cost analysis. We already
had the inputs for straight run from a submission from Vendor A. Next we needed to estimate
the inputs for light cycle oil and for the other cracked stocks. We used some of the information
we obtained from our discussions with the vendors to make these estimates. Since we need to
estimate costs for both a revamp and a grassroots units for each refinery, it was necessary to
develop costs for both. These costs are presented in Table V.C-8 for a revamped unit, and
further below in Table V.C-9 for a grassroots unit. The methodology for developing those costs
are described below each Table.
Individual Blendstock Process Projections for a Revamp
These process projections are for revamping an existing desulfurization unit with
additional hardware enabling the combined older and new unit to meet the 15 ppm sulfur cap
standard. If a refiner decides to replace their existing highway diesel fuel desulfurization unit
with a new grassroots unit, we assume that the operating costs of the new unit would still be the
same as a revamped unit because the refiner has already been incurring the operating cost for
producing 350 ppm highway diesel fuel. We assume the refiner would, however, incur all the
capital cost of the new unit.
V-82
-------
Chapter V: Economic Impact
Table V.C-8. Process Projections for Revamping an Existing Highway Diesel Hydrotreater
for Further Desulfurizing Diesel Fuel Blendstocks to Meet a 15 ppm Cap Standard
Capacity BPSD
(bbl/day)
Capital Cost (ISBL)
(MM$)
LHSV (Liquid Hour Space
Velocity (Hf ')
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/BPSD)
Yield Loss (wt%) Diesel
Naphtha
LPG
Fuel Gas
Straight Run
25,000
16
1.25
96
0.4
-
40
0.2
1.0
-0.7
-0.04
-0.04
Other Cracked
Stocks
25,000
19
0.7
230
0.7
-
70
0.4
1.9
-1.3
-0.07
-0.11
Light Cycle Oil
25,000
22
0.6
375
0.8
-
80
0.5
2.2
-1.5
-0.08
-0.13
The information in Table V.C-8 was derived from the Tables V.C-4-7 above, from Table V.C-9
below, and using other inputs and assumptions as described below.
Capital Costs
The inside battery limits (ISBL) capital costs for revamping a hydrotreater to handle
straight run was estimated by averaging the values for Vendors A and B from Table V.D-7. A
$1 million sum was added to that sum to account for improvements to existing heat exchangers
such as welding the tubes to the tubesheets, and for some refiners to replace their heat exchangers
altogether.
V-83
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The ISBL capital cost of treating coker and other cracked stocks is based on the need to
have more catalyst and reactor volume and probably a higher pressure than straight run to treat a
greater volume of sterically hindered compounds. The difficulty in treating coker distillate was
presumed to be similar to treating 1/3 LCO, 1/3 coker distillate and 1/3 straight run, because the
volume of sterically hindered compounds is similar to that combination of blendstocks. This is a
useful comparison to make because in their submission to us, Vendor A provided a capital cost
estimate for treating such a mix down to 10 ppm. Vendor A presumed that refiners would need
to invest $19 million, which is at the high end of the range given by Vendor A for achieving 10
ppm for a range of feeds, of which this particular blend of diesel stocks was the worst. This
value was increased by $1 million to achieve 7 ppm and another $1 million to revamp or replace
the heat exchangers, which increased the sum to $21 million. Like the case with 23 percent LCO
in the diesel fuel, Vendor B's capital costs were presumed to be $4 million less than Vendor A's
capital costs, which would still include the $1 million for improvements to existing heat
exchangers. On average, treating coker distillate is estimated to cost $19 in capital costs.
The ISBL capital cost for a revamp to an existing diesel hydrotreater for treating LCO can
be estimated from some assumptions on the relative difficulty of treating the sterically hindered
compounds contained in LCO. LCO contains proportionally more sterically hindered
compounds than what the other cracked stocks are estimated to contain relative to straight run
(coker distillate contains slightly more than twice the percentage of sterically hindered
compounds as straight run, and LCO contains a little more than twice the percentage of sterically
hindered compounds as the other cracked streams).52 Based on this observation and assuming
that the increased reactor volume and higher pressure needed to treat LCO is proportionally
higher than treating other cracked stocks compared to straight run distillate, we presume that the
capital costs are proportionally higher as well. Thus, the capital cost was increased by the same
amount over the other cracked stocks as the difference between the other cracked stocks and
straight run, which is $3 million more. Then the same $1 million increase was assumed for
improving the heat exchangers. Thus, hydrotreating LCO is estimated to cost $22 million in
capital costs.
Hydrogen Consumption
The hydrogen consumption for treating straight run, other cracked stocks and LCO was
calculated from the values in Table V.C-8 for desulfurizing these untreated distillate streams in a
grassroots hydrotreating unit down to 7 ppm. Based on the relative hydrogen consumption for
revamped units versus grassroots units from Vendor A and B for a typical feed, the revamped
hydrogen consumption is estimated to be about one-third of the hydrogen consumption of the
grassroots unit for straight run and LCO. However, because of the high olefm content of the
other cracked stocks which consumes a significant amount of hydrogen in a first stage, a revamp
would only be expected to require one-fourth of the estimated amount of hydrogen consumed in
a grassroots unit. These factors are applied to the hydrogen consumption values without losses,
V-84
-------
Chapter V: Economic Impact
and the losses are added back after multiplication by the various factors. For treating straight run
and other cracked stocks, the losses for a grassroots unit are small and assumed to not be lower
for a revamped unit. However, the larger losses for treating LCO are assumed to decrease to 25
scf/bbl from the 50 scf/bbl assumed for the grassroots unit. Based on these factors, hydrogen
consumption, including losses, for a revamped highway diesel fuel desulfurization unit for
meeting the 15 ppm cap standard is 96 scf/bbl for straight run, 230 scf/bbl for other cracked
stocks, and 375 scf/bbl for LCO
Space Velocity and Other Operating Costs
The estimated space velocity for a revamped unit treating straight run, other cracked
stocks and LCO was calculated from the space velocity values for a grassroots unit summarized
below. According to Vendor A, who estimated the space velocity for both a grassroots unit and a
revamp for desulfurizing an average blend of diesel fuel down to an average of 10 ppm, a
revamped unit's space velocity is 50 percent higher than a grassroots unit. This factor was
applied to the space velocities for a grassroots unit listed in Table V.C-8.
The utilities, the catalyst cost and the yield loss were costed out using the space velocity
as the cost factor. This calculation was implemented by using the reciprocal of the space
velocity, which is the residence time, and multiplying it times each of these operating cost inputs.
The catalyst volume would correlate exactly with this relationship, and a less than perfect, but
reasonable, correlation would be expected with yield loss and utility cost. The loss of diesel
mass was estimated with this approach, however, the cost was ultimately calculated outside of
these equations as described below.
Individual Blendstock Process Projections for a Grassroots Unit
Similar process projections are provided for a grassroots unit in this section. It is
important to note that a refinery only producing, or predominantly producing, non-highway diesel
fuel would be faced with these estimated costs. However, as stated above, if a refinery has an
existing hydrotreater for desulfurizing their highway diesel fuel and they install a grassroots unit
instead of revamping their existing hydrotreater, they would incur the capital costs outlined here,
but their operating costs would be based on a revamp as described above.
V-85
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-9. Process Projections for Installing a New Grassroots Unit for Desulfurizing
Untreated Diesel Fuel Blendstocks to Meet a 15 ppm Cap Standard
Capacity BPSD
(bbl/day)
Capital Cost (ISBL)
(MM$)
LHSV (Liquid Hour Space
Velocity (Hr1)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/BPSD)
Yield Loss (%) Diesel
Naphtha
LPG
Fuel Gas
Straight Run
25,000
31
0.8
240
0.6
-
60
0.3
1.5
1.1
0.06
0.06
Coker Distillate
25,000
37
0.5
850
1.1
-
105
0.6
2.9
2.0
0.11
0.17
Light Cycle Oil
25,000
42
0.4
1100
1.2
-
120
0.8
3.3
2.3
0.12
0.20
The information in Table V.C-9 was derived from Tables V.C-4 through Table V.C-7 above for
desulfurizing highway diesel fuel down to 7 ppm, and using other inputs and assumptions as
described here.
Capital Costs
The capital costs for a grassroots hydrotreater was calculated simply by increasing the
cost of a revamp by a factor two. This same calculation was used for straight run, coker distillate
and light cycle oil. The basis for this calculation is that Vendor A's information provided for
both a revamp and a grassroots unit for desulfurizing a typical feed to meet a stringent sulfur
standard showed that the grassroots unit's ISBL investment cost is projected to cost two times
higher than a revamp. The $1 million sum which was added to the revamped case to account for
V-86
-------
Chapter V: Economic Impact
improvements to existing heat exchangers was not included in the grassroots capital cost since
the grassroots unit includes new heat exchangers.
Hydrogen Consumption
The hydrogen consumption rate for straight run, coker distillate and light cycle oil were
estimated by applying certain factors used by vendors for estimating hydrogen consumption.
One such factor is that about 25 standard cubic feet per barrel (scf/bbl) of hydrogen is consumed
for each volume percent of polynuclear aromatics saturated to monoaromatics.53 54 As described
in Chapter IV, many of the polynuclear aromatics (PNAs) are saturated to monoaromatics to
enable desulfurization of the sterically hindered sulfur compounds. On a molecular level, four
hydrogen atoms are consumed for each PNA saturated to a monoaromatic. According to
Mathpro, about half the total amount of aromatics in a diesel blend are PNAs: straight run
contains about 8 volume percent PNAs, coker distillate contains about 20 volume percent PNAs ,
and LCO contains about 55 volume percent PNAs.55 However, these values are typical values
within a range of values which can vary depending on the type of crude oil processed by each
refinery and operating conditions of the unit producing the individual blendstock. Since we do
not know these variables for each refinery producing highway diesel fuel, we used the typical
values listed here. In a submission from Vendor A, which was based on feed qualities from
Mathpro, 5 volume percent of the PNAs are estimated to be saturated to monoaromatics to
achieve an average of 10 ppm sulfur. The conversion of this 5 volume percent represents about
two thirds of the total volume of PNAs shown to be typical for straight run by Mathpro. Thus, if
a similar fraction of PNAs are saturated for each blendstock, 12 percent of the PNAs in coker
(2/3 of 20) and 34 percent of the PNAs in LCO (2/3 of 55) would be converted to
monoaromatics. Since we don't have other information on which we can base our estimate of
the hydrogen consumption for the saturation of PNAs in LCO and other cracked stocks, we used
this factor for estimating this form of hydrogen consumption. As an example of how to apply the
factor described above, to estimate the hydrogen consumed due to the saturation of PNAs when
desulfurizing straight run down from uncontrolled levels of sulfur to 10 ppm, we would multiply
the 25 scf/bbl factor times the 5 volume percent of PNAs saturated, thus, 125 scf/bbl of hydrogen
would be consumed.
Of course the sulfur in each of these different blendstocks must be hydrotreated out of the
sulfur-containing hydrocarbon compounds. For most of the sulfur, four hydrogen atoms are
consumed to remove each sulfur atom. According to Vendor B, removing sulfur from diesel fuel
consumes 125 scf/bbl for each weight percent of sulfur removed.56 According to Mathpro,
typical straight run, LCO, and coker distillate contain on the order of 0.7, 1.3 and 3 percent
sulfur, respectively. As an example, removing the sulfur from a typical straight run feedstock
would consume 85 scf/bbl of hydrogen (0.7 multiplied times 125 scf/bbl) to desulfurize each
barrel of untreated straight run diesel fuel down to 10 ppm suflur.
V-87
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
During the hydrotreating process, the hydrocarbons which are olefins are very readily and
completely saturated to paraffins which consumes two additonal atoms of hydrogen for each
olefm. Coker distillate, and to a lesser degree, LCO contain some olefins which are readily
saturated at the top of any hydrotreater. One vendor we spoke to estimated that coker distillate
contain 30 volume percent olefins, which consumes on the order of 6 scf/bbl of hydrogen per
each volume percent of olefins saturated.57 We do not have an estimate for the olefm content of
LCO, however, we believe that LCO does contain some so we presume that it is about one-fifth
as much as coker distillate, or about 6 volume percent. As an example, saturating the olefins in
coker would consume 180 scf/bbl of hydrogen (30 times 6 scf/bbl) per each barrel of coker
distillate hydrotreated
Since the level of conversion of polyaromatics to monoaromatics was consistent with
achieving 10 ppm sulfur, this value must be increased to be consistent with achieving 7 ppm
sulfur. According to Vendor A, about another 20 scf/bbl are consumed to make up the difference
between 7 ppm and 10 ppm for a typical feed which, as described above, is comprised of 69
percent straight run, 8 percent coker and 23 percent LCO. Allocating this increased hydrogen
consumption to each blendstock we estimate that straight run will consume 8 scf/bbl more
hydrogen, other cracked stocks would consume about 15 scf/bbl more hydrogen and LCO would
consume about 50 scf/bbl more hydrogen. This allocation is based on the relative concentrations
of PNAs contained in each of these blendstocks.
The estimated amount of hydrogen consumption for each blendstock is summarized in the
following table.
Table V.C-10. Estimated Hydrogen Consumption to Desulfurize Nontreated Distillate,
Stocks to Meet the 15 ppm Highway Diesel Fuel Sulfur Cap
Straight Run
Other Cracked
Stocks
LCO
Conversion of
Polynuclear
Aromatics to
Monoaromatics
133
325
900
Sulfur Removal
85
375
165
Saturation of
Olefins
0
180
35
Total Hydrogen
Consumption
223
875
1100
After deriving these hydrogen consumption estimates for each blendstock, we compared
these estimates to the estimated amount of hydrogen consumed by Vendors A and B for
V-88
-------
Chapter V: Economic Impact
desulfurizing three different feeds down to 10 ppm. Vendor A provided hydrogen consumption
estimates for straight run, a blend of 69 percent straight run, 8 percent coker and 23 percent LCO,
and a blend of 1/3 straight run, 1/3 coker, and 1/3 LCO (not summarized above, but was
submitted to the docket). Vendor B provided hydrogen consumption estimates for a blend of 65
percent straight run, 10 percent coker and 25 percent LCO. This comparison is summarized in
Table V.C-11 below.
As shown in Table V.C-11, our estimated hydrogen consumption values seem to agree
fairly well with those provided by the vendors. The straight run and 1/3-1/3-1/3 feedstock are
both quite close. However, the estimated hydrogen consumption for a typical feed, which is
either 69 or 65 percent SR, 8 or 10 percent coker, and 23 or 25 percent LCO is between 20 to 30
percent high, with the highest discrepancy with Vendor B's estimated hydrogen demand. This 69
or 65 percent SR feed is probably the most important since it really represents the average of
diesel fuel today. The 1/3 SR, 1/3 other cracked and 1/3 LCO, stock feed is heavier than average
diesel fuel today. Because we are only modelling the average endpoint, we would be expected to
estimate a lower hydrogen consumption value compared to heavier feeds. For these reasons, we
recalculated the hydrogen consumption adjusting it downward by 5 percent. These recalculated
values are summarized in the last column in Table V.C-11. This recalculation reduces the
estimated hydrogen consumption values of straight run from 223 to 213 scf/bbl, other cracked
stocks from 875 to 830 scf/bbl, and LCO from 1100 to 1045 scf/bbl.
Table V.C-11. Comparison of Calculated Hydrogen Consumption with the Hydrogen
Consumption provided by Vendors A and B for Specific Distillate Feeds
Straight Run
69 % straight run, 8 %
coker, 23% LCO
65 % SR, 8% coker, 23%
LCO
1/3 straight run, 1/3
coker, and 1/3 LCO
Vendor A
233
395
730
Vendor B
395
Calculated
Hydrogen
Consumption
223
476
507
732
Recalculated
Hydrogen
Consumption
212
450
480
695
The hydrogen consumption values summarized in Table V.C-11 are only meant to
represent the chemical consumption of the hydrogen consumption, which is the hydrogen which
reacts with the hydrocarbon. Additional hydrogen is lost through entrainment in the diesel fuel
V-89
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
and other losses. When hydrogen becomes entrained in the diesel fuel and it is not recovered for
reuse, it is called solution losses. Other losses can occur through leaks from the unit or perhaps
due to flaring in cases of unit overpressure or due to a constant purge to prevent accumulation of
inerts in the hydrogen stream. To account for these losses, we added 25 scf/bbl for straight run
and the other cracked stocks, and 50 scf/bbl for LCO. Accounting for hydrogen losses, our
hydrogen consumption values increase to about 240 scf/bbl for straight run, 850 scf/bbl for other
cracked stocks, and 1100 scf/bbl for LCO.
Space Velocity and Other Operating Costs
The space velocity for a grassroots hydrotreater was calculated by multiplying the space
velocity of a revamp by a factor of 0.66 (a fifty percent increase in residence time). This same
adjustment was used for straight run, coker distillate and light cycle oil. The information
provided by Vendor A was the basis for using this adjustment factor as the space velocity for a
grassroots diesel hydrotreater treating a typical blend of straight run, coker distillate and light
cycle oil was two thirds the space velocity of a revamp. In terms of residence time, a grassroots
unit requires about 50 percent more residence time compared to a unit which is a revamp to an
existing diesel hydrotreater.
The utilities, the catalyst cost and the yield loss were costed out using the space velocity
as the cost factor. This calculation was made by using the reciprocal of the space velocity, which
is the residence time, and multiplying it times each of these operating cost inputs. The catalyst
volume correlates exactly with this relationship, and a reasonable correlation would be expected
with yield loss and utility cost as well. The loss of diesel mass was estimated with this approach,
however, the cost was ultimately calculated outside of these equations as described below.
Hydrocrackate Processing and Tankage Costs
We believe that refineries with hydrocrackers will have to invest some capital and incur
some operating costs to ensure that recombination reactions at the exit of the second stage of
their hydrocracker does not cause the diesel fuel being produced by their hydrocracker to exceed
the cap standard. The hydrocracker is a very severe hydrotreating unit capable of hydrotreating
its product from thousands of ppm sulfur to essentially zero ppm sulfur, however, hydrogen
sulfide recombination reactions which occur at the end of the cracking stage, and fluctuations in
unit operations, such as temperature and catalyst life, can result in the hydrocracker diesel
product having up to 30 ppm sulfur in its product stream.58 59 Thus, we assume that refiners will
need to install a finishing reactor for the diesel stream produced by the hydrocracker. According
to vendors, this finishing reactor is a low temperature, low pressure hydrotreater which can
desulfurize the simple sulfur compounds which are formed in the cracking stage of the
hydrocracker. The finishing reactor adds about 0.25 c/gal to the cost of desulfurizing diesel fuel
for those typical refineries with distillate hydrocrackers.
V-90
-------
Chapter V: Economic Impact
Additionally, since the diesel sulfur standard is a cap standard, we are taking into account
tankage costs that would be incurred due to the cap standard. We believe that refiners could
store high sulfur batches of highway diesel fuel during a shutdown of the highway diesel
hydrotreater. Highway diesel production would cease in the short term, but the rest of the
refinery could remain operative. To account for this, we provided for the installation of a tank
that would store 10 days of highway diesel production sufficient for a 10 day emergency
turnaround which is typical for the industry, which would be about 3 million dollars for a
270,000 barrel storage tank.60 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, have the capability to send the untreated
blendstock to a nearby refinery which had spare capacity for treating this high sulfur blendstock,
or would downgrade the high sulfur highway diesel batch to the high sulfur diesel pool (there is
already a significant amount of highway diesel fuel sold as off-highway diesel fuel).k Adding
such a storage tank to the typical refinery adds about 0.17 c/gal to the cost of desulfurizing diesel
fuel for that refinery.
The cost inputs for the storage tank and the finishing reactor are summarized in Table
V.C-12.
k Presuming that half of refineries will add a storage tank is reasonable, because some refineries will not
need to add a storage tank due to blendstock shifting and downgrading options to them, and that some will have to
install such a tank since they will not have such options available to them.
V-91
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-12. Process Operations Information for Additional Units used in the
Desulfurization Cost Analysis
Capacity
Capital Cost
(MM$)
Electricity
(KwH/bbl
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Cooling
Water
(Gal/bbl)
Operating
Cost
($/bbl)
Diesel Storage
Tank
50,000 bbls
0.75
—
—
—
—
none*
Distillate
Hydrocracker Post
Treat Reactor
25,000 (bbl/day)
5.761
0.98
4.2
18
5
—
* 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.
Refiners will also likely invest in a diesel fuel sulfur analyzer.62 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 diesel fuel sulfur analyzer would be about $50,000, and the installation
cost would be another $5000.63 Compared to the capital and operating cost of desulfurizing
diesel fuel, the cost for this instrumentation is far below 1 percent of the total cost of this
program.
V-92
-------
Chapter V: Economic Impact
/'. Capital Cost Adjustment Factors
Capital costs are the one-time costs incurred by purchasing and installing new hardware
in refineries. The capital costs supplied by the vendors, as discussed above, were designated to
apply for a particular volumetric capacity in 1999 dollars. These costs are adjusted to match the
volume of the particular case being analyzed using the "sixth tenths rule."1 According to this rule
commonly used in the refining industry, the capital cost of a smaller or larger piece of equipment
varies in proportion to the ratio of the smaller or larger capacity to the base capacity taken to
some power, typically 0.6.
The calendar day volume is increased by 20 percent to size the hydrotreating unit for
stream days which are the days which the unit is operating. This 20 percent calendar day to
stream day factor is used to size the new hydrotreater to account for changes in day-to-day
operations, for the difference in diesel fuel production throughout the year, and for treating
offspec batches.
The capital costs are adjusted further to account for the off site costs and differences in
labor costs relative to the Gulf Coast. The factors for calculating the offsite costs and accounting
for differences in labor costs is taken from Gary and Handewerk.64 The offsite and labor factors
from Gary and Handewerk are provided for different refinery sizes and different parts of the
country, respectively. For the Tier 2 gasoline sulfur rule they were calculated for each PADD
and we summarized those cost factors in Table V.C-13. The offsite factor provided by Gary and
Handewerk is for a new desulfurization unit, but offsite costs are much lower for a revamped
unit. We cut those factors in half to account for those units which are revamps of existing units.65
The PADD-specific and national average cost factors are summarized in Table V.C-13 below.
1 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
0.65 for desulfurization units (Peters and Timmerhaus, 1991).
V-93
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-13. Offsite and Location Factors Used for Estimating Capital Costs
Offsite Factor
- New Unit
- Revamped
Unit
Location Factor
PADD1
1.26
1.13
1.5
PADD2
1.26
1.13
1.3
PADD3
1.20
1.10
1
PADD4
1.30
1.15
1.4
PADD5
1.30
1.15
1.2
To account for other capital costs not accounted for by this cost estimate, such as some
refiners having to debottleneck the amine and sulfur plants to address the additional sulfur
removed and for other contingencies, capital costs were increased by 15 percent, a typical factor
used for this type of analysis.66 In addition, we modified this contingency factor based on
comments which we received since the NPRM. The Association of Automobile Manufacturers
provided comments on a cost study by the Department of Energy which estimated the cost of
desulfurizing diesel fuel. These comments, made by an oil industry consultant, provided
information on typical oil industry cost factors for starting up and operating new units in
refineries percent.67 One such cost factor is that the oil industry incurs a cost to start up a new
unit which corresponds to about 3 percent of total capital costs. This factor was incorporated
into our analysis by increasing our contingency factor from 15 to 18 percent.
The economic assumptions used to amortize capital costs over the production volume of
low sulfur highway diesel fuel are summarized below in Table V.C-14.68 These capital
amortization cost factors are used in the following section on the cost of desulfurizing diesel fuel
to convert the capital cost to an equivalent per-gallon cost."1
m The capital amortization factor is applied to a one time capital cost to create an amortized annual capital
cost which occurs each and every year for the 15 years of the economic and project life of the unit.
V-94
-------
Chapter V: Economic Impact
Table V.C-14. 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
/'/'. Fixed Operating Costs
Operating costs which are based on the cost of capital are called fixed operating costs.
These are fixed because these costs are normally incurred whether or not the unit is operating or
shutdown. Fixed operating costs normally include maintenance needed to keep the unit
operating, buildings costs for the control room and any support staff, supplies stored such as
catalyst, and insurance. The comments from the oil industry consultant referred to above were
useful here for updating this portion of our analysis.
Various fixed operating cost factors were estimated based on comments which we
received from the American Automobile Manufactures consultant referred to above.69
Maintenance costs are estimated to be 3 percent of final capital costs. Other fixed operating
costs are 1.5 percent of capital costs for buildings, 0.2 percent for land, one percent for supplies
which must be inventoried such as catalyst, and 1 percent for insurance. These other fixed
operating cost factors sum to 3.7 percent and, when combined with the 3 percent maintenance
cost factor, sum to 6.7 percent. This total fixed cost factor of 6.7 percent is applied to the final
capital cost (after including offsite costs and adjusting for location factor) to generate an annual
fixed operating cost.
Annual labor costs are also estimated using the cost equation in the Oak Ridge National
Laboratory (ORNL) refinery model. Labor cost is very small, on the order of one thousandth of a
cent per gallon.
/'/'/'. Utility and Fuel Costs
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. Here, variable operating costs are determined using
annual average diesel fuel production volumes instead of refinery specific production volumes to
V-95
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
avoid over- and under-counting of production when specific units are processing stored distillate
after a shutdown or downgrading product when a unit is shutdown. The operating cost demands
(utilities, hydrogen, and yield loss) are based on estimates from the desulfurization technology
licensors described above. The basis for the values is 98 percent desulfurization (340 ppm sulfur
reduced to 7 ppm sulfur on average) of the highway pool.
The utility cost inputs for our refinery model are from 1999 Energy Information
Administration (EIA) information for each of the five Petroleum Administrative Districts for
Defense (PADDs).70
Yield loss is based on the volume of diesel volume lost times its market price offset by
the additional volume of other products produced times their sales for resale market prices. A
representative refinery price for diesel fuel after the desulfurization programs begins is derived
by adding the estimated cost of desulfurizing diesel fuel for the highest cost producer to the
resale price for diesel fuel from EIA. These cost factors are summarized in Table V.D-15.
Fuel gas is consumed in running furnaces for heating up streams including the reboilers
used in distillation. Fuel gas cost is based on an estimation factor which is three dollars per
million British thermal units (BTU) for PADD 3,71 one quarter higher than that for PADDs 1, 2
and 5, and half higher for PADD 4. Steam demand is converted to BTU demand on the basis
that it is 300 pound per square inch (psi) steam, and that demand is presumed to be met with fuel
gas, however, we increase the cost by a factor of two which is consistent with published cost
estimation methodology.72 Producing steam is presumed to demand 809 BTU per pound of
steam required.
Hydrogen costs are assumed to vary by PADD. The cost of hydrogen supply was
estimated for PADD 3, and then increased for other PADDs that typically have higher costs.
Hydrogen cost for PADD 3 is based on an average of refiners putting in their own hydrogen
plants, which could cost as much as three dollars per thousand standard cubic foot (MSCF), and
purchasing hydrogen as a commodity from a large hydrogen plant at a little more than one dollar
per MSCF.73 Based on this range of possible cost, PADD 3 would be expected to have access to
hydrogen supplied at a cost of about two dollars per MSCF. PADD 4 is assumed to have to pay
the more conservative cost of three dollars per MSCF, and the other PADDs are assumed to incur
a cost between PADDs 3 and 4, which would be $2.5 per MSCF. This analysis does not consider
numerous other possibilities of providing hydrogen at a reduced cost by using hydrogen recovery
technology (which would recover hydrogen from plant gas), or by increasing hydrogen
production from the reformer by converting high pressure reformers to low or ultra low pressure
reformers.
V-96
-------
Chapter V: Economic Impact
Table V.C-15. Summary of Costs From EIA Information Tables for 1999,* and Other Cost
Factors
Electricity (c/KwH)*
LPG ($/Bbl)*
Highway Diesel (c/gal)*
Nonhighway Diesel (c/gal)*
Gasoline ($/Bbl)*
Fuel Gas ($/MMbtu)
Hydrogen Cost ($/MSCF)
PADD1
8.35
17.09
53.1
49.3
27.0
3.75
2.5
PADD2
6.40
14.11
55.9
55.7
25.9
3.75
2.5
PADD3
6.66
14.49
51.5
48.6
24.9
3.0
2.0
PADD4
5.4
14.53
62.4
60.4
28.9
4.5
2.0
PADD5
7.18
17.05
64.0
58.9
30.0
3.75
2.5
* c/KwH is cents per kilowatt-hour, $/Bbl is dollars per barrel, c/gal is cents per gallon, $/MMbtu is dollars per
million British Thermal Units (Btu), $/MSCF is dollars per thousand standard cubic feet.
Similar to the capital costs, we added a 10 percent operating cost safety factor to account
for other operating costs which are beyond the operating cost of the desulfurization unit.74 This
factor accounts for the operating cost of processing additional hydrogen sulfide in the amine
plant, additional sulfur in the sulfur plant, and other costs which may be incurred but not
explicitly accounted for in our cost analysis. We then increased this factor by 2 percent to
account for reprocessing of offspec material. For estimating capital costs, we estimated that 5
percent of the batches would be offspec and could not be blended down with lower sulfur
product. However, since this material was desulfurized once already, the operating costs for
reprocessing it would be much lower the second time around.
We also believe that refinery managers will have to place a greater emphasis on the
proper operation of other units within their refineries not just the new diesel fuel desulfurization
unit, to consistently deliver very low sulfur highway diesel fuel under the proposed cap standard.
For example, meeting a stringent sulfur requirement will require that the existing diesel
hydrotreater and hydrocracker units operate as expected. Also, the purity and volume of
hydrogen coming off the reformer and the hydrogen plant would be important for effective
desulfurization. Finally, the main fractionator of the FCC unit would have to be carefully
controlled to avoid significant increases in the distillation endpoint, as a significant volume
increase in sterically hindered compounds could be sent to the diesel hydrotreater with an
increase in endpoint. The diesel hydrotreater may not be designed to desulfurize a significant
increase in sterically hindered compounds. Improved operations management to control each of
these units or situations could involve enhancements to the computer systems which control the
V-97
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
refinery operations, as well as improved maintenance practices.75 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 below for diesel fuel
desulfurization.76 77 No costs were included in the cost analysis for these potential issues.
f. Future Diesel Fuel Volumes
The volume of diesel fuel produced in future years is expected to increase consistent with
projected future increases in diesel fuel demand. Estimating this increase is important as both the
per-gallon costs and the aggregate costs are affected by the increase. Ignoring inflation and
assuming that the prices of raw materials and products stay the same as in 1999, per-gallon costs
would decrease somewhat with slightly improved economies of scale. However the aggregate
capital and operating costs would increase as production volumes increase, although this increase
is slower than the rate of increase in demand due to economies of scale.
To project future diesel fuel consumption, we relied on projections from the Energy
Information Administration (EIA). EIA projects consumption of refined products into the future
based both on historical production trends and on market factors likely to affect future demand.
In the year 2000 Annual Energy Outlook, EIA projects that in 2006, highway diesel fuel
consumption will be 39.5 billion gallons per year, with imports of 2.0 billion gallons per year.
This level of diesel fuel consumption is 12.6 percent higher than today's consumption volume.
Since our analysis is performed on a refinery-by-refinery basis, it is important to project
how each refinery's production of highway diesel fuel will change as consumption increases.
Refiners tend to invest capital dollars in their refineries periodically for increasing the production
volume of their products. This process of increasing refinery throughput is called
debottlenecking. However, we have no way to project which refiners will invest to debottleneck
their refineries for increased production, thus we cannot assign increases to specific refineries.
Instead, we assume that each refinery will increase their production of highway diesel fuel by the
same 12.6 percent between now and 2006. While highway diesel fuel consumption would be
expected to increase again between 2006 and 2010, the change is modest, so we assumed that the
2006 volumes would apply in 2010 as well.
We made no changes in the volumes of diesel fuel processed to account for changes in
wintertime blending of kerosene. Our cost projections are based on the volume of highway
diesel fuel consumed today projected to the year 2006 and this assumes no changes in that
volume in our final rule." Thus, our cost projections include hydrotreating that volume of
n Actually, we assume that the total energy consumed in the form of diesel fuel remains
constant. Diesel fuel volume consumed increases slightly because of a small decrease in the
energy content of diesel fuel after additional hydrotreating.
V-98
-------
Chapter V: Economic Impact
kerosene which is currently blended into winter diesel fuel. Some of the kerosene which is
blended into winter diesel fuel is blended at the refinery. This kerosene should be able to be
added prior to the hydrotreater and desulfurized along with the rest of the highway diesel fuel
pool. The rest of this kerosene is added at terminals or at other points within the distribution
system. If this practice were to continue, then the kerosene distributed to these points would also
have to meet the sulfur cap. Given this would likely involve hydrotreating more kerosene than
actually needed to winterize diesel fuel, we believe that this practice would become much less
common. Instead, we believe that cold flow additives would be used in greater amounts in lieu
of kerosene blending downstream of the refinery. Cold flow improving additives are commonly
used today in economic competition with kerosene blending and we believe that the cost
differential between desulfurizing kerosine and blending in cold flow additives to achieve the
same effect is negligible. Thus, assuming that the difference in cost of cold flow additives and
kerosene blending is negligible, we expect that diesel fuel suppliers would reduce the current
amount of kerosene blending and increase additive use at no additional cost and avoid the need to
hydrotreat kerosene which may be used in other applications than highway diesel engines to less
than 15 ppm sulfur.
2. Projected Refinery Costs of Meeting the 15 ppm Sulfur Cap
For each of 121 refineries currently producing highway diesel fuel, the capital and
operating cost inputs described above were combined together in our refinery model along with
the fractions of the various blendstocks for each refinery to estimate the cost of desulfurizing
highway diesel fuel from a base sulfur level of 340 ppm to an average of 7 ppm sulfur to meet the
15 ppm cap standard.0
The per-refinery capital and operating costs, and the per-gallon cost for refineries were
classified into small and non-small refinery categories and are summarized in Table V.C-16
below.
0 Grass roots capital costs were determined based on new equipment required while grass
roots operating costs were assumed to be the same as a revamped unit.
V-99
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-16. Estimated Per-Refinery Capital, Operating and Per-Gallon Cost for FullA
Implementation of Desulfurizing Highway Diesel Fuel to Meet a 15 ppm Cap Standard
(1999 Dollars, 7% ROI before taxes)
Capital Cost
(SMillion)
Operating Cost
($Million/yr)
Per-Gallon Cost
(c/gal)
Average of
Nonsmall
Refineries
52
9.6
4.2
Average of
Small
Refineries
14
0.5
5.0
National
Average
44
7.9
4.3
A Based on the assumption that each refineries costs will be comprised of; 80% for revamping a refiner's existing
hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs were determined based
on new equipment required while grass roots operating cost were assumed to be the same as a revamped unit.
National average refinery costs includes refineries classified as small. Capital costs are total aggregate per refinery
in each category.
Table V.C-16 shows that, on average for full implementation of the 15 ppm highway
diesel fuel sulfur cap standard, non small refineries would incur initial capital cost of $52 million
to meet the proposed sulfur cap. In addition, these refineries would incur an average of $9.6
million per year in operating costs. The capital and operating cost for typical small refineries
would be much lower, $14 million and $0.5 million per year per refinery, respectively, but due to
poorer economies of scale their installed capital costs would be higher on a per-gallon basis. Our
cost estimates bear this out as the per-gallon cost to the average small refinery is about 20 percent
higher (about 1.0 cents per gallon) than the per-gallon cost of the average nonsmall refinery, thus,
our analysis projects that small refineries are more challenged than the refineries which treat a
larger volume of diesel fuel. The per-gallon cost for all of the refineries participating varied and
can be viewed in Figure V.C-1. Inspection of the graph reveals that for the 121 refineries, only
four to five volume percent of the total highway pool have high costs that exceed 5 cents per
gallon.
V-100
-------
Chapter V: Economic Impact
Figure V.C-1. Refinery Specific Costs for Fully Implemented 15 ppm Sulfur Cap Standard
20
15
c
o
CO
S"
-I—*
c
CD
O
5
0
(
<
:
i
j
O0o coo oo ooo oo o^o«*> «><> «oo<«> oo«x»«*><*> o ^^^^^
) 10 20 30 40 50 60 70 80 90 100
Cumulative Production (%) of Highway Diesel Pool
a Costs per treated volume of highway diesel for 121 refineries, 1999 dollars and capital is amortized 7% ROI before
taxes.
Refineries with LCO and coker gas oils had higher costs than those processing straight
run diesel. LCO feed stocks had the highest hydrotreater costs with an average feedstock based
incremental cost of 6.55 cents per gallon treated. Likewise, coker gas oil and straight run diesel
had average incremental feedstock costs of 4.72 and 3.47 cents per gallon, respectively. The costs
for LCO and coker feed stocks were higher due to the increased capital and operating cost
associated with treating these feed stocks, see Table V.C-17.
V-101
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table V.C-17. Costs for Treating LCO, Coker, and Straight Run Diesel Feedstocks
(1999 Dollars and 7% before tax ROI)
Cost to Treat Feedstock
c/gal
LCO"
6.55
Coker "
4.72
Straight Run
Diesel "
3.47
a Based on the assumption that each refinery's costs will be comprised of; 80% for revamping a refiner's existing
hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs were determined based
on new equipment required while grass roots operating costs were assumed to be the same as a revamped unit.
In Chapter 4, we discussed the temporary compliance option and small refinery hardship
provisions with respect to refineries initiating compliance to the new highway diesel sulfur cap
standard in either year 2006 or 2010. The refining industry is expected to take advantage of the
temporary compliance option with the lowest cost producers complying during 2006-2009 and
the highest cost producers complying starting in 2010. In each PADD for year 2006, the lowest
cost refineries were added to the 2006 year pool until the volume requirement was meet for
producing 80% of the respective PADDs' 15 ppm temporary compliance sulfur diesel pool. In
addition, for each PADD, small refineries with costs that placed them in the 80% low cost
temporary compliance pool were considered to enter the market in year 2006. Cost for 2006 also
included small refineries that were projected to select the potion that allows extending the
implementation date of the Tier 2 gasoline sulfur requirement. All remaining refineries which
were not classified as being in the 2006 year pool were considered to comply in year 2010.
V-102
-------
Chapter V: Economic Impact
Table V.C-18. Overall Estimated Per-Refinery Capital, Operating and Per-Gallon Cost for
Years 2006 and 2010 for Implementation of Desulfurizing Highway Diesel Fuel to Meet a
15 ppm Cap Standard (1999 Dollars, 7% ROI before taxes)
Capital Cost
(SMillion)
Operating Cost
($Million/yr)
Per-Gallon Cost
(c/gal)
Year 2006 Average Refinery "
61
11.8
4.1
Year 2010 Average Refinery "
24
6.41
5.0
a Based on the assumption that each refinery's costs will be comprised of 80% of the cost for revamping the
refinery's existing hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs
were determined based on new equipment required while grass roots operating costs were assumed to be the same
as a revamped unit.
Our analysis of the average refinery capital, operating costs and average per gallon cost is
summarized in Table V.C-18. On average, the 63 refineries entering the year 2006 pool would
have capital costs of $61 million per refinery. The average capital costs for refineries that newly
enter the 15 ppm highway pool in year 2010 are $24 million per refinery. These costs reflect
that the large refineries have lower overall costs due to economies of scale and will enter the
highway diesel market in year 2006. By delaying the revamp costs for the highest cost diesel
hydrotreater units until 2010 the refinery industry will be able to defer $1.4 billion dollars over a
four year period.
Table V.C-19 shows the aggregate capital and operating costs for the U.S. refining
industry that were developed for 2006-2030. To calculate the aggregate capital cost, the total
capital cost for each of the 121 refineries which we estimated in our refinery model was summed
together. With the temporary compliance option and small refinery hardships provisions, capital
costs for the years of 2006 and 2010 were $3.9 and $1.4 billion, respectively. Capital costs for
complying in years 2006 and 2010 were spread to reflect project installation according to the
following; one third of the capital costs assigned to the one year period before the compliance
date with the remaining two thirds costs assigned to the two year period before the compliance
date. Capital costs which are estimated to total $5.3 billion are presumed to be incurred in 2004,
2005, 2006, and 2008, 2009, and 2010 as the desulfurization units are installed in the refineries.
To maintain future program compliance requirements, a second round of capital cost investments
is assumed to occur 15 years later as the desulfurization units installed are replaced at the
presumed end of their useful life. Aggregate capital costs increase for the 2nd round of
investment in 2019 - 2025 relative to 2004 - 2010 due to increased fuel production volumes
required to meet growth in diesel demand. We then calculated the yearly aggregate operating
V-103
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
costs based on the projected diesel consumption in 2006-2030 shown in Table V.C-19. The
aggregate operating cost is calculated by simply multiplying the average per-gallon operating cost
and the aggregate volumetric consumption together. The aggregate operating costs increase each
year due to the constant increase in growth in diesel demand. These costs are summarized in
Table V.C-19.
V-104
-------
Chapter V: Economic Impact
Table V.C-19. Projected U.S. Aggregate Operating and Capital Cost of Desulfurizing
Highway Diesel Fuel to Meet a 15 ppm Cap Standard (1999 Dollars, 7% ROI before taxes)
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
Projected 7 ppm
Diesel Fuel
Production11
(Billion Gals)
-
-
39.5*0.58
40.1
40.7
41.3
41.9
42.6
43.2
43.8
44.5
45.2
45.8
46.5
47.2
47.9
48.7
49.4
50.1
50.9
51.6
52.4
53.2
54.0
54.8
55.6
56.5
Projected Aggregate
Operating Cost
(SBillion)
0.64
1.04
1.05
1.07
1.11
1.13
1.15
1.17
1.18
1.20
1.22
1.24
1.26
1.27
1.30
1.31
1.33
1.35
1.37
1.39
1.42
1.44
1.46
1.48
1.50
Projected Aggregate
Capital Cost
(SBillion) a
1.3
1.9
0.7
-
0.5
0.7
0.2
1.5
2.2
0.8
0.6
0.8
0.3
Projected Total
Aggregate Cost
(SBillion)
1.30
1.90
1.34
0.75
1.55
1.77
1.31
1.02
1.04
1.05
1.07
1.09
1.10
1.12
1.14
2.77
3.5
2.11
1.20
1.95
2.17
1.69
1.28
1.30
1.32
1.34
1.36
For U.S. refiners only.
V-105
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-19 shows that the aggregate capital cost for complying with the proposed 15
ppm highway diesel sulfur cap is expected to total about $5.3 billion spread out over seven years.
This level of capital expenditure is estimated to be slightly more than the capital expenditures
expected to be made by the U.S. refining industry for complying with gasoline sulfur standards,
(see Section B of Chapter IV). We believe that these costs are not excessive. For example,
during the early nineties the U.S. refining industry invested over twenty billion dollars in capital
for environmental controls for their refining and marketing operations; 85 this cost represented
about one half of the total capital expenditures made by refiners for the downstream operations of
their refineries. Considering the effects of inflation we believe that a program requiring the
refining industry to spend about $5.3 billion 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 also estimated the per-gallon cost of this program based on different
capital cost amortization premises. In Table V.C-20 below, projected average per-gallon costs of
complying with the proposed sulfur cap for small refineries and non-small refineries are shown
based on various rates of return on investment (ROI) before taxes. The first row of costs shown
are our estimates of the costs to society, which utilize a seven percent before tax ROI. We then
present two additional cost estimates which are based on six and ten percent after tax ROIs.
These latter rates of return are indicative of the economic performance of the refining industry
over the past 10-15 years.
Table V.C-20. Per-Gallon Cost for Desulfurizing Highway Diesel Fuel to Meet a 15 ppm
Cap Standard Based on Different Capital Amortization Rates (1999 Dollars)
Societal Cost
7% ROI before Taxes
Capital Payback
(6% ROI, after
Taxes)
Capital Payback
(10% ROI, after
Taxes)
Average Cost of
Non Small
Refineries "
(c/gal)
4.2
4.3
4.6
Average Cost of
Small Refineries
(c/gal)
5.0
5.2
5.8
U.S. Average Cost
(c/gal)
4.3
4.4
4.7
1 Average refinery costs excludes refineries classified as small.
V-106
-------
Chapter V: Economic Impact
In Chapter 4, we addressed the ability of the refining industry to produce adequate
supplies of highway diesel fuel to avoid shortages under the 15 ppm highway diesel fuel cap
standard. First, the temporary compliance option and small refinery hardship provisions
substantially enhances supplies of highway diesel fuel by allowing roughly 22% of highway
diesel fuel to continue to meet the 500 ppm cap. This gives roughly 58 refineries four more years
before needing to invest in desulfurization equipment to meet the 15 ppm standard. By the time
these refiners need to decide on a desulfurization technology, those units built in 2006 will have
been operating for 1-2 years, providing commercial data upon which to conduct a comparison.
This data will help these refiners to borrow money, if necessary, to pay for the new equipment.
The other factor easing highway diesel supplies is the ability of a number of refiners to
economically produce 15 ppm fuel from current nonhighway diesel fuel blendstocks. To
quantify this factor, we developed a model to estimate the cost to each refinery of desulfurizing
all their existing nonhighway diesel fuel to an average sulfur level of 7 ppm (i.e., that needed to
ensure compliance with the 15 ppm cap). These costs were developed for all U.S. refineries that
currently produce nonhighway diesel. Especially in cases where grass roots refinery
modifications are necessary to process current highway diesel fuel to 15 ppm sulfur, there are no
competitive disadvantages, and in some cases improved economies of scale by investing to
convert current nonhighway diesel to highway diesel. As was the case when estimating each
refinery's cost to produce 15 ppm fuel from its highway diesel blendstocks, the cost for
processing nonhighway diesel blendstocks were based on volume throughput and feedstock
compositions. Again, as was done for their highway diesel blendstocks, each refinery's
nonhighway blendstock composition was estimated from distillate pool information taken from
the data provided by EIA for 1998 and 1999. These processing costs were reduced by using the
average price differential between highway and nonhighway diesel fuel of EIA83 and Muse
Stancil & Go's 84 product pricing data. The EIA data was based on historical price difference
between highway and nonhighway diesel fuel at the refinery gate while Muse Stancil & Go's
pricing data was based on the historical price difference between low and high sulfur No. 2 Oil of
batches being transported by pipeline to market. Using this average for credit is appropriate,
since the highway diesel fuel produced from nonhighway diesel blendstocks would command the
price of highway diesel fuel under the new sulfur cap, compared not to the price of highway
diesel fuel prior to the cap, but to the price of nonhighway diesel fuel prior to the cap. See Table
V.C-21.
V-107
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-21. PADD-Average Price Difference Between 500 ppm Highway and Non-
Highway Diesel (1999 Dollars, 7% ROI before taxes)
PADD 1
PADD 2
PADD 3
PADD 4
PADD 5
Muse, Stancil's a
Delta Price
Between Low and
High Sulfur No. 2
Oil (c/gal)
2.0
0.0
2.8
2.1
3.9
EIA a Delta Price
Between Highway
and Nonhighway
Diesel
(c/gal)
1.6
1.8
1.6
5.0
Average of EIA
and Muse Stancil
& Go's data
(c/gal)
1.8
0.9
2.2
2.1
4.5
a EIA data based on 1995-1999 average price difference between low and high sulfur diesel fuel. Muse,
Stancil & Co. prices from Alternate Markets for Highway Diesel Fuel Components, September 2000 and
are based on 1995-1999 average price difference between low and high sulfur No.2 Oil. Overall volume
weighted highway and nonhighway diesel cost adjustment for USA PADD regions is 2.2 c/gal.
Through this analysis, we found that a number of refineries could produce highway diesel
fuel from nonhighway diesel blendstocks in separate hydrotreating units at a cost which was
competitive with other refineries in their PADD. In these cases, the volume of nonhighway
diesel fuel was large so, regardless if the refineries are producing highway or not, we assumed
that these would be new grassroots units. In our model, the nonhighway diesel blendstocks are
processed in a new grass roots unit, while the highway diesel blendstocks are processed in either
a revamped or grassroots unit, according to an 80:20 ratio. In reality, a refinery deciding to
process both its highway and nonhighway diesel blendstocks to meet a 15 ppm cap would likely
do so in a single grassroots unit sized to process both current products. We compared the cost of
such a single larger grassroots unit to the two unit approach for a few refineries and found that
the single grassroots unit would be less costly. Thus, the costs used in this analysis, which
assume that the refinery would process its nonhighway diesel blendstocks in a separate unit, are
likely to be slightly overestimated. For hydrotreater highway units with large volumes of
highway and small volumes of nonhighway diesel, combining the two production streams as feed
for revamping the existing hydrotreater would provide economies of scale and would reduce the
overall costs in generating 15 ppm sulfur cap highway fuel. The costs used in this analysis did
not consider this option. Figure V.C-2 illustrates that additional distillate volume would be
available as feedstock to convert to highway diesel. Number 2 Oil in this figure is the
summation of highway and nonhighway diesel fuel per refinery.
V-108
-------
Chapter V: Economic Impact
Figure V.C-2. Refinery Specific Production Rates of Highway Diesel versus No. 2
Distillate Poola
Oil
CD
L"- CO
"CD Q
CO !_
CD CD
Q °-
^ w
CO CD
J CO
0) DQ
±
ou,uuu
70,000
60,000
50,000
40,000
30,000
20,000
10,000
1
i
_ u
m
rj-\
— 1 — ' H
\\ H
l~nU 1 1
• ' ' '
n IV n • ^
n U t4 U _ • _
^ ^
•-H— • " -• n
rCCfl n
r| jl | |1 1P 1 ' 1 | |
rfTOTlSf ^ °
0 20,000 40,000 60,000 80,000
10,000 30,000 50,000 70,000
No. 2 Oil Distillate
Barrels Per day
a Per Annum Refinery Specific plot of Highway Diesel Production volume versus total No. 2 Distillate volume
produced by the refinery. Based on EIA refinery production data for 1998/1999.
V-109
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Figure V.C-3. Lowest Refinery Costs for Converting NonHighway to 15 PPM Highway
Diesel Fuel A
5
4.5
c
o
(0
1 4
O
3.5
3
(
.
•
) 1 2 3 4 5 6 7 8
Cumulative Number of Converted Off Highway Diesel Units
Costs per treated volume of nonhighway diesel, 1999 dollars, and 7% ROI before taxes.
Overall, we found 2 refineries which produce essentially no highway diesel fuel today which
could meet the new 15 ppm standard for less than 5.0 cents per gallon. Production from these
refineries would increase nationwide highway diesel fuel production by 2 percent. We also
found that 4 other refineries could increase production of highway diesel fuel from their
nonhighway diesel fuel blendstocks for less than 5.0 cents per gallon. Production from these 4
refineries would increase highway diesel fuel production by an additional 5 percent. See Figure
V.C-3 for plot of the cost of these nonhighway diesel fuel converted units.
A sensitivity analysis was then performed to estimate the cost of meeting the 15 ppm
sulfur cap if some of the blendstocks currently being used to produce nonhighway diesel fuel
were used to produce 15 ppm diesel pool and some of the refineries currently producing highway
diesel fuel shifted their fuel to the nonhighway diesel fuel market.
We imposed a number of restrictions on such shifts. First, 15 ppm diesel fuel produced
from nonhighway blendstocks used in PADDs 3, 4 and 5 had to be produced in those PADDs,
V-110
-------
Chapter V: Economic Impact
with the further restriction that no such fuel could be transported to either Hawaii or Alaska from
outside of those states. Second, 15 ppm diesel fuel produced from nonhighway blendstocks used
in PADD 2 had to either come from within the PADD or could come from PADD 3 if it
displaced higher cost highway diesel fuel in the southern portion of PADD 2. Practically, this
limited any additional transfers of 15 ppm fuel from PADD 3 to PADD 2 to a very small amount
(0.05 percent of current PADD 2 highway diesel fuel production). Finally, 15 ppm diesel fuel
produced from nonhighway blendstocks in PADD 3 was allowed to displace current highway
diesel fuel produced in PADD 1. PADD 3 currently sends sizeable amounts of both highway and
nonhighway diesel fuel to PADD 1. The relative amount of highway diesel fuel produced in
PADD 3 could therefore easily increase and the amount produced in PADD 1 decrease without
changing the total volume of diesel fuel transported. We found that about 14% of current PADD
1 highway diesel fuel production could be made in compliance more economically from
nonhighway diesel blendstocks in PADD 3. After considering these restrictions in the
substitution of nonhighway to highway diesel fuel, only 5 percent of the total 15 ppm highway
production volume is shifted to replace the high cost highway producers. This is less than the 7
percent of nonhighway diesel fuel which we found available with estimated costs less than 5
cents per gallon. Table V.C-22 highlights the cost difference between the nonhighway
hydrotreaters and the highway producers which were supplanted by the nonhighway producers.
Table V.C-22. Costs Under Nonhighway Production Shift Scenario
(1999 Dollars, 7% ROI before taxes)
Number of Refineries
Capital Cost, Per Refinery
(SMillion)
Operating Cost, Per Refinery
($Million/yr)
Per-Gallon Cost
(c/gal)
Higher Cost
Highway Units'1
17
12
1.5
6.3
15 ppm Diesel
from
NonHighway
6
29
2.5
4.5
a Based on the assumption that each refinery's costs will be comprised of 80% of the cost for revamping the
refinery's existing hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs were
determined based on new equipment required while grass roots operating costs were assumed to be the same as a
revamped unit.
V-lll
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The effect of this shift on average costs and total capital cost are very small. These are
shown in Table V.C-23. The effect of this shift on the maximum cost in each PADD is more
significant, particularly in PADDs 1 and 5. In these PADDs, it would be very expensive to bring
a very small percent of current highway diesel fuel production into compliance with the 15 ppm
cap, primarily because of poor economies of scale. Refer to Section IV, Table IV. A-7 which
compiles the PADD specific reduction in maximum costs attributed to using nonhighway to
make 15 ppm sulfur cap highway diesel. With supplemental 15 ppm fuel from current
nonhighway blendstocks, these small quantities of current highway fuel can be shifted to the
nonhighway diesel fuel market with no loss of supply of highway diesel fuel or flooding of the
nonhighway markets. Figures V.C-4 and V.C-5 illustrate the use of supplemental nonhighway
to reduce maximum costs in each PADD. Figure V.C-4 represents the distribution of refinery
cost by PADD for the case where production shifts were not presumed to occur between
nonhighway and highway diesel producers. Whereas Figure V.C-5 represents a similar plot
where production shifts are allowed. Comparing the two figures demonstrates that production
shifts from nonhighway to highway would eliminate the highest cost producers. Both figures
reveal that, for each PADD, costs are relatively constant for highway production volumes from 0
to 80 percent with the costs escalating for volumes greater than 80 percent. Inspection of the
Figures also show that PADD 4 has the highest costs while PADD 3 has the lowest costs for
producing highway diesel fuel which meets the 15 ppm sulfur cap standard.
V-112
-------
Chapter V: Economic Impact
Table V.C-23. Estimated Costs of Nonhighway Production Shift Scenario versus Current
Highway Producer Scenario to Meet 15 ppm Highway Diesel Fuel Cap Standard (1999
dollars, 7% ROI before taxes)
U.S. Aggregate Capital Cost
(SBillion)
U.S. Aggregate Operating Cost
($Million/yr)
Average Refinery Capital Cost
(SMillion)
Average Refinery Operating Cost
($Million/yr)
Average Per-Gallon Cost
(c/gal)
Nonhighway
Units Shift
Scenario a
5.4
970
51
9.1
4.2
Current
Highway Units
Scenario
5.3
960
44
7.9
4.3
a Based on the assumption that each refinery's costs will be comprised of 80% of the cost for revamping the
refinery's existing hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs
were determined based on new equipment required while grass roots operating costs were assumed to be the same
as a revamped unit.
V-113
-------
Figure V.C-4. Refinery Costs per PADD for Current Highway Units Scenario for Meeting
the 15 ppm Sulfur Highway Diesel Fuel Cap Standard a
(1999 dollars, 7% ROI before taxes)
ro
CD
O
•» PADD 1
x PADD2
•*• PADD 3
•* PADD 4
— PADD5
10 20 30 40 50 60 70 80 90 100
Cumulative Production (%) of Highway Diesel Pool
a Costs excludes Hawaiian, Alaskan, and small refineries projected to take the gasoline extension option. Based on
the assumption that each refinery's costs will be comprised of 80% of the cost for revamping the refinery's existing
hydrotreater unit and 20% for building a new grassroots unit. Grass roots units capital costs were determined based
on new equipment required while grass roots operating costs were assumed to be the same as a revamped unit.
Figure V.C-5. Refinery Costs per PADD under Converted NonHighway Units Shift
Scenario for Meeting the 15 ppm Highway Diesel Fuel Cap Standard"
(1999 dollars, 7% ROI before taxes)
1 9
11
m
c
0 o
CO
CD ,
^n '
-§->
c c
O
2
i
^- ^- •*- • •*- • - *~
**" ' ~~ >(,,i»f ^-*~ ' • ~*t-£-if*
^«r--^~™— -**~
D 10 20 30 40 50 60 70 8
Cumulative Production (%) of Highway Diesel Pool
•\
/>
J-
•^"£^
Jx—'X.'C*^ltl
-£--f-jfo
t+**^^ r
0 90 1C
* PADD 1
x PADD 2
•^ DARR •*
•^ PADD 4
— PADD 5
)0
Costs excludes Hawaiian, Alaskan, and small refineries projected to take the gasoline extension option.
V-114
-------
Chapter V: Economic Impact
a. Other Cost Estimates for Desulfurizing Highway Diesel Fuel
A number of cost estimates of the 15 ppm highway diesel fuel sulfur standard were
submitted as part of the comments on the proposed rulemaking. Mathpro used a notional
refinery model to estimate the national average costs of the proposed standard for the Engine
Manufacturers Association (EMA). For the American Petroleum Institute (API), Charles River
Associates, along with Baker and O'Brien, used the Prism refinery model to estimate the cost of
U.S. refineries to produce highway fuel in the U.S. EnSys used the Oak Ridge National
Laboratory PADD 3 refinery model to estimate costs for the Department of Energy(DOE).
Finally, the National Petroleum Council (NPC) used the Mathpro refinery modeling work to
estimate a cost for meeting a less stringent standard. The cost estimates from each of these
studies is presented in the respective sections and, if appropriate, compared to our cost analysis.
Mathpro's Cost Analysis for EMA
In a study conducted for the EMA,78 MathPro, Inc. estimated the cost of desulfurizing
diesel fuel to meet a 15 ppm highway diesel fuel sulfur cap standard. MathPro assumed that
desulfurization would occur entirely through severe conventional hydrotreating, and refining
operations and costs were modeled using their ARMS modeling system with technical and cost
data provided by Criterion Catalyst Company LP, Akzo-Nobel Chemicals Inc., and Haldor
Topsoe, Inc. The resulting cost estimates were created based on what Mathpro terms a
"notional" refinery. The notional refinery is configured to be typical of the refineries producing
highway diesel fuel for PADDs 1, 2, and 3, and also represent the desulfurization cost for those
three PADDs based on the inputs used in the refinery model. The Mathpro notional refinery
model maintained production of highway diesel fuel at their base levels.
Mathpro made a number of estimates in their study to size their diesel desulfurization
units for estimating the capital cost, and these estimates were similar to those included in our
methodology. The calendar day volume was adjusted to stream day volume using a 10 percent
factor to account for variances in day-to-day operations, and another 10 percent to account for
variance in seasonal demand. In addition, Mathpro applied a factor which falls somewhere in the
range of 1 - 8 percent for reprocessing off-spec material to meet a number of different sulfur
targets. Since meeting a 15 ppm cap standard is a relatively stringent sulfur standard compared
to the sulfur levels studied, Mathpro likely assumed the desulfurization unit would be sized
larger by 5 - 8 percent. Onsite investment was adjusted to include offsite investment using a
factor of 1.4. In the final report, capital costs were amortized at a 10 percent after tax rate of
return.
There are several differences between our cost analysis and the cost analysis made by
Mathpro. First, the MathPro costs are based on a 10 percent ROI after taxes. As stated above,
our costs are calculated based on a 7 percent rate of return on investment (ROI) before taxes, so
V-115
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
to compare our cost analysis with the cost analysis made by Mathpro, we adjusted the Mathpro
costs to reflect the rate of return on capital investment which we use. Second, the Mathpro study
did not attempt to project how much of highway diesel fuel will be produced by revamping
existing diesel hydrotreaters versus installing new grassroots units. Instead, Mathpro provided
cost estimates for both revamped and grassroots units. This range of costs is presented here, and
we include a cost which represents 80 percent revamp and 20 percent grassroots units. Third, the
MathPro estimate includes a cost add-on (called an ancillary cost) for reblending and
reprocessing offspec diesel fuel or for storing nontreated diesel fuel. While this is conceptually
an appropriate adjustment, it appears that some of the reblending costs in the MathPro study
appear to be transfer payments,11 not costs. Fourth, MathPro assumed that all new hydrogen
demand is met with new hydrogen plants installed in the refinery, which does not consider the
advantage of hydrogen purchased from a third party which can be produced cheaper in many
cases. As a result, their hydrogen cost may be exaggerated, which would tend to increase costs.
Finally, it should be noted that the MathPro study did take into consideration the need for
lubricity additives, but did not address costs that might be incurred in the distribution system.
Thus, in a comparison of our costs with Mathpro's, we will include our cost estimate for adding
the appropriate amount of lubricity. A comparison of Mathpro's cost and our cost to desulfurize
highway diesel fuel to meet a 15ppm sulfur cap standard is shown below in Table V.C-24.
Table V.C-24. Comparison of Mathpro's and EPA's Costs for Meeting a 15 ppm Highway
Diesel Fuel Sulfur Cap Standard (7% ROI before taxes)
Per-gallon Cost
Capital Cost
Mathpro's Cost
4.2-6.1 (4.6)
3.4-6.1 (3.9)
EPA's Cost
4.3
5.3
Cost assumes the addition of lubricity additives, but no distribution costs.
Lower end of the range in per-gallon costs assumes 100 percent revamped equipment; upper end assumes
all new equipment; EPA costs and the Mathpro costs in parentheses assume 80 percent revamps and 20
percent new units.
Charles River and Baker and O'Brien Study for API
Charles River Associates and Baker and O'Brien (heretofore referred to as CRA), in a
study for API, analyzed the impacts of a 15 ppm highway diesel fuel sulfur cap standard on the
U.S. oil industry. Nonroad diesel fuel was also reduced to 350 ppm, probably to meet an
assumed future 500 ppm cap standard. CRA used the Prism refinery model along with their own
p A transfer payment is when money changes hands, but no real resources (labor, natural resources,
manufacturing etc.) are consumed.
V-116
-------
Chapter V: Economic Impact
estimates of hydrotreating costs to estimate the cost to each refinery of meeting the cap standard
taking into account the estimated fractions of the various blendstocks which comprise highway
diesel fuel and the quality of crude oil used by each refinery. CRA based their cost analysis on
desulfurization technology (not on ring opening technology, and hydrogen consumption was
similar to Mathpro's), but estimated that 40 percent of refiners would build new hydrotreating
units with the balance of refiners revamping their existing units.
CRA surveyed the major refiners which produce about half of the total amount of
highway diesel fuel produced in the U.S. asking if they anticipated producing highway diesel
under a 15 ppm sulfur cap standard. Refiners responded with a range of responses. Some said
that they would increase or maintain their highway diesel fuel production, while others said that
they would decrease their production. CRA concluded from their analysis of the survey
responses that highway diesel production would decrease by 9 to 11 percent. Since this was an
estimated shortfall in domestic highway diesel fuel production associated with a lack on
investment by a large number of refineries, only imports were presumed to be available to make
up the difference.
CRA's estimates for sizing their diesel desulfurization units are summarized here. First,
each unit size is increased by 20 percent to account for sizing a unit's calendar day volume to a
stream day volume, which addresses variances in daily or seasonal highway diesel production
output, and unit downtime. Then, CRA assumed, based on a study by Baker and O'Brien, that 10
percent of the highway diesel fuel being produced would be downgraded to nonhighway diesel
due to contamination in the distribution system. To make up for that loss in volume, each
refinery's diesel desulfurization unit size and the operating costs were increased by 10 percent to
account for this projected volume shortfall. The unit size was increased by another 10 percent to
account for reprocessing of offspec batches. Thus, after consolidating all these factors, each
refinery unit was sized 40 percent larger than calendar day volume. Then, the calculated capital
costs were adjusted upward by 20 percent to cover contingencies. In estimating per-gallon costs,
CRA amortized the capital costs at a 10 after tax percent rate of return.
CRA did not directly provide an average cost estimate for their analysis, estimate an
average cost from CRA's report, we examined CRA's cost curve which plots individual cost for
each refinery in the U.S., which CRA assumes are continuing to produce highway diesel fuel,
against cumulative highway diesel fuel production. The average cost for the U.S. refineries is
about 6.2 c/gal. CRA did not attempt to determine a diesel desulfurization cost for the balance of
the highway diesel fuel which would have to be made up by imports.
We have a couple of observations and comments on the analysis by CRT. First, the study
incurred costs for desulfurizing nonroad diesel fuel to meet a 500 ppm cap standard, however, the
study's report did not provide the reader with information to determine what impact desulfurizing
nonhighway might of had on the per-gallon cost of desulfurizing highway diesel fuel. CRA
assumed that this 500 ppm fuel would be produced by blending 8 ppm sulfur highway diesel fuel
V-117
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
and 3000 ppm nonroad diesel or heating oil. While, much of this production was assumed to
occur due to mixing in the distribution system, an unknown amount of 500 ppm fuel was
produced at refineries. Desulfurization costs are not linear, as shown by CRA's own study.
Thus, any blending of 15 ppm sulfur highway diesel fuel with non-desulfurized heating oil at
refineries was much more costly than simply hydrotreating nonroad diesel fuel to 500 ppm.
Second, the cost study conservatively assumed that refiners would build their diesel
desulfurization units 40 percent larger than their calendar day production volume. Our analysis
assumed that the revamped or grassroots units would be sized 20 percent larger than the calendar
day diesel fuel volume being desulfurized, and Mathpro assumed that the revamped and
grassroots units would be sized 25 percent larger. Finally, the analysis did not attempt to
estimate the likelyhood and did not estimate the cost of nonhighway diesel
On a more fundamental level, we doubt that the perspective of whether to invest or not
held by the surveyed refiners might have had earlier this year, or even now, will necessarily be
the perspective that they will have several years from now when construction of the new units
will have to begin. For example, many of these refiners haven't had the chance to test their
diesel fuel to really understand what their cost would be for desulfurizing their highway diesel
fuel. As the development of catalysts progresses which vendors expect to occur over the next
two or so years,q refiners may see that the difficulty and cost of meeting the cap standard is not as
high as they once thought. Furthermore, these refiners would likely not make a firm decision on
how they will invest at this point in time because they would need to better understand the plans
of the rest of the refining industry. The temporary compliance option will give refiners insight
on who will participate in the program and what their likely market share will be for distillate
products. If refiners do not consider the intended actions of their fellow refiners, there is
significant economic risk. Using this analysis as an example, if refiners invest in a way that
would result in a shortfall of 12 percent in highway diesel fuel capacity, we estimate that there
would be overproduction of nonroad diesel fuel by 20 percent. Those refiners which choose not
to produce highway diesel fuel would see the price of nonroad diesel drop through the floor and
their profits suffer accordingly. We do not believe that refiners would put themselves at that kind
of risk.
Ensysfor the Department of Energy
Ensys estimated the cost of desulfurizing highway diesel fuel for the Department of
Energy (DOE). Ensys studied various levels of desulfurization, however, we will discuss the
case which estimated the cost of averaging 8 ppm, which is about the level of sulfur control
needed to meet the 15 ppm cap standard. Ensys only studied the cost of meeting the highway
diesel fuel sulfur requirement in PADD 3. EnSys did not estimate how many refiners would
q Two vendors have announced higher activity desulfurization catalysts since the point in time that the CRI
survey was completed.
V-118
-------
Chapter V: Economic Impact
build new desulfurization units and how many would modify their current hydrotreaters, but
presented costs for doing either. Thus, the lower limits of the ranges shown in Table V.C-24
assume refiners will modify their current hydrotreaters, while the upper limits assume that
refiners would build new units. EnSys also projected costs for two separate sets of technologies.
One set was considered conservative and relied upon technology that is already in commercial
use. EnSys' costs using the conservative technology are higher than our estimates. This is due to
the fact that this technology involves greater capital investment and greater consumption of
hydrogen, because this technology is not just designed to reduce sulfur, but to reduce aromatic
content, increase cetane levels and perform some cracking. The second technology analyzed by
EnSys was labeled as optimistic. We believe the technology assumed to be used in the optimistic
case was similar to that projected to be used by EPA (as well as CRA and Mathpro) since Ensys
developed these costs after we shared vendor information with Ensys. Ensys reported their costs
based on a 10 percent after tax return on investment, however, in Table V.C-25 below, we
adjusted the Ensys costs to a 7 percent ROI before taxes.
Ensys made the following estimates for sizing their diesel desulfurization units. Unit size
based on calendar throughput was increased by 5 percent to account for unit downtime, then an
additional 15 percent calendar to stream day factor was added on. Unit size was adjusted upward
by another 15 percent as a "redundancy" factor to cover the processing of off-spec batches. The
offsite battery limit capital costs for new units were 40 percent of the onsite battery limit costs,
while revamp unit inside and offsite capital costs were 50 percent of new unit onsite and offsite
costs. Ensys received comments on their modeling study by a refining industry consultant with
Pricewaterhouse Coopers retained by the Association of Automobile Manufacturers. The
consultant commented on a series of cost factors used in Ensys' refinery modeling study. Ensys
estimated maintenance costs to be 3.5 percent of total capital costs, while the consultant
explained that the maintenance cost typically is 2.5 - 3 percent of the refinery's replacement
value. Ensys estimated taxes insurance and overhead to be 2% of total investment, while the
refining industry typically experiences 0.5 to 0.7 percent for taxes and insurance. The consultant
also recommended that three other factors, 3 percent for buildings, 7 percent for environmental
and 10 percent for startup, be reduced by 50 percent.
V-119
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table V.C-25. Comparison of DOE and EPA Refining Costs for Meeting a 15 ppm
Highway Diesel Fuel Sulfur Cap Standard (7% ROI before taxes)
Per-Gallon Cost (c/gal)
Capital Cost ($MM)
DOE Conservative
Technology
5.1 -6.0(5.3)
3.9-6.5(4.4)
DOE Optimistic
Technology
4.2 - 4.4 (4.2)
2.7-4.5(3.1)
EPA
4.3
5.3
Lower end of the range in per-gallon costs assumes 100 percent revamped equipment; upper end assumes all new
equipment; EPA costs and DOE costs in parentheses assume 80 percent revamps and 20 percent new units.
DOE costs are only for the Gulf Coast refining region, which have slightly lower per-gallon costs than the entire
U.S., and about half the capital costs.
National Petroleum Council Study
At the request of the Secretary of Energy, the National Petroleum Council (NPC) studied
the impact of various possible fuel programs on the industry's capability to continue to produce
and distribute refined products, and maintain the viability of its refineries. The fuel programs
studied by the NPC include desulfurizing gasoline, desulfurizing diesel fuel, eliminating MTBE
from gasoline, and reducing the driveability index of gasoline. To carry out the study, the NPC
established a committee comprised primarily of representatives of the oil industry, but
representatives of the pipeline companies, engineering contractors, the Department of Energy,
and the EPA participated as well. An important part of the study was to estimate the cost of the
fuel programs being studied. The NPC estimated the cost for desulfurizing diesel fuel to meet an
average sulfur standard of 30 ppm. Since the NPC did not study the cost of a 15 ppm cap
standard, we cannot compare the NPC costs with our costs. However, it would still be useful to
summarize those costs to get some indication of how an NPC cost for 15 ppm would compare to
ours.
The NPC did not fund its own refinery modeling work. Instead, NPC relied upon the
EMA-funded Mathpro cost analysis as the basis for its cost analysis. NPC concluded that it does
not believe that the Mathpro study adequately captured the costs of achieving the very low sulfur
levels included in some of the Mathpro study cases. While NPC admits that it could not review
the vendor submissions on which Mathpro based its analysis, nevertheless, NPC concluded that
the vendor data used was optimistic about achieving very low sulfur levels treating typical
feedstocks which are eventually blended into highway diesel fuel. Consistent with its conclusion
that the Mathpro analysis was optimistic, the NPC made a number of adjustments to the Mathpro
cost analysis to provide its own cost analysis. Capital investments were increased by 20 percent.
However, how hydrogen consumption was handled was less clear as early on in the report,
hydrogen consumption and other operating costs were increased by 15 percent, but later on in the
report the study described the adjustment for hydrogen consumption to be 20 percent. Also, the
report stated that the offsite factor for the diesel desulfrization units were reduced from 1.5 to
V-120
-------
Chapter V: Economic Impact
1.4. Thus, assuming that both adjustments applied, there was a net increase in the investment
costs of 10 percent. Mathpro modeled various refiner investment strategies which included the
construction of a new unit and a revamp with another reactor in series. To meet a 30 ppm
average sulfur standard, NPC assumed that half of highway diesel fuel would be desulfurized by
a revamped unit, while the other half would be desulfurized with a new unit. After making these
adjustments, NPC estimated that desulfurizing highway diesel fuel down to 30 ppm on average
would cost 5.8 c/gal with capital costs amortized at a 10 percent after tax rate of return.
We have several comments on NPC's diesel cost analysis. First, NPC applied cost
adjustment factors to increase the Mathpro cost analysis without having seen the vendor
submissions. Also NPC adjusted Mathpro's cost estimates based on its assertion that the
vendor's costs are overly optimistic. Even if NPC's adjustments factors correctly account for
overoptimism in the vendor's estimate, they don't consider expected reductions in operating
costs, and perhaps even capital costs, likely to occur as diesel desulfurization technology
improves over time. Second, the considerations voiced above concerning Mathpro's modeled
source and cost of hydrogen still apply for the NPC costs as well. Finally, NPC assumed 50-50
mix for revamps and new units which seems conservative for a moderate decrease in sulfur.
NPCs mix of revamp and new units is much more conservative than the Charles River and Baker
and O'Brien analysis for API. The analysis for API assumed a 40-60 mix for revamps and new
units, respectively, however, for meeting a much more stringent 15 ppm cap sulfur standard.
3. The Added Cost of Distributing Low-Sulfur Fuel
a. Summary
Please refer to section IV.D. in this RIA for a detailed discussion of the changes that will
need to take place in the highway diesel fuel distribution system as a result of our program. This
section addresses the costs of these changes. The majority of the increase in distribution costs to
adequately limit sulfur contamination during the distribution of 15 ppm diesel fuel are associated
with an increase in the volume of highway diesel fuel that must be downgraded to a lower value
product during transport by pipeline. There are also substantial costs associated with the need for
additional storage tanks to handle two grades of highway diesel fuel during the initial period of
our sulfur program when two grades of highway diesel fuel are allowed to be sold (15 ppm and
500 ppm sulfur cap highway diesel fuels).
We estimate that as a result of our sulfur program, distribution costs will increase by 0.5
cents per gallon of highway diesel fuel supplied when the sulfur requirements are fully effective
beginning in the year 2010. During the initial years of our sulfur program (2006 through mid-
2010) we estimate that the increase in distribution costs will be 1.1 cents per gallon of highway
diesel fuel supplied. This estimate includes 0.7 cents per gallon for new storage tanks to handle
two grades of highway diesel fuel (500 ppm and 15 ppm) during the initial years. For the sake of
simplicity and to allow a comparison with distribution costs once the sulfur program is fully
V-121
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
effective, the distribution costs during the initial years are also expressed in terms of the total
volume of highway diesel fuel supplied. This includes 500 ppm as well as 15 ppm highway
diesel fuel.
In the proposed rule, we estimated that distribution costs would increase by 0.2 cents per
gallon if the proposed requirement that the entire highway diesel fuel pool meet a 15 ppm sulfur
cap beginning in 2006 be adopted. This cost was comprised of roughly 0.1 cents per gallon due
to an increased volume of highway diesel fuel downgraded to a lower value product during
shipment by pipeline and additional terminal testing costs, and 0.1 cents per gallon for
distributing the additional volume of highway diesel fuel needed due to an anticipated decrease in
fuel energy density as a side effect of reducing the sulfur content to the proposed 15 ppm cap.
The case evaluated in the Notice of Proposed Rulemaking (NPRM) is most similar to that for the
fully effective sulfur program in the final rule.
We took advantage of additional information contained in the comments to the NPRM in
formulating a more comprehensive estimate of the distribution costs for the final rule. In some
cases this involved adjusting an estimate for a parameter that factored into our calculation of
costs in the NPRM. One important example is that we increased our estimate of the additional
volume of highway diesel shipped by pipeline that would need to be downgraded to a lower
value product. This downgrade volume is primarily the result of mixing that occurs between
highway diesel fuel and high sulfur products that are shipped in the pipeline adjacent to highway
diesel fuel. This mixture is referred to as interface when it can be blended into another product
and transmix when it must be returned to the refinery for reprocessing. In other cases, our
reevaluation of distribution costs included the consideration of parameters that did not factor into
the estimation of distribution costs in the proposed rule. For example, commenters to the NPRM
brought to our attention that there would be additional costs associated with changes in the
handling practices for interface volumes that result from shipments of highway diesel fuel and jet
fuel or kerosene which abut each other in the pipeline. We also attributed some cost to account
for the process of testing and optimizing the distribution system to limit sulfur contamination.
This includes the cost for testing to evaluate potential sources of contamination, and for
miscellaneous minor procedural and hardware changes that may be needed, but have yet to be
identified.
There are a number of common factors in the estimation of distribution costs during the
initial years of our program and after the sulfur requirements becomes fully effective, such as the
increase in interface volumes for pipeline shipments of highway diesel fuel. However, there are
other factors that are unique to the estimation of costs during the initial years of the program.
The factors that cause distribution costs to differ during the period when both 15 ppm and 500
ppm fuels are available for highway use:
Having a lesser volume of 15 ppm diesel fuel in the system during the initial years of the
program reduces some of the direct costs associated with distributing 15 ppm fuel.
V-122
-------
Chapter V: Economic Impact
Having an additional grade of highway diesel fuel in the system (500 ppm) during the
initial years of our program creates additional pipeline interface volumes, and additional
product downgrade costs. Having 500 ppm highway diesel fuel in the system during the
initial years of our program also allows some opportunity for the pipeline interface
volumes associated with the shipment of 15 ppm fuel and jet fuel or kerosene to be
downgraded to 500 ppm diesel fuel rather than off highway diesel fuel. This will reduce
the cost of making this downgrade.
The need for additional storage tanks to handle an additional grade of highway diesel fuel
when the optional compliance option program is available creates additional costs that
must be accounted for during the initial years of our program.
Table V.C-26 on the following page presents a summary of the distribution costs during
the initial years of our sulfur program and after the program becomes fully effective. The manner
in which these costs were estimated is discussed in the following sections.
V-123
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table V.C-26. Distribution Costs During the Initial Years of Our Sulfur Program and
After the Program Becomes Fully Effective
Cost Components
Cost to Distribute Additional Volume Needed to
Compensate for Reduced Energy Density of 15
ppm Sulfur Highway Diesel Fuel
Cost to Downgrade Additional Volume of 15 ppm
Sulfur Highway Diesel Fuel to a Lower Value
Product During Transport by Pipeline
Increased Cost for the Current Volume of
Highway Diesel Fuel that Must be Downgraded in
the Pipeline System
Increased Cost to Downgrade the Interface
Volume Between Pipeline Shipments of Highway
Diesel Fuel and Jet Fuel or Kerosene to Off
Highway Diesel Fuel
Cost of Increased Terminal Testing
Cost of Additional Tanks to Handle Pipeline
Interface Between Shipments of Jet Fuel and 15
ppm Sulfur Highway Diesel Fuel
Cost to Downgrade the Interface Volumes
Associated with Pipeline Shipments of 500 ppm
Fuel During the Initial Years of Our Program
Cost of Additional Tanks at Refineries, Terminals,
Bulk Plants, and Truck Stops to Handle Two
Grades of Highway Diesel Fuel During the Initial
Years of Our Program
Cost of Optimizing the Distribution System to
Limit Sulfur Contamination8
Total
Distribution Costs (cents per gallon of all
highway diesel fuel supplied) A
Fully Effective Sulfur
Program
(20 10 and later)
0.17
0.14
0.09
0.07
0.002
Completely amortized
during the initial years
of program
No Additional Cost
Completely amortized
during the initial years
of program
0.025
0.5
Initial Period
(2006-2010)
0.14
0.10
0.08
0.03
0.002
0.009
0.004
0.7
0.027
1.1
A During the initial years of our program, "all highway diesel fuel" includes 500 ppm highway diesel fuel as well as
15 ppm highway diesel fuel.
Cost amortized over the first 15 years of our sulfur program (through 2020).
V-124
-------
Chapter V: Economic Impact
There were some instances where we recognized that the rule would cause some change
to current industry practice, but we concluded that the associated costs would not be significant.
In one such case, we acknowledged that tank-truck operators and other distributors of highway
diesel fuel downstream of the pipeline may need to be more careful in their observance of current
industry practices used to limit product contamination, but we estimated that this would not
result in a significant increase in costs (see Section V.C.3.L). In another such case, we
recognized that the use of diesel fuel additives with a sulfur content above 15 ppm would likely
be phased out gradually by marketplace forces resulting from our diesel sulfur program, but
concluded that this would be accomplished without a significant burden (see Section V.C.3.J.).
Our response to the public comments on the NPRM related to the costs of our sulfur
control program are contained in a separate Response to Comments (RTC) document.
b. Cost of Distributing the Additional Volume of Highway Diesel Fuel Needed
to Compensate for a Reduction in Energy Density
The energy density of highway diesel fuel is expected to decrease as a side effect of
reducing the sulfur content to meet the proposed 15 ppm cap. As a result of this reduction in
energy density, an increased volume of diesel fuel will need to move through the distribution
system to meet the same level of consumer demand. The cost of producing this additional
volume is included in the calculation of refinery costs (see Section V.C. 1.). The cost of
distributing the additional volume of highway diesel fuel needed to compensate for the lower
energy density of highway diesel fuel that meets a 15 ppm sulfur cap is estimated at 0.17 cents
per gallon of highway diesel fuel supplied under the fully effective program. During the initial
years of our program, this cost is estimated at 0.14 cents per gallon. This cost is 20 percent lower
during the period when the temporary compliance option is available because approximately 80
percent of the highway diesel fuel pool is required to meet a 15 ppm sulfur cap during this
period/
In the NPRM, we estimated that the cost of distributing highway diesel fuel was equal to
the difference in price at the refinery rack and the retail price. For the final rule, we based our
estimate of distribution cost on a PADD by PADD evaluation of the difference in the price of
highway diesel fuel at the refiner rack versus the retail price. The price differential for each
PADD was weighted by the additional volume of fuel we anticipate will need to be produced in
each PADD to arrive at an estimate of distributing the additional volume needed for the nation as
a whole. Table V.C-27 provides a summary of the PADD-based values used in this calculation.
1 See section V.C.S.k. in this RIA for a discussion of how the relative volumes of 15 ppm and 500 ppm
highway diesel fuel vary over the period when the temporary compliance program is available.
V-125
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table V.C-27. Data Used to Calculate the Cost of Distributing the Additional Volume of
Highway Diesel Fuel Needed to Compensate for a Reduction in Energy Density
PADD
1
2
3
4
5
National
Average
Price at the
Retail Pump1
(cents / gallon)
68.8
68.6
65.5
75.8
80.0
71.7
Price at the
Refinery RackA
(cents per gallon)
55.5
56.9
54.0
66.7
62.9
59.2
Retail -Refinery
Rack Price
(cents per gallon)
13.3
11.7
11.5
9.1
17.1
12.5
Additional Volume
Needed
(fraction of supply)8
0.034
0.035
0.035
0.034
0.033
0.034
A Average price, excluding taxes, over the five year period from 1995 -1999. Energy Information Administration
(EIA), Petroleum Supply Annual (PSA), 1995-1999. Five year average costs were used for the purpose of this
calculation to provide an estimate of the typical difference between the price at the refinery rack and at the retail
pump.
B Based on our estimate of the changes refiners will make to meet the 15 ppm sulfur cap for highway diesel fuel.
See Section IV.A.
We believe the approach outlined above provides a more accurate estimate of costs.
Since the difference in price at the refiner rack versus that at retail also includes some profit for
the distributor and retailer, its use provides a conservatively high estimate of distribution costs.
The fact that a slightly less dense (lighter, less viscous) fuel would require slightly less energy to
be distributed also indicates that this estimate is conservative.
c. Cost of Downgrading an Increased Volume of Highway Diesel Fuel to a
Lower Value Product During Shipment by Pipeline
We estimated that the volume of highway diesel fuel that is currently downgraded to a
lower value product during shipment by pipeline is 2.2 percent of the total volume of highway
diesel fuel supplied and that this volume would double to 4.4 percent due to the implementation
of our sulfur control program. Please see section IV.D.2.a. for a discussion of how we arrived at
this estimate. This section addresses the cost of the additional downgrade volume (2.2 percent)
caused by our sulfur program. The cost to produce this additional volume is discussed in section
V.C.2.
V-126
-------
Chapter V: Economic Impact
The cost of downgrading the increased volume of highway diesel fuel to a lower value
product is based on the difference in the cost of 15 ppm sulfur diesel fuel and the product to
which the interface is downgraded. When our program is fully effective, this downgrade will be
made into the off highway diesel pool. The cost of this increased volume of downgrade when the
program is fully effective is estimated at approximately 0.14 cents per gallon of highway diesel
supplied under the fully effective program. The cost of this additional downgrade is somewhat
less during the initial years of our sulfur program because of the ability to downgrade 40 percent
of the additional downgrade volume to 500 ppm diesel fuel in those pipelines that we expect will
carry 500 ppm diesel fuel8. The cost of the additional downgrade during the initial years of our
program is estimated at 0.1 cents per gallon of highway diesel fuel supplied.
Following is a discussion of how we arrived at the above estimates.
There are two factors which influence the cost of making the downgrade of highway
diesel fuel discussed above. The first is the volume of the amount of highway diesel fuel that
must be downgraded. The second is the cost of making the downgrade based on the difference
between the cost of highway diesel fuel and the product that it is being downgraded to.
When our sulfur program is fully effective, the cost of downgrading the additional 2.2
percent of highway diesel fuel to a lower value product is the 6.5 cents / gallon difference in the
cost of producing a gallon of 15 ppm highway diesel fuel and that of producing a gallon of off
highway diesel fuel.1 To derive an estimate of the cost of this additional downgrade in terms of
the total volume of highway diesel fuel supplied, 6.5 cents / gallon was multiplied by the
additional fraction of the highway diesel pool that will need to be downgraded (0.022) to arrive at
result 0.14 cents per gallon.
During the initial years of our program, there will be a smaller additional volume of
highway diesel fuel that must be downgraded to a lower value product because some of the
highway diesel pool will continue to be 500 ppm fuel. We estimated that approximately 80
percent of the highway diesel pool will be 15 ppm fuel during the initial years of our program."
This reduces the cost associated with the additional downgrade. The cost of the additional
downgrade is also reduced during the initial years of our program because 40 percent can be
downgraded to 500 ppm highway diesel fuel which is a higher value product than off highway
diesel fuel. This is based on our estimate that 40 percent of the pipeline systems that carry
s See section V.C.S.k. in this RIA for additional discussion regarding the extent to which we anticipate 500
ppm diesel fuel will be present in the distribution system..
4 See section table V.C-20 and attending text in section V.C.2. for a discussion on the difference in the cost
of producing 15 ppm highway diesel fuel and that of producing off highway diesel fuel.
u See section V.C.S.k. for a discussion on the relative volumes of 15 ppm and 500 ppm highway diesel fuel
during the initial years of our program.
V-127
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
highway diesel fuel will carry both 15 ppm and 500 ppm highway diesel fuel.v We used our
estimate of the cost of producing 15 ppm highway diesel fuel under the fully effective program
relative to the cost to produce 500 ppm diesel fuel today (4.1 cents per gallon) in calculating the
cost of downgrading 15 ppm highway diesel fuel to 500 ppm highway diesel fuel. This provides
a conservatively high estimate, since production costs for 15 ppm are somewhat lower during the
start up of the program.™
Based on the above inputs, we estimate that the cost of the additional downgrade will be
0.11 cents per gallon for the 40 percent of fuel distributed using the part of the system that
handles both grades of highway diesel fuel, and 0.08 cents per gallon for the 60 percent of fuel
that is distributed using the part of the system that carries only 15 ppm highway diesel fuel. By
weighting these two results, we arrived at our over-all estimate of the cost of the additional
volume of highway diesel fuel that will be downgraded to a lower value product of 0.1 cents per
gallon of highway diesel fuel supplied.
d. Increased Cost of Downgrading the Current Interface Volumes Associated
with Pipeline Shipments of Highway Diesel Fuel
We identified that there would also be an increase in the economic impact for the existing
volume of interface currently associated with pipeline shipments of highway diesel fuel. This is
because the cost of downgrading the existing interface volume would be determined by the
difference between the cost of 15 ppm sulfur fuel and off highway diesel fuel rather than the
difference in cost between current 500 ppm diesel fuel and off highway diesel fuel as it is today.
We estimate that the increase in the cost of downgrading the existing highway diesel interface
would be 0.09 cents per gallon of highway diesel fuel supplied during the fully effective
program. During the initial years of our program, we estimate this cost at 0.08 cents per gallon
of highway diesel fuel supplied. Following is a discussion of how we arrived at these estimates.
When our sulfur program is fully effective, all of the volume of highway diesel fuel
shipped by pipeline that must be downgraded to a lower value product must be downgraded to
off highway diesel fuel. Therefore, the additional cost of downgrading the current volume of
highway diesel fuel that must be downgraded will be based on the difference in the cost of
producing 15 ppm highway diesel fuel and off highway diesel fuel (6.5 cents per gallon)
compared to the current difference in cost between 500 ppm highway diesel fuel and off highway
v See section V.C.S.k. regarding the extent that we expect the distribution system will carry 500 ppm
highway diesel fuel during the initial years of our program when the temporary compliance option is available.
w See table V.C-20 and the associated text in section V.C.2. for a discussion on the difference in the cost
of producing 15 ppm highway diesel fuel and that of producing 500 ppm highway diesel fuel.
V-128
-------
Chapter V: Economic Impact
diesel fuel (2.2 cents per gallon)." Our estimate of the additional cost (0.09 cents per gallon of
highway diesel fuel supplied) after our sulfur program is fully effective was calculated by
multiplying the 4.3 cents per gallon price differential by the fraction of the highway diesel pool
that is currently downgraded to a lower value product (0.022)7 Costs during the initial years of
our program are reduced by 20 percent at (to 0.08 cents per gallon of highway diesel fuel
supplied) because on average only 80 percent of the highway diesel fuel pool will be 15 ppm fuel
during the initial years of our program when the temporary compliance option is available.2
e. Increased Cost of Downgrading the Interface Between Pipeline Shipments of
Highway Diesel Fuel and Jet Fuel or Kerosene
Please refer to section IV.D.2.a in this RIA for a more thorough discussion of the change
that will need to take place in the handling practices for the interface volumes between adjacent
pipeline shipments of highway diesel fuel and jet fuel or kerosene. This section addresses our
estimation of the costs of this change.
Expressed in terms of the volume of highway diesel fuel supplied, we estimate the
increased cost of handling these interface volumes will be 0.07 cents per gallon after our
becomes sulfur program fully effective. This cost arises from the fact that all of the interface
volume between adjacent batches of highway diesel fuel and jet fuel or kerosene will need to be
downgraded to off highway diesel fuel once our program becomes fully effective. During the
initial years of our program, we estimate that the cost will be 0.03 cents per gallon. The costs is
somewhat less during the initial years of our program because there is some opportunity to make
the downgrade to 500 ppm highway diesel fuel rather than off highway diesel fuel. Additional
storage tanks will be needed at those terminals that currently do not handle off highway diesel
fuel. The cost of these tanks has been fully accounted for in the calculation of costs during the
initial years of our program as discussed below (0.009 cents per gallon).
Since a clean interface cut is already made between batches of highway diesel fuel and jet
fuel or kerosene, there will be no increase in the volume of product downgraded under our
program. However, the entire interface volume between highway diesel fuel and jet fuel or
kerosene will need to be directed into a storage tank containing off highway diesel fuel when the
15 ppm cap on the sulfur content of highway diesel fuel is implemented. The current practice is
to cut all of the interface volume associated with adjacent batches of highway diesel fuel and jet
x See table V.C-20 and the associated text in section V.C.2. for a discussion on the relative cost of
producing these different types of diesel fuel.
y See section IV.D.2. for a discussion in our estimation of the current downgrade volume.
z See section V.C.S.k. for a discussion on the relative volumes of 15 ppm and 500 ppm highway diesel fuel
during the initial years of our sulfur program.
V-129
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
fuel or kerosene into the batch of highway diesel fuel. When our sulfur program is fully
effective, the increased cost associated with this downgrade will be based on the difference in
cost between 500 ppm highway diesel fuel and off highway diesel fuel (2.2 cents per gallon).
This is because the downgrade will be made to off highway diesel fuel rather than 500 ppm
highway diesel fuel as it is today.
To account for the fact that not all batches of highway diesel fuel are shipped by pipeline
adjacent to a batch of jet fuel, 2.2 cents per gallon was multiplied by the ratio of the volume of jet
fuel and kerosene supplied to the volume of highway diesel fuel supplied (0.72). For the
purpose of this calculation, we assumed that 72 percent of the highway diesel fuel batches
shipped by pipeline abut a shipment of jet fuel or kerosene. We derived the ratio of the volume
of jet fuel and kerosene supplied to the volume of highway diesel fuel supplied using the
following data from the Energy Information Administration (EIA):79
Jet Fuel and kerosene supplied in 1999 = 637,123,000 barrels = 26,759,166,000 gallons
500 ppm diesel supplied in 1999 = 887,355,000 barrels = 37,268,910,000 gallons
(jet fuel + kerosene) / 500 ppm diesel = 0.72
To arrive at our estimate of the additional cost of the downgrade associated with batches
of highway diesel fuel that abut batches of jet fuel or kerosene during shipment by pipeline when
our program is fully effective, we multiplied the volume of the downgrade by the difference
between the cost of 500 ppm highway diesel fuel and off highway diesel fuel (2.2 cents per
gallon).
During the initial years of our program, 40 percent of pipeline systems will carry both 500
ppm and 15 ppm highway diesel fuel. In such systems the downgrade can continue to be made to
500 ppm diesel fuel rather than off highway diesel fuel. Therefore, there will be no additional
cost associated with this downgrade. Consequently, the additional cost of the downgrade is
reduced by 40 percent during the initial years of our sulfur program (0.07 x 40 percent = 0.03
cents per gallon).
Following is a discussion of how we arrived at our estimate of 0.009 cents per gallon of
highway diesel fuel produced for the storage tanks that will be needed to accommodate the
interface between pipeline shipments of highway diesel fuel and jet fuel/kerosene at terminals
that do not already have a storage tank that contains off highway diesel fuel. We estimated that
approximately 60 percent of terminals will not have such a tank (588 terminals). At such
terminals, we estimate that a single 4,000 gallon above ground tank will be installed at a cost of
$20,000 per tank. The total tank cost will be $11,760,000. This cost was amortized (at 7 percent
per annum) over the initial years of the sulfur program to arrive at our estimate of 0.009 cents per
V-130
-------
Chapter V: Economic Impact
gallon of highway diesel fuel supplied. We used our estimate of the total volume of highway
diesel fuel supplied in 2006aa (39,504,000,000 gallons) to arrive at this per gallon estimate.
f. Cost of Additional Quality Control Testing at Petroleum Terminals
The additional quality control testing at the terminal level needed to ensure compliance
with the 15 ppm sulfur cap would be the same during the initial years of our program as after the
requirements are fully implemented. We estimate the cost of such additional quality assurance
measures will be $100 for each batch. This estimate includes the cost of sampling and testing
each batch for its sulfur content. A typical pipeline batch of highway diesel fuel shipped by
pipeline is 100,000 barrels. By dividing the estimated cost per batch by the average size of a
batch, we arrived at an estimate of 0.002 cents per gallon of highway diesel fuel supplied for the
cost of the additional quality control measures needed at terminal facilities.
g. Cost of Downgrading the Additional Pipeline Interface Volumes Associated
with the Shipment of Highway Diesel Fuel that Meets a 500 ppm Sulfur Cap
During the Initial Years of Our Sulfur Program
The presence of two grades of highway diesel fuel (500 ppm and 15 ppm) during the
initial years of our program will cause the generation of additional pipeline interface volumes and
associated downgrade costs. This is because there will be more batches of highway diesel fuel
shipped in the 40 percent of pipelines that carry both grades of highway diesel fuel.bb We
estimate the additional cost during the initial years of our program will be 0.004 cents per gallon
of the total volume of highway diesel fuel supplied (500 and 15 ppm sulfur cap fuel).
We arrived at this estimate by multiplying the following factors:
The volume of the downgrade associated with pipeline shipments of 500 ppm highway
diesel fuel (2.2 percent)00
aa The estimate of highway diesel fuel supplied in 2006 was derived by growing the estimate of highway
diesel fuel supply in 1999 from the Energy Information Administration (EIA) Petroleum Supply Annual, June 2000,
by 1.5 percent each year. See docket item IV-A-07 for a discussion of our use of the 1.5 percent growth factor.
bb See section V.C.S.k. in this RIA regarding the extent that we expect the distribution system will carry
500 ppm highway diesel fuel during the period when the temporary compliance option is available.
00 See section IV.D.2. in this RIA for a discussion of our estimate of the volume downgraded.
V-131
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The fraction of the highway diesel pool we expect to be 500 ppm fuel (approximately 20
percent during the initial years of our program)dd
The fraction of the pipeline system that will carry 500 ppm fuel (40 percent)66
The cost of the downgrade (2.2 cents per gallon)
h. Cost of Optimizing the Distribution System to Distribute 15 ppm Highway
Diesel Fuel
As more fully discussed in section IV.D, we expect that the distribution industry will
conduct various tests to evaluate potential sources of contamination prior to the implementation
of our sulfur control program. During this evaluation, we anticipate that minor procedural and
equipment changes may be identified in addition to those which we have specifically assigned a
cost to. Such additional changes may include:
Testing the system to evaluate sources of contamination
Valve replacements
Moving pipeline batch monitoring systems upstream and/or speeding the means to
make batch changes
Education programs for tank truck, tank wagon, and rail car operators on practices
to limit contamination
We believe that the costs associated with such optimization practices will be relatively
minor and readily accommodated by the distribution industry. Such costs will only occur once
and the associated situations will be the exception rather than the rule. Since commenters did not
provide an estimate of the frequency when such instances might arise, it is difficult to estimate a
cost. Based on engineering judgement, having reviewed the information in the comments and
the potential cost for a range of potential activities, we estimate that the fuel distribution industry
will invest another $100,000,000 to optimize its ability to limit sulfur contamination in addition
to the costs that we have specifically identified (e.g. downgrade, additional tanks). We estimate
that this investment will be made almost entirely by pipeline and terminal operators. We
amortized this cost at 7 percent per annum over the period from the program's start-up in 2006
through 2020. During the initial years of our program, this results in a cost of 0.027 cents per
dd For the purpose of this calculation, we used 20 percent for the entire 4 year period when the temporary
compliance option is available. See section V.C.S.k. in this RIA for a discussion of how the relative volumes of 15
ppm and 500 ppm highway diesel fuel vary over the period when the temporary compliance program is available.
For example, during the first year of our sulfur program, we estimate that 22 percent of highway diesel fuel will
meet a 500 ppm sulfur cap.
ee See section V.C.S.k. in this RIA regarding the extent that we expect the distribution system will carry
500 ppm highway diesel fuel during the period when the temporary compliance option is available.
V-132
-------
Chapter V: Economic Impact
gallon of highway diesel fuel supplied. When our program is fully effective, we estimate the cost
at 0.025 cents be gallon through the year 2020.
i. Additional Measures by Tank Truck, Tank Wagon, and Rail Car Operators
to Limit Contamination
As discussed in the section on the feasibility of distributing 15 ppm highway diesel fuel
(section IV.D.), we continue to believe that there will only be negligible costs to tank truck, tank
wagon, and rail car operators associated with limiting contamination during the distribution of 15
ppm highway diesel fuel. Given the such potential cost would be very small, we believe they are
sufficiently well accounted within the costs we have attributed to the optimization of the
distribution system to limit contamination (V.C.S.h.).
j. Potential Costs Associated with the Voluntary Phase Out of High Sulfur
Diesel Additives
As discussed in the section on the feasibility of distributing 15 ppm diesel fuel (section
IV.D.), we believe that the allowance for the continued use of diesel fuel additives which exceed
15 ppm in sulfur content will prevent any significant cost impacts from our program related to
the use of diesel fuel additives.
k. Costs During the Initial Years of Our Program Due to the Need for
Additional Storage Tanks to Handle Two Grades of Highway Diesel Fuel
The most substantial additional costs associated with the temporary compliance option
are due to the need to handle an additional grade of highway diesel fuel in the distribution
system. During the initial years of our program when the temporary compliance option is
available, we expect that the production of 500 ppm sulfur fuel will be much less than that of 15
ppm fuel. At the same time, most of the diesel vehicle fleet can burn 500 ppm fuel during this
period. Because of its greater volume and the need to distribute it everywhere in the country, we
expect that essentially all pipelines and terminals will handle 15 ppm fuel. In contrast,
distribution of 500 ppm fuel will concentrate on those areas nearest the refineries producing that
fuel, plus a few major pipelines serving major refining areas.
Regarding distribution to the final user, we expect that nearly all truck stops in areas
where 500 ppm fuel is available will invest in piping and tankage to handle a second fuel.
Because of the significant expense involved in adding a second tank, in these areas, we expect
service stations will only carry one fuel or the other, as market demands dictate. Likewise, we
expect that centrally fueled fleets and card locks will only handle 15 ppm fuel. Under this
scenario, sales of 500 ppm fuel are limited to only those vehicles which refuel at truck stops and
service stations. This is somewhat conservative since some centrally fueled fleets may have the
flexibility to inexpensively handle two fuels. Likewise, some card locks in a given area may be
V-133
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
able to carry 15 ppm fuel and others 500 ppm fuel and still serve their clients at little extra cost.
Still, given the above assumptions, we project that the 500 ppm fuel will have to be distributed to
areas representing about 50% of the national diesel fuel demand. Also, as the fleet turns over to
2007 and later vehicles, the amount of 500 ppm fuel produced under the temporary compliance
option will gradually decrease from roughly 22 percent in 2007 to about 16 percent in 2010.
The tankage cost at refineries, terminals, pipelines and bulk plants handling both fuels is
estimated to be $0.81 billion. We estimate that 11 refineries will produce both fuels. These are
refineries with hydrocrackers which are not projected to invest in new hydrotreating equipment in
2006. Thus, these refineries will produce a small amount of 15 ppm fuel from its hydrocrackate
and 500 ppm fuel from the rest of its current highway diesel fuel blendstocks. At $1 million per
tank, this totals to $11 million.
We estimate that there are 853 terminals which currently carry highway diesel fuel,
excluding tanks at refineries. We assume that 40% of these terminals would build a new tank in
order to distribute two fuels to 50% of the U.S. market and keep tank truck driving distances at
current levels. We estimate only 40% of these terminals would need an extra tank, rather than
50%, because 56 refineries will be producing the 500 ppm fuel and will distribute this fuel
directly to their local areas. At a cost of $1 million per tank, terminal tankage will cost a total of
$340 million.
Likewise, we estimate that there are 9200 bulk plants which currently carry highway
diesel fuel, excluding tanks at refineries. We estimate that a new tank at these facilities would
cost $125,000. Again assuming that 40% of these bulk plants would build a new tank in order to
distribute two fuels to 50% of the U.S. market, this tankage would cost a total of $460 million.
Finally, we estimate that 50% of the nation's truck stops would also build a new tank or
otherwise provide for a second fuel. There are 4800 truck stops currently operating in the U.S.80
The National Association of Truck Stop Operators (NATSO) surveyed their members regarding
the expected cost to handle a second grade of highway diesel fuel.81 We weighted the responses
to this survey to arrive at our estimate that it would cost $100,000 per truck stop on average to
handle a second fuel. This totals to $240 million. Thus, the total cost for new tankage at all of
these facilities is $1.05 billion.
We then amortized these one time costs over the 15 ppm fuel produced during the initial
years of our program at 7 percent per annum. We estimated that, with the small refiner option,
the total percentage of 15 ppm fuel produced during the first year of our program would be 78%
(though it could be as low as 75% if all small refiners chose to delay production of 15 ppm fuel).
This continued through 2008. However, in 2008, the limitation of distributing 500 ppm fuel only
through truck stops and service stations and only in 70% of the U.S. diesel fuel market, as well as
the turnover of the vehicle fleet to 2007 and later vehicles, began to be controlling. We estimate
that truck stops and service stations distribute 61% of all highway diesel fuel in the U.S. We
V-134
-------
Chapter V: Economic Impact
assumed that these outlets sold 15 ppm and 500 ppm in proportion to the in-use vehicle fleet (i.e.,
2007 and later vehicles used 15 ppm fuel and earlier vehicles used 500 ppm fuel). Thus, in 2009
and 2010, we estimated that 81%, and 83.5% of all highway diesel fuel would meet the 15 ppm
standard. The last figure was assumed to apply through mid- 2010, based on the use of banked
credits from earlier periods. Amortizing the tankage cost over the 15 ppm fuel produced over
this period, the cost per gallon is 0.9 cents.
4. What is the Cost of Lubricity Additives?
Adoption of the cap on diesel fuel sulfur could result in a decrease in the lubricity of
highway diesel fuel produced by some refiners. This could necessitate the use of additional
quantities of lubricity-improver additives to maintain in-use lubricity performance (see Section
IV.C.).
A study by MathPro Inc. (MathPro)82 in 1999, sponsored by the Engine Manufacturers
Association to estimate the costs of diesel fuel desulfurization under sulfur standards that we
were likely to require, received estimates from lubricity additive suppliers indicating that the
costs of lubricity additives would average 0.1 to 0.5 cents per gallon. The lower the sulfur
standard, typically the higher the lubricity cost. We independently contacted some producers and
distributors of lubricity additives, which also provided estimated average costs in the range of 0.1
to 0.5 cents per gallon for large volumes of treated fuel. Again, the estimates varied depending
on the sulfur standard, ranging from a cap of 5 to 50 ppm. MathPro utilized vendor cost
estimates to derive lubricity additive cost estimates under a number of possible diesel fuel sulfur
control scenarios. These estimates ranged from 0.1 to 0.3 cents per gallon depending on the
control case (see Table V.C-28).
V-135
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-28. MathPro Lubricity Additive Cost Estimates
Sulfur Control Case (avg. sulfur standard)
Highway Diesel
ISOppm
ISOppm
50 ppm
20ppm
20 ppm
2 ppm
2 ppm
Off Highway Diesel
uncontrolled (3500 ppm)
ISOppm
50 ppm
350 ppm
20 ppm
3 50 ppm
2 ppm
Estimated Lubricity Additive Cost
(cents/gallon)
0.1
0.1
0.1
0.1
0.2
0.2
0.3
Unfortunately, MathPro did not provide costs for a case consistent with the 15 ppm
sulfur standard. In addition, MathPro cases included control of off highway diesel fuel.
Nevertheless, the cases evaluated in the MathPro study can be used to approximate the cost of
lubricity additives under the 15 ppm cap for highway diesel fuel. Of the cases evaluated by
MathPro, we believe its highway/off-highway 20 ppm average scenario most closely matches our
highway-only 15 ppm cap case with respect to the potential impact on lubricity additive cost.
While our projected refinery average sulfur level of 7 ppm is closer to 2 ppm than 20 ppm, we
believe that Mathpro's 2 ppm case, which includes the desulfurization of both highway and non-
highway diesel fuel to this level, is much more severe with respect to lubricity changes than a 7
ppm level for highway diesel fuel only. Thus, using the vendor-supplied cost estimates, coupled
with the estimates for the various scenarios evaluated by MathPro, we estimate that the cost of
lubricity additives under the 15 ppm sulfur cap would be in the range of 0.2 cents per gallon.
In considering the comments, we have found no basis in today's action to use a different
average cost estimate to treat low sulfur diesel fuel for lubricity than that which was used in the
proposal. Of the two comments we received on this issue, one supported our cost estimate of 0.2
cents per gallon. The other was submitted by DOD, which indicated it has experienced lubricity
additive costs from one to five cents per gallon. We believe that DOD's experience with lubricity
properties and lubricity additives is not typical of commercial users for several reasons. First,
DOD commented that, due to harsher operating conditions, engines used in DOD vehicles,
especially tactical vehicles, are more vulnerable to lubricity problems than the same engines
operated in commercial vehicles. Also, the fuel DOD uses at its facilities is purchased under
V-136
-------
Chapter V: Economic Impact
contract usually for a year or longer. Thus, the DOD fuel generally is from a single supplier and
does not have the beneficial effect of blending or mixing different batches of fuel or fuel from
different suppliers, such as that which occurs in the commercial market. As discussed in Section
IV, blending or mixing different batches of diesel fuel minimizes the effect of isolated poor
lubricity fuels. Consequently, DOD might be taking more aggressive action in responding to
lubricity concerns than might be needed for commercial applications. Second, DOD is using an
additive that is primarily a corrosion inhibitor. It is our understanding that DOD found that the
additive it uses to address a corrosion property in the fuel is also effective at improving lubricity,
and subsequently has been using that additive to also address its lubricity concerns. If DOD were
able to ignore its corrosion property concerns, it is possible that a formulation specifically for
lubricity might cost less, or that its treat rate could be less, than that of the corrosion inhibitor
formulation and treat rate it currently uses. Finally and most importantly, we believe that DOD's
experience is more reflective of the prices that might be experienced with specialty additives
supplied in relatively small quantities. With the 15 ppm standard, most, if not all, of the nation's
highway diesel fuel may need to be treated for lubricity. Economies of scale associated with
bulk production as opposed to more specialty products will drive down the unit cost of lubricity
additives considerably.
5. Benefits of 15 ppm Diesel Fuel for the New and Existing Diesel
Fleet
In addition to its role as a technology enabler, low sulfur diesel fuel gives benefits in the
form of reduced sulfur induced corrosion of vehicle components and slower acidification of
engine lubricating oil, leading to longer maintenance intervals and lower maintenance costs.
These benefits will apply to new vehicles and to the existing heavy-duty vehicle fleet beginning
in 2006 when the fuel is introduced. These benefits can offer significant cost savings to the
vehicle owner without the need for purchasing any new technologies. These benefits are
estimated here for new vehicles and for vehicles in the existing fleet (pre-2007 fleet).
The individual components of the engine system which might be expected to realize
benefits from the use of low sulfur diesel fuel are summarized in Table V.C-29 and are described
in more detail in the following sections.
V-137
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-29. Components Potentially Affected by Lower Sulfur Levels in Diesel Fuel
Affected Components
Piston Rings
Cylinder Liners
Oil Quality
Exhaust System
(tailpipe)
EGR
Affect of Lower Sulfur
Reduce corrosion wear
Reduce corrosion wear
Reduce deposits and less
need for alkaline additives
Reduces corrosion wear
Reduces corrosion wear
Potential Impact on Engine System
Extended engine life and less
frequent rebuilds
Extended engine life and less
frequent rebuilds
Reduce wear on piston ring and
cylinder liner and less frequent oil
changes
Less frequent part replacement
Less frequent part replacement
The actual value of these benefits over the life of the vehicle will depend upon the length
of time that the vehicle operates on low-sulfur diesel fuel. For a vehicle near the end of its life in
2007 the benefits will be quite small. However for vehicles produced in the years immediately
preceding the introduction of low-sulfur fuel the savings will be substantial. These savings are
estimated here for new and existing diesel vehicles beginning in 2006 and continuing through
2035. The costs are expressed in terms of dollars saved per mile or in terms of dollars saved in a
particular year (for rebuild savings).
These savings, due to the use of low sulfur diesel fuel, can also be expressed in terms of a
savings in cents per gallon of low sulfur diesel fuel. Taking the savings detailed in each of the
subsections below and expressing them in terms of cents per gallon gives an average savings of
approximately 1.4 cents/gallon for light heavy-duty diesels, 1 cent/gallon for medium heavy-duty
diesel engines and 0.7 cents/gallon for heavy heavy-duty diesel engines. The average savings
estimated across all weight classes is therefore approximately one cent per gallon. While there
may be uncertainty regarding the magnitude of this effect, this estimate may in fact be a
conservative estimate of the savings as there are likely to be other benefits not accounted for in
this analysis.
a. Methodology
Under contract from EPA, ICF Consulting provided surveys to nine engine manufacturers
seeking their input on expectations for cost savings which might be enabled through the use of
V-138
-------
Chapter V: Economic Impact
low sulfur diesel fuel and seeking their estimations of the cost and types of emission control
technologies which might be applied with low sulfur diesel fuel. In general, the respondents to
the survey gave qualitative rather than precise quantitative estimates of the benefits of low sulfur
diesel fuel. While all respondents agreed that savings will occur, their estimates were often
based on rough approximations of future engine characteristics. Based on responses to this
survey, EPA estimated cost savings to the current and future fleets through the use of low sulfur
diesel fuel.83
For new vehicles we have estimated the value of these benefits in terms of a net present
value in the year of vehicle sale. This allows for us to calculate a per vehicle cost of control and
a per vehicle cost effectiveness for the program. In order to calculate aggregate benefits for the
new fleet and for the existing fleet this approach is not appropriate as each vehicle in the fleet
will accrue benefits at different rates over different periods, depending upon their year of
introduction and their technology mix. Additionally, it is more telling to describe the cost
savings as an aggregate benefit to the fleet, just as fuel costs are shown as an aggregate cost to
the fleet. Therefore, where possible, we have estimated the benefits of low sulfur diesel fuel to
the new and existing heavy-duty vehicle fleets in terms of dollars per vehicle mile traveled. In
the one case, where the savings are related to a discrete event (engine rebuilds), we have applied
a single savings estimated to a specific fraction of the existing fleet as described below. These
savings are then accumulated over the entire pre-2007 heavy-duty fleet and over the new fleet of
vehicles introduced in 2007 in each year from 2006 through 2035, and are reported as an
aggregate savings.
If refiners avail themselves of the temporary compliance option and hardship provisions
available to them in the early years of the program, some fraction of the existing fleet would
continue to operate on current 500 ppm sulfur diesel fuel. In order to account for this possibility
in our analysis, we have assumed that 22 percent of the total fuel consumption during the
transition period will be today's 500 ppm sulfur fuel. The analysis also assumes that the new
vehicles will be fueled exclusively on the new low sulfur diesel fuel and that only the fraction of
the existing fleet operating on the remaining fraction of the low sulfur diesel fuel will realize a
benefit.
b. Extended Oil Change Intervals
Sulfur in diesel fuel leads to acidification of engine lubricating oils, directly causing
increased corrosion and increased rates of engine wear. Lubricating oils use alkaline additives to
neutralize the acidifying nature of sulfur compounds formed in the engine from sulfur in diesel
fuel. These basic compounds are consumed over time leading to a loss of pH control in the oil.
Oil change intervals are often determined based upon the period of time required for the basic
compounds in the oil to be consumed. The use of low sulfur diesel fuel will decrease this rate of
oil acidification leading to extended periods between required oil change maintenance intervals.
While it is difficult to quantify a precise benefit, most observers agree that use of very low sulfur
V-139
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
fuel will probably extend oil drain intervals. Based on information from some engine
manufacturers and others, we have assumed that engine oil change intervals will be extended by
ten percent due to the use of low sulfur diesel fuel. Based on this benefit the per mile savings can
be estimated as shown in Table V.C-30.
Table V.C-30. Cost Savings to the Existing Fleet from Extend Oil Change Intervals
Made Possible by Low Sulfur Diesel Fuel
Base Oil Change Interval*
Low Sulfur Oil Change Interval*
Cost Per Oil Change*
Base Oil Change Cost per Mile
Low Sulfur Oil Change Cost per Mile
Oil Change Cost Difference per Mile
Average Fuel Economy
Cost Savings Per Gallon Fuel
Units
miles
miles
$
$/mile
$/mile
$/mile
miles/gallon
$/gallon
LED
8,000
8,800
$100
$0.0125
$0.0114
$0.0011
11.8
$0.0134
MHD
11,000
12,100
$150
$0.0136
$0.0124
$0.0012
8.0
$0.0099
HHD
18,000
19,800
$200
$0.0111
$0.0101
$0.0010
5.9
$0.0060
*Oil change intervals for vehicles operating on low sulfur diesel fuel are assumed to increase by ten percent,
average oil change intervals, and costs for oil changes from ICF Consulting report.84
For vehicles produced after the introduction of the low sulfur diesel fuel in 2006 these
benefits can also be expressed in terms of an average cost savings over the life of the vehicle.
The cost savings are estimated using typical mileage accumulation rates given in each year of a
vehicles life from our inventory emissions model and the typical oil change interval and costs
described above. These savings are then expressed in terms of a net present value in the year of
the vehicle sale. The savings realized for extended oil change intervals on vehicles fueled
exclusively on low sulfur diesel fuel are estimated to be $153 for light heavy-duty vehicles, $249
for medium heavy-duty vehicles and $559 for heavy heavy-duty vehicles.
c. Extended EGR System Life
In the RIA for the 2004 heavy-duty engine standards, we estimated that exhaust gas
recirculation (EGR) systems, particularly EGR valves, will require service or replacement as part
of the engine rebuild process. This estimate was based primarily upon our concern for the
V-140
-------
Chapter V: Economic Impact
detrimental effects of sulfur in diesel fuel on EGR system durability. The use of low sulfur diesel
fuel mitigates this concern and leads us to conclude that the EGR valve used in these systems can
be expected to last the life of the engine. Eliminating the replacement of the EGR valve on
heavy heavy-duty diesel engines represents a cost savings to vehicles built with EGR systems of
$115 in the year of the engine rebuild. These savings are only estimated for vehicles built after
2004, because vehicles built prior to that date will have operated primarily on current high sulfur
diesel fuel. Savings for light and medium heavy duty vehicles are not estimated because engines
in these vehicle classes are less likely to be rebuilt. The analysis also assumes that vehicles with
EGR systems will be operated primarily on the new low sulfur diesel fuel when it becomes
available in 2006. Although some fraction of the existing fleet may be operating on high sulfur
diesel fuel during that period, we believe that owners of vehicles with EGR systems will
preferentially choose the low sulfur diesel fuel for the maintenance benefits it provides. The
aggregate savings for vehicles sold in 2004-2006 and rebuilt in 2009-2011 are shown in Table
V.C-31. The aggregate savings for vehicles built beginning in 2007 and rebuilt beginning in
2012 are presented in Table V.C-32. These savings can also be expressed in terms of a net
present value in the year of vehicle sale of $51.
Table V.C-31. Cost Savings to the Existing Fleet for Reduced EGR System Replacement
Made Possible by Low Sulfur Diesel Fuel*
Year Rebuilt
(7th year of life)
2010
2011
2012
Model
Year
2004
2005
2006
Calendar Yr
Sales
259,600
264,000
268,400
Surviving in
Year 7
185,874
189,024
192,174
Number
Rebuilt
176,580
179,573
182,566
Aggregate
Savings
$20,306,691
$20,650,872
$20,995,053
* $115 per vehicle cost savings if the EGR valve is not replaced when the engine rebuild occurs. The table
assumes that only Heavy Heavy-Duty engines are rebuilt, that 95 percent of vehicles reaching 560,000 miles are
rebuilt, and that 72 percent of heavy heavy-duty vehicles reach 560,000 miles (on average in year 7 of their life).
V-141
-------
Table V.C-32. Cost Savings to the New Fleet (2007 and later) for Reduced EGR System
Replacement Made Possible by Low Sulfur Diesel Fuel*
Year Rebuilt (7th
year of life)
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
Model Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
Calendar Yr
Sales
272,800
277,200
281,600
286,000
290,400
294,800
299,200
303,600
308,000
312,400
316,800
321,200
325,600
330,000
334,400
338,800
343,200
347,600
352,000
356,400
360,800
365,200
369,600
Surviving in
Year?
195,325
198,475
201,625
204,775
207,925
211,075
214,225
217,375
220,525
223,675
226,825
229,975
233,125
236,275
239,425
242,575
245,725
248,875
252,025
255,175
258,325
261,475
264,625
Number
Rebuilt
185,559
188,551
191,543
194,535
197,527
200,519
203,511
206,503
209,495
212,487
215,479
218,471
221,463
224,455
227,447
230,439
233,431
236,423
239,415
242,407
245,399
248,391
251,383
Aggregate
Savings
$21,339,234
$21,683,416
$22,027,598
$22,371,780
$22,715,962
$23,060,144
$23,404,326
$23,748,508
$24,092,690
$24,436,872
$24,781,054
$25,125,236
$25,469,418
$25,813,600
$26,157,782
$26,501,964
$26,846,146
$27,190,328
$27,534,510
$27,878,692
$28,222,874
$28,567,056
$28,911,238
* $115 per vehicle cost savings if the EGR valve is not replaced when the engine rebuild occurs. The table
assumes that only Heavy Heavy-Duty engines are rebuilt, that 95 percent of vehicles reaching 560,000 miles are
rebuilt, and that 72 percent of heavy heavy-duty vehicles reach 560,000 miles (on average in year 7 of their life).
V-142
-------
Chapter V: Economic Impact
d. Extended Exhaust System Life
Exhaust system components, specifically exhaust pipes and mufflers, typically fail due to
perforations caused by corrosion of the pipe walls. Corrosion rates are increased by sulfuric acid
present in diesel exhaust which can condense on the walls of the exhaust system. This sulfuric
acid is a by-product of combustion with sulfur in diesel fuel. When sulfur is removed from
diesel fuel the amount of sulfuric acid formed decreases proportionally, thereby reducing
corrosion rates due to sulfuric acid in diesel exhaust. The survey respondents acknowledged that
this may be a cost savings to the consumer, but were not able to quantify the savings or
determine the percent extended life. One manufacturer characterized the savings as marginal.
Based on this information, we have assumed that the reduction in sulfuric acid induced corrosion
may extend exhaust system component life by five percent, leading to a cost savings to the
existing vehicle fleet. Based on this estimate and estimates of average exhaust system life and
average exhaust system replacement costs, a per mile estimate of this cost savings can be
determined as shown in Table V.C-33. We have not applied this savings to estimates for the new
vehicle fleet because we do not anticipate the use of a muffler on vehicles equipped with diesel
PM filters.
Table V.C-33. Cost Savings to the Existing Fleet from Extend Exhaust System
Replacement Intervals Made Possible by Low Sulfur Diesel Fuel
Exhaust System Change Interval
Low Sulfur Exhaust Change Interval*
Exhaust Replacement Cost
Base Cost per Mile
Low Sulfur Cost per Mile
Cost Difference Per Mile
Average Fuel Economy
Cost Savings Per Gallon Fuel
Units
miles
miles
$
$/mile
$/mile
$/mile
miles/gallon
$/gallon
LED
110,000
115,500
$275
$0.0025
$0.0024
$0.0001
11.8
$0.0014
MHD
147,000
154,350
$379
$0.0026
$0.0025
$0.0001
8.0
$0.0010
HHD
334,000
350,700
$491
$0.0015
$0.0014
$0.0001
5.9
$0.0004
* Exhaust system life for vehicles operating on low sulfur diesel fuel are expected to increase by 5 percent.8
V-143
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
e. Extended Rebuild Intervals and Engine Life
Engine rebuilds and replacements often occur when excessive wear of the engine cylinder
kit (primarily the cylinder liner and engine piston rings) causes high oil consumption rates,
decreased engine performance and increased fuel consumption rates. Wear rates of these
components can increase due to corrosion caused by sulfur in diesel fuel. Therefore, in as much
as low sulfur diesel fuel can be expected to decrease corrosion, it can also be expected to
similarly decrease component wear rates, thereby leading to increased component life.
Extending engine life or the time between engine rebuilds, can lead to a direct savings to the
consumer.
Estimating an average extension of engine life is difficult due to the many factors that
affect engine wear and overall engine life. We believe the strong influence of sulfur in diesel
fuel on engine wear could lead to estimates of about five percent. However, because engine wear
rates are also linked to oil change intervals it may not be appropriate to claim full credit for both
extended oil change intervals and extended engine rebuild intervals. Therefore, in order to be
conservative in our estimates, we have not included these cost savings in our estimates of
aggregate cost savings realized through the use of low sulfur diesel fuel.
f. Aggregate Cost Savings for the New and Existing Diesel Fleet Realized from
Low Sulfur Diesel Fuel
By applying the cost savings described in the preceding sections to the predicted vehicle
miles traveled for each class of heavy-duty vehicle in the inventory calculation model described
in chapter 2 of this RIA, an estimated aggregate savings can be calculated. These savings are
shown for the fraction of the existing fleet (pre-2007 vehicles) operating on the new low sulfur
diesel fuel in Table V.C-34 beginning with the savings realized in 2006 from the introduction of
low sulfur diesel fuel in that year. As vehicles in the pre-2007 fleet are retired from service these
cost savings decrease as reflected in the table.
Aggregate savings for vehicles introduced beginning in 2007 are estimated in the same
manner except that they are assumed to always be operated on the required low sulfur diesel fuel
and are presented in Table V.C-35. As the number of new vehicles in the fleet increases the total
savings realized through the use of low sulfur diesel fuel increases in proportion as seen in the
table.
V-144
-------
Chapter V: Economic Impact
Table V.C-34. Aggregate Savings to the Existing Fleet (pre-2007 fleet) Made
Possible by Low Sulfur Diesel Fuel
Calendar Year
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
2035
Aggregate Savings
$80,431,146
$220,884,072
$182,897,940
$149,160,147
$150,798,213
$134,706,583
$120,588,873
$86,874,251
$75,690,690
$65,859,406
$63,293,317
$73,979,411
$64,024,039
$55,275,661
$47,592,312
$40,856,134
$34,971,949
$29,858,909
$25,419,320
$21,528,956
$18,117,665
$15,124,047
$12,494,644
$10,182,634
$8,147,249
$6,307,402
$4,685,085
$3,334,379
$2,061,983
$949.181
V-145
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table V.C-35. Aggregate Savings for the New Fleet (2007 and later) Made Possible
by Low Sulfur Diesel Fuel
Calendar Year
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
2035
Aggregate Savings
$0
$24,971,224
$66,505,419
$104,161,427
$138,377,022
$169,546,107
$198,021,885
$245,459,554
$269,806,304
$292,307,844
$313,186,610
$332,638,971
$350,838,377
$367,935,864
$384,061,052
$399,321,575
$413,804,401
$427,583,405
$440,747,742
$453,410,634
$465,636,121
$477,480,251
$488,991,818
$500,213,568
$511,182,712
$521,973,234
$532,565,116
$542,908,991
$553,181,391
$563.308.011
V-146
-------
Chapter V: Economic Impact
6. Per-Engine Life-Cycle Fuel Costs
The additional cost of diesel fuel meeting our 15 ppm cap is encountered by the average
engine owner each time the fuel tank is refilled. The impacts of the diesel sulfur standard on the
average engine owner can therefore be calculated as the increased fuel costs in cents per gallon,
multiplied by the total number of gallons used by an engine over a particular timeframe. Thus
we have calculated the in-use impact of our diesel sulfur standard on a per-engine basis for both a
single year and for an engine's entire lifetime.
Since we have introduced a temporary compliance option and small refiner hardship
provisions for the diesel sulfur standard that will apply in the initial years, both 15 ppm highway
fuel and 500 ppm highway fuel will be present in the distribution system at the same time during
these years. As discussed in Section V.C above, there are both refinery cost savings and
distribution system costs that occur as a result of these provisions. It is appropriate to consider
these costs and savings as applying to the entire highway diesel pool. In order for refiners to
continue producing 500 ppm fuel, we anticipate that they will have to purchase credits from
refiners producing 15 ppm fuel, in essence rasing the cost of producing 500 ppm fuel and
lowering the cost of producing 15 ppm fuel. Furthermore the distribution system costs are likely
to be recouped by the industry across both grades of highway diesel fuel. As a result, we have
concluded that the fuel costs associated with the program we are finalizing today should be
assigned equally to all gallons of highway diesel fuel, whether 15 ppm or 500 ppm sulfur fuel.
The total cost of 15 ppm diesel fuel is the sum of refinery desulfurization costs, addition
of a lubricity additive, and increases in distribution costs. Refinery desulfurization and
distribution costs are discussed earlier in this Chapter, and average 3.3 eVgal and 1.1 eVgal
respectively during the initial years of the program. Lubricity additives are discussed in Section
V.C.4, and average approximately 0.2 eVgal. Thus we estimate the total cost of diesel fuel
meeting our 15 ppm cap to be 4.5 eVgal during the initial years of the program. This cost will
increase to 5.0 eVgal after 2010.
In a single year, the average in-use heavy-duty engine travels approximately 30,000
milesff, though the mileage of any given engine varies by usage, age, and other factors. Applying
the average heavy-duty fuel economy, the cost for 15 ppm diesel fuel of 4.5 eVgal leads us to a
per-engine estimate of approximately $187. This is the additional cost that the average engine
owner will incur for fuel in the first year of our program, if the full social costs of meeting our
standards are passed onto consumers. However, fuel prices may be higher or lower depending on
market conditions. The costs for different engine classes will vary, of course, based on their
respective annual mileages and fuel economies.
ff Calculated from the annual miles traveled per heavy-duty engine for each year of a engine's life,
multiplied by a distribution of engine registrations by year. Estimate of 30,000 miles per year includes all HD
weight classes and urban buses.
V-147
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 _ EPA420-R-00-026
The per-engine cost of 15 ppm diesel fuel can also be calculated over the lifetime of a
engine. However, to calculate a lifetime cost for the average in-use engine, it is necessary to
account for the fact that individual engines experience different lifetimes in terms of years that
they remain operational. This distribution of lifetimes is the engine survival rate distribution, for
which we used registration data from an Arcadis report. The costs of 15 ppm diesel fuel incurred
over the lifetime of the average fleet engine 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 $/engine
(AVMT); = Annual engine miles travelled in year i of a engine's operational life86
(SURVIVE); = Fraction of engines still operating after i years of service87
C = Cost of 15 ppm diesel fuel,$0.045/gal in 2006 and $0.050/gal in 201 1
FE = Fuel economy in miles per gallon (Appendix VI-A)
i = Engine years of operation, counting from 1 to 30
We used the above equation to calculate lifetime fuel costs separately for LH, MH, HH, and
urban buses. We also weighted the per-engine costs for the individual engine classes by their
contribution to sales. The results are shown in Table V.C-36 as "undiscounted lifetime costs."
An alternative approach to calculating lifetime per-engine costs of 15 ppm diesel fuel is
to discount future year costs. This approach leads to "net present value" lifetime fuel costs, and
is a useful means for showing what the average engine owner would have to spend in the first
year in order to pay for all future year fuel costs. It also provides a means for comparing the
program's costs to its emission reductions in a cost-effectiveness analysis, as described in
Chapter VI.
Discounted lifetime fuel costs are calculated in an analogous manner to the undiscounted
values, except that each year of the summation is discounted at the average rate of 7 percent. The
equation given above can be modified to include this annual discount factor:
LFC = £ [{(AVMT); • (SURVIVE); • (C) - (FE)}/(1.07)M]
Once again, we used the above equation to calculate discounted lifetime fuel costs separately for
LH, MH, HH, and urban buses, then weighted the per-engine costs for the individual engine
classes by their contribution to sales. The results are shown in Table V.C-36 as "discounted
lifetime costs."
V-148
-------
Chapter V: Economic Impact
Table V.C-36. Fleet Average Per-Engine Lifecycle Costs Of Diesel Fuel ($)
First year
Undiscounted lifetime, near-term
Undiscounted lifetime, long-term
Discounted lifetime, near-term
Discounted lifetime, long-term
LH
58
801
837
576
609
MH
110
1497
1565
1077
1141
HH
390
5395
5654
3969
4209
UB
428
7426
7629
4772
4959
All
187
2583
2703
1881
1993
LH = Light heavy duty, MH = Medium heavy duty, HH = Heavy heavy duty,
UB = Urban buses, All = Consumption weighted average of all engine weight classes
D. Combined Total Annual Nationwide Costs
Figure V.D-1 and Table V.D-1 summarize EPA's estimates of total annual costs to the
nation for heavy-duty diesel engines, heavy-duty gasoline vehicles, and 15 ppm diesel fuel. 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 15 ppm diesel fuel consumed nationwide, including
that consumed in both highway and off-highway applications. Annual aggregate costs change as
our new standards are phased-in and projected per-vehicle costs and annual sales change over
time. The aggregate fuel costs change due to the temporary compliance option which applies
between 2006 and 2010, and as annual fuel consumption changes over time as predicted by the
Energy Information Administration. The methodology we used to derive the aggregate costs are
described in detail in the previous Sections of this chapter. As shown below, total annual costs
increase over the period of the temporary compliance option and peak at about $3.6 billion in
2010. Total annualized costs are projected to increase gradually after 2010 due to projected
growth in vehicle sales and fuel consumption.
V-149
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
5.0
4.0
3.0
in
O
= 2.0
m
1.0
0.0
Diesel
engine
Gasoline
vehicle
Diesel
fuel
Total
2005
2010
2015
2020
2025
2030
Figure V.D-1. Total annualized costs of heavy-duty diesel engines, heavy-duty gasoline
vehicles, and 15 ppm diesel fuel
In Figure V.D-1, aggregate engine costs exhibit notable drops in years 2009 and 2011.
The drop in year 2009 is due to the onset of the "learning curve" in the third year of the engine
manufacturer's production of engines meeting our new standards. In year 2010, the NOx phase-
in ends and the remaining 50 percent of new engines must meet the new standards. This change
causes a sudden increase in aggregate costs in 2010. In year 2011, a learning curve adjustment is
again made and costs drop once again. Finally, in 2012 the fixed costs expire and the costs drop
one last time. Thereafter, costs continue to increase due to growth in the fleet.
V-150
-------
Chapter V: Economic Impact
Table V.D-1. Total annualized costs of heavy-duty diesel engines, heavy-duty gasoline
vehicles, and 15 ppm diesel fuel (Smillion)
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
Diesel
engines
(80)
1,266
1,321
1,072
1,520
1,225
1,133
1,157
1,180
1,141
1,156
1,159
1,182
1,205
1,226
1,247
1,268
1,288
1,307
1,326
1,344
1,362
1,380
1,398
1 415
Gasoline
vehicles
0
0
46
80
81
82
83
78
79
80
82
83
84
85
86
87
89
90
91
92
93
94
95
97
98
Diesel fuel
880
1,786
1,809
1,904
2,014
2,128
2,160
2,192
2,225
2,258
2,292
2,327
2,362
2,397
2,433
2,469
2,506
2,544
2,582
2,621
2,660
2,700
2,741
2,782
2824
Total
799
3,052
3,177
3,056
3,615
3,434
3,376
3,427
3,484
3,480
3,530
3,568
3,628
3,687
3,746
3,804
3,863
3,921
3,980
4,039
4,098
4,157
4,217
4,276
4.337
In support of the this rulemaking, the Agency is preparing both a benefit-cost analysis
(BCA) and a cost-effectiveness analysis. The BCA presents and compares the social benefits
(e.g., avoided adverse health effects) and social costs (e.g., direct compliance expenditures) of
the program. Since many of the benefits and costs are manifest in future years, we apply
discounting methods to adjust the dollar values of these effects to reflect the finding that society
V-151
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
as a whole typically values the realization (or avoidance) of a given effect differently depending
on when the effect occurs. Because the BCA reflects the value of benefits and costs from the
perspective of society as a whole, we use a 3 percent rate to discount future year effects in our
primary analysis. The 3 percent rate is in the 2 to 3 percent range recommended by the Science
Advisory Board's Environmental Economics Advisory Committee for use in EPA social benefit-
cost analyses, a recommendation incorporated in EPA's new Guidelines for Preparing Economic
Analyses (November 2000). OMB Circular A-94 requires us to generate benefit and cost
estimates reflecting a 7 percent rate, and results based on OMB's preferred 7 percent rate are also
presented to demonstrate the sensitivity of our results to the discount rate assumption.
The BCA focuses on calender year 2030 in its comparison of costs and benefits. The
2030 total program cost shown in Table V.D-1 above was based on a discount rate of 7 percent.
Since the BCA requires a 2030 cost which based on a 3 percent discount rate, we developed this
separately. Thus the total program cost in calender year 2030 using a discount rate of 3 percent is
$4.2 billion. Note that since the discount rate only affects the return on capital investments in
any given calender year, and since all engine and vehicle capital investments have been recovered
by 2030, the only effect of the discount rate in year 2030 is for fuel costs. As a result, the total
program cost under the 3 percent assumption is very close to that under a 7 percent assumption.
V-152
-------
Chapter V: Economic Impact
Chapter V. References
1. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced Maintenance"
ICF Consulting, December 9, 1999. Copy available in EPA Air Docket A-99-06, docket
item number II-A-28.
2. "Fleet Characterization Data for MOBILE6: Development and Use of Age Distributions,
Average Annual Mileage Accumulation Rates and Projected Vehicle Counts for Use in
MOBILE6." EPA420-P-99-011, April 1999.
3. "Update of EPA's Motor Vehicle Emission Control Equipment Retail Price Equivalent
(RPE) Calculation Formula," Jack Faucett Associates, Report No. JACKFAU-85-322-3,
September 1985. Copy available in EPA Air Docket A-99-06, docket item number II-A-
31.
4. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Revised Final Report, Engine, Fuel, and
Emissions Engineering, Incorporated, December 15, 1999. Copy available in EPA Air
Docket A-99-06, docket item number II-A-27.
5. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February
23, 1990, Vol. 247, pp. 920-924. Copy available in EPA Air Docket A-98-32, docket
item number II-D-14.
6. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
Incorporated, December 15, 1999 EPA Docket A-99-06.
7. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
Incorporated, December 15, 1999 EPA Docket A-99-06.
8. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
Incorporated, December 15, 1999 EPA Docket A-99-06.
9. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
Incorporated, December 15, 1999 EPA Docket A-99-06.
10. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering,
V-153
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Incorporated, December 15, 1999 EPA Docket A-99-06.
11. "Estimated Economic Impact of New Emission Standards for Heavy-Duty Highway
Engines," Acurex Environmental Corporation. March 31, 1997. Copy available in EPA
Air Docket A-98-32, docket item number IV-A-05.
12. "Estimating Closed Crankcase System Costs for Turbocharged Heavy-Duty Engines,"
Memo from Byron Bunker to Air Docket A-99-06.
13. See the Regulatory Impact Analysis for the proposed 2004 Heavy-Duty rule. Copy
available in EPA Air Docket A-98-32, docket item number HI-B-Ol.
14. "Final Regulatory Impact Analysis: Control of Emissions from Marine Diesel Engine",
EPA420-R-99-026, November 1999, page 87.
15. Hawker, P. et al, Experience with a New Paniculate Trap Technology in Europe, SAE
970182. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
16. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Revised Final Report, Engine, Fuel, and
Emissions Engineering, Incorporated, December 15, 1999. Copy available in EPA Air
Docket A-99-06, docket item number II-A-27.
17. Dou, D., et al.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE 982594.
Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
18. Herzog, P. et al, NOx Reduction Strategies for DI Diesel Engines, SAE 920470, Society
of Automotive Engineers 1992 (from Figure 1).
19. Zelenka, P., et al., "Cooled EGR - A Key Technology for Future Efficient HD Diesels",
SAE 980190. Copy available in EPA Air Docket A-99-06, docket item IV-G-131.
20. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February
23, 1990, Vol. 247, pp. 920-924. Copy available in EPA Air Docket A-98-32, docket
item number II-D-14.
21. "Treating Progress Functions As Managerial Opportunity", J.M Dutton and A. Thomas,
Academy of Management Review, Rev. 9, 235, 1984. Copy available in EPA Air Docket
A-98-32, docket item number H-D-13.
22. See the Regulatory Impact Analysis for the proposed 2004 Heavy-Duty rule. Copy
available in EPA Air Docket A-98-32, docket item number HI-B-Ol.
23. See 65 FR 59896, October 6, 2000.
V-154
-------
Chapter V: Economic Impact
24. See 65 FR 6698, February 10, 2000.
25. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Final Report, Arcadis Geraghty &
Miller, September 30, 1998. Copy available in EPA Air Docket A-99-06.
26. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by
Reduction of Diesel Fuel Sulfur Content," Revised Final Report, Engine, Fuel, and
Emissions Engineering, Incorporated, December 15, 1999.
27. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Final Report, Arcadis Geraghty &
Miller, September 30, 1998. Copy available in EPA Air Docket A-99-06.
28. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Final Report, Arcadis Geraghty &
Miller, September 30, 1998. Copy available in EPA Air Docket A-99-06.
29. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Final Report, Arcadis Geraghty &
Miller, September 30, 1998. Copy available in EPA Air Docket A-99-06.
30. See Chapter V of the final Tier 2 final RIA, contained in Air Docket A-97-10.
31. 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.
32. "Final Regulatory Impact Analysis: Control of Emissions from Marine Diesel Engine,"
EPA420-R-99-026, November 1999.
33. See the Regulatory Impact Analysis for the Tier 2 Final Rulemaking, Chapter V,
contained in Air Docket A-97-10.
34. Worldwide Refining/Worldwide Production, Oil and Gas Journal, vol. 97, No. 51,
December 20, 1999.
35. Final Report: 1996 American Petroleum Institute/National Petroleum Refiners
Association Survey of Refining Operations and Product Quality, July 1997.
36. Conversation with Steve Mayo, Hydroprocessing Specialist, Akzo Nobel, February 1999.
37. Final Report: 1996 American Petroleum Institute/National Petroleum Refiners
Association Survey of Refining Operations and Product Quality, July 1997.
38. American Automobile Manufacturers Association Diesel Fuel Survey, Summer 1997.
39. Diesel Fuel, Specifications and Demand for the 21st Century, UOP, 1998.
V-155
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
40. Refining Economics Report, Octane Week, various publications.
41. Confidential Information Submission from Diesel Desulfurization Vendor A, August
1999.
42. UOP Information Submission to the National Petroleum Council, August 1999.
43. "The Lower it Goes, The Tougher it Gets," Bjorklund, Bradford L., UOP, Presentation at
the National Petroleum Council Annual Meeting, March 2000.
44. U.S. Petroleum Refining Draft Report, Appendix H, National Petroleum Council, March
30, 2000.
45. Conversation with Cal Hodge, A Second Opinion, February 2000.
46. Recent Advances in Biodesulfurization of Diesel Fuel, Pacheco, Michael A., Presentation
made to the National Petrochemical and Refiners Association Annual Meeting, March
1999.
47. Desulfurization by Selective Oxidation and Extraction of Sulfur-Containing Compounds
to Economically Achieve Ultra-Low Proposed Diesel Fuel Sulfur Requirements,
Chapados, Doug, Presentation made to the National Petrochemical and Refiners
Association Annual Meeting, March 2000.
48. Conversation with Brad Bjorklund, Senior Manager R&D Hydroprocessing, UOP, March
2000.
49. Reverdy, Fracois R., Packinox: Economic and Environmental Advantages of Welded
Plate Feed-Effluent Heat Exchangers in Refining, Presentation at the National Petroleum
Council Annual Meeting, March 1995.
50. Conversation with Art Suchanek and Woody Shiflett, Criterion Catalysts, March 2000.
51. Conversation with Tom W. Tippett et al, Refining Technology Division, Haldor Topsoe,
March 2000.
52. Hirshfeld, David, Analyzing the Refining Economics of Diesel Fuel Sulfur Standards,
Presentation made by Mathpro to EPA, October 12, 1999.
53. Conversation with Dave Dicamillo, Criterion Catalysts, February 2000.
54. Conversation with Jim Kennedy, UOP, November 2000.
V-156
-------
Chapter V: Economic Impact
55. Hirshfeld, David, Analyzing the Refining Economics of Diesel Fuel Sulfur Standards,
Presentation made by Mathpro to EPA, October 12, 1999.
56. Conversation with Jim Kennedy, Manager Project Sales, Distillate and Resid
Technologies, UOP, November 2000.
57. Conversation with Tom W. Tippett et al, Refining Technology Division, Haldor Topsoe,
March 2000.
58. Conversation with Tim Heckel, Manager of Distillate Technologies Sales, UOP, March
2000.
59. Conversation with Tom W. Tippett et al, Refining Technology Division, Haldor Topsoe,
March 2000.
60. Very-Low-Sulfur Diesel Distribution Cost, Engine Manufacturers Association, August
1999.
61. Moncrief, T. I, Mongomery, W. D., Ross, M.T., Charles River Associates, Ory, R. E.,
Carney, J. T., Baker and O'Brien Inc., Ann Assessment of the Potential Impacts of
Proposed Environmental Regulations on U.S. Refinery Supply of Diesel Fuel, Charles
River and Baker and O'Brien for the American Petroleum Institute, August 2000.
62. Christie, David A., Advanced Controls Improve Reformulated Fuel Yield and Quality,
Fuel Reformulation, July/August 1993.
63. Personal conversation with Debbie Pack, ABB Process Analytics Inc., November 1998.
64. Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).
65. Conversation with Lyondel-Citgo refinery staff, April 2000.
66. Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical
Engineers, Third Edition, McGraw Hill Book Company, 1980.
67. Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for
Discussion, Price-Waterhouse Coopers for the American Automobile Manufacturers,
October 5, 2000.
68. U.S. Petroleum Refining, Meeting Requirements for Cleaner Fuels and Refineries,
Volume V - Refining Capability Appendix, National Petroleum Council, 1993.
V-157
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
69. Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for
Discussion, Price-Waterhouse Coopers for the American Automobile Manufacturers,
October 5, 2000.
70. Petroleum Marketing Annual, Energy Information Administration, 1999.
71. Personal conversation with A. Judzis of Equilon.
72. Perry, Robert H., Chilton, Cecil H., Chemical Engineer's Handbook, McGraw Hill 1973.
73. Presentation by Air Products and Chemicals, Hart' s/IRI World Fuels Conference, San
Francisco, California, March 1998.
74. Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical
Engineers, Third Edition, McGraw Hill Book Company, 1980.
75. Jena, Rabi, Take the PC-Based Approach to Process Control, Fuel Reformulation,
November/December 1995.
76. Sutton, IS., Integrated Management Systems Improve Plant Reliabiliy, Hydrocarbon
Processing, January 1995.
77. King, M. J., Evans, H. N., Assessing your Competitors' Application of CEVI/CIP,
Hydrocarbon Processing, July 1993.
78. "Refining economics of diesel fuel sulfur standards," study performanced for the Engine
Manufactuers Association by MathPro, Inc. October 5, 1999. See EPA Air Docket A-99-
06.
79. Energy Information Administration (EIA), Petroleum Supply Annual (PSA), Department
of Energy, 1999.
80. National Association of Truck Stop Operators (NATSO) fact sheet,
http ://www.natso. com.
81. Letter from Jason Lynn of the National Association of Truck Stop Operators (NATSO)
regarding the results of a survey of its members, September 15, 1999, docket item II-D-
85.
82. "Refining Economics of Diesel Fuel Sulfur Standards," study performed by
MathPro, Inc. for the Engine Manufacturers Association, October 5, 1999. EPA Air
Docket A-99-06.
V-158
-------
Chapter V: Economic Impact
83. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced Maintenance"
ICF Consulting, December 9, 1999.
84. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced Maintenance,"
ICF Consulting, December 9, 1999.
85. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced Maintenance,"
ICF Consulting, December 9, 1999.
86. Table 6, Agency Report Number EPA420-P-99-011, "Fleet characterization data for
MOBILE6: development and use of age distributions, average annual mileage
accumulation rates and projected vehicle counts for use in MOBILE6," April 1999.
87. See "Update of fleet characterization data for use in MOBILE6," Arcadis Inc., April27,
1998, prepared for U.S. EPA. Registration distribution from Table 4-7 was used to
represent survival distribution.
83. Energy Information Administration (EIA), Petroleum Supply Annuals (PSA), Department
of Energy, 1995-99.
84. "Alternate Markets for Highway Diesel Fuel Components" Muse, Stancil & Co.
September 2000
85. API Reported Refining & Marketing Capital Investments, 1990-1998.
V-159
-------
Chapter VI: Cost-Effectiveness
Chapter VI: Cost-Effectiveness
This Chapter will present the cost-effectiveness analysis we completed for our new
heavy-duty gasoline vehicle, heavy-duty diesel engine, and diesel fuel sulfur standards. Under
Clean Air Act Section 202(a)(3), we are required to promulgate standards which reflect the
greatest degree of emission reduction achievable, giving appropriate consideration to cost,
energy, and safety factors. The standards we set are not premised on the need to promulgate the
most cost-effective standards. However, we have determined that cost-effectiveness is a useful
tool in evaluating the appropriateness of our standards.
The cost-effectiveness analysis described in this Section relies in part on cost information
from Chapter V and emissions information from Chapter II to estimate the dollars per ton of
emission reductions produced from our standards. We have calculated the cost effectiveness
using two different approaches, a per-vehicle approach that considers the costs incurred and
emission reductions produced for a single vehicle or engine, and a 30-year net present value
approach that accounts for all costs and emission reductions over a 30 year period beginning in
2006. The comparative merits and drawbacks of both approaches are described in Sections VIA
and VIE. Finally, this Chapter compares the cost-effectiveness of the new provisions with the
cost-effectiveness of other control strategies from previous and potential future EPA programs.
Sections VI.A, VLB and VI.C describe the per-vehicle calculations for our combined
heavy-duty diesel engine and diesel fuel sulfur standards, while Section VI.D describes the per-
vehicle calculations for heavy-duty gasoline vehicles. Section VIE describes the 30-year net
present value cost effectiveness analysis. The results of all cost-effectiveness calculations are
given in Section VI.F. Comments we received in response to our Notice of Proposed
Rulemaking on the subject of cost effectiveness, along with our responses to those comments,
can be found in Issue 5.9 of the Response To Comments document.
A. Overview of the Per-vehicle Analysis
The per-vehicle cost-effectiveness analysis conducted for our standards focused on the
costs and emission reductions associated with a single engine (or vehicle, in the case of heavy-
duty gasoline vehicle standards) meeting the 2007 model year standards, and operating on low
sulfur fuel. Both costs and emission reductions were calculated over the life of the engine and
then discounted at a rate of seven percent. Costs and emission reductions were measured relative
to a baseline consisting of the 2004 certification standards and average diesel sulfur levels falling
under the current 500 ppm cap. The calculations were performed separately for each engine class
and the results weighted according to the expected fleet mix. Details on the per-vehicle approach
VI-1
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
to cost-effectiveness follow. The presentation of the 30-year net present value cost effectiveness
calculations can be found in Section VIE. This latter approach includes the fuels costs incurred
by the pre-2007MY fleet which are not accounted for in the per-vehicle analysis. Note that many
of the issues discussed in this Section VIA also apply to the calculation of 30-year net present
value cost-effectiveness.
1. Temporal and Geographic Applicability
Our per-engine approach to our cost-effectiveness calculations produces $/ton values
representing any controlled engine, no matter where that engine operates. In effect, this means
that emission reductions in both attainment and nonattainment areas are included in our cost-
effectiveness analysis. The same holds true for our 30-year net present value analysis. Although
this may limit the usefulness of comparisons to stationary source controls, we believe that our
approach is appropriate. Both the engine and diesel sulfur programs are to apply nationwide, so
the same emission reductions will occur regardless of where the engine operates. Attainment
area emission reductions also produce health benefits. In general, the benefits of NMHC
reductions in ozone attainment areas include reductions in emissions of air toxics, reductions in
the contribution from NMHC emissions to the formation of fine particulate matter, and
reductions in damage to agricultural crops, forests, and ecosystems from ozone exposure.
Emission reductions in attainment areas help to maintain clean air as the economy grows and
new pollution sources come into existence. Also, ozone health benefits can result from
reductions in attainment areas, although the most certain health effects from ozone exposure
below the NAAQS appear to be both transient and reversible. The closure letter from the Clean
Air Science Advisory Committee (CAS AC) for the recent review of the ozone NAAQS states
that there is no apparent threshold for biological responses to ozone exposure.1
In the Regulatory Impact Analysis for a recent rulemaking for highway heavy-duty diesel
engine standards,2 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.
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-engine 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
VI-2
-------
Chapter VI: Cost-Effectiveness
program will be higher immediately after it is implemented than they will be after several years,
since engine and vehicle manufacturers can take advantage of decreasing capital and operating
costs over time, and will learn how to produce their products more efficiently as time goes on.
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-engine costs are given in
Section VLB.2 for diesel engines and in Section VI.D.l for heavy-duty gasoline vehicles.
We are also proposing that our combined engine/sulfur program (hereafter, this includes
our standards for heavy-duty diesel engines, heavy-duty gasoline vehicles, and diesel sulfur) be
an annual program. Since cost-effectiveness only has relevance when compared to alternative
strategies, we must use an approach to calculating the cost-effectiveness of our annual program
that is consistent with the approaches taken for other rulemakings. For programs that generate
emission reductions outside of the ozone season, we generally include those "winter season"
reductions in the cost-effectiveness calculations. Thus our cost-effectiveness estimates will
include all the emission reductions produced as a result of our standards, no matter where or
when those reductions occur. This is consistent with the methodology used in prior rulemakings
and allows for an apples-to-apples comparison.
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.
As stated above, we must use an approach to cost-effectiveness that is consistent with the
approach taken in other rulemakings in order to provide an apples-to-apples comparison. Most
other mobile source rulemakings use average cost-effectiveness, including our recently
VI-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
promulgated standards for Tier 2 vehicles and gasoline sulfur. Therefore, we have chosen to
calculate cost-effectiveness on an average rather than an incremental basis for our standards.
Since today's program includes both fuel standards and engine standards, it was necessary
for us to define a baseline for both fuels and engines. For highway diesel fuel, the previous
standard was set in 1990, limiting the sulfur content to a maximum of 500 ppm starting in 1993.
However, the average sulfur level has been significantly less than 500 ppm, closer to 340 ppm.3
Therefore we have determined that the sulfur baseline should be 340 ppm.
For heavy-duty diesel engines, the previous set of standards was originally set in 1997
and applies to the 2004 model year.1 These standards included a 2.4 g/bhp-hr cap for
NOx+NMHC or 2.5 g/bhp-hr with a 0.5 g/bhp-hr cap on NMHC. For the purposes of analysis
we have assumed that manufacturers will met this standard with 2.3 g/bhp-hr NOx and 0.2
g/bhp-hr NMHC. However, unlike the PM standards we are proposing today, 2004 model year
urban buses are required to meet a different PM standard (0.05 g/bhp-hr) than other heavy-duty
engines (0.1 g/bhp-hr). Thus we have used two different baselines for PM, one for urban buses
and another for other heavy-duty engines. Despite this, we are calculating only a single set of
cost-effectiveness values for all engines since we are proposing that a single set of standards
apply to urban buses and other heavy-duty engines.
For heavy-duty gasoline vehicles, the previous set of standards applies to the 2005 model
year. For incompletes, these include a 1.0 g/bhp-hr NOx+NMHC standard, which we assume
separates practically into a 0.8 g/bhp-hr standard for NOx and a 0.2 g/bhp-hr standard for
NMHC. For Class 2b completes, the 2005 standards include 0.9 g/mile for NOx and 0.28 g/mile
for NMHC. Finally, for Class 3 completes, the 2005 standards include 1.0 g/mi for NOx and
0.33 g/mi for NMHC.
B. Diesel Costs
The costs used in our cost-effectiveness calculations are the sum of the added costs of
compliance with the 2007 engine and diesel sulfur standards on a per-engine basis, in comparison
to the engine and fuel baselines. Costs result from discounting over the lifetime of an engine at a
seven percent discount rate. In addition, all costs represent the fleet-weighted average of all
light, medium, and heavy-heavy engines, as well as urban buses.
a Under a consent decree, many manufacturers will be complying with these heavy-duty standards as early
as 2002. Standards were finalized in the Federal Register at 65 FR 59896, October 6, 2000.
VI-4
-------
Chapter VI: Cost-Effectiveness
1. Near and Long-Term Cost Accounting
Since the costs of complying with the 2007 engine standards will vary over time, we
believe 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 engines that
will meet the 2007 standards would generally be 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 (the so-called "learning curve"), and as a result their operating costs decrease over time. The
implications of this learning curve on engine costs is discussed in Section V.A.I.
Thus near-term costs represent the highest costs of the program, as they include all capital
costs and no cost savings due to the manufacturer's learning curve. Long-term costs, on the other
hand, represent the lowest costs of the program which occur after a portion of capital cost
amortizations have ended and all learning curve cost savings have been accounted for. For the
purposes of this rulemaking, therefore, we will present cost-effectiveness of our program on both
a near-term and long-term basis. Details about the calculation of near and long-term engine costs
can be found in Section V.A.
2. Diesel Engine and Fuel Costs
The per-engine costs used in our cost-effectiveness calculations were derived and
presented in Section V.A. Engine hardware costs were presented in Section V.A for the four
engine categories affected by our standards. For the purposes of calculating cost-effectiveness,
we weighted the costs for those four individual engine categories by the expected fleet fractions
(see Table VI.C-2 below) to obtain fleet-average costs for our emissions standards. Also, we
treated first-year production costs as the "near-term" costs, and sixth-year production costs as the
"long-term" costs. For low sulfur diesel, we used the discounted lifetime costs presented in
Table V.C-36 which include costs for desulfurization, lubricity additives, and distribution costs.
The costs used in our cost-effectiveness calculations are shown in Table VI.B-1.
Table VI.B-1. Fleet-average, Per-engine Costs for HDDE
Near-term
Long-term
NOx adsorber, PM trap,
and oxy catalyst ($)
2457
1332
Fuel cost ($)
1881
1993
Total costs ($)
4338
3325
VI-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Note that the total costs in Table VI.B-1 were used for establishing "uncredited" cost-
effectiveness values. As described in Section VLB.4, the costs from Table VI.B-1 were also
adjusted to produce "credited" cost-effectiveness values.
3. Methodology for assigning costs to NOx, NMHC, and PM
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 the primary purpose of our standards is to reduce emissions of the ozone
precursors hydrocarbons and oxides of nitrogen, as well as emissions of particulate matter.
Therefore, consistent with our approach in previous rulemakings such as the recently finalized
standards for Tier 2 vehicles and gasoline sulfur, we have calculated cost-effectiveness on the
basis of total NOx + NMHC emissions.
However, since we are also proposing that a new standard be set for PM, we must
develop a separate cost-effectiveness value for that pollutant. We do not think it appropriate to
combine NOx, NMHC, and PM all into a single cost-effectiveness value, since there are separate
NAAQS for ozone and PM, and these two pollutants do not have identical effects on human
health and the environment. We must therefore determine a reasonable way to split the costs of
compliance with our combined engine/diesel sulfur program between NOx+NMHC and PM.
As described in Section HI. A and in our Draft RIA, we expect that manufacturers will use
both NOx adsorbers and PM traps to comply with our engine standards. However, since
publication of the NPRM we have determined that NMHC emissions may not be sufficiently
controlled by the NOx adsorber without the use of a clean-up oxidation catalyst downstream of
the adsorber. See Section ni.A.4 for a more detailed discussion of this issue. The NOx adsorber
and oxidation catalyst will together enable heavy-duty diesel engines to meet our new NOx and
NMHC standards. As a result, we believe that the total hardware costs associated with the NOx
adsorber and oxidation catalyst should be applied to the calculation of NOx+HC cost-
effectiveness. The PM trap will continue to provide reductions in both PM and HC as well as
pre-conditioning the engine-out exhaust stream for introduction to the NOx adsorber. As a
result, for the purposes of calculating cost-effectiveness, we believe that the hardware costs of the
PM trap should be divided equally between PM and NOx+HC, consistent with the approach
taken in the NPRM.
In order to divide the fuel costs appropriately between NOx+HC and PM, we have taken
an approach consistent with that described in the Draft RIA. The diesel fuel sulfur cap of 15 ppm
has been implemented in order to enable the two aftertreatment components of PM trap and NOx
adsorber (+ oxidation catalyst) to operate properly. Since the fuel sulfur standard applies equally
to both components of the aftertreatment, we believe it is appropriate to divide fuel costs evenly
between the PM trap and the NOx adsorber (+ oxidation catalyst).
VI-6
-------
Chapter VI: Cost-Effectiveness
However, as described above, the PM trap will continue to provide reductions in both PM
and HC, pre-conditioning the engine-out exhaust stream for introduction to the NOx adsorber.
We therefore believe it is appropriate to divide the fuel costs applicable to the trap, calculated as
half of total fuel costs, equally between PM and HC. As a result, 25 percent of total fuel costs
would apply to the calculation of PM cost effectiveness, while the remaining 75 percent would
apply to the calculation of cost effectiveness for NOx+NMHC. Likewise, half of the hardware
costs for the PM trap would be included in the calculation of cost effectiveness for NOx+NMHC.
This approach is consistent with that taken in the NPRM
4. Cost Crediting for SO2
The reduction in diesel sulfur levels that would result from our standards will necessarily
result in reductions in sulfur-containing compounds that exit the engine. These compounds are
limited to sulfur dioxide (SO2) and sulfate particulate matter. The latter will be taken into
account as manufacturers seek to comply with our new PM standard, and thus will be
automatically represented in our cost-effectiveness estimates of $/ton PM. However, there is no
engine standard for SO2. Since reductions in emissions of SO2 are beneficial and represent a true
value of our program, we believe it is appropriate to account for them in our cost-effectiveness
analysis.
The primary benefit of reductions in SO2 emissions is a reduction in secondary PM,
formed when SO2 reacts with water and ammonia in the atmosphere to form ammonium sulfate.
Therefore, we believe that any crediting for reductions in SO2 should be applied to our PM costs.
To account for reductions in emissions of SO2 in our cost-effectiveness calculations, we
have calculated a second set of $/ton values in which we credit some of the costs to SO2, with the
remaining costs being used to calculate $/ton PM. As a result, we have produced both "credited"
and "uncredited" $/ton PM values; the former takes into account the SO2 emission reductions
associated with our standards, while the latter does not.
Cost-effectiveness values for the control of SO2 represent conservative estimates of the
cost of measures that would need to be implemented in the future in order for all areas to reach
attainment. Such cost-effectiveness values are therefore an appropriate source for estimating the
amount of the costs to credit to SO2. As a result, we credited some costs to SO2 through the
application of cost-effectiveness ($/ton) values for this pollutant 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 SO2 However, in this rulemaking, we chose to
simplify by using more conservative approaches to establish crediting values for SO2. The
potential future programs evaluated as part of the NAAQS revisions rulemaking (discussed in
more detail in Section VI.F below) provided a reasonable source for identifying the value of SO2
VI-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
in terms of its cost-effectiveness. In this process we did not make a distinction between SO2
emissions from mobile or stationary sources since there is little data to suggest that a tons of SO2
from one source contributes differently to PM or acid rain problems than a ton of SO2 from
another source.
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 cost crediting was applied after all costs associated with compliance with our
standards were calculated and summed. The per-engine tons reduced of SO2 was multiplied by
the representative cost-effectiveness value of $4800/ton (see Section VI.C.2 below for SO2 tons
calculations). As a result, $446 of the total costs were apportioned to SO2 in the calculation of
PM cost-effectiveness. This amount is independent of whether we are considering a near-term or
long-term cost-effectiveness calculation, since the lifetime tons reduced for this compound is the
same, on a per-engine basis, in any year of the program. A summary of the costs used in our
cost-effectiveness calculations is given below in Table VI.B-2, including all engine, fuel, and
fuel economy costs.
Table VI.B-2. Fleet Average Per-Engine Costs for HDDE Used in Cost-effectiveness
Total uncredited costs
SO2 credit allocation
Total credited costs
Near-term costs ($)
NOx+NMHC
3381
n/a
3381
PM
956
-446
510
Long-term costs ($)
NOx+NMHC
2563
n/a
2563
PM
762
-446
316
C. Emission Reductions from Diesel Engines
In order to determine the overall cost-effectiveness of the standards we are proposing, it
was necessary to calculate the lifetime tons of each pollutant reduced on a per engine basis. This
section will describe the steps involved in these calculations. In general, emission reductions
VI-8
-------
Chapter VI: Cost-Effectiveness
were calculated for NOx, NMHC, PM, and SO2 in a manner analogous to the discounted lifetime
fuel costs described in Section V.C.6.
1. NOx, NMHC, and PM
The discounted lifetime tonnage numbers for NOx, NMHC, and PM for our combined
diesel engine and diesel fuel standards were based on the difference between emissions produced
by engines meeting our 2004 and 2007 standards, as described in Section II.B.l. These in-use
emission levels were expressed in terms of average g/bhp-hr emissions for each year in a
engine's life, up to 30 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 engine's life to generate total mass emissions produced in
each year by that engine (this step included the use of bhp-hr/mile conversion factors)
2) The resultant mass emissions were multiplied by the probability of survival in the
appropriate year, known as the "survival" rate.
3) A seven percent annual discount factor, compounded from the first year of the engine's
life, was then applied for each year to allow calculation of net present value lifetime
emissions.
Converting to tons and summing across all years results in the total discounted lifetime
per-engine tons. This calculation can be described mathematically as follows:
LE = [{(AVMT); • (SURVIVE); • (ER); • (CF) • (K)}/(1.07)M]
Where:
LE = Discounted lifetime emissions in tons/engine
(AVMT); = Annual miles traveled in year i of a engine's operational life4
(SURVIVE); = Probability of engine survival after i years of service
(ER); = Emission rate, g/bhp-hr in year i of an engine's operational life
CF = Heavy-duty engine conversion factor, bhp-hr/mile (see Appendix VI-A)
K = Mass conversion factor, 1 . 102 x 10"6 tons/gram
i = Engine years of operation, counting from 1 to 30
For NOx, NMHC, and PM, we generated discounted lifetime tonnage values for each
engine class (LH, MH, HH, and urban buses) using the above equation. This was done
separately for the baseline and control cases. The baseline case included the 2004 model year
VI-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
engine standards and the in-use diesel sulfur level of 340 ppm. The control case entailed our
2007 model year engines standards and 7 ppm diesel sulfur. The tonnage values that we
calculated according to this procedure are shown in Table VI.C-1.
Table VI.C-1. Per-engine Discounted Lifetime Tons for HDDE
Baseline
LH
MH
HH
Urban bus
Control
LH
MH
HH
Urban bus
NOx
0.409
0.970
3.661
4.300
0.035
0.084
0.320
0.357
NMHC
0.037
0.086
0.325
0.174
0.025
0.059
0.224
0.155
PM
0.017
0.041
0.157
0.097
0.001
0.002
0.009
0.010
The final step before calculating the cost-effectiveness of our program was to weight the
discounted lifetime tonnage values for each engine class by their respective fraction of the HDDE
fleet. These fractions were based on engine count projections for use in MOBILE6 for the year
2020 for diesel-powered heavy-duty engines (see Appendix VI-A), which in turn were based on
current sales fractions for new vehicles. Table VI.C-2 presents the final weighting factors we
used to develop fleet-average tonnage values.
Table VI.C-2. Engine Class Sales Weighting Factors for HDDE
Light-heavy duty
Medium-heavy duty
Heavy-heavy duty
Urban buses
0.342
0.323
0.326
0.009
VI-10
-------
Chapter VI: Cost-Effectiveness
The final discounted lifetime tonnage values for an average fleet engine meeting either the 2004
or 2007 standards are shown in Table VI.C-3. It is these values that were used directly in
calculating the cost-effectiveness of our program.
Table VI.C-3. Fleet average, Per-engine Discounted Lifetime Tons for HDDE
Baseline: 2004 standards
with 340 ppm fuel
Control: 2007 standards
with 7 ppm fuel
Reduction
NOx + NMHC
1.8329
0.2490
1.5839
PM
0.07117
0.00399
0.06718
2. Sulfur Dioxide
The sulfur contained in diesel fuel exits the tailpipe as either sulfuric acid, a sulfate which
is a component of primary particulate matter, or as sulfur dioxide (SO2). Sulfur dioxide is
formed in the engine, and its conversion into sulfuric acid is a function of the type of
aftertreatment and temperature in the tailpipe. If there is no aftertreatment (as is expected to be
the case for engines meeting the 2004 standards), only about 2 percent of sulfur ends up being
converted into sulfuric acid, with the remaining 98 percent being retained as SO2. A large
percentage of the SO2 exiting the tailpipe is converted to sulfate (primarily ammonium sulfate) in
the atmosphere. For engines meeting our 2007 standards, however, we expect the conversion
rate of SO2 to sulfuric acid to be much higher, closer to 30 percent, due to the use of particulate
traps. Thus the calculation of tons of SO2 reduced due to our program compares a baseline of
340 ppm and 98 percent SO2 retention to a control of 7 ppm and 70 percent SO2 retention.
Discounted lifetime tons of SO2 reduced is calculated as the difference between tons of
SO2 for the baseline minus tons of SO2 for our program, where tons are calculated according to
the following equation:
LE = £ [{(AVMT); • (SURVIVE); - (FE) • (D) • (SUL) • (F) • (MC) • (CF) • (K)}/(1.07)M]
VI-11
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Where:
LE = Discounted lifetime emissions of SO2 in tons/engine for either the baseline or
our control program
(AVMT); = Annual engine miles traveled in year i of a engine's operational life
(SURVIVE); = Fraction of engines still operating after i years of service
FE = Fuel economy by engine class (see Appendix VI-A)
D = Density of diesel, 7.1 Ib/gal
SUL = Diesel sulfur concentration, 3.4 xlO"4 Ib sulfur/lb fuel (340 ppm) for the baseline
and 0.07 xlO"4 Ib sulfur/lb fuel (7 ppm) for our program
F = Fraction of total sulfur which exits the tailpipe as SO2
(0.98 for baseline case and 0.70 for control case)
MC = Molar conversion factor, 2 Ib SO2 per Ib sulfur
CF = Heavy-duty engine conversion factor, bhp-hr/mile
K = Mass conversion factor, 5.0 x 10"4 tons/lb
i = Engine years of operation, counting from 1 to 30
After applying the above equation separately for each engine class and weighting the
resulting tonnage values according to the factors presented in Table VI.C-2, we determined that
the fleet-average, per-engine discounted lifetime tons of SO2 reduced is 0.0929. 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.4 and summarized in Table VI.B-2.
D. Costs and Emission Reductions for Heavy-duty Gasoline
Vehicles
Since we are also proposing new standards for heavy-duty gasoline vehicles (HDGV), we
have calculated the costs and tons reduced for these standards as well. We did this on a per-
vehicle basis, consistent with our approach for diesel engines described above. However, unlike
for our diesel engine standards, our HDGV standards are not associated with new gasoline
specifications, since a standard of 30 ppm sulfur has already been set in the preceding Tier
2/gasoline sulfur rulemaking.
1. Gasoline Vehicle Costs
The impact of our standards for HDGV was discussed in Section ni.B and the associated
compliance costs were discussed in Section V.B.5. We have made use of the per-vehicle costs
shown in Table V.B-5 in our cost-effectiveness analysis, assuming that near-term costs are
represented by the 2008-2009 values, and long-term costs are represented by the 2013+ values.
We weighted the costs for the incompletes, Class 2b completes, and Class 3 completes by their
VI-12
-------
Chapter VI: Cost-Effectiveness
respective contributions to the 2020 fleet (see Table VI.D-3). The fleet-average costs are
repeated in Table VI.D-1 below.
Table VI.D-1. Fleet-average, Per-vehicle Costs for HDGV Used in Cost-effectiveness
Near-term
Long-term
Total costs ($)
198
167
2. Emission Reductions from Gasoline Vehicles
The discounted lifetime tonnage numbers for NOx and NMHC for our HDGV standards
were based on the difference between emissions produced by vehicles meeting our 2005 and
2007 standards. Section II.B describes the base emission factors, conversions, and adjustments
used to calculate the in-use emissions in grams/mile produced by HDGVs for each year of a
vehicle's life. 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.
3) A seven percent annual discount factor, compounded from the first year of the engine's
life, was then applied for each year to allow calculation of net present value lifetime
emissions.
Converting to tons and summing across all years results in the total discounted lifetime
per-vehicle tons. This calculation can be described mathematically as follows:
LE = £ [{(AVMT); • (SURVIVE); • (ER); •(K)}/(1.07)M]
VI-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Where:
LE
(AVMT);
(SURVIVE);
(ER);
K
= Discounted lifetime emissions in tons/vehicle
= Annual miles traveled in year i of a HDGV's operational life
= Probability of survival after i years of service
= Emission rate, g/mi in year i of a vehicle's operational life
= Mass conversion factor, 1.102 x 10"6 tons/gram
= Vehicle years of operation, counting from 1 to 24
For NOx and NMHC, we generated discounted lifetime tonnage values for each vehicle
class (incompletes, Class 2B completes, and Class 3 completes) using the above equation. This
was done separately for the baseline and control cases. The baseline case included the 2005
model year vehicle standards, while the control case entailed our 2007 model year vehicle
standards. The tonnage values that we calculated according to this procedure are shown in Table
VI.D-2.
Table VI.D-2. Per-vehicle Discounted Lifetime Tons for HDGV
Baseline
Incompletes
Class 2B completes
Class 3 completes
Control
Incompletes
Class 2B completes
Class 3 completes
NOx + NMHC
0.261
0.271
0.269
0.170
0.166
0.192
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
HDGV fleet. These fractions were based on vehicle count projections for 2020 for gasoline-
powered heavy-duty vehicles, which in turn were based on current sales of new vehicles. Table
VI.D-3 presents the final weighting factors we used to develop fleet-average tonnage values.
VI-14
-------
Chapter VI: Cost-Effectiveness
Table VI.D-3. Vehicle Class Sales Weighting Factors for HDGV
Incompletes
Class 2B completes
Class 3 completes
0.288
0.692
0.020
The final discounted lifetime tonnage values for an average fleet engine meeting either the 2005
or 2007 standards are shown in Table VI.D-4. It is these values that were used directly in
calculating the cost-effectiveness of our program.
Table VI.D-4. Fleet average, Per-vehicle Discounted Lifetime Tons for HDGV
Baseline: 2004 standards
Control: 2007 standards
Baseline - control
NOx + NMHC
0.268
0.167
0.100
Note that although we are proposing new PM standards for HDGV in order to establish
consistency with the HDDE PM standards, current HDGV are believed to already meet this PM
standard. Therefore, there are no PM emission reductions associated with HDGV.
Since we are calculating a single set of cost-effectiveness values for both diesel engines
and gasoline vehicles, it was necessary for us to weight the costs and emission reductions for
HDDE and HDGV by the fraction of new diesel-powered and gasoline-powered heavy-duty
vehicles in the fleet. These fractions are based on current sales of new vehicles, or the
corresponding estimates of in-use vehicle counts far into the future. We have chosen 2020 to
represent the far future for the purposes of this analysis. The According to projections for
MOBILE6, in year 2020 the in-use heavy-duty fleet will be composed of approximately 50
percent diesel-powered and 50 percent gasoline-powered vehicles. We applied this weighting to
the NOx+NMHC costs from Tables VI.B-2 and VI.D-1 to obtain per-vehicle costs representing
all heavy-duty vehicles (PM reductions are only produced by our HDDE standards, so the PM
cost-effectiveness values represent only HDDE). We likewise applied the 50:50 weighting to the
NOx+NMHC tons reduced from Tables VI.C-3 and VI.D-4. Final costs and tons reduced for the
entire heavy-duty fleet on a per-vehicle basis are given in Table VI.F-1 below.
VI-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
E. 30-year Net Present Value Cost-Effectiveness
The per-vehicle approach described in the preceding sections is designed to show the
cost-effectiveness of our program for 2007 and later model year engines complying with our new
standards. It presumes that all delays and the Temporary Compliance Option have been
completed and the fleet has fully turned over to engines meeting our standards. However, the
per-vehicle approach does not account for costs and emission reductions associated with the
existing (pre-2007 model year) fleet due to operation on diesel fuel meeting our 15 ppm cap, nor
does it take into account phased-in engine or temporary fuel provisions at the start of the
program.
We have also calculated the cost effectiveness of our program using a "30-year net
present value" approach that includes all nationwide emission reductions and costs for a 30 year
period. This timeframe captures both the early period of the program when very few
vehicles/engines meeting our standards will be in the fleet, and the later period when essentially
all vehicles/engines in the fleet will meet our standards. The 30-year net present value approach
also accounts for cost and emission impacts of our 15 ppm sulfur cap on engines manufactured
before model year 2007. The 30-year net present value approach does have one important
drawback in that it includes the engine costs for engines sold 30 years after the program goes into
effect, but includes almost none of the emission benefits from those engines. Thus the 30-year
net present value approach does not necessarily match all costs with all the emission reductions
that those costs are intended to produce. It is presented here, nevertheless, as a reasonable
measure of the cost effectiveness of this combined vehicle-fuel program.
We have calculated this "30-year net present value" cost-effectiveness using the net
present value of the annual emission reductions and costs described in Sections n and V,
respectively. The calculation of 30-year net present value cost-effectiveness follows the pattern
described above for the per-engine analysis:
i-2006
Where:
DNAE = Reduction in nationwide 30-year net present value emissions in tons
(NE); = Reduction in nationwide emissions in tons for year i of the program
i = Year of the program, counting from 2006 to 2035
and
i-2006
VI- 16
-------
Chapter VI: Cost-Effectiveness
Where:
DNAC
(NC);
= Nationwide 30-year net present value costs in dollars
= Nationwide costs in dollars for year i of the program
= Year of the program, counting from 2006 to 2035
The 30-year net present value cost-effectiveness is produced by dividing DNAC by DNAE. The
nationwide reductions in emissions for each year are given in Section n. The nationwide costs
are given in Table V.D-1. The results are given in Table VI.F-2 below.
F. Results
We calculated the cost-effectiveness of our standards using two different approaches.
The first divides the total per-vehicle, discounted lifetime costs by the total per-vehicle,
discounted lifetime tons reduced for our HDDE standards, diesel sulfur standard, and HDGV
standards. The results are given in Table VI.F-1.
Table VI.F-1. Per-vehicle Cost-effectiveness of the Standards
Pollutants
Near-term costs
NOx + NMHC
PM
Long-term costs
NOx + NMHC
PM
Discounted
lifetime
vehicle &fuel
costs
1789
956
1365
762
Discounted
lifetime emission
reductions (tons)
0.8421
0.0672
0.8421
0.0672
Discounted
lifetime cost
effectiveness
per ton
2,125
14,237
1,621
11,340
Discounted lifetime
cost effectiveness per
ton with SO 2 credit*
2,125
7,599
1,621
4,701
* $446 credited to SO2 (at $4800/ton) for PM cost effectiveness
We also calculated the cost-effectiveness of our program on a 30-year net present value
basis for our diesel engine, diesel fuel sulfur, and gasoline vehicle standards. To do this, we
summed net present value of total costs from Section V.D, and divided by the sum of the net
present value of tons reduced from Sections II.B.2 and II.C. These costs and emission reductions
are repeated in Appendices VI-B and VI-C. The results are given in Table VI.F-2.
VI-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VI.F-2. 30-year Net Present Value Cost-effectiveness of the Standards
NOx + NMHC
PM
30-year n.p.v.
engine, vehicle,
&fuel costs
$34.9 billion
$10.3 billion
30-year
n.p.v.
reduction
(tons)
16.2 million
0.8 million
30-year n.p.v.
cost
effectiveness
per ton
$2,149
$13,607
3 0-year n.p.v. cost
effectiveness per ton
with SO 2 credit*
$2,149
$4,195
$7.1 billion credited to SO2 (at $4800/ton)
The values in Tables VI.F-1 and VI.F-2 differ from those in the NPRM for several
reasons. First, our estimate of costs increased for HDDE, HDGV, and diesel fuel sulfur as
described in SectionV. Second, the NMHC benefits associated with HDDE were reduced due to
our re-evaluation of the means through which manufacturers would meet our new standards, as
described in Section II.B. Third, our final program includes a phase-in for the engine standards
and a Temporary Compliance Option for the fuel sulfur standards, which reduced both the costs
and emission reductions in the first few years of the program.
Since many of the benefits and costs are manifest in future years, we apply discounting
methods to adjust the dollar values of these effects to reflect the finding that society as a whole
typically values the realization (or avoidance) of a given effect differently depending on when the
effect occurs. In the discounting calculations used to produce the net present values that were
used in our cost-effectiveness calculations, we used a discount rate of 7 percent, consistent with
the 7 percent rate reflected in the cost-effectiveness analyses for other recent mobile source
programs. OMB Circular A-94 requires us to generate benefit and cost estimates reflecting a 7
percent rate.
However, we anticipate that the primary cost and cost-effectiveness estimates for future
proposed mobile source programs will reflect a 3 percent rate. The 3 percent rate is in the 2 to 3
percent range recommended by the Science Advisory Board's Environmental Economics
Advisory Committee for use in EPA social benefit-cost analyses, a recommendation incorporated
in EPA's new Guidelines for Preparing Economic Analyses (November 2000). This
recommendation was published after the current program was proposed. Therefore, we have also
calculated the overall cost-effectiveness of today's rule based on a 3 percent rate to facilitate
comparison of the cost-effectiveness of this rule with future proposed rules which use the 3
percent rate. The results are shown in Tables VI.F-3 and VI.F-4.
VI-18
-------
Chapter VI: Cost-Effectiveness
Table VI.F-3. Per-vehicle Cost-effectiveness of the Standards Using 3 Percent ROI and
Discount Rate
Pollutants
Near-term costs
NOx + NMHC
PM
Long-term costs
NOx + NMHC
PM
Discounted
lifetime
vehicle &fuel
costs
1860
1008
1452
821
Discounted
lifetime emission
reductions (tons)
0.9961
0.0786
0.9961
0.0786
Discounted
lifetime cost
effectiveness
per ton
1,867
12,817
1,458
10,439
Discounted lifetime
cost effectiveness per
ton with SO 2 credit*
1,867
6,168
1,458
3,790
* $523 credited to SO2 (at $4800/ton) for PM cost effectiveness.
Table VI.F-4. 30-year Net Present Value Cost-effectiveness of the Standards Using 3
Percent ROI and Discount Rate
NOx + NMHC
PM
30-year n.p.v.
engine, vehicle,
&fuel costs
$54.6 billion
$16.0 billion
30-year
n.p.v.
reduction
(tons)
30.6 million
1.4 million
30-year n.p.v.
cost
effectiveness
per ton
$1,784
$11,791
3 0-year n.p.v. cost
effectiveness per ton
with SO 2 credit*
$1,784
$3,384
* $11.4 billion credited to SO2 (at $4800/ton)
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.F-1 and VI.F-2 and
the cost-effectiveness of other programs. Table VI.F-5 summarizes the cost effectiveness of
several recent EPA actions for controlled emissions from mobile sources for NOx and NMHC,
while Table VI.F-6 does the same for PM. The programs shown in these tables are those for
which cost-effectiveness was calculated in a similar manner allowing for an apples-to-apples
comparison.
VI-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VI.F-5. Cost-effectiveness of Previous Mobile Source Programs for NOx + NMHC
Program
Tier 2 vehicle/gasoline sulfur
2004 Highway HD diesel
Off-highway diesel engine
Tier 1 vehicle
NLEV
Marine SI engines
On-board diagnostics
Marine CI engines
$/ton
1,340-2,260
212-414
425 - 675
2,054 - 2,792
1,930
1,171 - 1,846
2,313
24 - 176
Note: costs adjusted to 1999 dollars.
Table VI.F-6. Cost-effectiveness of Previous Mobile Source Programs for PM
Program
Marine CI engines
1996 urban bus
Urban bus retrofit/rebuild
1994 highway HD diesel
$/ton
5222-3881
12,264 - 19,622
30,251
20,900 - 24,467
Note: costs adjusted to 1999 dollars.
By comparing values from Tables VI.F-1 and VI.F-2 for NOx+NMHC to those in Table
VI.F-5 we can see that the cost-effectiveness of our engine/diesel sulfur standards falls within the
range of these other programs. Our program overlaps the range of the recently promulgated
standards for Tier 2 light-duty vehicles and gasoline sulfur shown in Table VI.F-5. Our program
also overlaps the cost-effectiveness of past programs for PM. It is true that some previous
programs have been more cost efficient than the program we are proposing today. However, it
should be expected that the next generation of standards will be more expensive than the last,
since the least costly means for reducing emissions is generally pursued first.
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. Therefore, in evaluating the
cost-effectiveness of our engine/diesel sulfur program, we also considered whether our proposal
VI-20
-------
Chapter VI: Cost-Effectiveness
is cost-effective in comparison with potential future means of controlling emissions. In the
context of the Agency's rulemaking which would have revised the ozone and PM NAAQS2, the
Agency compiled a list of additional known technologies that could be considered in devising
new emission reductions strategies.5 Through this broad review, over 50 technologies were
identified that could reduce NOx, VOC, or PM. The cost-effectiveness of these technologies
averaged approximately $5,000/ton for VOC, $13,000/ton for NOx, and $40,000/ton for PM.
Although a $10,000/ton limit was actually used in the air quality analysis presented in the
NAAQS revisions rule, these values clearly indicate that, not only are future emission control
strategies likely to be more expensive (less cost-effective) than past strategies, but the cost-
effectiveness of our engine/diesel sulfur program falls within the range of potential future
strategies.
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 and PM emission
reductions indicates that our engine/diesel sulfur program 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.
b 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-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
APPENDIX VI - A: Factors Used in Diesel Engine Calculations for
Cost-effectiveness
MOBILE6
engine class
Class 2B
Class 3
Class 4
Class 5
Class 6
Class 7
Class 8A
Class 8B
School buses
Urban transit buses
Weight
category4
LH
LH
LH
LH
MH
MH
HH
HH
MH
HH
Sales Conversion factors,
weighting8 bhp-hr/miB
0.199
0.060
0.056
0.027
0.115
0.164
0.098
0.227
0.044
0.009
1.09
1.25
1.458
1.573
1.942
2.409
2.763
3.031
2.989
4.679
Fuel economy,
miles/galD
12.96
11.66
10.2
9.88
8.71
7.53
6.59
6.3
6.18
3.79
LH = Light heavy duty, MH = Medium heavy duty, HH = Heavy heavy duty.
B Based on 2020 heavy-duty diesel engine count, Tables 17 & 18 from EPA Report Number EPA420-P-99-
011, April 1999, "Fleet characterization data for MOBILE6: development and use of age distributions,
average annual mileage accumulation rates and projected vehicle counts for use in MOBILE6."
c Tables 28 and 30 from EPA Report Number EPA420-P-98-015, May 1998, "Update heavy-duty engine
emission conversion factors for MOBILE6: Analysis of BSFCs and calculation of heavy-duty engine
emission conversion factors."
D Tables 14 and 15 from EPA Report Number EPA420-P-98-014, May 1998, "Update heavy-duty engine
emission conversion factors forMOBILE6: Analysis of fuel economy, non-engine fuel economy
improvements, and fuel densities.
VI-22
-------
Chapter VI: Cost-Effectiveness
APPENDIX VI - B: Costs used in 30-year Net Present Value Cost
Effectiveness Analysis ($millions)
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
2035
Diesel NOx adsorber
+ oxv catalvst
(49)
765
798
648
918
740
684
699
713
689
698
700
714
728
741
753
766
778
789
801
812
823
834
844
855
865
876
886
896
906
Diesel PM
trap
(32)
501
523
425
602
485
449
458
467
452
458
459
468
477
486
494
502
510
518
525
532
540
547
554
561
567
574
581
588
594
Gasoline
vehicle
0
0
46
80
81
82
83
78
79
80
82
83
84
85
86
87
89
90
91
92
93
94
95
97
98
99
100
101
102
104
Diesel sulfur
880
1,786
1,809
1,904
2,014
2,128
2,160
2,192
2,225
2,258
2,292
2,327
2,362
2,397
2,433
2,469
2,506
2,544
2,582
2,621
2,660
2,700
2,741
2,782
2,824
2,866
2,909
2,953
2,997
3.042
VI-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
APPENDIX VI - C: Emission Reductions Used in 30-year Net
Present Value Cost Effectiveness Analysis (thousand tons)
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
2035
Diesel NOx
0
58
160
255
406
599
776
939
1,090
1,228
1,356
1,473
1,581
1,680
1,772
1,857
1,939
2,017
2,091
2,163
2,232
2,299
2,364
2,428
2,490
2,552
2,615
2,677
2,739
2.801
Diesel VOC
0
2
7
12
17
22
27
32
37
43
47
52
56
60
64
66
69
71
74
76
78
80
83
85
87
89
91
93
94
96
Diesel PM
5
11
19
27
35
41
46
51
56
61
65
69
73
77
82
85
88
91
93
96
99
101
104
106
109
111
113
116
118
120
Diesel SOx
78
79
80
82
107
109
111
113
115
117
119
121
122
124
126
128
129
131
133
134
136
137
139
140
142
143
144
146
147
149
Gasoline
NOx
0
0
2
7
13
19
24
29
34
38
43
48
51
55
58
63
67
70
72
75
78
80
83
86
88
91
94
96
99
102
Gasoline
VOC
0
0
1
2
5
7
9
10
12
13
15
16
18
20
21
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
VI-24
-------
Chapter VI: Cost-Effectiveness
Chapter VI. References
1. U.S. EPA; Review of NAAQS for Ozone, Assessment of Scientific and Technical
Information, Office of Air Quality Planning and Standards Staff Paper; document
number: EPA-452VR-96-007.
2. "Final Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway
Heavy-Duty Engines." September 16, 1997. Alan Stout, U.S. EPA, OAR/OMS/EPCD.
3. "A Review of Current and Historical Nonroad Diesel Fuel Sulfur Levels", Memorandum
from David J. Korotney to EPA Air Docket A-97-10, March 3, 1998, Docket Item
II-B-01.
4. Table 6, Agency Report Number EPA420-P-99-011, "Fleet characterization data for
MOBILE6: development and use of age distributions, average annual mileage
accumulation rates and projected vehicle counts for use in MOBILE6," April 1999.
5. "Regulatory Impact Analyses for the Particulate Matter and Ozone National Ambient Air
Quality Standards and Regional Haze Rule," Appendix B, "Summary of control measures
in the PM, regional haze, and ozone partial attainment analyses," Innovative Strategies
and Economics Group, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC, July 17, 1997.
VI-25
-------
Chapter VII: Benefit-Cost Analysis
Chapter VII: Benefit-Cost Analysis
This chapter reports EPA's analysis of the economic benefits of the final HD
Engine/Diesel Fuel rule. EPA is required by Executive Order 12866 to estimate the benefits of
major new pollution control regulations. Accordingly, the analysis presented here attempts to
answer three questions: 1) what are the physical health and welfare effects of changes in ambient
air quality resulting from reductions in nitrogen oxides (NOx), sulfur dioxide (SO2), non-
methane hydrocarbons (NMHC), carbon monoxide (CO) and direct diesel particulate matter
(PM) 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 the relative merits of this regulation.
The benefit-cost analysis 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 standards will have on the nationwide
inventories for NOx, non-methane hydrocarbons (NMHC), SO2, and PM
emissions;
2. Air quality modeling to determine the changes in ambient concentrations of ozone
and PM that will result from the changes in nationwide inventories of precursor
pollutants;
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; and
4. Comparison of the costs of the standards to the monetized benefits.
It is important to note that there are significant categories of benefits which can not be monetized
(or in many cases even quantified), resulting in a significant limitation to this analysis. Also,
EPA currently does not have appropriate tools for modeling changes in ambient concentrations of
CO or air toxics for input into a national benefits analysis. They have been linked to numerous
health effects; however, we are unable to quantify the CO- or air toxics-related health or welfare
benefits of the HD Engine/Diesel Fuel rule at this time.
EPA has used the best available information and tools of analysis to quantify the expected
changes in public health, environmental and economic benefits of the final HD Engine/Diesel
Fuel rule, given the constraints on time and resources available for the analysis. In general, we
vn-i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
follow the same general methodology used in the benefits analysis of the Tier 2/Gasoline Sulfur
rulemaking. However, we have updated some aspects of the analysis in response to public
comment and to reflect advances in modeling and the literature for economics and health effects.
EPA also relies heavily on the advice of its independent Science Advisory Board (SAB) in
determining the health and welfare effects considered in the benefits analysis and in establishing
the most scientifically valid measurement and valuation techniques. Since the publication of the
final Tier 2/Gasoline Sulfur RIA, we have updated some of the assumptions and methods used in
our analysis to reflect SAB recommendations. Changes to the methodology include the
following:
• Using Regulatory Model System for Aerosols and Deposition (REMSAD) to model
baseline and post-control ambient particulate matter;
• Updating concentration-response (C-R) functions for PM-related premature mortality;
• Updating C-R functions for PM-related hospital admissions;
• Presenting chronic asthma as an alternative calculation;
• Reporting asthma attacks as a separate endpoint and adjusting minor restricted activity
days to remove the possibility of double-counting of asthma attacks;
Relying only on the value of statistical life method to value reductions in the risk of
premature mortality in the primary estimate; and
• Adjusting benefits to reflect the expected growth in willingness-to-pay (WTP) for health
and environmental benefits as real income grows over time.
All of these changes are expected to improve the quality of the benefits estimation. These
changes reflect the latest peer-reviewed scientific literature and most of the improvements have
been approved by the SAB in its review of EPA methods in other analyses. A detailed discussion
of each change is included in the body of this chapter.
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 pay particular
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. Some welfare endpoints, for instance, are
quantified to some extent but no dollar value can be reliably assigned. The omission of these
items from the total of monetary benefits reflects our inability to measure them. It does not
vn-2
-------
Chapter VII: Benefit-Cost Analysis
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
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 on social welfare that result from the final rule. We include both benefits and
environmental costs in this analysis. 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 the value of changes in air quality 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 HD Engine/Diesel Fuel
rule occurring in 2030. The predicted emissions reductions that will result from the rule have yet
to occur, and therefore 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 HD Engine/Diesel Fuel 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 heavy duty vehicles
and fuels 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
required vehicle and fuel changes are then calculated, along with predictions of emissions for
other industrial sectors in the baseline. The predicted emissions described in Chapter in are then
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) used as the basis for this benefits analysis.
Changes in ambient concentrations will lead to new levels of environmental quality in the
U.S., reflected both in human health and in non-health welfare effects. Thus, the predicted
changes in ambient air quality serve as inputs into functions that 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
vn-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
changes in air quality as they related to ozone, PM, CO, and NMHC.a 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 changes in CO and NMHC are not
directly quantified for this analysis due to a lack of appropriate air quality models for these
pollutants. For changes in risks to human health from changes in 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
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 for (agriculture and forestry), 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. The majority of the
analytical assumptions used to develop our estimates have been reviewed and approved by the
SAB. However, like all estimates, they also contain unavoidable uncertainty, as does any
prediction of the future. In Section C and its subsections on health and welfare endpoints, this
uncertainty is discussed and characterized.
This chapter proceeds as follows: in Sections A and B, we summarize emissions and air
quality results and discuss the way emissions and air quality changes are used as inputs to the
benefits analysis. In Section C, we introduce the categories of benefits that are estimated, present
the techniques that are used, and provide 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 welfare effects and report the results of the
analysis for welfare effects. In Section F, we report our estimates of total monetized benefits and
alternative calculations of benefits. Finally, in Section G, we summarize annual cost results and
in Section H, we present a comparison of monetized benefits and costs.
a The NMHC listed in Table VII-1 are also listed as hazardous air pollutants in the Clean Air Act. We are
not able to quantify their direct effects. To the extent that they are precursors to ozone or PM, they are included in
our quantitative results.
vn-4
-------
Figure VII-1. Steps in the Heavy Duty Engine/Diesel Fuel Benefits Analysis
HD Engine/Diesel Fuel Standards
Evaluate Changes in Vehicles and
Fuels
I
Estimate Expected Reductions in
Pollutant Emissions
I
Model Changes in Ambient
Concentrations of Ozone and PM
Estimate Expected Changes in
Visibility, Agricultural Yields and
Other Welfare Effects
Estimate Expected Changes in
Human Health Outcomes
Estimate Changes in Monetary
Value of Visibility and Other
Welfare Effects
Estimate Monetary Value of
Changes in Human Health
Outcomes
Account for Income Growth
and Calculate Total Benefits
vn-s
-------
Table VII-1. Human Health and Welfare Effects of Pollutants Affected by the HD Engine/Diesel Fuel Rule
Pollutant/Effect
Primary Quantified and Monetized
Effects*
Alternative Quantified and/or
Monetized Effects8
Unquantified Effects
Ozone/Health
Minor restricted activity days
Hospital admissions - respiratory and
cardiovascular
Emergency room visits for asthma
Asthma attacks
Chronic Asthma0
Premature mortality13
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 (selected species)
Decreased Eastern commercial forest
productivity (selected
species)
Decreased Western commercial forest productivity
Decreased Eastern commercial forest productivity
(other species)
Decreased yields for fruits and vegetables
Decreased yields for other commercial and
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
Asthma attacks
Lower and upper respiratory illness
Minor restricted activity days
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
vn-6
-------
Pollutant/Effect
Primary Quantified and Monetized
Effects*
Alternative Quantified and/or
Monetized Effects8
Unquantified Effects
PM/Welfare
Visibility in California, Southwestern,
and Southeastern Class I areas
Visibility in Northeastern,
Northwestern, and Midwestern Class I
areas
Visibility in residential and non-Class I
areas
Household soiling
Nitrogen and
Sulfate
Deposition/
Welfare
Costs of nitrogen controls to reduce
eutrophication in selected
eastern estuaries
Impacts of acidic sulfate and nitrate deposition on
commercial forests
Impacts of acidic deposition on commercial
freshwater fishing
Impacts of acidic deposition on recreation in
terrestrial ecosystems
Impacts of nitrogen deposition on commercial
fishing, agriculture, and forests
Impacts of nitrogen deposition on recreation in
estuarine ecosystems
Reduced existence values for currently healthy
ecosystems
SO,/Health
Hospital admissions for respiratory and cardiac
diseases
Respiratory symptoms in asthmatics
NOx/Health
Lung irritation
Lowered resistance to respiratory infection
Hospital Admissions for respiratory and cardiac
diseases
vn-7
-------
Pollutant/Effect
Primary Quantified and Monetized
Effects*
Alternative Quantified and/or
Monetized Effects8
Unquantified Effects
CO/Health
Premature mortality8
Behavioral effects
Hospital admissions - respiratory, cardiovascular,
and other
Other cardiovascular effects
Developmental effects
Decreased time to onset of angina
Non-asthma respiratory ER visits
NMHCs E
Health
Cancer (diesel PM, 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 mucous membranes
(formaldehyde)
Respiratory and respiratory tract
Asthma attacks in asthmatics (formaldehyde)
Asthma-like symptoms in non-asthmatics
(formaldehyde)
Irritation of the eyes, skin, and respiratory tract
(acetaldehyde)
Upper respiratory tract irritation & congestion
(acrolein)
vn-s
-------
Pollutant/Effect
NMHCs E
Welfare
Primary Quantified and Monetized
Effects*
Alternative Quantified and/or
Monetized Effects8
Unquantified Effects
Direct toxic effects to animals
Bioaccumlation in the food chain
A Primary quantified and monetized effects are those included when determining the primary estimate of total monetized benefits of the HD Engine/Diesel Fuel rule. See Section
C-2 for a more complete discussion of presentation of benefits estimates.
B Alternative quantified and/or monetized effects are those presented as alternatives to the primary estimates or in addition to the primary estimates, but not included in the
primary estimate of total monetized benefits.
c While no causal mechanism has been identified linking new incidences of chronic asthma to ozone exposure, an epidemiological study shows a statistical association between
long-term exposure to ozone and incidences of chronic asthma in some non-smoking men (McDonnell, et al., 1999).
D Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the American Cancer Society (ACS)/ Krewski, et al., 2000 C-R function
we use for premature mortality captures both PM mortality benefits and any mortality benefits associated with other air pollutants (ACS/ Krewski, et al., 2000).
E All non-methane hydrocarbons (NMHCs) listed in the table are also hazardous air pollutants listed in the Clean Air Act.
vn-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
A. Emissions Inventory Implications
This section explains why 2030 emission inventories were developed and what the
possible implications are for this benefit-cost analysis of uncertainties associated with the
inventories. The national inventories for NOx, NMHC, SO2, and PM have already been
presented and discussed in Chapter in 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 HD Engine/Diesel Fuel program has various cost and emission related components,
as described earlier in this RIA. These components would begin at various times and in some
cases would phase in over time. This means that during the early years of the program there
would not be a consistent match between cost and benefits. This is especially true for the vehicle
control portions and initial fuel changes required by the program, where the full vehicle cost
would be incurred at the time of vehicle purchase, while the fuel cost along with the emission
reductions and benefits resulting from all these costs would 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 program, our analysis uses a future year when the fleet is nearly fully
turned over (2030). Consequently, we developed emission inventories for 2030 baseline
conditions and a 2030 HD Engine/Diesel Fuel control scenario.
In the years before 2030, the benefits from the HD Engine/Diesel Fuel program will be
less than those estimated here, because the compliant heavy-duty fleet will not be fully phased in.
Moreover, to the extent that a lower ratio of benefits to costs early in the program is the result of
the mismatch of costs and benefits in time, a simple analysis of an individual year would be
misleading. A more appropriate means of capturing the impacts of timing differences in benefits
and costs would be to produce a net present value comparison of the costs and benefits over
some period of years. Unfortunately, while this is relatively straight-forward for the costs, it is
currently not feasible to do a multi-year analysis of the benefits as this would require a significant
amount of air quality modeling to capture each year. We did not have the resources for such an
extensive analysis. Instead, for the purpose of the benefit calculations, we assume that 2030 is a
representative year for the fully implemented rule to consider in comparison with the costs. The
resulting analysis represents a snapshot of benefits and costs in a future year in which the heavy
duty fleet consists almost entirely of vehicles and fuels meeting the HD Engine/Diesel Fuel
standards.
In addition, there is uncertainty in any prediction, and the emissions inventory growth
factors can add uncertainty because they are applied for a 30-year period and propagate through
the entire analysis. This uncertainty may be more important for welfare effects such as ozone-
related crop damage where the predicted 2030 baseline may be an important factor. These
VII-10
-------
Chapter VII: Benefit-Cost Analysis
exposure metrics for crop damage and forestry impacts are a cumulative measure above a certain
level (i.e., 0.06 ppm). Thus, the accuracy of the emissions inventory growth rates can affect the
magnitude of the benefits (For discussion see Section E). This is less of an issue for exposure
metrics that rely on changes in air quality (e.g., the health endpoints). Nevertheless, the
inventory is a crucial building block on which the analysis rests.
B. Air Quality Impacts
This section summarizes the methods for and results of estimating air quality for the 2030
base case and HD Engine/Diesel Fuel control scenario for the purposes of benefit-cost analyses.
EPA has focused on the health, welfare, and ecological effects that have been linked to air quality
changes. These air quality changes include the following:
• Ambient ozone-as estimated using a regional-scale version of the Urban Airshed
Model-Variable Grid (UAM-V);
Ambient paniculate matter (PM10 and PM2 5)-as estimated using a national-scale
version of the Regulatory Modeling System for Aerosols and Deposition
(REMSAD);
Visibility degradation (i.e., regional haze), as developed using empirical estimates
of light extinction coefficients and efficiencies in combination with REMSAD
modeled reductions in pollutant concentrations; and
Airborne nitrogen deposition to estuaries-as predicted using local and regional
coefficients of nitrogen deposition for selected estuaries from the Regional Acid
Deposition Model (RADM) in combination with modeled reductions in NOx
emissions.
The air quality estimates in this section are based on the emission changes discussed in Chapter
in. These air quality results are in turn associated with human populations and ecosystems to
estimate changes in health and welfare effects.
In Section B-l, we describe the estimation of ozone air quality using UAM-V, and in
Section B-2, we cover the estimation of PM air quality using REMSAD. In Section B-3, we
discuss the estimation of visibility degradation. Lastly, in Section B-4 we describe the
estimation of nitrogen deposition.
VII-11
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1. Ozone Air Quality Estimates
We use the emissions inputs described in Section A with a regional-scale version of
UAM-V to estimate ozone air quality in the Eastern U.S. 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 HD Engine/Diesel Fuel rule on U.S. ozone concentrations. As
described fully in the air quality technical support document, the model performance in the
Western U.S. was not acceptable for including those results as inputs to the benefits analysis (US
EPA, 2000). Comparisons of base year model output data against ambient observations in the
Western U.S. indicated that the model was significantly underestimating (by 30-50 percent) the
observed levels of ozone in most areas of the West. Given that model performance was degraded
to the extent that the directional response of the model to controls may be questionable, it was
determined that this application of the model should not be used in assessing the impacts of the
emissions control strategy in the Western U.S.
Thus, our analysis applies the modeling system to the Eastern U.S. for a base-year of
1996 and for two future-year scenarios: a 2030 base case and a 2030 HD Engine/Diesel Fuel
control scenario. As discussed in the technical support document, we use the two separate years
because the relative model predictions are used with ambient air quality observations from 1996
to determine the expected change in 2030 ozone concentrations due to the rule (Abt Associates,
2000).
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. As applied to the Eastern region of the continental U.S., the model segments the
area within the 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 to be used as inputs to the health and welfare 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.
VII-12
-------
Chapter VII: Benefit-Cost Analysis
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" (US 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 nine vertical layers. The unshaded
area of Figure VII-2 has less horizontal resolution, as it uses a 36 km grid with the same nine
vertical layers. The vertical height of the modeling domain is 4,000 meters above ground level,
for both the shaded and unshaded regions. The split between Eastern and Western counties is
made at the 100th degree longitude (which runs through North and South Dakota, Nebraska,
Kansas, Oklahoma, and Texas).
b.
Simulation Periods
For use in this benefits analysis, the simulation periods modeled by UAM-V included
several multi-day periods when ambient measurements recorded high ozone concentrations. A
simulation period, or episode, consists of meteorological data characterized over a block of days
Figure VII-2. UAM-V Modeling Domain for Eastern U.S.
Note: The shaded section represents fine grid modeling (12 km) and the other portions represent coarse
grid modeling (36 km).
VII-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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. We focused on the summer of 1995 for selecting the episodes to model in the East
because 1995 is a recent time period for which we had model-ready meteorological inputs and
this timeframe contained several periods of elevated ozone over the Eastern U.S. As detailed in
the technical support document for this modeling, this analysis used three multi-day simulation
periods to prepare the future-year ozone profiles for the Eastern U.S. ozone analysis: June 12-24,
July 5-15, and August 7-21, 1995 (US EPA, 2000). These episodes include a three day "ramp-
up" period to initialize the model, but the results for these days are not used in this analysis.
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 obtained from the
Aerometric Information Retrieval System (AIRS) to generate ozone concentrations for the entire
ozone season.b'c The predicted changes in ozone concentrations from the 2030 base case to 2030
HD Engine/Diesel Fuel control scenario serve as inputs to the health and welfare 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 contiguous U.S., full-
season ozone data are 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 8 km by 8 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. d e For the analysis of ozone impacts on agriculture and commercial forestry, we
use a similar approach except air quality is interpolated to county centroids as opposed to
b 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.
0 Based 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.
d The 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.
e This approach is a generalization of planar interpolation that is technically referred to as enhanced
Voronoi Neighbor Averaging (EVNA) spatial interpolation (See Abt Associates (2000) for a more detailed
description).
VII-14
-------
Chapter VII: Benefit-Cost Analysis
population grid-cells. We report ozone concentrations as a cumulative index called the SUM06.
The SUM06 is the sum of the ozone concentrations for every hour that exceeds 0.06 parts per
million (ppm) within a 12-hour period from 8 am to 8 pm in the months of May to September.
These methods are described in detail in the technical support document to the RIA (Abt
Associates, 2000).
d. Ozone Air Quality Results
This section provides a summary the predicted ambient ozone concentrations from the
UAM-V model for the 2030 base case and changes associated with the HD Engine control
scenario. In Table VII-2, we provide those ozone metrics for grid-cells in the Eastern U.S. that
enter the concentration response functions for health benefits endpoints. In addition to the
standard frequency statistics (e.g., minimum, maximum, average, median), Table VII-2 provides
the population-weighted average which better reflects the baseline levels and predicted changes
for more populated areas of the nation. This measure, therefore, will better reflect the potential
benefits of these predicted changes through exposure changes to these populations. As shown,
the rule results in reductions between 3 and 5 percent, or between 0.8 to 1.7 ppb, in the daily and
seasonal average ozone concentrations across Eastern U.S. population grid-cells. A similar
relative decline is predicted for the population-weighted average, which indicates rather uniform
reductions in these concentrations across urban and rural areas. Additionally, the daily maximum
ozone concentrations are predicted to decline between 3.5 and 5 percent, or in the neighborhood
of 1.5 ppb.
In Table VII-3, we provide the seasonal SUM06 ozone metric for counties in the Eastern
U.S. that enters the concentration response function for agriculture benefit end-points. This
metric is a cumulative threshold measure so that the increase in baseline NOx emissions from
Tier 2 post-control to this rulemaking have resulted in a larger number of rural counties
exceeding the hourly 0.06 ppm threshold. As a result, changes in ozone concentrations for these
counties are contributing to greater impacts of the HD Engine/Diesel Fuel rule on the seasonal
SUM06 ozone metric. Table "VTI-3 indicates that the average across all Eastern U.S. counties
declined by 78 percent, or almost 17 ppb. Similarly high percentage reductions are observed
across the other points on the distribution with the maximum declining by almost 30 ppb, or 55
percent, and the median declining by almost 20 ppb, or 83 percent.
An important factor to consider when interpreting the ozone air quality results presented
here is the omission of changes in the Western U.S. Over 22 percent of national NOx emission
reductions occur in the Western U.S., with over 10 percent of total NOx emissions occurring in
California alone. This suggests that ozone changes in the West may be substantial, and that our
estimate of Eastern ozone changes may underestimate populations across the nation that will
experience ozone-related benefits of the HD Engine/Diesel Fuel NOx reductions.
VII-15
-------
Table VII-2. Summary of UAM-V Derived Ozone Air Quality Metrics Due to HD
Engine/Diesel Fuel Standards for Health Benefits EndPoints: Eastern U.S.
Statistic A
2030 Base Case
Change B
Percent Change B
Seasonal Average 8-Hour Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average D
17.60
81.80
34.93
34.90
37.76
-1.20
-3.20
-1.64
-1.67
-1.43
-6.82%
-3.91%
-4.65%
-4.78%
-3.88%
Daily 1-Hour Maximum Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average D
22.11
108.27
44.15
43.94
49.69
-1.37
-3.66
-1.68
-1.57
-1.71
-6.20%
-3.38%
-3.81%
-3.57%
-3.44%
Daily 5 -Hour Maximum Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average D
18.21
84.43
34.96
34.98
37.69
-1.32
-3.27
-1.64
-2.13
-1.43
-7.25%
-3.87%
-4.69%
-6.09%
-3.79%
Daily 24-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average D
11.43
47.71
28.30
28.40
28.76
-0.59
-1.60
-0.82
-0.86
-0.72
-5.16%
-3.35%
-2.90%
-3.03%
-2.50%
Daily 12-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average D
16.49
75.90
34.46
34.52
36.97
-1.10
-2.89
-1.53
-1.13
-1.35
-6.67%
-3.81%
-4.44%
-3.27%
-3.65%
A These ozone metrics are calculated at the CAPMS grid-cell level for use in health effects estimates based on the results of enhanced spatial
interpolation. Except for the daily 24-hour average, these ozone metrics are calculated over relevant time periods during the daylight hours (7
am to 7 pm) of the "ozone season," i.e., May through September. For the 8-hour average, the relevant time period is 9 am to 5 pm, and, for
the 5-hour maximum, it is 10 am to 3 pm.
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," and then multiplied by 100 to convert the value to a percentage.
c The base case minimum (maximum) is the value for the CAPMS grid cell with the lowest (highest) value.
D Calculated by summing the product of the projected 2030 CAPMS grid-cell population and the estimated 2030 CAPMS grid-cell seasonal
ozone concentration, and then dividing by the total population. The resulting value is then multiplied by 100 to convert the value to a
percentage.
VII-16
-------
Chapter VII: Benefit-Cost Analysis
Table VII-3. Summary of UAM-V Derived Ozone Air Quality Metrics Due to HD
Engine/Diesel Fuel Standards for Welfare Benefits Endpoints: Eastern U.S.
Statistic A
Sum06 (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average D
2030 Base Case
0.00
53.36
21.66
23.44
23.19
Change B
0.00
-29.10
-16.91
-19.50
-11.19
Percent Change B
0.00%
-54.54%
-78.05%
-83.19%
-48.26%
A SUM06 is defined as the cumulative sum of hourly ozone concentrations over 0.06 ppm (or 60 ppb) that occur during daylight hours (from
Sam to 8pm) in the months of May through September. It is calculated at the county level for use in agricultural benefits based on the results
of enhanced spatial interpolation.
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," which is then multiplied by 100 to convert the value to a percentage.
c The base case minimum (maximum) is the value for the county level observation with the lowest (highest) concentration.
D Calculated by summing the product of the projected 2030 county population and the estimated 2030 county level ozone concentration, and
then dividing by the total population. The resulting value is then multiplied by 100 to convert the value to a percentage.
2. PM Air Quality Estimates
We use the previously described emissions inputs with a national-scale version of the
Regulatory Model System for Aerosols and Deposition (REMSAD) to estimate PM air quality in
the contiguous U.S. REMSAD was developed as an extension of the episodic UAM-V regional
model. Like UAM-V, REMSAD is a three-dimensional grid-based Eulerian air quality model
designed to estimate annual particulate concentrations and deposition over large spatial scales
(e.g., over the contiguous U.S.). Consideration of the different processes that affect primary
(directly emitted) and secondary (formed by atmospheric processes) PM at the regional scale in
different locations is fundamental to understanding and assessing the effects of proposed
pollution control measures that affect ozone, PM and deposition of pollutants to the surface/
Because it accounts for spatial and temporal variations as well as differences in the reactivity of
emissions, REMSAD is useful for evaluating the impacts of the HD Engine/Diesel Fuel rule on
U.S. PM concentrations. Our analysis applies the modeling system to the entire U.S. for a base-
year of 1996 and for two future-year scenarios: a 2030 base case and a 2030 HD Engine/Diesel
Fuel control scenario.
f Given the potential impact of the HD Engine/Diesel Fuel rule on secondarily formed particles it is
important to employ a Eulerian model such as REMSAD. The impact of secondarily formed pollutants typically
involves primary precursor emissions from a multitude of widely dispersed sources, and chemical and physical
processes of pollutants that are best addressed using an air quality model that employs an Eulerian grid model
design.
VII-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
REMSAD was peer-reviewed in 1999 for EPA as reported in "Scientific Peer-Review of
the Regulatory Modeling System for Aerosols and Deposition." Earlier versions of REMSAD
have been employed for the EPA's Prospective 812 Report to Congress and for EPA's Analysis
of the Acid Deposition and Ozone Control Act (Senate Bill 172). Version 4.1 of REMSAD was
employed for this analysis and is fully described in the air quality technical support documents
(US EPA, 2000).
REMSAD simulates every hour of every day of the year and, thus, requires a variety of
input files that contain information pertaining to the modeling domain and simulation period.
These include gridded, 3-hour average emissions estimates and meteorological fields, initial and
boundary conditions, and land-use information. As applied to the contiguous U.S., the model
segments the area within the region into square blocks called grids (roughly equal in size to
counties), each of which has several layers of air conditions. Using this data, REMSAD
generates predictions of 3-hour average PM concentrations for every grid. We then calculated
daily and seasonal PM air quality metrics as inputs to the health and welfare 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
As shown in Figure VH-3, the modeling domain encompasses the contiguous 48 States.
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 model contains horizontal grid-cells
across the model domain of roughly 36 km by 36 km. There are 8 vertical layers of atmospheric
conditions with the top of the modeling domain at roughly 16,000 meters. The 36 by 36 km
horizontal grid results in a 120 by 92 grid (or 10,080 grid-cells) for each vertical layer. Figure
VII-4 illustrates the horizontal grid-cells for Maryland and surrounding areas.
b. Simulation Periods
For use in this benefits analysis, the simulation periods modeled by REMSAD included
separate full-year application for 2030 base case and control scenarios with emissions inventories
described in Chapter HI.
c. Model Inputs
REMSAD requires a variety of input files that contain information pertaining to the
modeling domain and simulation period. These include gridded, 3-hour average emissions
estimates and meteorological fields, initial and boundary conditions, and land-use information.
Separate emissions inventories were prepared for the 1996 base-year and each of the 2030 future-
year base case and control scenarios. All other inputs were specified for the 1996 base-year
model application and remained unchanged for each future-year modeling scenario.
VII-18
-------
Chapter VII: Benefit-Cost Analysis
Figure VII-3. REMSAD Modeling Domain for Continental U.S.
Note: Gray markings define individual grid-cells in the REMSAD model.
VII-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Similar to UAM-V, REMSAD requires detailed emissions inventories containing
42
S3
104
Figure VII-4. Example of REMSAD 36 x 36km Grid-cells for Maryland Area
Note: Gray markings define individual grid-cells in the REMSAD model.
temporally allocated emissions for each grid-cell in the modeling domain for each species being
simulated. The previously described annual emission inventories reflecting 2030 base case and
control scenarios were preprocessed into model-ready inputs through the Emissions
Preprocessing System, Version 2.5 (EPS2.5). The core of EPS2.5 is a series of FORTRAN
modules that incorporate spatial, temporal, and chemical resolution into an emissions inventory
for use in a photochemical model. Meteorological inputs reflecting 1996 conditions across the
contiguous U.S. were derived from Version 5 of the Mesoscale Model (MM5). These inputs
included horizontal wind components (i.e., speed and direction), temperature, moisture, vertical
diffusion rates, and rainfall rates for each grid cell in each vertical layer. Details of the annual
1996 MM5 modeling are provided in MCNC (2000).
Initial species concentrations and lateral boundary conditions were specified to
approximate background concentrations of the species; for the lateral boundaries the
concentrations varied (decreased parabolically) with height. These background concentrations
are provided in the air quality modeling TSD (US EPA, 2000a). Land use information was
obtained from the U.S. Geological Survey database at 10 km resolution.
VII-20
-------
Chapter VII: Benefit-Cost Analysis
d. Converting REMSAD Outputs to Benefits Inputs and Model Performance
REMSAD generates predictions of hourly PM concentrations for every grid. The
participate matter species modeled by REMSAD include a primary coarse fraction
(corresponding to PM in the 2.5 to 10 micron size range), a primary fine fraction (corresponding
to PM less than 2.5 microns in diameter), and several secondary particles (e.g., sulfates, nitrates,
and organics). PM25 is calculated as the sum of the primary fine fraction and all of the
secondarily-formed particles. These hourly predictions form the basis for direct calculation of
daily and annual PM air quality metrics (i.e., annual mean PM concentration) as inputs to the
health and welfare C-R functions of the benefits analysis. In addition, the speciated predictions
from REMSAD are employed as inputs to a post-processing module that estimates atmospheric
visibility, as discussed later in Section B-3 (US EPA, 2000a).
We modeled 1996 and 2030 base and HD Engine/Diesel Fuel control scenarios. The
2030 modeling is used in this benefits assessment. The goal of the 1996 base year modeling was
to reproduce the atmospheric processes resulting in formation and dispersion of PM25 across the
U.S. and to evaluate operational model performance for PM25 and its related speciated
components (e.g., sulfate, nitrate, elemental carbon) in order to estimate the ability of the
modeling system to replicate base year concentrations.
This evaluation is comprised principally of statistical assessments of model versus
observed pairs. The robustness of any evaluation is directly proportional to the amount and
quality of the ambient data available for comparison. Unfortunately, there are few PM25
monitoring networks with available data for evaluation of the HD Engine/Diesel Fuel PM
modeling. Critical limitations of the existing databases are a lack of urban monitoring sites with
speciated measurements and poor geographic representation of ambient concentration in the East.
The largest available ambient database for 1996 comes from the Interagency Monitoring of
PROtected Visual Environments (IMPROVE) network. IMPROVE is a cooperative visibility
monitoring effort between EPA, federal land management agencies, and state air agencies. Data
is collected at Class I areas across the United States mostly at National Parks, National
Wilderness Areas, and other protected pristine areas (IMPROVE 2000). There were
approximately 60 IMPROVE sites across the nation that had complete annual data in 1996.
Forty two of these sites were in the Western U.S. and 18 sites were in the Eastern U.S.
A comparison of predicted versus observed annual average PM25 concentrations at the
IMPROVE sites indicates that PM2 5 is underpredicted by about 25% on a nationwide aggregated
basis. Most of the underprediction occurs at the Western sites where the overall underprediction
is about 35%. However, in the East, ambient PM2 5 is overpredicted by about 10%. In addition,
model performance was examined for the five component species of PM25 (sulfate, nitrate,
elemental carbon, organic carbon, and other (crustal) fine PM. The results indicate that the
performance for both sulfate and elemental carbon was similar to that of PM25. That is, sulfate
and elemental carbon were slightly overpredicted in the East and slightly underpredicted in the
West. The performance for nitrate, crustal PM, and organic aerosols was not as good as the
performance for the other species. Specifically, nitrate and crustal PM were overpredicted in the
East, and organic carbon was underpredicted domainwide.
VII-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
It should be noted that PM2 5 modeling is an evolving science. There have been few
regional or national scale model applications for primary and secondary PM. In fact, this is one
of the first nationwide applications of a full chemistry Eulerian grid model for the purpose of
estimating annual average concentrations of PM25 and its component species. Also, unlike ozone
modeling, there is essentially no database of past performance statistics against which to measure
the performance of the HD Engine/Diesel Fuel PM modeling. Given the state of the science
relative to PM modeling, it is inappropriate to judge PM model performance using criteria
derived for other pollutants, like ozone. Still, the performance of the HD Engine/Diesel Fuel PM
modeling is very encouraging, especially considering that the results may be limited by our
current knowledge of PM science and chemistry, and by the emissions inventories for primary
PM and secondary PM precursor pollutants. Further details of the model performance for PM
can be found in the air quality modeling Technical Support Document (US EPA 2000).
e. PM Air Quality Results
Table VII-4 provides a summary of the predicted ambient PM10 and PM2 5 concentrations
from REMSAD for the 2030 base case and changes associated with HD Engine/Diesel Fuel
control scenario. The REMSAD results indicate that the predicted change in PM concentrations
is composed almost entirely of reductions in fine particulates (PM25) with little or no reduction in
coarse particles (PM10less PM25). Therefore, the observed changes in PM10are composed
primarily of changes in PM2 5. In addition to the standard frequency statistics (e.g., minimum,
maximum, average, median), Table VII-4 provides the population-weighted average which better
reflects the baseline levels and predicted changes for more populated areas of the nation. This
measure, therefore, will better reflect the potential benefits of these predicted changes through
exposure changes to these populations. As shown, the average annual mean concentrations of
PM25 across all U.S. grid-cells declines by roughly 3.1 percent, or 0.27 //g/m3. The population-
weighted average mean concentration declined by 4.4 percent, or 0.65 //g/m3, which is much
larger in absolute terms than the spatial average. This indicates the HD Engine/Diesel Fuel rule
generates greater absolute air quality improvements in more populated, urban areas.
Table VII-5 provides information on the 2030 populations that will experience improved
PM air quality. There are significant populations that live in areas with meaningful reductions in
annual mean PM2 5 concentrations resulting from the HD Engine/Diesel Fuel rule. As shown,
just over 15 percent of the 2030 U.S. population are predicted to experience reductions of greater
than 1 //g/m3. Furthermore, almost 33 percent of the 2030 U.S. population will benefit from
reductions in annual mean PM25 concentrations of greater than 0.75 //g/m3 and slightly over 60
percent will live in areas with reductions of greater than 0.5 //g/m3. This information indicates
how widespread the improvements in PM air quality are expected to be and the large populations
that will benefit from these improvements.
VII-22
-------
Chapter VII: Benefit-Cost Analysis
Table VII-4. Summary of 2030 Base Case PM Air Quality and Changes Due to HD
Engine/Diesel Fuel Standards
Statistic
PM10
Minimum Annual Mean (,ug/m3) B
Maximum Annual Mean (,ug/m3) B
Average Annual Mean (,ug/m3)
Median Annual Mean (,ug/m3)
Population- Weighted Average Annual Mean (,ug/m3) c
2030 Base Case
1.52
65.68
10.31
8.15
21.70
Change^
-0.03
-1.39
-0.28
-0.18
-0.66
Percent Change
-2.0%
-2.1%
-2.4%
-2.3%
-3.1%
PM25
Minimum Annual Mean (,ug/m3) B
Maximum Annual Mean (,ug/m3) B
Average Annual Mean (,ug/m3)
Median Annual Mean (,ug/m3)
Population- Weighted Average Annual Mean (,ug/m3) c
1.19
39.55
7.87
5.96
14.85
-0.03
-1.35
-0.27
-0.17
-0.65
-2.4%
-3.4%
-3.1%
-3.0%
-4.4%
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 populated grid-cell with the lowest (highest) annual average. The change relative to
the base case is the observed change for the populated grid-cell with the lowest (highest) annual average in the base case.
c Calculated by summing the product of the projected 2030 grid-cell population and the estimated 2030 PM concentration, for that grid-cell
and then dividing by the total population in the 48 contiguous States.
VII-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-5. Distribution of PM2 5 Air Quality Improvements Over 2030 Population Due to
HD Engine/Diesel Fuel Standards
Change in Annual Mean PM25 Concentrations
(tUg/m1)
0 > A PM25 Cone < 0.25
0.25>APM25Conc < 0.5
0.5>APM25Conc < 0.75
Q.15> APM25Conc < 1.0
1.0> APM25Conc < 1.25
\.25>APM25Conc < 1.5
\.5>APM25Conc < 1.75
APM25Conc>\.75
2030 Population
Number (millions) Percent (%)
43.0
95.0
94.9
60.5
23.4
20.9
2.9
5.2
11.2%
27.5%
27.5%
17.5%
6.8%
6.1%
0.9%
1.5%
The change is defined as the control case value minus the base case value.
Table VII-6 provides additional insights on the changes in PM air quality resulting from
the HD Engine/Diesel Fuel standards. The information presented previously in Table VII-4
illustrated the absolute and relative changes for different points along the distribution of baseline
2030 PM concentration levels, e.g., the change reflects the lowering of the minimum predicted
baseline concentration rather than the minimum predicted change for 2030. The latter is the
focus of Table VII-6 as it presents the distribution of predicted changes in both absolute terms
(i.e., //g/m3) and relative terms (i.e., percent) across individual grid-cells. As shown, the absolute
reduction in annual mean PM10 concentration ranged from a low of 0.02 //g/m3 to a high of 2.18
//g/m3, while the relative reduction ranged from a low of 0.2 percent to a high of 9.9 percent.
Alternatively, for mean PM25, the absolute reduction ranged from 0.02 to 2.13 //g/m3, while the
relative reduction ranged from 0.4 to 13.1 percent.
VII-24
-------
Chapter VII: Benefit-Cost Analysis
Table VII-6. Summary of Absolute and Relative Changes in PM Air Quality Due to HD
Engine/Diesel Fuel Standards
Statistic
PM10 Annual Mean
PM25 Annual Mean
Absolute Change from 2030 Base Case (^g/m3)A
Minimum
Maximum
Average
Median
Population- Weighted Average c
Relative Change from 2030 Base Case (%)B
Minimum
Maximum
Average
Median
Population- Weighted Average c
-0.02
-2.18
-0.28
-0.18
-0.66
-0.17%
-9.87%
-2.35%
-2.29%
-3.08%
-0.02
-2.13
-0.27
-0.17
-0.65
-0.38%
-13.14%
-3.08%
-2.97%
-4.36%
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 gridcell. The information reported in this section does not necessarily reflect the same gridcell as is portrayed in
the absolute change section.
c Calculated by summing the product of the projected 2030 gridcell population and the estimated 2030 gridcell 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,
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. The higher the deciview value, the worse the visibility. Thus, an improvement in
visibility is a decrease in deciview value.
VII-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-7 provides the distribution of visibility improvements across 2030 population
resulting from the HD Engine/Diesel Fuel rule. The majority of the 2030 U.S. population live in
areas with predicted improvement in annual average visibility of between 0.4 to 0.6 deciviews
resulting from the HD Engine/Diesel Fuel rule. As shown, almost 20 percent of the 2030 U.S.
population are predicted to experience improved annual average visibility of greater than 0.6
deciviews. Furthermore, roughly 70 percent of the 2030 U.S. population will benefit from
reductions in annual average visibility of greater than 0.4 deciviews. The information provided
in Table "VTi-7 indicates how widespread the improvements in visibility are expected to be and
the share of populations that will benefit from these improvements.
Because the visibility benefits analysis 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 estimated using outputs
from REMSAD for the 2030 base case and FID Engine/Diesel Fuel control scenarios.
Table VII-7. Distribution of Populations Experiencing Visibility Improvements in 2030
Due to HD Engine/Diesel Fuel Standards
Improvements in Visibility A
(annual average deciviews)
0 > A Deciview < 0.2
0.2 > A Deciview < 0.4
0.4 > A Deciview < 0.6
0.6> A Deciview < 0.8
0.8> A Deciview < 1.0
A Deciview > 1.0
2030 Population
Number (millions) Percent (%)
12.1
87.4
179.7
54.5
10.7
1.5
3.5%
25.3%
51.9%
15.8%
3.1%
0.4%
L The change is defined as the control case deciview level minus the base case deciview level.
a. Residential Visibility Improvements
Air quality modeling results predict that the FID Engine/Diesel Fuel rule will create
improvements in visibility through the country. In Table VII-8, we summarize residential
visibility improvements across the Eastern and Western U.S. in 2030. The baseline annual
average visibility for all U.S. counties is 14.8 deciviews. The mean improvement across all U.S.
counties is 0.28 deciviews, or almost 2 percent. In urban areas with a population of 250,000 or
more (i.e., 1,209 out of 5,147 counties), the mean improvement in annual visibility was 0.39
VII-26
-------
Chapter VII: Benefit-Cost Analysis
deciviews and ranged from 0.05 to 1.08 deciviews. In rural areas (i.e., 3,938 counties), the mean
improvement in visibility was 0.25 deciviews in 2030 and ranged from 0.02 to 0.94 deciviews.
On average, the Eastern U.S. experienced slightly larger absolute but smaller relative
improvements in visibility than the Western U.S. from the HD Engine/Diesel Fuel reductions. In
Eastern U.S., the mean improvement was 0.34 deciviews from an average baseline of 19.32
deciviews. Western counties experienced a mean improvement of 0.21 deciviews from an
average baseline of 9.75 deciviews projected in 2030. Overall, the data suggest that the HD
Engine/Diesel Fuel rule has the potential to provide widespread improvements in visibility for
2030.
Table VII-8. Summary of 2030 Baseline Visibility and Changes by Region: Residential
(Annual Average Deciviews)
Regions^
Eastern U.S.
Urban
Rural
Western U.S.
Urban
Rural
National, all counties
Urban
Rural
2030 Base Case
19.32
20.88
18.70
9.75
10.58
9.57
14.77
17.12
14.06
Change8
-0.34
-0.40
-0.32
-0.21
-0.37
-0.18
-0.28
-0.39
-0.25
Percent Change
-1.7%
-1.9%
-1.7%
-2.1%
-3.5%
-1.9%
-1.9%
-2.3%
-1.8%
A Eastern and Western regions are separated by 100 degrees north longitude. Background visibility conditions differ by
region.
B An improvement in visibility is a decrease in deciview value. The change is defined as the HD Engine/Diesel Fuel
control case deciview level minus the basecase deciview level.
b. Recreational Visibility Improvements
In Table VII-9, we summarize recreational visibility improvements by region in 2030 in
Federal Class I areas. These recreational visibility regions are shown in Figure VII-5. As shown,
the national improvement in visibility for these areas is 2.4 percent, or 0.34 deciviews. Predicted
relative visibility improvements are the largest in the Western U.S. as shown for California
(4.9%), and the Southwest (2.4%), the Northwest (2.3%), and the Rocky Mountain (1.9%).
Although Federal Class I areas in the Southeast region are predicted to have the second largest
absolute improvement of 0.42 deciviews, it reflects only a 1.6 percent change from 2030 baseline
visibility of 25.44 deciviews. The Northeast/Midwest region was predicted to have the smallest
relative visibility improvement at 1.2 percent, or 0.25 deciview decline from a baseline of 21.25
deciviews.
VII-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Study Region
Transfer Region
Figure VII-5. Recreational Visibility Regions for Continental U.S.
Table VII-9. Summary of 2030 Baseline Visibility and Changes by Region: Recreational
(Annual Average Deciviews)
Class I Visibility Regions*
Southeast
Southwest
California
Northeast/Midwest
Rocky Mountain
Northwest
National Average (unweighted)
2030 Base Case
25.44
8.90
12.21
21.25
12.54
15.80
14.38
Change11
-0.42
-0.21
-0.60
-0.25
-0.24
-0.36
-0.34
Percent Change
-1.6%
-2.4%
-4.9%
-1.2%
-1.9%
-2.3%
-2.4%
A Regions are pictured in Figure VI-5 and are defined in the technical support document (see Abt Associates, 2000).
B An improvement in visibility is a decrease in deciview value. The change is defined as the HD Engine/Diesel Fuel
control case deciview level minus the basecase deciview level.
Note: Study regions were represented in the Chestnut and Rowe (1990a, 1990b) studies used in evaluating
the benefits of visibility improvements, while transfer regions used extrapolated study results.
VII-28
-------
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 HD Engine/Diesel Fuel
rule. 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:
n. 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;
o. Deposition from atmospheric emissions were divided into local and regional areas
that contribute to airborne nitrogen deposition;
p. 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
q. 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 twelve 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 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.8 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 D 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 deposition 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
deposition outputs from RADM as employed for the final NOx SIP Call (US EPA, 1998b).
More detail on this approach and results may be found in Pechan-Avanti (2000).
B This assumption is consistent with reported case studies such as Valiela et al., 1997. These authors
report that 89 percent of atmospherically deposited nitrogen was retained by the watershed of Waquoit Bay,
suggesting an 11 percent pass through factor.
VII-29
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-10 provides a summary of the baseline deposition and change in nitrogen
deposition estimates for the selected estuaries as a result of the HD Engine/Diesel Fuel rule. As
shown, implementation results in roughly a 21 percent reduction in the average annual deposition
across these estuaries. These predicted reductions range from a low of 17.2 percent for Delaware
Inland Bay to highs of 21.6 percent for Long Island Sound and 24 percent for Tampa Bay.
Table VII-10. Summary of 2030 Nitrogen Deposition in Selected Estuaries and Changes
Due to HD Engine/Diesel Fuel Rule (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
7.66
2.98
12.04
2.56
0.32
0.90
3.07
4.51
1.03
0.89
0.24
1.46
37.64
Change*
-1.64
-0.61
-2.46
-0.49
-0.05
-0.19
-0.61
-0.97
-0.21
-0.18
-0.05
-0.35
-7.82
Percent Change
-21.4%
-20.4%
-20.5%
-19.4%
-17.2%
-20.8%
-19.9%
-21.6%
-20.3%
-20.5%
-20.6%
-24.0%
-20.8%
Change is defined here as the emissions level after implementing the HD Engine/Diesel Fuel rule minus the base case
emissions.
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 we selected to monetize the benefits included in this HD Engine/Diesel Fuel RIA.
VII-30
-------
Chapter VII: Benefit-Cost Analysis
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 (WTP) 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.
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, but might relate to existence
values and bequest values. 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 HD Engine/Diesel
Fuel 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
VII-31
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 lower bound
estimate (under most conditions) 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 because no WTP values were
available in the published literature.
The most direct way to measure the economic value of air quality changes is in cases
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 (CV) method has been employed in the economics
literature to value endpoint changes for both visibility and ecosystem functions (Chestnut and
Dennis, 1997). The CV method 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 CV. EPA believes that well-designed and well-executed CV studies are valid for
estimating the benefits of air quality regulation.11
hConcerns about the reliability of value estimates from CV studies arose because research has shown that
bias can be introduced easily into these studies if they are not carefully conducted. Accurately measuring WTP 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
VII-32
-------
Chapter VII: Benefit-Cost Analysis
Our analysis accounts for expected growth in real income over time. Economic theory
argues that WTP for most goods (such as environmental protection) will increase if real incomes
increase. There is substantial empirical evidence that the income elasticity1 of WTP for health
risk reductions is positive, although there is uncertainty about its exact value. Thus, as real
income increases the WTP for environmental improvements also increases. While many
analyses assume that the income elasticity of WTP is unit elastic (i.e., ten percent higher real
income level implies a ten percent higher WTP to reduce risk changes), empirical evidence
suggests that income elasticity is substantially less than one and thus relatively inelastic. As real
income rises, the WTP value also rises but at a slower rate than real income.
The effects of real income changes on WTP estimates can influence benefit estimates in
two different ways: (1) through real income growth between the year a WTP study was
conducted and the year for which benefits are estimated, and (2) through differences in income
between study populations and the affected populations at a particular time. Empirical evidence
of the effect of real income on WTP gathered to date is based on studies examining the former.
The Environmental Economics Advisory Committee (EEAC) of the SAB advised EPA to adjust
WTP for increases in real income over time, but not to adjust WTP to account for cross-sectional
income differences "because of the sensitivity of making such distinctions, and because of
insufficient evidence available at present" (EPA-SAB-EEAC-00-013).
Based on a review of the available income elasticity literature, we adjust the valuation of
human health benefits upward to account for projected growth in real U.S. income. Faced with a
dearth of estimates of income elasticities derived from time-series studies, we applied estimates
derived from cross-sectional studies in our analysis. Details of the procedure can be found in
Kleckner and Neumann (1999). An abbreviated description of the procedure we used to account
for WTP for real income growth between 1990 and 2030 is presented below.
Reported income elasticities suggest that the severity of a health effect is a primary
determinant of the strength of the relationship between changes in real income and WTP. As
such, we use different elasticity estimates to adjust the WTP for minor health effects, severe and
chronic health effects, and premature mortality. We also expect that the WTP for improved
visibility in Class I areas would increase with growth in real income. The elasticity values used
to adjust the primary estimate of benefits are presented in Table VII-11. In addition to the
primary estimate, we also present the impacts of using different assumed elasticities in Table VII-
25.
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.
'Income elasticity is a common economic measure equal to the percentage change in WTP for a one percent
change in income.
VII-33
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-11. Elasticity Values Used to Account for Projected Real Income Growth^
Benefit Category
Minor Health Effect
Severe and Chronic
Health Effects
Premature Mortality
Visibility8
Lower Sensitivity
Bound
0.04
0.25
0.08
-
Primary
0.14
0.45
0.40
0.90
Upper Sensitivity
Bound
0.30
0.60
1.00
-
A Derivation of these ranges can be found in Kleckner and Neumann (1999) and Chestnut (1997). Cost of Illness (COI) estimates
are assigned an adjustment factor of 1.0.
B No range was applied for visibility because no ranges were available in the current published literature.
Accounting for real income growth over time requires projections of both real gross
domestic product (GDP) and populations. For consistency with the emissions and benefits
modeling, we use population estimates for the years 2015, 2020, and 2030 as described in
Davidson (1999). These population estimates are based on 1990 U.S. Census data and Bureau of
Economic Analysis growth projections/" For the years between 1990 and 2010, we use
population estimates provided in Kleckner and Neumann (1999), which were obtained from the
US Bureau of Census.15 We use projections of real GDP provided in Kleckner and Neumann
(1999) for the years 1990 to 2010.1 We use projections of real GDP (in chained 1996 dollars)
provided by Standard and Poor's for the years 2010 to 2024.m The Standard and Poor's database
only provides estimates of real GDP between 1990 and 2024. We were unable to find reliable
projections of GDP beyond 2024. As such, we assume that per capita GDP remains constant
between 2024 and 2030. This assumption will lead us to under-predict benefits because at least
some level of income growth would be projected to occur between the years 2024 and 2030.
Using the method outlined in Kleckner and Neumann (1999), and the population and
income data described above, we calculate income growth factors for each of the elasticity
J US Bureau of Census. Annual Projections of the Total Resident Population, Middle Series, 1999-2010.
(Available on the internet at http://www.census.gov/population/projections/nation/summary/np-tl.txt)
k US Bureau of Census. Historic National Population Estimates. (Available on the internet at
http://www.census.gov/population/estimates/nation/poplockest.txt) and US Bureau of Census. Resident Population
Projections of the U.S.; Middle Series. (Available on the internet at
http://www.census.gov/population/estimates/nation/npaltsrs.txt)
1 US Bureau of Economic Analysis, Table 2A (1992$). (Available on the internet at
http://www.bea.doc.gov/bea/dn/0897nip2/tab2a.htm) and US Bureau of Economic Analysis, Economics and
Budget Outlook. Note that projections for 2007 to 2010 are based on average GDP growth rates between 1999 and
2007.
m Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter 2000.
VII-34
-------
Chapter VII: Benefit-Cost Analysis
estimates listed in Table VII-11. Benefits for each of the categories (minor health effects, severe
and chronic health effects, premature mortality, and visibility) will be adjusted by multiplying the
unadjusted benefits by the appropriate adjustment factor. In Table VII-12 we list the estimated
factors. Note that for premature mortality, we apply the income adjustment factor ex post to the
present discounted value of the stream of avoided mortalities occurring over the lag period. Also
note that no adjustments will be made to benefits based on the cost-of-illness approach or to
work loss days and worker productivity. This assumption will also lead us to under predict
benefits since it is likely that increases in real U.S. income would also result in increased cost-of-
illness (due, for example, to increases in wages paid to medical workers) and increased cost of
work loss days and lost worker productivity (reflecting that if worker incomes are higher, the
losses resulting from reduced worker production would also be higher). No adjustments are
needed for agricultural and commercial forestry benefits, as these models are based on
projections of supply and demand in future years and should already incorporate future changes
in real income. The results are presented in section F.
Table VII-12. Adjustment Factors Used to Account for Projected Real Income GrowthA
Benefit Category
Minor Health Effect
Severe and Chronic
Health Effects
Premature Mortality
Visibility8
Lower Sensitivity
Bound
1.026
1.176
1.053
-
Primary
1.095
1.341
1.297
1.821
Upper Sensitivity
Bound
1.214
1.482
1.956
-
A Based on elasticity values reported in Table VTI-11, US Census population projections, and projections of real gross domestic
product per capita.
B No range was applied for visibility because no ranges were available in the current published literature.
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 uncertainty." 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
n It should be recognized that in addition to uncertainty, the annual benefit estimates for the final HD
Engine/Diesel Fuel 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-35
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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) (Abt Associates, 2000).
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; and
• Biases due to omissions or other research limitations.
Some of the key uncertainties in the benefits analysis are presented in Table VII-13.
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, alternative calculations, and bounding
exercises. 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, 2000). But, this information should
be interpreted within the context of the larger uncertainty surrounding the entire analysis.
Our approach to characterizing model uncertainty is to present a primary estimate of the
benefits, 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
VII-36
-------
Chapter VII: Benefit-Cost Analysis
would only add to the uncertainty of the analysis or present a false picture about the precision of
the results. Instead, the reader is invited to examine the impact of applying the 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.0 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-25.
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 estimated range of total benefits should be viewed as an approximate result because
of the sources of uncertainty discussed above (see Table VII-13). The total benefits estimate may
understate or overstate actual benefits of the rule.
In considering the monetized benefits estimates, the reader should remain aware of the
many limitations of 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
serious 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.
0 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-37
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Table VII-13. 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.
Application of C-R relationships only to those subpopulations matching the original study population.
2. Uncertainties Associated With Ozone and PM Concentrations
Responsiveness of the models to changes in precursor emissions resulting from the control policy.
Projections of future levels of precursor emissions, especially ammonia and crustal materials.
Model chemistry for the formation of ambient nitrate concentrations.
Lack of ozone monitors in rural areas requires extrapolation of observed ozone data from urban to rural areas.
Use of separate air quality models for ozone and PM does not allow for a fully integrated analysis of pollutants and
their interactions.
Full ozone season air quality distributions are extrapolated from a limited number of simulation days.
VI. Comparison of model predictions of particulate nitrate with observed rural monitored nitrate levels indicates that
REMSAD overpredicts nitrate in some parts of the Eastern US and underpredicts nitrate in parts of the Western US.
3. Uncertainties Associated with PM Mortality Risk
No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
vii. Direct causal agents within the complex mixture of PM 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.
ii The extent to which effects reported in the long-term exposure 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
The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels
would occur in a single year is uncertain as well as the portion that might occur in subsequent years.
5. Uncertainties Associated With Baseline Incidence Rates
9. 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 approximate well baseline incidence rates in 2030.
Projected population and demographics may not represent well 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.
xi. Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to differences
in income or other factors.
Future markets for agricultural and forestry products are uncertain.
7. Uncertainties Associated With Aggregation of Monetized Benefits
ii Health and welfare benefits estimates are limited to the available C-R functions. Thus, unquantified or
unmonetized benefits are not included.
VII-38
-------
Chapter VII: Benefit-Cost Analysis
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 (US EPA, 1996a and 1996b). This section describes individual
effects and the methods used to quantify and monetize changes in the expected number of
incidences of various health effects.
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 address 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 HD Engine/Diesel Fuel rule 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 10 |ig/m3 decrease in daily PM25 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.
The baseline incidence used in our analyses needs to match the specific population
studied. For example, because some mortality studies 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) (US Centers for Disease Control, 1999a). An additional
adjustment was necessary to provide baseline incidences for adults 30 and older for use in the
Krewski, et al. (2000) and Pope, et al. (1995) PM mortality C-R functions. 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. We applied this same procedure to develop baseline incidences for
adults 25 and older for use in alternative premature mortality estimates based on Harvard Six-
City/Krewski, et al. (2000).
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
VII-39
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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
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 this one. In the benefits analyses for some recent RIAs (see the
PM NAAQS RIA, the Regional Haze RIA, and the NOx SIP Call RIA), the low-end estimate of
benefits assumed a threshold in PM health effects at 15 |ig/m3. However, the SAB subsequently
advised EPA 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, 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 no adequate scientific evidence exists
to support such a calculation. Although not included in the primary analysis, the potential impact
of a health effects threshold on avoided incidences of PM-related premature mortality is explored
as a key sensitivity analysis and is presented in Appendix VII-A.
3. Quantifying and Valuing Individual Health Endpoints
Quantifiable health benefits of the final HD Engine/Diesel Fuel rule may be related to
ozone only, PM only, or both pollutants. Decreased worker productivity is the only health
endpoint related to ozone but not PM.P PM-only health effects include premature mortality,
chronic bronchitis, acute bronchitis, upper and lower respiratory symptoms, and work loss days.q
p In the benefits analysis for the recent Tier 2/Gasoline Sulfur rule, based on our interpretation of the
advice from the SAB (EPA-SAB-COUNQL-ADV-00-001), we included avoided incidences of chronic asthma in
adult males as a primary health endpoint associated with ozone. Recent advice from asthma experts both within and
outside the Agency has led us to conclude that while the McDonnell, et al. (1999) study raises concerns about the
possibility of a connection, the scientific evidence supporting the relationship between ozone and new incidences of
asthma is not sufficient to support its inclusion in our primary analysis. We do, however, include this important
endpoint as an alternative calculation in Table VII-25.
q Some evidence has been found linking both PM and ozone exposures 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).
Because of concern about overstating of benefits and because the evidence associating mortality with
exposure to PM is currently stronger than for ozone, only the benefits related to the long-term exposure study
VII-40
-------
Chapter VII: Benefit-Cost Analysis
Health effects related to both PM and ozone include hospital admissions, asthma attacks, and
minor restricted activity days.
For this analysis, we rely on C-R functions estimated in published epidemiological
studies relating serious health effects to ambient air quality. The specific studies from which
C-R functions are drawn are included in Table VII-14. A complete discussion of the C-R
functions used for this analysis and information about each endpoint are contained in the benefits
TSD for this RIA (Abt Associates, 2000).
While a broad range of serious 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 (US 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
single C-R function may be chosen over other potential functions because the underlying
epidemiological study used superior methods, data or techniques, or because the C-R function is
more generalized and comprehensive.
(ACS/Krewkski, et al, 2000) of mortality are included in the total primary 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-41
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-14. Endpoints and Studies Included in the Primary Analysis
Endpoint
Premature Mortality
Long-term exposure
Chronic Illness
Chronic Bronchitis (pooled estimate)
Hospital Admissions
All Respiratory
COPD
Pneumonia
Asthma
Total Cardiovascular
Cardiac Dysrythmias
Asthma-Related ER Visits
Asthma-Related ER Visits
Other Illness
Asthma Attacks
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Work Loss Days
Minor Restricted Activity Days (minus
asthma attacks)
Pollutant
PM25
PM2.5
PM10
Ozone
PM
PM
PM
PM
Ozone
Ozone
PM
PM, Ozone
PM
PM
PM
PM
PM, Ozone
Study
Krewski, etal.(2000)A
Abbey, et al. (1995)
Schwartz, etal. (1993)
Pooled estimate (8 studies)
Samet, etal. (2000)
Samet, etal. (2000)
Sheppard, et al. (1999)
Samet, etal. (2000)
Burnett, etal. (1999)
Pooled estimate (3 studies)
Schwartz, etal. (1993)
Whittemore and Korn (1980)
Dockeryetal. (1996)
Pope etal. (1991)
Schwartz et al. (1994)
Ostro(1987)
Ostro and Rothschild (1989)
Study Population
Adults, 30 and older
> 26 years
> 29 years
All ages
> 64 years
> 64 years
< 65 years
> 64 years
All ages
All ages
All ages
Asthmatics, all ages
Children, 8-12 years
Asthmatic children, 9-11
Children, 7-14 years
Adults, 18-65 years
Adults, 18-65 years
1 Estimate derived from Table 31, PM2.5(DC), All Causes Model (Relative Risk =1.12 for a 24.5 ug/m3 increase in mean PM25).
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, 2000). For example, pooled C-R functions are used to estimate
incidences of chronic bronchitis related to PM exposure and to estimate hospital admissions for
all respiratory causes and asthma-related emergency room visits related to 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 to 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-
VII-42
-------
Chapter VII: Benefit-Cost Analysis
specific C-R functions are generally not available. A single function applied everywhere may
result in overestimates of incidence changes in some locations and underestimates in other
locations, but 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 relative risks 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 a primary
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
illness (COI) estimates generally understate the true value of reductions in risk of a health effect.
They tend to reflect the direct expenditures related to treatment but not the value of avoided pain
and suffering from the health effect. Table VII-15 summarizes the value estimates per health
effect that we used in this analysis. Note that the unit values for hospital admissions are the
weighted averages of the ICD-9 code-specific values for the group of ICD-9 codes included in
the hospital admission categories. Details of the derivation of values for hospital admissions can
be found in the benefits TSD for this RIA (Abt Associates, 2000).
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 HD Engine/Diesel Fuel 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-43
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VII-15. Unit Values Used for Economic Valuation of Health End points
Health or Welfare
Endpoint
Premature Mortality
Chronic Bronchitis (CB)
Estimated Value
Per Incidence
(1999$)
Central Estimate
$6 million per
statistical life
$331,000
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 Costs and Benefits of the Clean
Air Act, 1990-2010 (US EPA, 1999).
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.
Hospital Admissions
Chronic Obstructive
Pulmonary Disease (COPD)
(ICD codes 490-492, 494-496)
Pneumonia
(ICD codes 480-487)
Asthma admissions
All Cardiovascular
(ICD codes 390-429)
Emergency room visits for
asthma
$12,378
$14,693
$6,634
$18,387
$299
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total COPD category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total pneumonia category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total asthma category illnesses) reported in
Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total cardiovascular illnesses)
reported in Elixhauser (1993).
COI estimate based on data reported by Smith, et al. (1997).
VII-44
-------
Chapter VII: Benefit-Cost Analysis
Table VII-15. Unit Values Used for Economic Valuation of Health En (I points
Health or Welfare
Endpoint
Estimated Value
Per Incidence
(1999$)
Central Estimate
Derivation of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
Lower Respiratory Symptoms
(LRS)
Acute Bronchitis
$24
$15
$57
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 1 1 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 1 1 different types of LRS.
Average of low and high values recommended for use in
Section 812 analysis (Neumann, et al. 1994)
Restricted Activity and Work Loss Days
Work Loss Days (WLDs)
Minor Restricted Activity
Days (MRADs)
Variable
$48
Regionally adjusted median weekly wage for 1990 divided by
5 (adjusted to 1999$) (US Bureau of the Census, 1992).
Median WTP estimate to avoid one MRAD 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 endpoints quantified in this analysis. Although these endpoints
account for over 90 percent of the total monetized benefits, 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
VII-45
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
summarized in the 1996 PM Criteria Document (US EPA, 1996a). There is much about this
relationship that is still uncertain. As stated in preamble to the 1997 PM National Ambient Air
Quality Standards (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 (NAS, 1998), indicates 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 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 (US EPA, 1996a).
In addition, 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 the Tier 2/Gasoline Sulfur
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 exposure) 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., one or more years) exposure to PM
and annual mortality rates. Researchers have found significant associations using both types of
studies (US EPA, 1996a); however, for this analysis, we follow SAB advice (EPA-SAB-
COUNCIL-ADV-99-005, 1999), and we rely exclusively on long-term exposure studies to
quantify mortality effects.
VII-46
-------
Chapter VII: Benefit-Cost Analysis
Following advice from the SAB (EPA-SAB-COUNCIL-ADV-99-005, 1999), we prefer
to use long-term exposure 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 exposure study design is preferred, they are expensive to conduct and
consequently there are relatively few well designed long-term exposure studies. For PM, there
have been only a few, and the SAB has explicitly recommended use of only one — the American
Cancer Society (ACS) Study, as reported in Pope, et al. (1995) (EPA-SAB-COUNCIL-ADV-99-
005, 1999). The data from this study were reanalyzed and we used a C-R function from the HEI
reanalysis (ACS/Krewski et al., 2000).
The ACS/Pope, et al. study used a prospective cohort design to estimate the risk of
premature mortality from long-term exposures to ambient PM concentrations. The ACS/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 and uses the longer exposure intervals,
and more locations (50 cities) in the U.S. than other studies. Recently, the Health Effects
Institute (HEI), a non-profit, independent research organization commissioned an extensive
reanalysis of the data used in the ACS/Pope, et al. (1995) study/
The HEI reanalysis, as reported in Krewski, et al. (2000) and mentioned above, confirmed
the general findings of the ACS/Pope, et al (1995) study. In addition, the reanalysis tested a
number of alternative model specifications, some of which may be preferred to the original
ACS/Pope, et al. (1995) specification. One important alternative specification examines the
relationship between relative risk of premature mortality and mean PM2 5 levels rather than
median levels used in the Pope, et al. (1995) analysis (Table 31, "PM2 5(DC)" model). For
policy analysis purposes, functions based on the mean air quality levels may be preferable to
functions based on the median air quality levels because changes in the mean more accurately
reflect changes in peak values than do changes in the median. Policies which affect peak PM
days more than average PM days will result in a larger change in the mean than in the median. In
these cases, all else being equal, C-R functions based on median PM2 5 will lead to lower
estimates of avoided incidences of premature mortality than C-R functions based on mean PM25.
In addition to specifying a preference for the ACS study based on the larger set of cities
examined, the SAB has also noted a preference for applying mean PM2 5 in premature mortality
functions (US EPA-SAB, 1999). For these reasons, we have selected the C-R function based on
the relative risk of 1.12 from the "PM2.5(DC), All Causes" model reported in Table 31 of the
HEI report.8
'Additional information on the Health Effects Institute and the reanalysis of the Harvard Six Cities and
American Cancer Society Studies can be obtained at http://www.healtheffects.org.
s Note that in several recent RIAs, we erroneously applied the ACS/Pope et al. C-R function to a baseline
of non-accidental mortality. The correct baseline, matching the mortality measured in the ACS/Pope et al. and
Krewski et al. studies is all-cause mortality. This correction results in a slight increase in the estimated mortality
reductions resulting from a reduction in PM2 5.
VII-47
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Although we use the Krewski, et al. (2000) mean-based ("PM2.5(DC), All Causes")
model exclusively to derive our primary estimates of avoided premature mortality, we also
examine the impacts of selecting alternative C-R functions for premature mortality. There are
several candidates for alternative C-R functions, some from the Krewski, et al. study, and others
from the original ACS study by Pope, et al. or from the "Harvard Six-City Study" by Dockery, et
al. (1993).
Commentary by an independent review panel noted that "a major contribution of the
[HEI] Reanalysis Project is the recognition that both pollutant variables and mortality appear to
be spatially correlated in the ACS data set. If not identified and modeled correctly, spatial
correlation could cause substantial errors in both the regression coefficients and their standard
errors (HEI, 2000)." The HEI reanalysis provides results for several models which control for
spatial correlations in the data. These models are based on the original ACS air quality dataset,
which contained only median PM2 5 concentrations. Ideally, our primary C-R function for
premature mortality would be both based on the mean and adjusted for regional variability.
Unfortunately, Krewski, et al. do not provide such an estimate. As such, we have chosen to use
the mean-based relative risk in our primary analysis and to use the median-based regionally
adjusted relative risks to provide alternative estimates exploring the impact of adjustments for
spatial correlations (see Table VII-16).
Krewski, et al. (2000) also reanalyzed the data from another prospective cohort study (the
Harvard "Six Cities Study") authored by Dockery, et al. (1993). The Dockery, et al.(1993) study
used a smaller sample of individuals from fewer cities than the study by Pope, et al.; however, 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 SAB has noted that
"the [Harvard Six Cities] study had better monitoring with less measurement error than did most
other studies" (EPA-SAB-COUNCIL-ADV-99-012, 1999). The Dockery, et al. (1993) study
finds a larger effect of PM on premature mortality relative to the Pope, et al. (1995) study. To
provide a more complete picture of the range of possible premature mortality risks that may be
associated with long-term exposures to fine particles, we also present alternative estimates based
on the Krewski, et al. (2000) reanalysis of the Dockery, et al. (1993) data and the original study
estimates. The HEI commentary notes that "the inherent limitations of using only six cities,
understood by the original investigators, should be taken into account when interpreting the
results of the Six Cities Study." We emphasize, that based on our understanding of the relative
merits of the two datasets, the Krewski, et al. (2000) ACS model based on mean PM2 5 levels in
63 cities is the most appropriate model for analyzing the premature mortality impacts of the HD
Engine/Diesel Fuel rule. It is thus used for our primary estimate of this important health effect.
Table VII-16 summarizes the alternative C-R functions for PM-related premature
mortality. Note that the right most column provides a standardized estimate of the incidences of
premature mortality that would be reduced by a one microgram reduction in PM2 5 applied to a
population of one million. Note that the relative magnitude of the values will not necessarily
correlate with the estimates of avoided incidences that will result from application of the HD
Engine/Diesel Fuel reductions in PM25 to 2030 national populations. This is because some of the
functions are based on changes in mean PM2 5 concentrations while others are based on median
VII-48
-------
Chapter VII: Benefit-Cost Analysis
PM2 5 concentrations. Estimated reductions in premature mortality will depend on both the size
of the C-R coefficient and the change in the relevant PM25 metric (mean or median).
Table VII-16. Alternative Concentration-Response Models Relating Premature Mortality
and Chronic Exposure to Fine Particulates
Model Description
(as listed in the study)
PM2.5(DC), All Causes
Source: Table 31, Krewski, et
al. (2000)
Fine Particles Alone, Random
Effects, Regional Adjustment
Source: Table 46, Krewski, et
al. (2000)
Fine Particles Alone, Random
Effects, Independent Cities
Source: Table 46, Krewski, et
al. (2000)
All Combined, All Cause, Fine
Particles
Source: Table 3, Pope, et al.
(1995)°
All Causes, Extended, Age
Time Axis: Table 3, Krewski,
et al. (2000)
All Subjects
Source: Table 3, Dockery, et al.
(1993)
# of Cities
63
50
50
50
6
6
PM Metric
Mean
Median
Median
Median
Mean
Mean
Reported Relative
Risl^
(95% Confidence
Interval)
1.12
(1.06-1.19)
1.16
(0.99-1.37)
1.29
(1.12-1.48)
1.17
(1.09-1.26)
1.27
(1.09-1.48)
1.26
(1.08-1.47)
Avoided
Incidences of
Premature
Mortality per
Million Population
for a 1 ug/m3
Decrease in PM2 5B
68
89
152
90
173
153
A Reported relative risks for the Pope, et al. (1995) and Dockery, et al. (1993) studies are comparisons of mortality rates between most polluted
and least polluted cities. For the Pope et al. study the relative risk is based on a difference in median PM2.5 levels of 24.5 ug/m3. For the
Dockery, etal. study, the relative risk is based on a difference of 18.6 ug/m3. The Krewski, etal. reanalysis of the Pope, etal. study reports all
relative risks based on a 24.5 ug/m3 difference for comparability with the Pope, et al. (1995) results, rather than comparing the means or medians
of the most polluted and least polluted studies. Likewise, the Krewski, et al. reanalysis of the Dockery, et al. Harvard Six Cities study reports all
relative risks based on a 18.6 ug/m3 difference for comparability with the Dockery, et al. (1993) study.
BAssumes national all-cause mortality rate of 0.0147 per person for adults aged 30 and older and 0.0131 per person for adults aged 25 and older.
(US Centers for Disease Control. 2000 National Vital Statistics Reports 48(11): Table 8).
0 The Pope, et al. estimate of the relative risk of premature mortality from fine particle exposure is the basis for the estimates of premature
mortality found in the final Tier 2/Gasoline Sulfur rule.
VII-49
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 published 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 $6
million in 1999 dollars. This represents an intermediate value from a variety of estimates that
appear in the economics literature, and it 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 $6 million estimate is
consistent with Viscusi's conclusion (updated to 1999$) 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. As indicated in the previous section on
quantification of premature mortality benefits, we assume for 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 three percent discount rate to the value of premature
mortality occurring in future years.1
The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. 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. Regardless of the theoretical economic considerations, 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.
Following the advice of the EEAC of the SAB, EPA currently uses the VSL approach in
calculating the primary estimate of mortality benefits, because we believe this calculation to
provide the most reasonable single estimate of an individual's willingness to trade off money for
4 The choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within
the federal government. EPA adopted a 3 percent discount rate for its primary analysis in this case to reflect
reliance on a "social rate of time preference" discounting concept. We have also calculated benefits and costs using
a 7 percent rate consistent with an "opportunity cost of capital" concept to reflect the time value of resources
directed to meet regulatory requirements. In this case, the benefit and cost estimates were not significantly affected
by the choice of discount rate. Further discussion of this topic appears in EPA's Guidelines for Preparing
Economic Analyses (in press).
VII-50
-------
Chapter VII: Benefit-Cost Analysis
reductions in mortality risk (EPA-SAB-EEAC-00-013). 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. Table VII-17 lists some of these
differences and the expected effect on the VSL estimate for air pollution-related mortality. For
example, adjusting for age differences may imply the need to adjust the $6 million VSL
downward, 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 $6 million value while acknowledging the
significant limitations and uncertainties in the available literature.
Some economists emphasize that the value of a statistical life is not a single number
relevant for all situations. Indeed, the VSL estimate of $6 million (1999 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,
some economists prefer to adjust the VSL estimate to reflect those differences. Some of the
alternative approaches that have been proposed for valuing reductions in mortality risk are
discussed in Figure VII-6.
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 individual's attitudes towards risk, and the health
status of the individual. While the empirical basis for adjusting the $6 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, 2000). EPA recognizes the need for investigation by the
scientific community to develop additional empirical support for adjustments to VSL for the
factors mentioned above.
Table VII-17. Expected Impact on Estimated Benefits of Premature Mortality Reductions of
Differences Between Factors Used in Developing Applied VSL and Theoretically Appropriate VSL
Attribute
Age
Attitudes toward risk
Income
Voluntary vs. Involuntary
Catastrophic vs. Protracted Death
Expected Direction of Bias
Uncertain, perhaps overestimate
Underestimate
Uncertain
Uncertain, perhaps underestimate
Uncertain, perhaps underestimate
VII-51
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Figure VII-6. 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 (VSLY) - 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. Similar to VSLY, adjustment may not reflect how individuals consider life-years, and
typically assumes an equal value for all remaining life-years despite evidence to the contrary.
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 Figure VII-6 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. In the recent Tier
2/Gasoline Sulfur benefits analysis, we employed a value of statistical life years (VSLY)
approach developed for the Section 812 studies in exploring the impact of age on VSL. Since the
VSLY alternative calculation was introduced in the Section 812 studies, the SAB raised new and
additional concerns about the merits of the VSLY approach. Specifically, they note in their
recent report that "inferring the value of a statistical life year, however, requires assumptions
about the discount rate and about the time path of expected utility of consumption" (EPA-SAB-
EEAC-00-013). In considering the merits of age-based adjustments, the Committee also notes
that "the theoretically appropriate method is to calculate WTP for individuals whose ages
correspond to those of the affected population, and that it is preferable to base these calculations
on empirical estimates of WTP by age." Several studies conducted by Jones-Lee, et al. (1985,
1989, 1993) found a significant effect of age on the value of mortality risk reductions expressed
VII-52
-------
Chapter VII: Benefit-Cost Analysis
by citizens in the United Kingdom. Using the results of the Jones-Lee et al. analysis, U.S. EPA
(2000b) calculated ratios of the value of life for different age groups to the mean value of life
estimated by Jones-Lee, et al. (1989, 1993). The Jones-Lee-based analysis suggests a U-shaped
relationship between age and VSL, peaking around age 40, and declining to between 60 and 90
percent of the mean VSL value for individuals over the age of 70, and declining further as
individuals age. This finding has been supported by two recent analyses conducted by Krupnick,
et al. (2000a, 2000b), which asked samples of Canadian and U.S. residents their values for
reductions in mortality risk. We apply the ratios based on the Jones-Lee, et al. (1989, 1993)
studies to the estimated premature mortalities within the appropriate age groups to provide an
alternative age-adjusted estimate of the value of avoided premature mortalities. However, we
have not attempted in this analysis to provide a consistent treatment of age-dependence between
the underlying wage-risk studies and the present calculation. Therefore, the downward
adjustment for age relative to our primary benefit estimate may be significantly overestimated,
implying a significant underestimation of age-adjusted total benefits.
The SAB-EEAC advised in their recent report that the EPA "continue to use a wage-risk-
based VSL as its primary estimate, including appropriate sensitivity analyses to reflect the
uncertainty of these estimates," and that "the only risk characteristic for which adjustments to the
VSL can be made is the timing of the risk"(EPA-SAB-EEAC-00-013). In developing our
primary estimate of the benefits of premature mortality reductions, we have discounted over the
lag period between exposure and premature mortality. However, in accordance with the SAB
advice, we use the VSL in our primary estimate and present the Jones-Lee calculations in the
table of alternative calculations, Table VII-25.
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 new incidences of chronic
bronchitis. Schwartz (1993) and Abbey, et al.(1995) provide evidence that long-term PM
exposure gives rise to the development of chronic bronchitis in the U.S. Following the same
approaches, the Section 812 Prospective Report (US EPA, 1999a), our analysis pools the
estimates from these studies to develop a C-R function linking PM to chronic bronchitis. The
Schwartz (1993) study examined the relationship between exposure to PM10 and prevalence of
chronic bronchitis. The Abbey, et al. (1995) study examined the relationship between PM25 and
new incidences of chronic bronchitis. Both studies have strengths and weaknesses which suggest
that pooling the effect estimates from each study may provide a better estimate of the expected
change in incidences of chronic bronchitis than using either study alone. However, the HD
Engine/Diesel Fuel rule is expected to result in reductions in both the fine and coarse fractions of
PM10. As such, reliance on the Abbey, et al. (1995) estimate will result in an underestimate of
the change in chronic bronchitis incidences if both the fine and coarse fractions of PM10 are
associated with chronic bronchitis. To address this problem, we apply the C-R functions from
both Schwartz (1993) and Abbey, et al. (1995) to generate the changes in chronic bronchitis
incidences associated with the change in PM2 5 and then pool the incidence estimates to obtain a
VII-53
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
primary estimate of avoided PM2 5 related chronic bronchitis incidences. We then apply the
Schwartz (1993) C-R function to the change in coarse PM (PM2 5_10) to obtain a primary estimate
of avoided incidences of chronic bronchitis due to the change in coarse fraction PM. The
primary estimate of total avoided incidences is then the sum of the avoided incidences from
changes in PM25 and PM25_10U
It should be noted that Schwartz used data on the prevalence of chronic bronchitis, not its
incidence. Following the Section 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.
d. Chronic Bronchitis: Valuation
The best available estimate of WTP to avoid a case of chronic bronchitis (CB) comes
from Viscusi, et al. (1991)v 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 the Viscusi, et al. (1991) estimate of the WTP to avoid a severe
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, 2000).
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 uncertainly analysis techniques. The expected value (i.e., mean) of this distribution,
11 This assumption implies that the observed relationship between chronic bronchitis and PM10 in the
Schwartz (1993) study is equally attributable to the fine and coarse fractions of PM10. If the relationship is due
primarily to the fine fraction, then the estimate of avoided incidences associated with coarse fraction PM changes
will be overstated. However, if this is the case then the estimate of avoided incidences associated with fine fraction
will be somewhat understated. The net effect on avoided incidences of chronic bronchitis is ambiguous.
vThe 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 SAB (EPA-SAB-COUNCIL-ADV-00-002, 1999) has indicated that the severity-adjusted values from this study
provide reasonable estimates of the WTP for avoidance of 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-54
-------
Chapter VII: Benefit-Cost Analysis
which is about $331,000 (1999$), is taken as the central tendency estimate of WTP to avoid a
PM-related case of CB.
e. Hospital and Emergency Room 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 ER 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 are 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. are asthma-related visits.
/'. PM-related Hospital Admissions
To estimate avoided incidences of hospital admissions associated with PM, we use a
study by Samet, et al. (2000) which examined the relationship between PM10 and admissions for
pneumonia, chronic obstructive pulmonary disease (COPD), and cardiovascular disease in
fourteen U.S. cities. In previous analyses, we have pooled estimates from a number of studies in
different cities. However, Samet, et al. (2000) represents a comprehensive analysis of the
relationship between hospital admissions and air pollution conducted under the auspices of the
Health Effects Institute as part of the National Morbidity, Mortality, and Air Pollution study.
This extensive analysis by the HEI was intended to provide a consistent, comparable set of
effects estimates over a wide range of cities. As such, the pooled estimates of relative risk for
pneumonia, COPD, and cardiovascular disease provided by the study (Table 14, "Unconstrained
distributed lag, Random effects estimate"), which covers most of the studies included
individually in previous benefits analyses, represents the most up-to-date estimate of the
relationship between PM air pollution and hospital admissions. One study (Moolgavkar, 1997)
found a much lower effect of PM on hospital admissions for pneumonia and COPD. The effect
of using Moolgavkar (1997) instead of Samet, et al. (2000) is presented as a alternative
calculation in Table "VTI-25.
The Samet, et al. (2000) HEI analysis estimated separate C-R functions for pneumonia
and COPD hospital admissions for people 65 years and older. In addition, Sheppard, et al.
(1999) estimated a C-R function for asthma hospital admissions for people under age 65. These
three estimates can be combined to calculate total avoided incidences of PM-related respiratory-
related hospital admissions.
To estimate the effects of PM air pollution reductions on asthma-related ER visits, we use
the C-R function based on a study of Seattle residents by Schwartz, et al. (1993). Because we are
VII-55
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 ER 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 percent of the resulting change in
asthma ER visits associated with a given change in pollutant concentrations is counted in the ER
visit incidence change.
/'/'. Ozone-related Hospital Admissions
To estimate avoided incidences of hospital admissions associated with ozone, we use a
number of studies examining hospital admissions for a range of respiratory illnesses and one
study examining hospital admissions for cardiac dysrythmias. Hospital admissions for
respiratory diseases studied include admissions for pneumonia, COPD, asthma, and a number of
other respiratory illnesses. Hospital admissions for cardiac dysrythmias are estimated using a C-
R function derived from Burnett, et al. (1999).
f. 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.
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 although they underestimate the
benefits. 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 advice offered by the SAB, the COI
values used in this benefits analysis will not be adjusted (EPA-SAB-COUNCIL-ADV-98-003,
1998).
For the valuation of avoided respiratory and cardiovascular 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. Using the methods developed for the Section 812
reports, ICD-code-specific COI estimates were generated based on estimated hospital charges
and the estimated opportunity cost of time spent in the hospital (estimated as the value of the lost
VII-56
-------
Chapter VII: Benefit-Cost Analysis
daily wage, regardless of whether or not the individual is in the workforce). The value of an
avoided asthma-related ER visit is based on data reported in Smith, et al. (1997). The average
cost per ER visit reported in this study (1999$) is $298.62.
g. Asthma Attacks: Quantification
Asthma is the most prevalent chronic disease among children in the U.S., affecting over
seven percent of children under 18 years old (US CDC, 1998). Among adults, it currently affects
over six percent of the U.S. population (US CDC, 1998). Asthma attacks are a serious health
effect for people with asthma. During an attack, muscles around the airways constrict, the
airways become inflamed, and less air passes in and out of the lungs. The attack is also called an
episode or exacerbation and can include coughing, chest tightness, wheezing, and difficulty
breathing (Jack, Boss, and Millington, 2000). The literature supports a direct relationship
between air pollution and increased incidence and severity of asthma-related respiratory
symptoms. Studies have documented this relationship for both PM (Yu, et al., 2000; McConnell
et al., 1999; Delfino et al., 1998; Delfino et al., 1997; US EPA, 1996a; Ostro et al., 1995;
Whittemore and Korn, 1980) and ozone (Delfino et al., 1998; Thurston et al., 1997; US EPA,
1996b; Delfino et al., 1996; Ostro et al., 1995; US EPA, 1986; Whittemore and Korn, 1980).
There are a number of these studies showing a relationship between PM and/or ozone
levels and asthma-related respiratory symptoms such as wheezing, coughing, acute bronchitis and
shortness of breath. However, only one study (Whittemore and Korn, 1980) estimated the
relationship between asthma attacks and photochemical air pollutant concentrations. The likely
reason for the emphasis of most studies on particular asthma symptoms is the subjective
definition of an asthma attack and the subsequent lack of specificity in measuring an asthma
attack occurrence. In this analysis, the endpoint "asthma attack" is a better match for the
economic valuation studies and avoids potential overprediction (as one attack may involve some
combination of symptoms). Accordingly, an asthma attack is an endpoint that summarizes the
collection of symptoms, so potential double-counting may occur if individual asthma symptoms
estimated from other studies are summed. An asthma attack, as measured by Whittemore and
Korn (1980), is based on subjective reporting by study participants and likely consisting of one or
more of the respiratory symptoms listed above occurring at varying levels of severity. For
example, a subject reporting an asthma attack in the Whittemore and Korn (1980) study may
have shortness of breath and wheezing. This is accounted for as one attack, while using
individual symptom studies would record this as two separate symptom occurrences.
Conversely, a participant may experience symptoms but not consider the symptoms to be "an
attack." Thus, the use of "asthma attacks" as an indicator may understate symptoms.
In addition, a limited number of economic studies have been conducted on the value of
reduced asthma symptoms. One valuation study by Rowe and Chestnut (1986) calculated the
value of reduction in "bad asthma days," which we interpret as equivalent to a day with an
asthma attack. By using the Whittemore and Korn (1980) asthma attack C-R function in
combination with the Rowe and Chestnut (1986) valuation study, we are able to provide a
quantified and monetized estimate of asthma-related symptoms that is representative of the full
spectrum of impacts of air pollution reductions on asthma sufferers.
VII-57
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Although the Whittemore and Korn (1980) study had a number of methodological flaws,
including omission of some potentially confounding variables and use of proxies for ozone and
PMW, we believe that the more recent literature supports the general magnitude of the
relationship. As such, we use Whittemore and Korn in our primary analysis to estimate the
effects of air pollution on asthma symptoms, recognizing that the Whittemore and Korn based
estimate represents symptoms examined in other studies, though perhaps undercounting the
frequency of symptom occurrence. Other analyses of the impacts of air pollution reductions on
asthma symptoms have used collections of asthma symptom studies (Kunzli et al., 2000).
However, we believe it is more illustrative to provide a single endpoint that represents a
combination of symptoms. The Whittemore and Korn study was also previously used to estimate
asthma attacks in the Section 812 analysis (although it was not included in the primary estimate
of total benefits), which was reviewed and accepted by the EPA SAB. Table VII-18 provides a
summary of the more recent studies of air pollution and respiratory symptoms in asthmatics.
Also, several asthma-related endpoints are provided as supplementary calculations in Appendix
VII-A to this chapter.
Note that the estimated number of avoided asthma attacks is the total change over the full
population of asthmatics, potentially including multiple avoided attacks for a single individual.
Also, because our estimate of asthma attacks is based on both the incidence of asthma attacks
and the prevalence of asthma in the population, to the extent that asthma incidence rates are
increasing (or decreasing), the number of asthma attacks avoided will also be increasing (or
decreasing). The prevalence of asthma, especially among children, has been increasing over the
past two decades (Pew Environmental Health Commission, 2000), suggesting that if current
trends continue, the impact on asthma symptoms of reductions in air pollution will be greater
than we estimated in this analysis.
h. Asthma Attacks: Valuation
In the primary analysis, we do not present a monetized value. As an alternative, asthma
attacks are valued at $41 per incidence (1999$), based on the mean of average WTP estimates for
the four severity definitions of a "bad asthma day," described in Rowe and Chestnut (1986). This
study surveyed asthmatics to estimate WTP for avoidance of a "bad asthma day," as defined by
the subjects. For purposes of valuation, an asthma attack is assumed to be equivalent to a day in
which asthma is moderate or worse as reported in the Rowe and Chestnut (1986) study. To the
extent that an asthma attack differs from a "bad asthma day" as defined by Rowe and Chestnut
(1986), the value of an asthma attack may be over or underestimated. Recent evidence from the
United Kingdom (Hoskins et al., 2000) suggests that our value for avoided asthma attacks may
understate true benefits by a significant amount. Hoskins et al. used a very specific definition of
an asthma attack that is likely to be more severe than at least some of the asthma attacks reported
by subjects in the Whittemore and Korn (1980) study. Using this definition, however, they found
that asthmatics who suffered at least one asthma attack in a year had increased asthma-related
costs of £273, or around $450US (1999$).
"Whittemore and Korn used oxidants instead of ozone and TSP instead of PM10.
VII-58
-------
Chapter VII: Benefit-Cost Analysis
Table VII-18. Recent Studies on the Effects of Air Pollution on Asthma Symptoms
Study
McConnell, et
al. (1999)
Delfino, et al.
(1997)
Ostro, et al.
(1995)
Thurston, et al.
(1997)
Delfino, et al.
(1998)
Delfino, et al.
(1996)
Location/
Date
Southern CA,
1993
Southern CA,
1994
Los Angeles,
1992
CT, 1991-
1993
Southern CA,
1995
San Diego,
CA, 1993
Asthmatic
Study
PopulationA
493 asthmatic
children, ages
9-15
22 asthmatics,
ages 9-46
83 African-
American
asthmatic
children, ages
7-12
166 asthmatic
children, ages
7-13
25 asthmatic
children, ages
9-17
12 asthmatic
children, ages
9-16
Symptoms8
Bronchitis,
Phlegm,
Cough
Symptom
severity,
PEFR,
inhaler use
Shortness
of breath
Chest
symptoms,
PEFR,
inhaler use
Asthma
symptom
score
Asthma
symptom
score,
inhaler use
Pollutants
PM10,
PM25,
NO2,
Ozone,
Acid
vapor
pollen,
fungi,
Ozone,
PM10
Ozone,
PM10,
SO2, NO2,
pollen,
fungi
Ozone,
S04,
Hydrogen
ion
PM10,
ozone,
fungi
Ozone,
PM25,
fungi
Main Findings for Ozone and
PM Exposures c
Significant effects of PM10 on
bronchitis (OR=1.4, 95% CI
= 1.1, 1.4 ) and phlegm
(OR=2.1, 95%CI=1.4, 3.3).
Significant effect of PM2 5 on
phlegm (OR=2.6, 95%
CI=1.2, 5.4)
Significant effect of PM10 on
inhaler use (0.15 inhaler
puffs/10 ug/m3, p<0.02)
Significant effect of PM10 on
shortness of breath (OR=1.6,
95%CI=1.1,2.4). Significant
effect of ozone on shortness
of breath (OR=1.4, 95%
CI=1.0, 1.8)
Significant effect0 of ozone
on chest symptoms (OR=1.4)
and inhaler use (OR= 1 .4)
Significant effect of 24-hr
mean PM10 on asthma
symptoms (OR=1.7, 95%
CI=1.0, 2.7)
Significant effect of ozone on
inhaler use (1.1 puffs/100
ppb, p<0.03) and symptom
scores
A Study population is not the only measure of the power of a statistical analysis. For some studies, such as the Delfino, et al. (1996) analysis, the
relatively small number of subjects were followed for a period of time. Thus, the number of person-days in these studies is a better indicator of
statistical power than the number of study subjects.
B PEFR is peak expiratory flow rate, a measure of lung function.
c OR is the odds ratios.
D No 95% confidence interval was reported for the odds ratios in the Thurston, et al. (1997) study.
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
VII-59
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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.
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
Dockery, et al. (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 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). Research on the effects of air pollution
on upper respiratory symptoms have thus focused on effects in asthmatics. Incidences of upper
respiratory symptoms in asthmatic children aged nine to eleven are estimated using a C-R
function developed from Pope, et al. (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) is estimated using a C-R function
derived from Ostro and Rothschild (1989). Because MRADs are characterized by many of the
same symptoms as those which define an asthma attack and the study population in Ostro and
Rothschild did not exclude asthmatics, we reduce the estimated number of avoided MRAD
incidences by the estimated number of avoided asthma attacks to prevent double-counting of
asthma attacks. This simple subtraction may result in an underestimate of non-asthma attack
related MRADs, since asthma attacks are estimated for asthmatics of all ages and MRADs are
estimated only for ages 18 to 65. However, without further information on the percent of
MRADs that are related to asthma attacks, we have chosen to provide a conservative estimate of
MRAD benefits.
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
(US EPA, 1996b), increased emergency room visits for non-asthma respiratory causes (US EPA,
1996a; 1996b), impaired airway responsiveness (US EPA, 1996a), increased susceptibility to
respiratory infection (US EPA, 1996a), acute inflammation and respiratory cell damage (US
EPA, 1996a), premature aging of the lungs and chronic respiratory damage (US 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 are not available for
VII-60
-------
Chapter VII: Benefit-Cost Analysis
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 reduction
strategies 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 percent 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.
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
VII-61
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 it provides some flexibility in constructing estimates
to represent a variety of health effects.
Valuation estimates for individual minor health effects are listed in Table VII-16
Derivation of the individual valuation estimates is provided in the benefits TSD for this RIA.
Mean estimates range from $15 for an avoided day of lower respiratory symptoms to $57 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 representative value of a
work loss day is $106 (1999$). However, depending on where an affected individual lives, the
value of work loss day may be higher or lower than $106.
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; US 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 VH-19. All of the monetary benefits are in constant 1999
dollars.
Not all known PM- and ozone-related health effects could be quantified or monetized.
These unmonetized benefits are indicated by place holders, labeled El and B2. In addition,
unmonetized benefits associated with CO and NMHC reductions are indicated by the
placeholders B3 and B4. Unquantified physical effects are indicated by Ul through U4. 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.
VII-62
-------
Chapter VII: Benefit-Cost Analysis
An important factor to consider when interpreting the ozone-related benefits in Table VII-
19 is the omission of ozone-related benefits in the Western U.S.X Over 22 percent of national
NOx emission reductions occur in the Western U.S., with over 10 percent of total NOx emissions
occurring in California alone. This suggests that ozone benefits in the West may be substantial,
and that our estimate of Eastern ozone benefits may significantly underestimate national ozone-
related benefits of the HD Engine/Diesel Fuel NOx reductions.
The largest monetized health benefit is associated with reductions in the risk of premature
mortality, which accounts for over $60 billion, which is over 90 percent of total monetized health
benefits. 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 $10 million each, however, they represent a large
number of avoided incidences affecting many individuals. Alternative calculations for premature
mortality incidences and valuation are presented in Tables VII-24 and VII-25, respectively. An
alternative calculation is also provided in Table VII-25 for chronic bronchitis incidences and for
chronic asthma incidences.
x We define the Western U.S. as west of 100 degrees longitude.
VII-63
-------
Table VII-19. Primary Estimate of Annual Health Benefits Associated With Air Quality
Changes Resulting from the HP Engine/Diesel Fuel Rule in 2030
Endpoint
PM-related Endpointsc
Premature mortality13 (adults, 30 and over)
Chronic bronchitis (adults, 26 and over)
Hospital Admissions - Pneumonia (adults, over 64)
Hospital Admissions - COPD (adults, 64 and over)
Hospital Admissions - Asthma (65 and younger)
Hospital Admissions - Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)E
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Other PM-related health effects13
Ozone-related Endpoints (Eastern U.S. only)F
Hospital Admissions - Respiratory Causes (all ages)
Hospital Admissions - Cardiac Dysrhymias (all ages)
Emergency Room Visits for Asthma (all ages)
Asthma Attacks (asthmatics, all ages)E
Minor restricted activity days (adults, age 18-65)
Decreased worker productivity (adult working
population)
Other ozone-related health effectsE
CO and NMHC -related health effects13
Monetized Total Health-related Benefits0
Avoided
Incidence*
(cases/year)
8,300
5,500
1,100
900
900
2,700
2,100
175,900
17,600
192,900
193,400
1,539,400
7,990,400
u,
1,200
300
300
185,500
1,848,100
—
U2
U3+U4
—
Monetary Benefits8
(millions 1999$, not
adjusted for growth
in real income)
$48,250
$1,810
$20
$10
$10
$50
<$5
Ba
<$5
<$5
$10
$160
$390
B,
$20
<$5
<$1
Ba
$100
$140
B2
B3+B4
$50,980+BH
Monetary Benefits8
(millions 1999$,
adjusted for growth
in real income)
$62,580
$2,430
$20
$10
$10
$50
<$5
Ba
<$5
<$5
$10
$160
$430
B,
$20
<$5
<$1
Ba
$100
$140
B2
B3+B4
$65,970+BH
A Incidences are rounded to the nearest 100.
B Dollar values are rounded to the nearest 10 million.
c PM-related benefits are based on the assumption that Eastern U.S. nitrate reductions are equal to one-fifth the nitrate reductions predicted by
REMSAD (see Chapter II fora discussion of REMSAD and model performance).
D Premature mortality associated with ozone is not separately included in this analysis (also note that the estimated value for PM-related
premature mortality assumes the 5 year distributed lag structure described in Section D-3.
E A detailed listing of unquantified PM, ozone, CO, and NMHC related health effects is provided in Table VII-1. For some endpoints such as
asthma attacks, we are able to quantify the reduction in incidence, but we present the monetization as an alternative calculation.
F Ozone-related benefits are only calculated for the Eastern U.S. due to unavailability of reliable modeled ozone concentrations in the Western
U.S. See Section C-3 for a detailed discussion of the UAM-V ozone model and model performance issues.
0 Bg is equal to the sum of all unmonetized categories, i.e. B.+Bl+B2+B1+B4.
VII-64
-------
Chapter VII: Benefit-Cost Analysis
E. Assessment of Human Welfare Benefits
PM 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 PM and ozone list numerous
physical and ecological effects known to be linked to ambient concentrations of these pollutants
(US EPA, 1996a; 1996b). This section describes individual effects and how we quantify and
monetize them. These effects include changes in commercial crop and forest 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 5, we summarize the monetized estimates for
welfare effects. A more detailed description of these analyses can be found in the benefits TSD
for this RIA (Abt Associates, 2000).
1. Visibility Benefits
Changes in the level of ambient parti culate matter caused by the final HD Engine/Diesel
Fuel 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
VII-65
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
models were used to predict the change in visibility, measured in deciviews, of the areas affected
by the final HD Engine/Diesel Fuel rule.y
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. Use values include the aesthetic benefits of better visibility, improved
road and air safety, and enhanced recreation in activities like hunting and birdwatching. 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 areas.2 For
the purposes of this analysis, recreational visibility improvements are defined as those that occur
specifically in federal Class I areas. A key distinction between recreational and residential
benefits 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.aa
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 CV
surveys in the past decade. In EPA's judgment, the Chestnut and Rowe study contains many of
the elements of a valid CV study and is sufficiently reliable to serve as the basis for monetary
estimates of the benefits of visibility changes in recreational areas.bb This study serves as an
y A change of less than 10 percent in the light extinction budget represents a measurable improvement in
visibility, but may not be perceptible to the eye in many cases. 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.
z The Clean Air Act designates 156 national parks and wilderness areas as Class I areas for visibility
protection.
aa For details of the visibility estimates discussed in this chapter, please refer to the benefits technical
support document for this RIA (Abt Associates 2000).
bb An SAB advisory letter indicates that "many members of the Council believe that the Chestnut and
Rowe study is the best available." (EPA-SAB-COUNCIL-ADV-00-002, 1999) However, the committee 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
VII-66
-------
Chapter VII: Benefit-Cost Analysis
essential input to our estimates of the benefits of recreational visibility improvements in the
primary benefits estimates. Consistent with SAB advice, EPA has designated the McClelland, et
al. study as significantly less reliable for regulatory benefit-cost analysis, although it does provide
useful estimates on the order of magnitude of residential visibility benefits (EPA-SAB-
COUNCIL-ADV-00-002, 1999). Residential visibility benefits are therefore only included as an
alternative calculation in Table VII-25. 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 WTP 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-25. 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,
2000).
The estimated relationship from the Chestnut and Rowe study is only directly applicable
to the populations represented by survey respondents. EPA used benefits transfer methodology
to extrapolate these results to the population affected by the final HD Engine/Diesel Fuel 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 HD Engine/Diesel Fuel 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 1 percent increase in income is associated with a 0.9 percent increase
in WTP for a given change in visibility.
(Chestnut and Dennis, 1997).
VII-67
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Using the methodology outlined above, EPA estimates that the total WTP for the
visibility improvements in California, Southwestern, and Southeastern Class I areas brought
about by the final HD Engine/Diesel Fuel rule is $3.3 billion. 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, and the value accounts for growth in real
income.
For the alternative calculation for residential visibility, the McClelland, et al. study's
results were used to calculate the parameter to measure the effect of deciview changes on WTP.
The WTP equation was then run for the population affected by the final FID Engine/Diesel Fuel
rule. The results indicate that improvements to residential visibility provide an economic benefit
of $ 2.1 billion dollars for the continental U.S.CC
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.
2. Agricultural and Forestry Benefits
The Ozone Criteria Document notes that "ozone affects vegetation throughout the United
States, impairing crops, native vegetation, and ecosystems more than any other air pollutant" (US
EPA, 1996). Reduced levels of ground-level ozone resulting from the final FID Engine/Diesel
Fuel 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 and silvicultural producers and to consumers. These techniques use models of
planting decisions, yield response functions, and agricultural and forest products supply and
demand. The resulting welfare measures are based on predicted changes in market prices and
production costs.
a. Agricultural Benefits
Laboratory and field experiments have shown reductions in yields for agronomic crops
exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and wheat).
The most extensive field experiments, conducted under the National Crop Loss Assessment
00 The McClelland, et al. (1993) 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-25.
VII-68
-------
Chapter VII: Benefit-Cost Analysis
Network (NCLAN) examined 15 species and numerous cultivars. The NCLAN results show that
"several economically important crop species are sensitive to ozone levels typical of those found
in the U.S." (US EPA, 1996). In addition, economic studies have shown a relationship between
observed ozone levels and crop yields (Garcia, et al., 1986). 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 HD Engine/Diesel Fuel rule) that
affect commodity crop yields or production costs.dd
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 basis.ee 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 HD Engine/Diesel Fuel rule.
The model employs biological exposure-response information derived from controlled
experiments conducted by the NCLAN (NCLAN, 1996). For the purpose of our analysis, we
analyze changes for the six most economically significant crops for which C-R functions are
available: corn, cotton, peanuts, sorghum, soybean, and winter wheat.ff For some crops there are
multiple C-R 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 HD Engine/Diesel Fuel rule is $1.1 billion (1999$).
b. Forestry Benefits
Ozone also has been shown conclusively to cause discernible injury to forest trees (US
EPA, 1996; Fox and Mickler, 1996). In this section, we describe methods for benefits we are able
to quantify and we present a qualitative description of benefits we are not able to quantify at this
time. For commercial forestry impacts, the effects of changes in ozone concentrations on tree
growth for a limited set of species are predicted. For future analyses, it would be helpful to use
ddAGSIM° 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 an 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.
ee 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.
ff The total value for these crops in 1998 was $47 billion.
VII-69
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
econometric models of forest product supply and demand to estimate changes in prices, producer
profits and consumer surplus. However, for this RIA we were not able to monetize the biological
changes we predicted for commercial tree species. For commercial forestry, well-developed
techniques are used to estimate biological and market changes. Limitations of the approach
presented here include: the lack of underlying forest inventory information which is not available
for the Western U.S., and the unavailability of parameterization data for all relevant species
nationally. Thus, we must assume that no ozone-related changes occur to forest inventories in
the Western U.S. or Canada, although as described earlier, it is likely that across the country, this
rulemaking could result in decreases in ozone and improvements in forest health compared to
baseline conditions. Therefore, using these assumptions will underestimate the commercial
forestry benefits associated with the program.
Similar to the agriculture analysis, assessing the forestry benefits couples air quality
modeling results, C-R functions derived from a biological model, forest inventory estimates, and
an economic model. Again, we are only able to quantify the physical effect, and further details
are contained in the technical support document (Hubbell et al., 2000).
Our analysis used species-specific C-R functions derived from the TREGRO model
(Laurence, et al., 2000). We developed ozone C-R functions for 6 species for which there were
parameterization data by climatic region of the Eastern U.S.: black cherry, loblolly pine, red oak,
red spruce, sugar maple, and tulip poplar. TREGRO is a model of tree physiological response to
environmental stresses (Weinstein and Yanai, 1994). It was developed to simulate the response
of sapling and mature trees to ozone and acidic precipitation stress in conjunction with other
stressors. The model has been used to evaluate long-term effects of pollutants on resource
availability.
The next step would be to use economic model such as the Timber Assessment Market
Model (TAMM)/Aggregated Timberland Assessment System (ATLAS). In brief, the approach
would be to use the biological inputs to modify the accumulation of inventory within ATLAS,
which then shifts the timber supply functions in TAMM. The economic value of yield changes
for commercial forests would be estimated using TAMM. This model is a US Forest Service
(Adams and Haynes, 1996) spatial model of the solid wood and timber inventory elements of the
U.S. forest products sector. The model provides projections of timber markets by geographic
region and wood type through the year 2050. Nine regions covering the continental U.S. are
included in the analysis; however, the effects of reduced O3 concentrations were only considered
for the Eastern U.S. The TAMM model perturbs timber market (spatial) equilibrium and yields
timber price, quantity and welfare effects. However, it is limited to sawtimber and does not
capture all relevant forest product markets (e.g., pulp wood). TAMM, in turn, would predict the
effect of these reductions on timber markets by changing the annual growth rates of commercial
forest growing-stock inventories. The model uses applied welfare economics to value changes in
ambient O3 concentrations. However, we were not able to complete this step for the RIA.
The six species we analyzed account for as much as 73 percent and as little as zero
percent of total growing stock volume depending on the region and forest type. The annual
change in growth adjustment factors ranged from zero to 0.009841. While the adjustment factor
VII-70
-------
Chapter VII: Benefit-Cost Analysis
may seem small on an absolute basis, when compounded over the lifetime of a tree, the effects
may be significant. The full set of adjustment factors are presented in the technical support
document (Hubbell et al., 2000).
c. Other Effects
An additional welfare benefit expected to accrue as a result of reductions in ambient
ozone concentrations in the U.S. 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 (US 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 HD Engine/Diesel Fuel 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,
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 (US EPA, 1993).
VII-71
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
On the other hand, there is evidence that forest ecosystems in some areas of the United
States are nitrogen saturated (US 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 HD Engine/Diesel Fuel 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 have been able to provide quantitative estimates of
household soiling damage. Consistent with SAB advice, we determined that the existing data
(based on consumer expenditures from the early 1970's) are too out of date to provide a reliable
enough estimate of current household soiling damages (EPA-SAB-Council-ADV-003, 1998).
An estimate is included in the alternative calculations presented in Table VII-25.
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
(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 (Bricker et al., 1999).
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 "VTi-20.
VII-72
-------
Chapter VII: Benefit-Cost Analysis
Table VII-20. Reduction Goals and 1998 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: US EPA, 1998.
Estimated reductions in deposition of atmospheric nitrogen to these four estuaries are
listed in Table VII-21, 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 HD Engine/Diesel Fuel rule will provide significant progress towards meeting nitrogen
reduction goals in several of these estuaries.
Table VII-21. Estimated Annual Reductions in Nitrogen Loadings in Selected Eastern
Estuaries for the Final HD Engine/Diesel Fuel Rule in 2030
(tons per year)
Estuary
Albemarle/Pamlico Sound
Chesapeake Bay
Long Island Sound
Tampa BayA
Change in Nitrogen Loadings
1,804
2,706
1,067
385
% of Estuary Nitrogen
Reduction Goal
23.7%
7.6%
3.4%
over 100%
A Tampa Bay had a very low nitrogen loadings reduction goal. As such, the HD Engine/Diesel Fuel 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.
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
VII-73
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
deposited nitrogen and the abandonment of a costly regulatory program. However, there are
currently no readily available alternatives to this approach.
Based on SAB advice, 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 Clean Air Act (EPA-SAB-COUNCTL-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 Clean Air Act, we are currently unable to
provide a meaningful primary estimate. Thus, the avoided cost estimate is presented as an
alternative calculation in Table VII-25.
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,
then CV studies can be designed to elicit individuals' WTP 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 and forestry
yields. These estimates are presented in Table Vn-22. All of the monetary benefits are in
constant 1999 dollars.
We are unable to provide primary monetized estimates of residential visibility, household
soiling, materials damage, and nitrogen deposition, in addition to the other welfare effects listed
in Table VII-1. These unmonetized benefits are indicated by placeholders, labeled B5 to B12. 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.
VII-74
-------
Chapter VII: Benefit-Cost Analysis
Table VII-22. Primary Estimate of Annual Monetary Values for Welfare Effects
Associated With Improved Air Quality Resulting from the HD Engine/Diesel Fuel 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 effects13
Ozone-related Endpoints
Commercial Agricultural Benefits (6 major crops) (Eastern
U.S. only)c
Commercial Forestry Benefits (Eastern U.S. only)c
Other ozone-related welfare effects5
CO-related welfare effects5
NMHC-related welfare effects13
Total Monetized Welfare-related Benefits0
Monetary Benefits
(millions 1999$,
Unadjusted for growth in
real income)A
$1,790
B5
B6
B7
B8
B9
$1,120
B9
BIO
Bn
B12
$2,910
Monetary Benefits
(millions 1999$,
Adjusted for Growth
in Real Income)A
$3,260
B5
B6
B7
B8
B9
$1,120
B9
BIO
Bn
B12
$4,380
A Rounded to the nearest 10 million and visibility benefits are adjusted to account for growth in real GDP per capita between 1990 and 2030.
See Section C. B A detailed listing of unquantified PM, ozone, CO, and NMHC related welfare effects is provided in Table VII-1.
c Ozone-related benefits are only calculated for the Eastern U.S. due to unavailability of reliable modeled ozone concentrations in the Western
U.S. This results in an underestimate of national ozone-related benefits. See Section D-3 for a detailed discussion of the UAM-V ozone model
and model performance issues. D BW is equal to the sum of all unmonetized welfare categories, i.e. B5+B6+...+B13.
Total monetized welfare-related benefits are around $4.4 billion. Monetized welfare
benefits are roughly l/20th the magnitude of monetized health benefits. However, due to the
difficulty in quantifying and monetizing welfare benefits, a higher proportion of welfare benefits
are not monetized. It is thus inappropriate to conclude that welfare benefits are unimportant just
by comparing the estimates of the monetized benefits. Also, as with health benefits, ozone-
related welfare benefits may be significantly underestimated due to the omission of ozone-related
benefits in the Western U.S.
Alternative calculations for recreational visibility, residential visibility, household soiling,
and nitrogen deposition are presented in Table VII-25 later in this chapter.
VII-75
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
F. Total Benefits
We provide our primary estimate of benefits for each health and welfare endpoint as well
as the resulting primary 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
primary estimate of the total benefits associated with the health and welfare effects is the sum of
the separate effects estimates. Total monetized benefits associated with the final HD
Engine/Diesel Fuel rule are listed in Table VII-23, 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. For example, there are over 40
times more asthma attacks than premature mortalities, yet these asthma attacks account for only a
very small fraction of total monetized benefits. This reflects the fact that many of the less severe
health effects, while more common, are valued at a lower level than the more severe health
effects. Also, some effects, such as asthma attacks, are valued using a proxy measure of WTP.
As such the true value of these effects may be higher than that reported in Table "VH-23.
Our primary estimate of total monetized benefits for the final HD Engine/Diesel Fuel rule
is $70.4 billion, of which $62.6 billion is the benefits of reduced premature mortality risk from
PM exposure. Total monetized benefits are dominated by the benefits of reduced mortality risk.
Mortality related benefits account for 89 percent of total monetized benefits followed by
recreational visibility (4.6 percent) and chronic bronchitis (3.5 percent). Health benefits account
for 94 percent of total benefits.
VII-76
-------
Table VII-23. Primary Estimate of Annual Monetized Benefits Associated With Improved
Air Quality Resulting from the HP Engine/Diesel Fuel Rule in 2030AB
Endpoint
Premature mortalityE (adults, 30 and over)
Chronic bronchitis (adults, 26 and over)
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper resp. symptoms (asthmatic children, 9-11)
Asthma attacks (asthmatics, all ages)F
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Other health effectsF'°
Decreased worker productivity
Recreational visibility (86 Class I Areas)
Residential visibility
Household soiling damage
Materials damage
Nitrogen Deposition to Estuaries
Agricultural crop damage (6 crops)
Commercial forest damage (6 species)
Other welfare effectsF>0
Pollutant
PM
PM
O3 and PM
O3 and PM
O3 and PM
PM
PM
PM
O3 and PM
PM
O3 and PM
O3, PM, CO, HAPs
03
PM
PM
PM
PM
Nitrogen
03
03
O3, PM, CO, HAPs
Avoided
Incidence0
(cases/year)
8,300
5,500
4,100
3,000
2,400
17,600
192,900
193,400
361,400
1,539,400
9,838,500
U!+U2+U3+U4
—
—
—
—
—
—
—
—
—
Monetized TotalH
Monetary Benefits"
(millions 1999$,
Adjusted for
Income Growth)
$62,580
$2,430
$60
$50
<$5
<$5
<$5
$10
Ba
$160
$530
B!+B2+B3+B4
$140
$3,260
B5
B6
B7
B8
$1,120
B9
Bio+Bn+B12+B13
$70,360+B
AMonetary benefits are adjusted to account for growth in real GDP per capita between 1990 and 2030. See Section
C. B Ozone-related benefits are only calculated for the Eastern U.S. due to unavailability of reliable modeled ozone
concentrations in the Western U.S. This results in an underestimate of national ozone-related benefits. See Section
D-3 for a detailed discussion of the UAM-V ozone model and model performance issues. c Incidences are rounded
to the nearest 100. D Dollar values are rounded to the nearest 10 million. E Premature mortality associated with
ozone is not separately included in this analysis. It is assumed that the Section D-3 ACS/Krewski et al., 2000 C-R
function for premature mortality captures both PM mortality benefits and any mortality benefits associated with
other air pollutants. Also note that the valuation assumes the 5 year distributed lag structure described earlier. F The
U; are the incidences for the unqualified category i, the Bj are the monetary values for the unqualified endpoint i.
For some categories such as asthma attacks, we were able to quantify the reduction in incidence, but we present the
monetization as an alternative calculation. G A detailed listing of unqualified PM, ozone, CO, and NMHC related
health and welfare effects is provided in Table VII-1. H B is equal to the sum of all unmonetized categories, i.e.
Ba+B1+B2+...+B13.
VII-77
-------
As discussed in Section C.I, we have adjusted our primary estimate of benefits to reflect the
projected growth in real income between base income in 1990 and the 2030 analytical year. We
account for real income growth by applying the primary adjustment factors from Table VII-12 to
the appropriate health and welfare endpoints in Tables VII-19 and VII-22.
In addition to the primary estimate in Table VII-23, in Tables VII-24 and VII-25 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. Specifically, in Table VII-24, we present the impact of different C-R functions
for PM-related premature mortality. In Table VII-25, we show the impact of alternative
assumptions about other parameters. For example, Table VII-25 can be used to answer questions
like "What would total benefits be if we were to value avoided incidences of premature mortality
using the VSLY approach rather than the VSL approach?" 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., possible recovery from chronic illnesses). 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. As discussed earlier, individual adjustments
in the table should not be added together without addressing potential issues of overlap and low
joint probability among the endpoints. Accompanying Table VII-25 is a brief discussion of each
of the alternative calculations.
While Tables VII-24 and VII-25 provide 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. This appendix also contains several illustrative endpoint calculations for which the
scientific uncertainty is too great to provide a reasonable estimate and if included, might lead to
double-counting of benefits. These include premature mortality associated with daily
fluctuations in PM, infant mortality associated with PM, and premature mortality associated with
daily fluctuations in ozone.
VII-78
-------
Chapter VII: Benefit-Cost Analysis
Table VII-24. Alternative Estimates of Premature Mortality Benefits
for the HD Engine/Diesel Fuel Rule in 2030
Model
1
2
3
4
5
Fine Particles Alone, Random
Effects, Regional Adjustment
Source: Table 46, Krewski, et al.
(2000) "ACS Study "
Fine Particles Alone, Random
Effects, Independent Cities
Source: Table 46, Krewski, et al.
(2000) "ACS Study"
All Combined, All Cause, Fine
Particles
Source: Table 3, Pope, et al.
(1995) "ACS Study"
All Causes, Extended, Age Time
Axis: Table 3, Krewski, et al.
(2000) "Harvard Six-city Study"
All Subjects
Source: Table 3, Dockery, et al.
(1993) "Harvard Six-city Study"
Avoided
Incidences
9,400
16,000
9,900
24,200
23,100
Value
Adjusted for
Growth in Real
Income
(million 1999$)
$69,940
$93,940
$75,360
$181,080
$173,450
Impact on Primary
Benefits Estimate
Adjusted for Growth
in Real Income
(million 1999$)
+$7,370 (+10.5%)
+$59,270 (+84.2%)
+12,780 (+18.2%)
+$118,500 (+168.4%)
+$110,874 (+157.6%)
A Reported relative risks for the Pope, et al. (1995) and Dockery, et al. (1993) studies are comparisons of mortality rates between most polluted
and least polluted cities. For the Pope et al. study the relative risk is based on a difference in median PM2.5 levels of 24.5 ug/m3. For the
Dockery et al. study, the relative risk is based on a difference of 18.6 ug/m3. The Krewski et al. reanalysis of the Pope et al. study reports all
relative risks based on a 24.5 ug/m3 difference for comparability with the Pope, et al. (1995) results, rather than comparing the means or medians
of the most polluted and least polluted studies. Likewise, the Krewski et al. reanalysis of the Dockery et al. Harvard Six Cities study reports all
relative risks based on a 18.6 ug/m3 difference for comparability with the Dockery, et al. (1993) study.
BAssumes national all-cause mortality rate of 0.0147 per person for adults aged 30 and older and 0.0131 per person for adults aged 25 and older.
(U.S. Centers for Disease Control. 2000 National Vital Statistics Reports 48(11): Table 8).
The first alternative C-R function (row 1 of Table VII-24) is based on the relative risk of
1.16 from the "Fine Particles Alone, Regional Adjustment Random Effects" model reported in
Table 46 of the HEI report. This C-R function is a reasonable specification to explore the impact
of adjustments for broad regional correlations. However, the HEI report noted that the spatial
adjustment methods "may have over adjusted the estimated effect for regional pollutants such as
fine particles and sulfate compared with the effect estimates for more local pollutants such as
sulfur dioxide." Thus, the estimates of avoided incidences of premature mortality based on this
C-R function may underestimate the true effect. Note that this C-R function is based on the
original air quality dataset used in the ACS study, covering 50 cities, and used the median PM2 5
levels rather than mean PM2 5 as the indicator of exposure.
VII-79
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Krewski, et al. (2000) also estimated a random effects model which accounts for between
city variation but "ignores possible regional patterns in mortality." The estimated avoided
incidences of premature mortality is based on the original 50 city air quality dataset used in the
ACS study and used the median PM2 5 levels rather than mean PM2 5 as the indicator of exposure
(row2ofTableVn-24).
For comparison with earlier benefits analyses, we also include estimates of avoided
incidences of premature mortality based on the original ACS/Pope et al. (1995) analysis (row 3
of Table VII-24) and the original "Harvard Six Cities" estimate as reported in Dockery, et al.
(1993) analysis (row 5 of Table VII-24).
The Krewski, et al. "Harvard Six Cities" estimate of the relationship between PM
exposure and premature mortality (row 4 of Table VII-24) is a plausible alternative to the
Krewski, et al. "ACS 50 City" primary estimate. The SAB has noted that "the [Harvard Six
Cities] study had better monitoring with less measurement error than did most other studies"
(EPA-SAB-COUNCIL-ADV-99-012, 1999). However, the Krewski-Harvard Six Cities study
had a more limited geographic scope (and a smaller study population) than the Krewski-ACS
study. The demographics of the ACS study population, i.e., largely white and middle-class, may
also produce a downward bias in the estimated 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 Krewski-Harvard Six Cities study also covered a broader age
category (25 and older compared to 30 and older in the ACS study) and followed the cohort for a
longer period (15 years compared to 8 years in the ACS study). For these reasons, the Krewski -
Harvard Six Cities study is considered to be a plausible alternative estimate of the avoided
premature mortality incidences associated with the final HD Engine/Diesel Fuel rule.
VII-80
-------
Table VII-25. Additional Alternative Benefits Calculations
for the HP Engine/Diesel Fuel Rule in 2030
Alternative
Calculation
1
2
3
4
5
6
7
8
9
10
Value of avoided
premature
mortality
incidences based
on age-specific
VSL
Chronic Asthma
Reversals in
chronic bronchitis
treated as lowest
severity cases
COPD and
pneumonia
hospital
admissions.
Value of visibility
changes in all
Class I areas
Value of visibility
changes in Eastern
U.S. residential
areas
Value of visibility
changes in
Western U.S.
residential areas
Household soiling
damage
Avoided costs of
reducing nitrogen
loadings in east
coast estuaries
Asthma attacks
Description of Estimate
Calculate the age distribution of avoided incidences of
premature mortality and apply age-adjusted VSL to the
incidences. Sources of age-adjustment ratios are Jones-Lee
(1989) and Jones-Lee et al. (1993)
Avoided incidences of chronic asthma are estimated using
the McDonnell, et al. (1999) C-R function. The number of
avoided incidences of chronic asthma is 820.
Instead of omitting 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 5,480 to 10,250
(87%).
Hospital admissions for Pneumonia and COPD estimated
using the Moolgavkar (1997) C-R function instead of the
Samet et al. (2000) pooled C-R function. The number of
hospital admissions for these two causes decreases from
2,0 10 to 600 (-70%)
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.
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).
Value of visibility changes outside of Class I areas are
estimated for the Western U.S. based on the reported
values for Chicago and Atlanta from McClelland et al.
(1990).
Value of decreases in expenditures on cleaning are
estimated using values derived from Manuel, et al. (1983).
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.
Avoided incidences of asthma attacks monetized using
Rowe and Chestnut (1986).
Impact on Primary
Benefit Estimate
Adjusted for Growth
in Real Income
(million 1999$)
Jones-Lee (1989)
-$28,5 10 (-40.5%)
Jones-Lee (1993)
-$6,820 (-10.0%)
+$40 (+
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
The age-specific VSL alternative calculation (row 1 of Table VII-25) 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 $6 million value. A complete discussion of
this issue can be found in section 3.b of this chapter. For this calculation, the method we use to
account for age differences is to adjust the VSL based on ratios of VSL's for specific ages to the
VSL for a 40 year old individual (row 2 of Table VII-25). There are several potential sources for
these ratios. Estimates from two Jones-Lee studies provide a reasonable low and high end for
this type of adjustment. The ratios based on Jones-Lee (1989), as summarized in U.S. EPA
(2000), suggest a steep inverted U shape between age and VSL, with the VSL for a 70 year old at
63 percent of that for a 40 year old, and the VSL for an 85 year old at 7 percent of that for 40 year
old. The ratios based on Jones-Lee (1993) and summarized in U.S. EPA (2000b), suggest a
much flatter inverted U shape, with the VSL for a 70 year old at 92 percent of that for a 40 year
old, and the VSL for an 85 year old at 82 percent of that for a 40 year old. The general U shaped
relationship is supported by recent analyses conducted in Canada and the U.S. by Krupnick et al.
(2000a, 2000b). Their results suggest a curvature somewhere between the two Jones-Lee
estimates. The wide range of age-adjustment ratios, especially at older ages demonstrates the
difficulty in making these kinds of adjustments. To calculate the age-adjusted VSL, we first
calculate the number of avoided premature mortalities in each age category, and then apply the
age adjusted VSL to the appropriate incidences in each age category88.
The alternative calculation for the development of chronic asthma (row 2 of Table VII-
25) is estimated using a recent study by McDonnell, et al. (1999) which found a statistical
association between ozone and the development of asthma in adult white, non-Hispanic males.
Other studies have not identified an association between air quality and the onset of asthma.
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 (US Centers for Disease Control and Prevention, 1999b).
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. 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.
gg The age categories and lower and upper end estimated age-adjustment ratios are: 30-39 (0.89, 0.98), 40-
59 (1.0, 1.0), 60-69 (0.86, 0.97), 70-79 (0.63, 0.92), 80-84 (0.28, 0.85), 85+ (0.07, 0.82).
VII-82
-------
Chapter VII: Benefit-Cost Analysis
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 the following: 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 SAB reviewed the study and deemed it appropriate for
quantification of changes in ozone concentrations in benefits analyses (EPA-SAB-COUNCIL-
ADV-00-001, 1999). EPA recognizes the need for further investigation by the scientific
community to confirm the statistical association identified in the McDonnell, et al. study.
Following SAB advice (EPA-SAB-COUNCIL-ADV-00-001, 1999) and consistent with
the Section 812 Prospective Report, we quantify this endpoint for the 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 alone are likely to measurably
change long-term risks of asthma.
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 who have
been diagnosed as asthmatics. The central estimate of lifetime WTP to avoid a case of chronic
asthma among adult males, approximately $25,000, 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, 2000).
Another important issue related to chronic conditions is the possible reversal in chronic
bronchitis incidences (row 3 of Table "VTI-25). 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 the lowest severity level. These cases thus get the lowest
value for chronic bronchitis.
VII-83
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
For this benefits analysis, we have adopted the C-R function for COPD and pneumonia
hospital admissions from Samet, et al. (2000). This estimate, while representing the state of the
art in epidemiological studies, is a good deal larger than the estimate from Moolgavkar (1997).
We explore the impact of using the Moolgavkar (1997) estimate instead of the Samet, et al.
(2000) estimate in row 4 of Table VII-25.
The alternative calculation for recreational visibility (row 5 of Table VII-25) is an
estimate of the full value of visibility in the entire region affected by the final HD Engine/Diesel
Fuel 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 regions.
The alternative calculations for residential visibility (rows 6 and 7 of Table VII-25) are
based on the McClelland, et al. study of WTP for visibility changes in Chicago and Atlanta. As
discussed in Section F-l, SAB advised EPA that the residential visibility estimates from the
available literature are 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-25) 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 are from a 1972 consumer
expenditure survey and as such may not accurately represent consumer preferences in 2030.
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 PM household soiling.
The alternative calculation for the avoided costs of reductions in nitrogen loadings (row 9
of Table VII-25) is constructed by examining the avoided costs to surrounding communities of
reduced nitrogen loadings for three case study estuaries (US EPA, 1998). 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
water pollution (US EPA, 1998). Valuation reflects water pollution control cost avoidance based
VII-84
-------
Chapter VII: Benefit-Cost Analysis
on the weighted average cost per pound of current non-point source water pollution controls for
nitrogen in the three case study estuaries. Taking the weighted cost per 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 12 estuaries represent only about
50 percent of the total watershed area in the Eastern U.S.; and 2) costs avoided are not good
proxies for WTP, understating true WTP under certain conditions.
We monetize the reduction of 361,400 asthma attacks (row 10 of Table VII-25) using
Rowe and Chestnut (1986).
G. Comparison of Costs to 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
policy alternative 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, as 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 the monitoring network for PM2 5 has produced only one year of data,
atmospheric models designed to capture the effects of alternative policies on PM25 have
not yet been 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.
• Another dimension adding to the uncertainty of this analysis is time. In our analysis we
are projecting over a 30 year time period, which can introduce significant uncertainty.
VII-85
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Projected growth in factors such as population, income, source-level emissions, and
vehicle miles traveled over the 30-year period have a significant effect on the benefits
estimates, as will changes in health baselines, technology, and other factors. In addition,
there is no clear way to predict future meteorological conditions compared to those used
in these analyses. Again, EPA believes that the assumptions used 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, the
benefits TSD, and 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 primary benefit estimate does not include the monetary value of health benefits from
ozone changes in the Western U.S. It also does not include the monetary value of several
known ozone and PM-related welfare effects, including residential visibility, recreational
visibility in over half of Federal Class I areas, agricultural and forestry benefits in the
Western U.S. and for many crops and species, 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 HD
Engine/Diesel Fuel rulemaking.
Our estimates of annual costs for this rulemaking are developed in Chapter V. As
described in that chapter, at a 7 percent discount rate, the total program cost in 2030 is
approximately $4.3 billion (1999$). If a discount rate of 3% is used instead, this cost estimate
drops to approximately $4.2 billion (1999$). This latter value is used in our comparison of costs
to benefits for calendar year 2030.
The primary estimate of monetized benefits is $70.4 billion (1999$). Comparing this
with costs of $4.2 billion (1999$), monetized net benefits are approximately $66.2 billion
(1999$). Therefore, implementation of the HD Engine/Diesel Fuel program will provide society
with a net gain in social welfare based on economic efficiency criteria. Table "VTI-26 summarizes
the costs, benefits, and net benefits for the HD Engine/Diesel Fuel rule. Note that the cost and
benefit estimates presented in Table VII-26 assume a 3 percent discount rate. Assuming a 7
percent discount rate does not materially alter the outcome. Net benefits are reduced by $3.9
billion to $62.3 billion, a reduction of 6 percent.
VII-86
-------
Chapter VII: Benefit-Cost Analysis
Table VII-26. 2030 Annual Monetized Costs, Benefits, and Net Benefits
for the Final HD Engine/Diesel Fuel Rule
Annual compliance costs
Monetized PM-related benefits8'0
Monetized Ozone-related benefits8'13
NMHC-related benefits
CO-related benefits
Total annual benefits
Monetized net benefitsE
Billions of 1999$
$4.2
$69.0 + BPM
$1.4 + B0zone
not monetized (BNMHC)
not monetized (Bco)
$70.4 +BPM + B0zon(, + BNMHC + Bco
$66.2 + B
A 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.
D Ozone-related benefits are only calculated for the Eastern U.S. due to unavailability of reliable modeled ozone concentrations in the Western
U.S. This results in an underestimate of national ozone-related benefits. See US EPA (2000a) for a detailed discussion of the UAM-V ozone
model and model performance issues.
E B is equal to the sum of all unmonetized benefits, including those associated with PM, ozone, CO, and NMHC.
VII-87
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Chapter VII. References
Abbey, D.E., B.L. Hwang, RJ. Burchette, T. Vancuren, and P.K. Mills. 1995a. "Estimated
Long-Term Ambient Concentrations of PM(10) and Development of Respiratory Symptoms in a
Nonsmoking Population." Archives of Environmental Health 50(2): 139-152.
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.
Abt Associates Inc., 2000.. Heavy Duty Engine/Diesel Fuel Rule: Air Quality Estimation,
Selected Health and Welfare Benefits Methods, and Benefit Analysis Results. Prepared for US
EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC. November.
Abt Associates, Inc., 1999a. Tier 2 Final Rule: Air Quality Estimation, Selected Health
and Welfare Benefits Methods, and Benefit Analysis Results, Prepared for the US Environmental
Protection Agency, Office of Air Quality Planning and Standards. Research Triangle Park, NC,
November.
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., 1998. Air Quality Estimation for the NOx SIP Call RIA. Prepared for US
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, K. and M. 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.L. Dennis. 1997. Economic 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. A New National Park Visibility Value Estimates. In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C. V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
VII-88
-------
Chapter VII: Benefit-Cost Analysis
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,
US Environmental Protection Agency, Research Triangle Park, NC and Air Quality Management
Division, National Park Service, Denver, CO.
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.
Delfmo, R.J., R.S. Zeiger, J.M. Seltzer, and D.H. Street. 1998. Symptoms in Pediatric
Asthmatics and Air Pollution: Differences in Effects by Symptom Severity, Anti-inflammatory
Medication Use and Particulate Averaging Time. Environmental Health Perspectives., 106(11):
751-761.
Delfmo, R.J., R.S. Zeiger, J.M. Seltzer, and D.H. Street, 1998. Symptoms in Pediatric
Asthmatics and Air Pollution: Differences in Effects by Symptom Severity. Anti-inflammatory
Medication Use and Particulate Averaging Time. Environmental Health Perspectives, 106(11):
751-761.
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, J.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, J.E., Dickie,
M.T., and R.E. Unsworth. 1994. Industrial Economics, Incorporated. Memorandum to Jim
DeMocker, US 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." New
England Journal of Medicine 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.
VII-89
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
E.H. Pechan & Associates, 1996. RegionalParticulate Control Strategies Phase II. Prepared for
the US Environmental Protection Agency, Office of Policy, Planning, and Evaluation.
Washington, DC. September.
E.H. Pechan & Associates, 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. for US 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, US
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, Health and Ecological Effects Subcommittee of the Advisory Council on Clean Air
Act Compliance Analysis. 1999. Summary Minutes of Public Meeting, April 21-22, 1999.
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-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.
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,
1999.
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. Effects Subcommittee on Initial
Assessments of Health and Ecological Effects: Part 2. October, 1999.
VII-90
-------
Chapter VII: Benefit-Cost Analysis
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.
Hoskins, G., C. McCowan, R.G. Neville, G.E. Thomas, B. Smith, and S. Silverman. 2000. "Risk
Factors and Costs Associated with and Asthma Attack." Thorax. 55: 19-24.
Hoskins, G., C. McCowan, R.G. Neville, G.E. Thomas, B. Smith, S. Silverman. 2000. Risk
Factors and Costs Associated with an Asthma Attack. Thorax, 55: 19-24.
Industrial Economics, Incorporated (TEc). 1994. Memorandum to Jim DeMocker, Office of Air
and Radiation, Office of Policy Analysis and Review, US Environmental Protection Agency,
March 31.
Jack, E., L.P. Boss, and W.M. Millington. 2000. Asthma - A Speaker's Kit for Public Health
Professionals. Centers for Disease Control and Prevention, National Center for Health Statistics,
Division of Environmental Hazards and Health Effects, Air Pollution and Respiratory Health
Branch. Available at: http://www.cdc.gov/nceh/asthma/speakit/cover.htm.
Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk. Oxford: Basil Blackwell.
Jones-Lee, M.W., G. Loonies, D. O'Reilly, and P.R. Phillips. 1993. The Value of Preventing
Non-fatal Road Injuries: Findings of a Willingness-to-pay National Sample Survey. TRY
Working Paper, WP SRC2.
Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M,
White WH. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer Society
Study of Particulate Air Pollution and Mortality. Special Report to the Health Effects Institute,
Cambridge MA, July 2000
Krupnick, A.J., W. Harrington and B. Ostro. 1990. "Ambient Ozone and Acute Health Effects -
Evidence From Daily Data." Journal of Environmental Economics and Management 18(1): 1-18.
Krupnick, A.J. and M.L. Cropper. 1992. "The Effect of Information on Health Risk
Valuations." Journal of Risk and Uncertainty 5(2): 29-48.
VII-91
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Krupnick, A.J., A. Alberini, M. Cropper, N. Simon, B. O'Brien, R. Goeree, and M. Heintzelman.
2000. Age, Health, and the Willingness to Pay for Mortality Risk Reductions: A Contingent
Valuation Survey of Ontario Residents. Resources for the Future Discussion Paper 00-37.
Krupnick, AJ. 2000. Personal Communication, December 5.
Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Kerry, F. Horak Jr., V.
Puybonnieux-Texier, P. Quenel, J. Schneider, R. Seethaler, J-C Vergnaud, and H. Sommer.
2000. Public-health Impact of Outdoor and Traffic-related Air Pollution: A European
Assessment. The Lancet, 356: 795-801.
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.
Laurence, J.A., W.A. Retzlaff, J.S. Kern, E.H.Lee, W.E. Hogsett, and D.A. Weinstein. 2000.
Predicting the regional impact of ozone and precipitation on the growth of loblolly pine and
yellow-poplar using linked TREGRO and ZELIG models. For Ecological Management. In
press.
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 SuspendedP articulates, Volumes I-IV. Prepared for US 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, US Environmental Protection Agency.
September.
McConnell, R., K. Berhane, F. Gilliland, S.J. London, H. Vora, E. Avol, WJ. Gauderman, H.G.
Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. Air Pollution and Bronchitic
Symptoms in Southern California Children with Asthma. Environmental Health Perspectives,
107(9): 757-760.
McConnell, R., K. Berhane, F. Gilliland, S.J. London, H. Vora, E. Avol, W.J. Gauderman, H.G.
Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. Air Pollution and Bronchitic
Symptoms in Southern California Children with Asthma. Environmental Health Perspectives.
107(9): 757-760.
VII-92
-------
Chapter VII: Benefit-Cost Analysis
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."
Environmental Research. 80(2 Pt 1): 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.
Ostro, B.D., MJ. Lipsett, J.K. Mann, H. Braxton-Owens, M.C. White, 1995. Air Pollution and
Asthma Exacerbations Among African-American Children in Los Angeles. Inhalation
Toxicology, 7:711-722.
Ostro, B.D. 1987. Air Pollution and Morbidity Revisited: a Specification Test. Journal of
Environmental Economics Management. 14: 87-98.
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, M.C. White. 1995. Air Pollution and
Asthma Exacerbations Among African-American Children in Los Angeles. Inhalation
Toxicology, 1:111-122.
Pew Environmental Health Commission. 2000. Attack Asthma: Why America Needs A Public
Health Defense System to Battle Environmental Threats. Johns Hopkins School of Public Health,
Baltimore, MD.
Pew Environmental Health Commission. 2000. Attack Asthma: Why America Needs A Public
Health Defense System to Battle Environmental Threats. John Hopkins School of Public Health,
Baltimore, MD.
Pope, C.A., JJI, 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." American Journal of Respiratory Critical Care Medicine 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" American Review of Respiratory Diseases
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 US EPA, Office of Policy Analysis. EPA-230-09-86-018. Washington, DC March.
VII-93
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A.
2000. The National Morbidity, Mortality and Air Pollution Study: Part II: Morbidity, Mortality
and Air Pollution in the United States. Research Report No. 94, Part n. Health Effects Institute,
Cambridge MA, June 2000.
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" American Journal of Respiratory Critical Care
Medicine 150: 1234-1242.
Schwartz, J. 1993. "Particulate Air Pollution and Chronic Respiratory Disease" Environmental
Research 62: 7-13.
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.
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." American Journal of Respiratory
Critical Care Medicine 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.
Taylor, C.R., K.H. Reichelderfer, and S.R. Johnson. 1993. Agricultural Sector Models for the
United States: Descriptions and Selected Policy Applications. Iowa State University Press:
Ames, IA.
Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. Summertime Haze Air
Pollution and Children with Asthma. American Journal of Respiratory Critical Care Medicine,
155: 654-660.
Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. Summertime Haze Air
Pollution and Children with Asthma. American Journal of Respiratory Critical Care Medicine,
155: 654-660.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks.
University of Chicago. Final Report for the US Environmental Protection Agency. January.
US 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, NC EPA report no. EPA/4521R-96-
013.
VII-94
-------
Chapter VII: Benefit-Cost Analysis
US Environmental Protection Agency, 1997e. Regulatory Impact Analyses for the Particulate
Matter and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule.
US 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).
US 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, NC December.
US 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.
US Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Health Statistics. 1999. National Vital Statistics Reports, 47(19).
US Environmental Protection Agency, 1997a. Regulatory Impact Analysis for P articulate
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, NC. July.
US Environmental Protection Agency, 1998a. Regulatory Impact Analysis for the NOx SIP
Call, FIP, and Section 126 Petitions. US EPA, Office of Air and Radiation. Washington, DC.
EPA-452/R-98-003. December. See EPA Air Docket A-96-56, VI-B-09.
US Environmental Protection Agency, 1998b. The Regional NOx SIP Call & Reduced
Atmospheric Deposition of Nitrogen: Benefits to Selected Estuaries, September.
US Environmental Protection Agency, 1999a. The Benefits and Costs of the Clean Air Act,
1990-2010. Prepared for US Congress by US EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC, November; EPA report no. EPA-410-R-99-001.
US Environmental Protection Agency, 1999b. 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, December.
US Environmental Protection Agency, 1999c. Regulatory Impact Analysis for the Final
Regional Haze Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
April 22, 1999.
US Environmental Protection Agency, 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).
VII-95
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
US Environmental Protection Agency, 1996b. 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, NC EPA report no. EPA/4521R-96-007.
US Environmental Protection Agency, 1996a. Proposed Methodology for PredictingPM2.5
from PM10 Values to Assess the Impact of Alternative Forms and Levels of the PMNAAQS.
Prepared by Terence Fitz-Simons, David Mintz and Miki Wayland (US 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).
US Department of Commerce, Economics and Statistics Administration. 1992. Statistical
Abstract of the United States, 1992: The National Data Book. 112th Edition, Washington, DC
US Environmental Protection Agency, 1994. Documentation for Oz-One Computer Model
(Version 2.0). Prepared by Mathtech, Inc., under Contract No. 68D30030, WA 1-29. Prepared
for US EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC.
August
US Environmental Protection Agency, 1993. External Draft, Air Quality Criteria for Ozone and
Related Photochemical Oxidants. Volume II. US EPA, Office of Health and Environmental
Assessment. Research Triangle Park, NC, EPA/600/AP-93/004b.3v.
US Department of Commerce, Bureau of Economic Analysis. BEA Regional Projections to
2045: Vol. 1, States. Washington, DC US Govt. Printing Office, July 1995.
US 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.
US Environmental Protection Agency, 2000a. Technical Support Document for the Heavy-Duty
Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements- Air
Quality Modeling Analyses. December 2000.
US Environmental Protection Agency, 2000b. Valuing Fatal Cancer Risk Reductions. White
Paper for Review by the EPA Science Advisory Board.
US 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/asthma/factsheets/asthma.htm
US Department of Health and Human Services, Centers for Disease Control and Prevention.
1999. "Forecasted State-Specific Estimates of Self-Reported Asthma Prevalence — United States,
1998." CDC-Morbidity and Mortality Weekly Report, 47: 1022-1025.
VII-96
-------
Chapter VII: Benefit-Cost Analysis
US Environmental Protection Agency, 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).
US Environmental Protection Agency, 1997b. The Benefits and Costs of the Clean Air Act,
1970 to 1990. Prepared for US Congress by US EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC
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.
Yu, O., L. Sheppard, T. Lumley, J. Q. Koenig, and G.G. Shapiro. 2000. Effects of Ambient Air
Pollution on Symptoms of Asthma in Seattle-Area Children Enrolled in the CAMP Study.
Environmental Health Perspectives, 108(12): 1209-1214.
Page A-97
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 HD Engine/Diesel Fuel rule using
the most comprehensive set of endpoints available. For some health endpoints, this meant using
a concentration-response (C-R) function that linked a larger set of effects to a change in
pollution, rather than using C-R functions for individual effects. For example, the minor
restricted activity day endpoint covers most of the symptoms used to characterize days of
moderate or worse asthma and shortness of breath. For premature mortality, we selected a C-R
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 C-R
function is expected to capture at least some of the mortality effects associated with exposure to
ozone. This ozone 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 HD Engine/Diesel Fuel 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, but illustrative of these issues and uncertainties. Supplemental
estimates included in this appendix include premature mortality associated with short-term
exposures to PM and ozone, acute respiratory symptoms in adults, shortness of breath in
asthmatic children, and occurrences of moderate or worse asthma symptoms in asthmatic adults.
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.
Tables "VTI-24 and VII-25 in Chapter Vn 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, four important parameters, the length and structure of
the potential lag in mortality effects, thresholds in PM health effects, discount rates, and the
income elasticity of WTP have been identified as key to the analysis, and are explored in this
appendix through the use of sensitivity analyses.
VII-98
-------
Chapter VII: Benefit-Cost Analysis
B. Supplementary Benefit Estimates
In the primary estimate, we use the Krewski, et al. (2000) study to provide the C-R
function relating premature mortality to long-term PM exposure. The primary analysis assumes
that 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 PM2 5 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.
Based on advice from the SAB (EPA-SAB-Council-ADV-99-012, 1999), we examine
ozone-related premature mortality as a supplemental estimate to avoid potential double-counting
of benefits captured by the Pope, et al. PM premature mortality endpoint.1* 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 HD Engine/Diesel Fuel 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, 2000). 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 SAB advice, 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.
tt While 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, et al. study does not include ozone in its model.
VII-99
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 (McDonell et al., 1998). 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). Incidences of shortness of breath (in African American asthmatics") are estimated using
a C-R function derived from Ostro, et al. (1995). The magnitude of these alternative calculations
confirms the magnitude of the asthma attack endpoint estimated from the Whittemore and Korn
(1980) study.
Occurrence of moderate or worse asthma symptoms are valued at $41 per incidence
(updated to 1999 dollars), 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. Incidences of shortness of breath are valued at $7 per incidence, based on the mean of
the median estimates from three studies of WTP to avoid a day of shortness of breath (Ostro et
al., 1995; Dickie et al., 1991; Loehman et al., 1979).
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, et al. 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 $80 million).
Estimates of the value for separate asthma endpoints are well under the estimate of the value of
all respiratory symptoms. All of these supplemental estimates support the set of endpoints and
assumptions chosen as the basis of the primary benefits estimate described in Chapter VII.
"Shortness of breath due to PM exposure is not necessarily limited to African-American asthmatics.
However, the Ostro et al. study was based on a sample of African-American children, who may be more sensitive to
air pollution than the general population so we chose not to extrapolate the findings to the general population.
VH-100
-------
Chapter VII: Benefit-Cost Analysis
Table VII-A-1. Supplemental Benefit Estimates for the Final HD Engine/Diesel Fuel Rule
for the 2030 Analysis YearA
Endpoint
Premature mortality (short-term exposures) (all ages)
Premature mortality (short-term exposures) (all ages)
Premature mortality in infant population
Any of 19 acute respiratory symptoms
Shortness-of-breath (African- American asthmatics, 7-12)
Moderate or Worse Asthma (adult asthmatics, 18-65)
Pollutant
PM
Ozone
PM
PMand
ozone
PM
PM
Avoided Incidence8
(cases/year)
2,600
500
30
4,987,600
39,000
182,500
Monetary Benefits0
(millions 1999$,
adjusted for growth in
real income)
$19,230
$3,430
$260
$790
<$1
$10
AOzone-related benefits estimated only for the Eastern U.S. due to ozone model performance issues (see chapter VII for details).
B Incidences are rounded to the nearest 100.
c 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 level of modeled PM25.
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
primary analysis, based on SAB advice, we assume that mortality occurs over a five year period,
vn-ioi
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
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 that alternative lag structures be explored as a
sensitivity analysis (EPA-SAB-COUNCIL-ADV-OO-OOl, 1999). 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, et al. and
Dockery, et al. studies, respectively'-'. However, neither the Pope, et al. or Dockery, et al studies
assumed any lag structure when estimating the relative risks from PM exposure. In fact, the
Pope, et al. and Dockery, et al. 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. $6 million per incidence, and are presented for both a 3 and 7 percent discount rate
over the lag period. The results using the primary 5-year lag are repeated here for comparison.
The table reveals that the length of the lag period is not as important as the distribution of
incidences within the lag period. A 15-year distributed lag with most of the incidences occurring
in the early years reduces monetary benefits less than an 8-year lag with all incidences occurring
at the eighth year. Even with an extreme lag assumption of 15 years, benefits are reduced by less
than half relative to the no lag and primary (5-year distributed lag) benefit estimates.
^ Although these studies were conducted for 8 and 15 years, respectively, the choice of the duration of the
study by the authors was not likely due to observations of a lag in effects, but is more likely due to the expense of
conducting long-term exposure studies or the amount of satisfactory data that could be collected during this time
period.
VII-102
-------
Chapter VII: Benefit-Cost Analysis
Table VII-A-2. Sensitivity Analysis of Alternative Lag Structures for PM-related
Premature Mortality
Lag
5 -year distributed
None
8-year
15 -year
15 -year distributed -
skewed early
15 -year distributed -
skewed late
Description
Primary estimate, incidences are
distributed with 25% in the 1st and 2nd
years, and 16.7% in the remaining 3
years.
Incidences all occur in the first year
Incidences all occur in the 8th year
Incidences all occur in the 15th year
Incidences are distributed with 30% in
the 1st year, 25% in the 2nd year, 15% in
the 3rd year, 6% in the 4th year, 4% in the
5th year, and the remainder 20%
distributed over the last 10 years.
Incidences are distributed with 4% in
the 1 1th year, 6% in the 12th year, 15% in
the 13th year, 25% in the 14th year, and
30% in the 15th year, with the remaining
20 % distributed over the first 10 years.
Monetary Benefit
Adjusted for
Growth in Real
Income (millions
1999$)
3%
discoun
t rate
$62,570
$65,820
$53,520
$43,510
$61,270
$47,200
7%
discoun
t rate
$58,770
$65,820
$40,990
$25,530
$56,530
$31,280
Percent of Primary
Estimate
3%
discoun
t rate
100%
105%
86%
70%
98%
75%
7%
discount
rate
94%
105%
66%
41%
90%
53%
2. PM Health Effect Threshold
The SAB advises 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). The most important health endpoint that would be
impacted by a PM threshold is premature mortality, as measured by the ACS/Krewski, et al.
(2000) C-R function. Krewski, et al. did not explicitly include a threshold in their analysis.
However, if the true mortality C-R relationship has a threshold, then Krewski, 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 threshold level. It is difficult to determine the size of the underestimate
without data on a likely threshold and without re-analyzing the Krewski, et al. data. Nevertheless,
it is illustrative to show at what threshold levels benefits are significantly affected.
VH-103
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
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 ACS/Krewski, et al. (2000) study.
The distribution of premature mortality incidences in Figure A-l indicate that
approximately 90 percent of the premature mortality related benefits of the final HD
Engine/Diesel Fuel rule are due to changes in PM concentrations occurring above 10 |ig/m3, and
around 80 percent are due to changes above 12 |ig/m3, the lowest observed level in the
ACS/Krewski, et al. study. Over 60 percent of avoided incidences are due to changes occurring
above 15 |ig/m3.
9,000
fO
8,000 -
7,000 -
S 6,000 -
| 5,000 -
js 4,000 -
"S 3,000 -
'3 2,000 H
^ 1,000 H
o
0 5 10 15 20 25 30 35 40
Assumed Effect Threshold (Annual Mean PM2.5 (ug/m3))
45
Figure VII-A-l. Impact of PM Health Effects Threshold on Avoided Incidences of
Premature Mortality Estimated with the American Cancer Society/Krewski, et al. (2000) C-
R Function
vn-104
-------
Chapter VII: Benefit-Cost Analysis
3. Income Elasticity of Willingness to Pay
As discussed in section C. 1 of Chapter VII, our primary estimate of monetized benefits
accounts for growth in real GDP per capita by adjusting the WTP for individual endpoints based
on the primary estimate of the adjustment factor for each of the categories (minor health effects,
severe and chronic health effects, premature mortality, and visibility). We examine how
sensitive the primary estimate of total benefits is to alternative estimates of the income
elasticities. The results of this sensitivity analysis are presented in Table VII-A-3. Note that the
alternative elasticities and adjustment factors on which this sensitivity analysis is based are
presented in Tables VII-11 and VII-12, respectively.
Consistent with the impact of mortality on total benefits, the adjustment factor for
mortality has the largest impact on total benefits. The value of mortality ranges from 81 percent
to 150 percent of the primary estimate based on the lower and upper sensitivity bounds on the
income adjustment factor. The effect on the value of minor and chronic health effects is much
less pronounced, ranging from 93 percent to 111 percent of the primary estimate for minor
effects and from 88 percent to 110 percent for chronic effects.
Table VII-A-3. Sensitivity Analysis of Alternative Income Elasticities
Benefit Category
Minor Health Effect
Severe and Chronic
Health Effects
Premature Mortality
VisibilityA
Total Benefits
Lower Sensitivity
Bound
$510
$2,120
$50,680
—
$56,980
Primary
$550
$2,420
$62,580
$3,260
$70,360
Upper Sensitivity
Bound
$610
$2,670
$94,140
—
$97,830
1 No range was applied for visibility because no ranges were available in the current published literature.
vn-ios
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Chapter VII. Appendix A 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 US Environmental
Protection Agency, Office of Air Quality Planning and Standards; Research Triangle Park, NC.,
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.
Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M,
White WH. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer Society
Study of Particulate Air Pollution and Mortality. Special Report to the Health Effects Institute,
Cambrid.ge MA, July 2000
Laurence, J.A., W.A. Retzlaff, J.S. Kern, E.H.Lee, W.E. Hogsett, and D.A. Weinstein. 2000.
Predicting the regional impact of ozone and precipitation on the growth of loblolly pine and
yellow-poplar using linked TREGRO and ZELIG models. For Ecological Management. In
press.
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." American Journal of Public Health 81(6): 694-702.
VH-106
-------
Chapter VII: Benefit-Cost Analysis
Pope, C.A., MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer and C.W.
Heath. 1995. " Particulate air pollution as a predictor of mortality in a prospective study of U.S.
adults.' American Journal of Respiratory Critical Care Medicine 151(3): 669-674.
Rowe, R.D. and L.G. Chestnut. 1986. Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis — Executive Summary. Prepared for US Environmental Protection Agency, Office of
Policy Analysis. Prepared by Energy and Resource Consultants, Inc. Washington, DC. EPA-
230-09-86-018. March.
Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A.
2000. The National Morbidity, Mortality and Air Pollution Study: Part II: Morbidity, Mortality
and Air Pollution in the United States. Research Report No. 94, Part n. Health Effects Institute,
Cambridge MA, June 2000.
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."
American Journal of 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.
VH-107
-------
Chapter VIM: Regulatory Flexibility Analysis
Chapter VIII: Regulatory Flexibility Analysis
This chapter presents our Final Regulatory Flexibility Analysis (FRFA) which evaluates
the impacts of the heavy-duty engine standards and diesel fuel sulfur standards on small
businesses. Prior to issuing our proposal last June, we analyzed the potential impacts of our
program on small businesses. As a part of this analysis, we convened a Small Business
Advocacy Review (SBAR) Panela, 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 the proposed engine and fuel standards. After the proposal
was published in the Federal Register, we held five public hearings around the country to gather
feedback on it. The small business provisions of today's action reflect revisions to the proposed
program based upon updated analyses as well as comments heard at the public hearings and those
submitted in writing during the public comment period.
A. Regulatory Flexibility Analysis
EPA has decided to prepare a Final Regulatory Flexibility Analysisb (FRFA) for today's
final rule. In accordance with section 603 of the RFA, EPA prepared an initial regulatory
flexibility analysis (IRFA) for the proposed rule and convened a Small Business Advocacy
Review Panel to obtain advice and recommendations of representatives of the regulated small
entities in accordance with section 609(b) of the RFA (see 65 FR 35541, June 2, 2000). A
detailed discussion of the Panel's advice and recommendations is found in the Panel Report
contained in the docket for this rulemaking. A summary of the Panel's recommendations is
presented at 65 FR 35541. The key elements of the FRFA include:
- the need for, and objectives of, the rule;
- the significant issues raised by public comments on the Initial RFA (IRFA), a summary
of the Agency's assessment of those issues, and a statement of any changes made to the
proposed rule as a result of those comments;
- the types and number of small entities to which the rule will apply;
a Including representatives from the Small Business Administration, White House Office of Management
and Budget, and EPA.
vm-i
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
- the reporting, recordkeeping, and other compliance requirements of the rule, including
the classes of small entities that will be affected and the type of professional skills
necessary to prepare the report or record;
- the steps taken to minimize the significant impact on small entities consistent with the
stated objectives of the applicable statute, including a statement of the factual, policy and
legal reasons why the Agency selected the alternatives we did, and why other significant
alternatives to the rule which affect the impact on small entities were rejected.
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.
In developing this rule, we concluded that the proposed heavy-duty engine and diesel fuel
sulfur standards would likely have a significant impact on a substantial number of small entities.
As discussed in more detail below, we identified several categories of small entities associated
with diesel fuel production or distribution. To our knowledge, no manufacturers of heavy-duty
engines meet the Small Business Administration (SB A) definition of a small business.
To comply with the requirements of the RFA, we quantified the economic impacts on the
identified small entities. Using the methodology discussed in Chapter V, we determined the
refinery costs for average size refineries and small refiners to produce low sulfur diesel fuel.
Chapter V also contains our estimation of diesel distribution costs for the entire distribution
system, including pipeline and tank wagon deliveries.
B. Need for and Objectives of the Rule
The preamble to this rule fully discusses the need for and objectives of this rule. As
discussed in detail in Chapter n of this RIA, emissions from heavy-duty vehicles contribute
greatly to a number of serious air pollution problems, and would have continued to do so into the
future absent further controls to reduce these emissions. Although the air quality problems
caused by diesel heavy-duty vehicles are challenging, we believe they can be resolved through
the application of high-efficiency emissions control technologies. Based on the Clean Air Act
requirements, we are setting stringent new emission standards that will result in the use of these
diesel exhaust emission control devices. We are also finalizing changes to diesel fuel sulfur
standards in order to enable these high-efficiency technologies. In consideration of the impacts
that sulfur has on the efficiency, reliability, and fuel economy impact of diesel engine exhaust
vm-2
-------
Chapter VIM: Regulatory Flexibility Analysis
emission control devices, we believe that controlling the sulfur content of highway diesel fuel to
the 15 ppm level is necessary, feasible and cost effective. The standards will result in substantial
benefits to public health and welfare and the environment through significant reductions in
emissions of nitrogen oxides, particulate matter, nonmethane hydrocarbons, carbon monoxide,
sulfur oxides, and air toxics.
C. Summary of Significant Public Comments on the IRFA
This FRFA addresses the issues raised by public comments on the IRFA, which was part
of the proposal of this rule. EPA received many comments from small refiners and others
pertaining to the options for hardship relief described in the NPRM. In general, many small
refiners commented on the financial difficulty their refinery would face in complying with the
proposed diesel sulfur program, and encouraged EPA to provide hardship relief. Many small
refiners acknowledged that there was not one single hardship relief option to best suit the needs
of all small refiners, and thus supported a menu of options. Section IV.C of the preamble
discusses the three hardship relief options available to small refiners under today's program.
These three options are based on concepts which were considered by the SBAR Panel and on
which we requested and received comment in the proposal. A summary of the comments
pertaining to regulatory alternatives for small refiners, and our response to them, is contained in
the Response to Comments document contained in the docket.
D. Types and Number of Small Entities To Which The Rule
Will Apply
Today's action will establish new heavy-duty engine standards and require low sulfur
highway diesel fuel. It will primarily affect manufacturers of heavy-duty engines, petroleum
refiners that produce diesel, and certain distributors of diesel fuel. As mentioned above, we are
not aware of any heavy-duty engine manufacturers that would be defined as a small business
under the SBA regulations ( 13 CFR Part 121). Although most refining companies are not
considered small businesses, we have identified several refining companies that do appear to
qualify under the applicable SBA definition. In addition, this rule may impact diesel fuel
distributors and marketers-of which several thousand appear to be small businesses. Table
Vin-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 (NAIC) systems. In this table, all the
industry categories listed below the "Petroleum Refiners" category have some role in either
distributing and/or marketing highway diesel fuel.
vm-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000
EPA420-R-00-026
Table VIII-1. Industries Containing Small Businesses Potentially Affected by the Low
Sulfur Diesel Fuel Rule
Industry
Petroleum Refiners
Refined Petroleum Pipelines
Petroleum Marketers and
Distributors
Other Terminals: Special
Warehousing and Storage
Fuel Oil Dealers
Petroleum Retailers
NAICS?
Codes
324110
486910
422710
422720
493110
493190
454311
447110
447190
SIC
Codes
2911
4613
5171
5172
4226
5983
5541
Defined by SBA as a Small
Business if:d
<1500 employees
corporate-wide
<1500 employees
corporate-wide
<100 employees
corporate-wide
<$18. 5 million
for the parent corporation
<$9 million
for the parent corporation
<6.5 million
for the parent corporation
The types and number of small entities to which the low sulfur diesel fuel rule will apply
are described in Table VIII-2 below. Under this rule, the only small entities that may be
significantly affected are small refiners, since they will have to invest in desulfurization
technology to produce low sulfur highway diesel fuel. We estimate that small refiners produce
approximately five percent of all highway diesel fuel in the U.S.
b North American Industry Classification System
0 Standard Industrial Classification System
d 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.
vm-4
-------
Chapter VIM: Regulatory Flexibility Analysis
Table VIII-2. Types and Number of Small Entities to Which the Diesel Sulfur Program
Will Apply
Type of Small Entity
Small Refiners
Small Diesel Marketers and
Distributors
Number of Companies Affected by
the Low Sulfur Diesel Fuel Rule
Approximately 24
Several Thousand
1. Small Refiners
We have identified several refiners that produce highway diesel fuel and meet the SBA
definition for a small petroleum refiner (Standard Industrial Classification (SIC) 2911), that is,
having 1500 or fewer employees corporate-wide. These refiners, approximately 24 out of the
approximately 124 refineries which produce highway diesel (there are about 158 refineries in the
U.S. today), operate 27 refineries (i.e., some small refiners own and operate more than one
refinery).
Some small refiners indicated that they will have greater difficulty than larger refiners in
complying with the diesel sulfur standard due to such factors as limited operational flexibility,
lack of access to alternate crude oil feedstocks, limited availability of new sulfur reduction
equipment, poorer economies of scale, or difficulty in raising capital to finance projects. Based
on these discussions and analyses, the Panel and we agree that small refiners would likely
experience a significant and disproportionate financial hardship in reaching the objectives of our
diesel fuel sulfur program. However, the Panel also noted that the burden imposed upon the
small refiners by our sulfur requirements varied from refiner to refiner and could not be
alleviated with a single provision. In addition, the small refiners strongly supported a "menu" of
compliance options. We agree with the Panel and are offering qualifying small refiners three
options to choose from in moving toward compliance with the low sulfur diesel fuel
requirements.
2. Small Distributors/Marketers of Highway Diesel Fuel
The low sulfur diesel fuel rule contains certain downstream compliance and enforcement
provisions, for all parties in the diesel fuel distribution system downstream of the refinery gate, to
prevent 1) contamination of highway diesel fuels with fuels containing higher levels of sulfur and
2) misfueling of motor vehicles with high sulfur fuels.
vm-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Under this rule, distributors and retailers may choose to handle 500 ppm diesel fuel, 15
ppm diesel fuel, or both (as permitted under the temporary compliance option and small refiner
hardship provisions described in the preamble). However, distributors and marketers will have
to segregate low sulfur diesel fuel from other distillates just as they do today with 500 ppm diesel
fuel. Retailers and wholesale purchaser-consumers will be responsible for ensuring that only low
sulfur diesel fuel is sold for use in model year 2007 and later heavy-duty vehicles. Under the
temporary compliance option and small refiner hardship provisions, where two grades of
highway diesel fuel are allowed for the initial years of the program, some distributors and
marketers may voluntarily decide (presumably based on economics) to add tankage or make
additional modifications to accommodate two grades of highway diesel fuel. We have taken
such costs into account in our diesel fuel cost analysis (described in more detail in Chapter V).
The low sulfur diesel fuel rule also includes a product downgrading restriction that is
designed to discourage the intentional downgrading of 15 ppm diesel fuel to 500 ppm diesel fuel
in the distribution system during the initial years of the program when the optional compliance
provision is in effect. All parties in the distribution system downstream of the refinery gate are
subject to this provision, except for those retailers that offer for sale and wholesale purchaser-
consumers that use 15 ppm fuel either as the only grade of diesel or in addition to 500 ppm diesel
(i.e., the only retailers and wholesale purchaser-consumers that are subject to this requirement are
those that offer for sale or use only 500 ppm diesel but not 15 ppm diesel). Under this
restriction, the volume of 15 ppm fuel that may be downgraded to 500 ppm highway diesel fuel
at each point in the distribution system (downstream of the refinery gate) is limited to not more
than 20 percent on an annual basis.6 Each party in the distribution system subject to this
provision will be required to meet this requirement separately, based on the amount of 15 ppm
fuel it receives and transfers/sells to the next party (or end user, in the case of retailers and
wholesale purchaser-consumers) on an annual basis.
However, this provision should have no meaningful burden on downstream entities. It is
only intended to prevent abuse and not intended to constrain any normal business operations.
Furthermore, it does not require the addition of any new recordkeeping or reporting requirements
beyond those required of the rest of the program.
E. Projected Costs of the Diesel Sulfur Standards
The average costs for a small refiner to produce low sulfur highway diesel fuel are
described below:
e The downgrading restriction applies only to 15 ppm downgraded to 500 ppm highway diesel fuel but not
to 15 ppm downgraded to off-highway diesel fuel.
vm-6
-------
Chapter VIM: Regulatory Flexibility Analysis
A. capital cost: $14 million
- operating cost: $500 thousand per year
per-gallon cost (assuming a seven percent ROI before taxes): 5.0 cents/gallon
In comparison, the average non-small refineries capital cost is $52 million with operating
costs of $9.6 million per year. The per-gallon cost of average non-small refineries is 4.2 cents
per treated diesel volume. Our methodology, including a comparison to recent industry
estimates, is described in Chapter V.
As discussed in more detail in Chapter V.C., above, we also analyzed the increase in
distribution costs associated with the low sulfur diesel fuel program. We estimate that
distribution costs will increase by 0.5 cents per gallon of highway diesel fuel supplied when the
sulfur requirements are fully effective beginning in the year 2010. During the initial years of the
program, we estimate that there will be an increase in distribution costs of roughly $1 billion for
new storage tanks to handle two grades of highway diesel fuel (500 ppm and 15 ppm). The total
distribution costs during the initial years of the low sulfur diesel fuel program equate to
approximately 1.1 cents per gallon of the total volume of all highway diesel fuel supplied.
F. Projected Reporting, Recordkeeping, and Other
Compliance Requirements of the Rule
The low sulfur diesel fuel program contains several hardship options to assist small
refiners in producing low sulfur diesel fuel. Under these options, small refiners may be subject
to new reporting and recordkeeping requirements to help ensure compliance with the options and
the integrity of the low sulfur diesel fuel as it moves from the refinery gate to the retail outlet.
However, we believe the benefits of the options will far outweigh any burdens imposed by their
associated recordkeeping and reporting requirements.
The low sulfur diesel fuel program does not impose any new reporting requirements for
small diesel marketers or distributors. However, this program does impose new record keeping
requirements for such parties, specifically product transfer documents that track transfers of
diesel fuel. Such transfer records are currently maintained by most parties for business and/or
tax reasons. In addition, the record keeping requirements for downstream parties are fairly
consistent with those in place today under other EPA fuel programs, including the current
highway diesel fuel program. Therefore, we expect that the new record keeping requirements for
downstream parties will not impose a significant burden.
vm-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
1. Registration Reports
Refiners that are either currently producing or supplying highway diesel fuel, or that
expect to do so by June 1, 2006, must register with us. The specific information required in the
registration is described in section VII of the preamble as well as the regulations.
In addition to the basic registration requirements above, a refiner seeking status as a small
refiner needs to apply for this status as a part of their registration and provide the average number
of employees for all pay periods from January 1, 1999 to January 1, 2000, for the company, all
parent companies, and all subsidiaries or joint ventures. The application also must include which
small refiner option the refiner expects to use at each of its refineries.
2. Pre-Compliance Reports
All refiners (including small refiners) and importers must report to us regarding their
progress toward compliance with the highway diesel fuel sulfur standards adopted today. These
pre-compliance reports are due each May 31 from 2003 through 2005. We will maintain the
confidentiality of information submitted in pre-compliance reports. We will present generalized
data from the reports on a PADD basis in annual reports following the receipt of each year's pre-
compliance reports. The specific information required in the compliance reports is described in
Section VII of the preamble as well as the regulations.
In addition to the information required for all refiners, small refiners must provide
additional information in their pre-compliance reports. The information required varies
according to which small refiner option the refiner will be using, as discussed in Section IV.C of
the preamble.
3. Annual Compliance Reports
After the highway diesel sulfur requirements begin June 1, 2006, refiners and importers
will be required to submit annual compliance reports that demonstrate compliance with the
requirements of this final rule. The first annual compliance report will be due by the end of
February 2007 (for the period of June 1, 2006 through December 31, 2006) and would be
required annually through February 2011. The specific information required in a refiner's annual
compliance report is described in section VU of the preamble as well as the regulations. As with
pre-compliance reports, in their annual compliance reports, small refiners must also supply
additional information related to the small refiner option they are using.
vm-8
-------
Chapter VIM: Regulatory Flexibility Analysis
4. Product Transfer Documents (PTDs)
In the low sulfur diesel fuel program, refiners as well as diesel distributors and marketers
will be required to keep records primarily consisting of product transfer documents (PTDs),
which document a party's diesel fuel transfers. Such records are already maintained by most
parties for business or tax purposes. We are adopting the proposed requirements that refiners,
importers, distributors, and marketers provide information on commercial PTDs that identifies
diesel fuel distributed for use in motor vehicles and states that the fuel is compliant with the 15
ppm sulfur standard. Since the low sulfur diesel fuel rule adopts provisions for the production
and sale of 500 ppm sulfur diesel fuel for use in pre-2007 model year vehicles, the rule also
adopts provisions which require PTDs to identify such fuel and state that its use in motor
vehicles is restricted to pre-2007 motor vehicles/ We believe this additional information on
commercial PTDs is necessary to 1) prevent the commingling of highway diesel fuel with high
sulfur distillate products, 2) avoid contamination of 15 ppm highway diesel fuel with 500 ppm
highway diesel fuel, and 3) prevent the misfueling of model year 2007 and later vehicles with any
fuel having a sulfur content greater than 15 ppm. To discourage the intentional downgrading of
15 ppm diesel fuel to 500 ppm diesel fuel in the distribution system during the initial years of the
program, PTDs must also include the volume of each fuel grade (15 ppm and 500 ppm)
delivered.8
5. Recordkeeping Requirements
Refiners that produce (or importers that import) both 500 ppm highway diesel fuel and 15
ppm highway diesel fuel under the temporary compliance option or any hardship program
(including small refiners), or that produce only 15 ppm sulfur content diesel fuel and that wish to
generate credits (including early credits), must maintain records for each batch of highway diesel
fuel produced, of the batch designations and the batch volumes. The refiner must maintain
records regarding credit generation, use, transfer, purchase, or termination. The specific
recordkeeping information required under the low sulfur diesel fuel program is described in
section VII of the preamble as well as the regulations.
Small refiners approved for temporary hardship relief due to extreme unforseen
circumstances or extreme financial hardship must include certain information in their application
for relief. The required information, and the factors we will consider in determining what relief,
if any, is appropriate, are discussed in Section IV.B.3 of the preamble. Such refiners will also
f Such fuel can also be used in nonroad vehicles, whose fuel is currently unregulated.
g As discussed in Section VII of the preamble, we are restricting the volume of 15 ppm fuel that can be
downgraded to 500 ppm highway diesel fuel at each point in the distribution system to not more than 20 percent on
an annual basis.
vm-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
have reasonable recordkeeping and reporting requirements, which will be determined on a
case-by-case basis.
6. Diesel Fuel Pump Labeling
The low sulfur diesel fuel rule also adopts pump labeling requirements for retailers and
wholesale purchaser-consumers similar to the requirements that we proposed, but with
modifications to account for the availability of diesel fuel subject to the 500 ppm sulfur standard
for use in pre-2007 motor vehicles. For any multiple-fuel program, like the low sulfur diesel fuel
program with its compliance flexibility option and hardship provisions, clearly labeling diesel
fuel pumps is vital for end users to distinguish between the two grades of fuel. We received
comments on the proposal that concurred with our assessment that pump labels, in conjunction
with vehicle labels, would help to prevent misfueling of motor vehicles with high sulfur diesel
fuel. The text of the labels appears in Section Vn of the preamble; the specific requirements for
label size and appearance are found in the regulatory language for this rule.
G. Regulatory Alternatives
For today's action, we have structured a selection of temporary flexibilities for qualifying
small refiners, both domestic and foreign, based on the factors described below. Generally, we
structured these provisions to address small refiner hardship while expeditiously achieving air
quality benefits and ensuring that the low sulfur diesel fuel coincides with the introduction of
2007 model year diesel vehicles. First, the compliance deadlines in the program, combined with
flexibility for small refiners, will quickly achieve the air quality benefits of the program, while
helping to ensure that small refiners will have adequate time to raise capital for new or revamped
equipment. Second, we believe that allowing time for refinery sulfur-reduction technologies to
be proven out by larger refiners before small refiners have to put them in place will likely allow
for lower costs of these improvements in desulfurization technology (e.g., better catalyst
technology or lower-pressure hydrotreater technology). Third, providing small refiners more
time to comply will increase the availability of engineering and construction resources. Since
most large and small refiners must install additional processing equipment to meet the sulfur
requirements, there will be a tremendous amount of competition for technology services,
engineering manpower, and construction management and labor. Finally, because the gasoline
and diesel sulfur requirements will occur in approximately the same time frame, small refiners
that produce both fuels will have a greater difficulty than most other refiners in securing the
necessary financing. Hence, any effort that increases small refiners' ability to stagger
investments for low sulfur gasoline and diesel will facilitate compliance with the two programs.
These factors are discussed further in Section IV.C of the preamble.
Providing these options to assist small refiners experiencing hardship circumstances
enables us to go forward with the 15 ppm sulfur standard beginning in 2006. Without this
VHI-10
-------
Chapter VIM: Regulatory Flexibility Analysis
flexibility, the benefits of the 15 ppm standard would possibly not be achieved as quickly. By
providing temporary relief to those refiners that need additional time, we are able to adopt a
program that expeditiously reduces diesel sulfur levels in a feasible manner for the industry as a
whole. In addition, we believe the volume of diesel that will be affected by this hardship
provision is marginal. We estimate that small refiners contribute approximately five percent of
all domestic highway diesel fuel production.
All refiners producing highway diesel fuel are able to take advantage of the temporary
compliance option discussed in Section IV. A of the preamble. Refiners that seek and are granted
small refiner status may choose from the following three options under the diesel sulfur program.
These three options have evolved from concepts on which we requested and received comment
in the proposal. In most cases, we believe that small refiners will find these options preferable to
either the broader diesel fuel temporary compliance option or the Geographic Phase-in Area
(GPA) provisions (as applicable) discussed in the preamble.
500 ppm Option. A small refiner may continue to produce and sell diesel fuel meeting the
current 500 ppm sulfur standard for four additional years, through May 31, 2010,
provided that it reasonably assures the existence of sufficient volumes of 15 ppm fuel in
the marketing area(s) that it serves.
Small Refiner Credit Option. A small refiner that chooses to produce 15 ppm fuel prior
to 2010 may generate and sell credits to other refiners under the broader diesel temporary
compliance option. Since a small refiner has no requirement to produce 15 ppm fuel
under this option, any volume of fuel it produced at or below 15 ppm sulfur will qualify
for generating credits.
Diesel/Gasoline Compliance Date Option. For small refiners that are also subject to the
Tier 2/Gasoline sulfur program (40 CFR Part 80), the refiner may choose to extend by
three years the duration of its applicable interim gasoline standards, provided that it also
produces all its highway diesel fuel at 15 ppm sulfur beginning June 1, 2006.
1. 500 ppm Option
The 500 ppm Option is available for any refiner that qualifies as a small refiner. Under
this option, small refiners could continue selling highway diesel fuel with sulfur levels meeting
the current 500 ppm standard for four additional years, provided that they supply information
showing that sufficient alternate sources of 15 ppm diesel fuel in their market area will exist for
fueling new heavy-duty highway vehicles. Under this option, small refiners could supply current
500 ppm highway diesel fuel to any markets for use only in vehicles with older (pre-2007)
technology through May 31, 2010. In other words, small refiners that choose this option may
delay production of highway diesel fuel meeting the 15 ppm standard for four years.
vm-ii
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
This 500 ppm option for small refiners is similar to the option provided to all refiners in
the temporary compliance option described in Section IV. A of the preamble in that it allows a
refiner to continue to produce and sell current 500 ppm fuel for a period of time. However, this
option differs from the broader temporary compliance option in that small refiners could produce
and sell 100 percent of their highway fuel at 500 ppm without needing to buy credits. By way of
contrast, under the broader program, refiners would need to buy credits in order to produce any
volume of 500 ppm fuel over 20 percent of their total highway diesel production.
2. Small Refiner Credit Option
We believe that the relative difficulty for small refiners to comply with today's program
warrants compliance flexibility for these refiners. At the same time, we want to encourage all
refiners to produce 15 ppm sulfur as early and in as many geographic areas as possible. As an
incentive for small refiners to invest in desulfurization capacity, those that choose to produce 15
ppm fuel earlier than they would need to under the 500 ppm option-before 2010-could generate
credits for each gallon of diesel fuel produced as meeting the 15 ppm standard. They could then
sell these credits to other refiners for use in the broader diesel fuel program described above in
Section IV. A, helping to offset some of the cost of producing the cleaner fuel.
Under this option, credits could be generated based on the volume of any diesel fuel
meeting the 15 ppm standard. The refiner could sell its remaining highway diesel fuel under the
500 ppm option above.
3. Diesel/Gasoline Compliance Date Option
The Tier 2/Gasoline Sulfur program included a special provision that applies for refiners
that qualify as small refiners (40 CFR Part 80, Subpart H). Under that program, each small
refiner is assigned an interim gasoline sulfur standard for each of its refineries. This interim
standard for each refinery is established based on the baseline sulfur level of that refinery. The
standards are designed to require each small refiner to either make a partial reduction in their
gasoline sulfur levels or, if they already produce low sulfur fuel, to maintain their current levels.
The gasoline interim program lasts for four years, 2004 through 2007, and the refiner can apply
for an extension of up to two years. After the interim program expires, small refiners need to
produce the same low sulfur gasoline as other refiners.
Today's diesel sulfur program takes effect in the same time frame as the small refiner
interim program for gasoline sulfur. To avoid the need for simultaneous investments in both
gasoline and diesel fuel desulfurization, several small refiners subject to both programs raised the
concept of allowing those investments to be staggered in time. Because of the relative difficulty
small refiners will face in financing desulfurization projects, especially for both diesel and
gasoline desulfurization in the same time frame, we agree that this concept has merit and have
vm-12
-------
Chapter VIM: Regulatory Flexibility Analysis
adopted it for this rule. Under this concept, small refiners could extend the duration of their
gasoline sulfur interim standards and, thus, potentially postpone some or all of their gasoline
desulfurization investments while they work to achieve the diesel sulfur standards "on time" in
2006. To the extent that small refiners choose this Diesel/Gasoline Compliance Date option, this
provision will benefit the overall diesel program because it will increase the availability of 15
ppm diesel fuel in the small refiner's market area.
Specifically, this option provides that a small refiner can receive a three-year extension of
a refinery's interim gasoline standard, until January 1, 2011, if it meets two criteria: 1) it
produces both gasoline and diesel fuel at a refinery and chooses to comply with the 15 ppm
diesel fuel sulfur standard by June 1, 2006 for all its highway diesel production at that same
refinery, and 2) it produces a minimum volume of 15 ppm fuel at that refinery that is at least 85
percent of the average volume of highway diesel fuel that it produced at that refinery during
calendar years 1998 and 1999. We believe that it is very important that the extension of a small
refiner's interim low sulfur gasoline standard be linked to a substantial environmental benefit
from the production of low sulfur diesel fuel in 2006. We have established a minimum volume
requirement to prevent the Diesel/Gasoline Compliance Date option from applying in situations
where a refiner changes its refinery product slate to produce very little highway diesel fuel-even
though this production is at a 15 ppm sulfur level-and yet receives an extension of its interim
gasoline sulfur standard. We believe the 85 percent level is sufficient to reflect a substantial
investment in desulfurization technology. At the same time the 85 percent level should allow
for any reasonable variation in production of highway diesel fuel that would be expected to occur
in typical situations between now and 2006, particularly given the continued growth of the
highway diesel market. Again, the three-year extension of the gasoline interim program is to
allow small refiners to stretch out their capital investments while increasing the quantity of 15
ppm fuel being produced. We expect that small refiners using this option will make a
substantive capital investment in diesel desulfurization and have thus set this minimum 15 ppm
diesel volume limit.
The Tier 2/Gasoline Sulfur program includes a general hardship provision for which
refiners may apply. (Today's program also includes a similar provision). Depending on the
nature of its hardship, a small refiner that applies for this general hardship provision under the
gasoline program could be granted a "tailor-made" interim gasoline sulfur program different
from the "default" program established in the rule. If such a small refiner were then to be
covered by today's diesel fuel requirements and chose this Diesel/Gasoline Compliance Date
option, we would allow it an extension of its special interim program for gasoline (as established
under the general hardship provision) for three years beyond the scheduled end date (although no
later than December 31, 2010) so long as it met the 15 ppm diesel fuel standard in 2006.
vm-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
4. Relationship of the Options to Each Other
By definition, since a small refiner must produce 100 percent of its highway diesel as 15
ppm under the Diesel/Gasoline Compliance Date option, that option is not compatible with either
the 500 ppm option or the Small Refiner Credit option. Thus a refiner choosing the
Diesel/Gasoline Compliance Date option may not choose either of the other two options.
However, the 500 ppm option and the Small Refiner Credit option are compatible with each
other, and so a refiner can choose either or both of these options.
H. Other Relevant Federal Rules Which May Duplicate,
Overlap, or Conflict with the Low Sulfur Diesel Fuel
Rule
The heavy-duty engine and diesel fuel sulfur standards that we are finalizing are similar,
in many respects, to existing regulations; in some cases, these regulations are replacing earlier
requirements with more stringent requirements for refiners and engine manufacturers. We also
note that more stringent diesel sulfur standards will likely require many refiners to obtain permits
from state and local air pollution control agencies under the CAA's New Source Review program
prior to constructing the desulfurization equipment needed to meet the standards. However, we
are not aware of any area where the new regulations will directly duplicate or conflict with
existing federal, state, or local regulations.
On the other hand, several small refiners commented that the low sulfur diesel fuel
program overlaps with the low sulfur gasoline program. Specifically, they indicated that they
will be making substantial investments to comply with the low sulfur gasoline program (which
phases in from January 1, 2004 through December 31, 2007) while concurrently planning and
investing to comply with the low sulfur diesel fuel program. One hardship option within the low
sulfur diesel fuel program was specifically designed for those small refiners that produce both
gasoline and highway diesel fuel. The intent of this option, which is described in more detail
below, is to help small refiners spread out their gasoline and diesel fuel desulfurization
investments.
vm-14
-------
Chapter IX: Sulfur Control in Alaska & Territories
Chapter IX: Sulfur Control in Alaska & Territories
A. What is the Authority For Exemptions?
Section 21 l(c) of the Clean Air Act allows EPA to regulate fuels where emission
products of the fuel either: 1) cause or contribute to air pollution that reasonably may be
anticipated to endanger public health or welfare, or 2) 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 such a regulation to be promulgated. EPA's authority to establish emissions
standards for heavy-duty engines and vehicles is based on section 202(a)(3), directing EPA to
establish regulations under section 202(a) that produce the greatest achievable reductions,
considering various factors such as technological feasibility, cost, and lead-time. Under section
21 l(c), EPA is required to consider available scientific and economic data, including a cost
benefit analysis comparing emission control devices or systems which are or will be in general
use and require the control or prohibition with emission control devices or systems which are or
will be in general use and do not require the control or prohibition.
Sections 21 l(i) and 21 l(g) of the Clean Air Act restrict the use of high-sulfur diesel fuel
in highway vehicles. Section 21 l(i)(l) prohibits the manufacture, sale, supply, offering for sale
or supply, dispensing, transport, or introduction into commerce of motor vehicle (highway) diesel
fuel which contains a concentration of sulfur in excess of 0.05 percent by weight (500 ppm), or
which fails to meet a minimum cetane index of 40, beginning October 1, 1993. Section 21 l(i)(2)
required EPA to promulgate regulations to implement and enforce the requirements of paragraph
(1), and authorized EPA to require that diesel fuel not intended for highway vehicles be dyed in
order to segregate that fuel from highway vehicle diesel fuel. Section 21 l(i)(4) provides that the
states of Alaska and Hawaii may seek an exemption from the requirements of subsection 21 l(i)
in the same manner as provided in section 325a of the Act. Section 21 l(g)(2) of the Act prohibits
the fueling of highway vehicles with diesel fuel containing sulfur in excess of 500 ppm or which
fails to meet a cetane index of 40 beginning October 1, 1993.
a Section 211(i)(4) mistakenly refers to exemptions under Section 324 of the Act ("Vapor Recovery for
Small Business Marketers of Petroleum Products"). The proper reference is to section 325, and Congress clearly
intended to refer to section 325, as shown by the language used in section 21 l(i)(4), and the United States Code
citation used in section 806 of the Clean Air Act Amendments of 1990, Public Law No. 101-549. Section 806 of
the Amendments, which added paragraph (i) to section 211 of the Act, used42U.S.C. 7625-1 as the United States
Code designation, the proper designation for section 325 of the Act. Also see 136 Cong. Rec. S17236 (daily ed.
October 26, 1990) (statement of Sen. Murkowski).
LX-1
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Section 325 of the Act provides that upon the request of Guam, American Samoa, the
Virgin Islands, or the Commonwealth of the Northern Mariana Islands, EPA may exempt any
person or source, or class of persons or sources, in that territory from any requirement of the
Clean Air Act, with some specific exceptions. The requested exemption could be granted if EPA
determines that compliance with such requirement is not feasible or is unreasonable due to
unique geographical, meteorological, or economic factors of the territory, or other local factors as
EPA considers significant.
The EPA highway vehicle diesel fuel regulation at 40 CFR 80.29 implements the sulfur
and cetane requirements of section 21 l(i) of the Clean Air Act. In addition, that regulation
establishes the requirement to dye diesel fuel that is not intended for highway vehicles. It
specifies that any diesel fuel that does not show visible evidence of the dye solvent red 164 is
considered to be available for use in highway vehicles and subject to the sulfur and cetane index
requirements.
B. Alaska Exemption From the 500 ppm Sulfur Standard
1. Why Are We Considering an Exemption for Alaska?
On February 12, 1993, Alaska submitted a petition13 under section 325 of the CAA to
exempt highway vehicle diesel fuel in Alaska from paragraphs (1) and (2) of section 21 l(i) of the
CAA, except for the minimum cetane index requirement. The petition requested that we
temporarily exempt highway vehicle diesel fuel in communities served by the Federal Aid
Highway System from meeting the sulfur content (500 ppm) specified in section 21 l(i) of the
CAA and the dye requirement for non-highway diesel fuel of 40 CFR 80.29, until October 1,
1996. The petition also requested a permanent exemption from those requirements for areas of
Alaska not reachable by the Federal Aid Highway System - the remote areas. On March 22,
1994, (59 FR 13610), we granted the petition based on geographical, meteorological, air quality,
and economic factors unique to Alaska.
On December 12, 1995 Alaska, submitted a petition for a permanent exemption for all
areas of the state served by the Federal Aid Highway System, that is, those areas covered only by
the temporary exemption. On August 19, 1996, we extended the temporary exemption until
October 1, 1998 (61 FR 42812), to give us time to consider comments to that petition that were
subsequently submitted by stakeholders. On April 28, 1998 (63 FR 23241) we proposed to grant
b Copies of information regarding Alaska's petition for exemption and subsequent requests by Alaska and
actions by EPA are available for inspection in public docket A-96-26 at the Air Docket of the EPA, first floor,
Waterside Mall, room M-1500, 401 M Street SW., Washington, D.C. 20460. A duplicate public docket is at EPA
Alaska Operations Office-Anchorage, Federal Building, Room 537, 222 W. Seventh Avenue, #19, Anchorage, AK
99513-7588.
rx-2
-------
Chapter IX: Sulfur Control in Alaska & Territories
the petition for permanent exemption. Substantial public comments and substantive new
information were submitted in response to the proposal. To give us time to consider those
comments and new information, we extended the temporary exemption for another nine months
until July 1, 1999 (September 16, 1998, 63 FR 49459). During this time period, we started work
on a nationwide rule to consider more stringent diesel fuel requirements, particularly for the
sulfur content (today's action). To coordinate the decision on Alaska's request for a permanent
exemption with the new nationwide rule on diesel fuel quality, we extended the temporary
exemption until January 1, 2004 (June 25, 1999 64 FR 34126).
In today's final rule, we are extending the temporary exemption from the 500 ppm sulfur
standard to the effective date for the new nationwide 15 ppm sulfur standard in 2006. While it is
important to implement in Alaska the cleaner diesel engines and fuel of the final rule, our goal is
to take action on the petition in a way that minimizes costs through Alaska's transition to the new
low sulfur program. The cost of compliance could be reduced if Alaska refiners were given the
flexibility to meet the low sulfur standard in one step (i.e., going straight from uncontrolled
levels to the new 15 ppm standard in 2006), rather than in two steps (i.e., going from
uncontrolled levels to the current 500 ppm standard in 2004, and then to the new 15 ppm
standard in 2006). We considered the prior public comments we received as a result of our
previous notices and actions regarding exemptions from the 500 ppm sulfur standard for highway
diesel fuel in Alaska.
2. Who Commented on the 1998 Proposal for a Permanent
Exemption?
Comments in support of the permanent exemption from the current 500 ppm sulfur
standard were submitted by the Alaska Department of Environmental Conservation, a fuel
producer association (Alaska Oil and Gas Association), individual fuel producers (MAPCO,
Alaska Petroleum, Petro Star, Arco Alaska, and BP Exploration), fuel distributors (Yutana Barge
Lines, Kodiac Oil Sales, and Petro Marine Services), a trade association (Resource Development
Council of Alaska), a utility association (Alaska Rural Cooperative Association), and some
individual businesses including the Anchorage International Airport and Alaska Railroad
Corporation.
Comments opposed to the permanent exemption from the current 500 ppm sulfur
standard were submitted by the Trustees for Alaska (Trustees) and individuals. Represented by
the Trustees, besides themselves, are the Cook Inlet Keeper, the Oregon Chapter Sierra Club, the
Alaska Chapter Sierra Club, the Alaska Clean Air Coalition including the Alaska Center for the
Environment, the American Lung Association of Alaska, the Alaska Health Project, the
Anchorage Audubon Society, and the League of Women Voters of Anchorage. Their comments
generally related to Alaska's geography, meteorology, economics and health and welfare. The
rx-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
specific comments submitted by individuals generally were the same as some of the specific
comments submitted by the Trustees.
Comments opposed to the permanent exemption from the current 500 ppm sulfur
standard were also submitted by the Engine Manufacturers Association (EMA). Its comments
related to engine manufacturer warranty and recall liability, and the impact of operating advanced
technology engines on Alaska's high-sulfur diesel fuel.
3. What are the Relevant Factors Unique to Alaska?
It should be noted that while the following section discusses factors unique to Alaska in
the context of the current 500 ppm sulfur standard (vis-a-vis the exemption), in general we
believe these factors are relevant to the new 15 ppm sulfur standard as well.
a. Geography, meteorology, and fuel production, distribution, usage
Alaska is about one-fifth as large as the combined area of the lower 48-states. Because of
its extreme northern location, rugged terrain and sparse population, Alaska relies on barges to
deliver a large percentage of its petroleum products. No other state relies on this type of delivery
system to the extent Alaska does. Only 35 percent of Alaska's communities are served by the
Federal Aid Highway System, which is a combination of road and marine highways.
Approximately 19 percent of these communities (approximately 18 communities) are part of the
Alaska Marine Highway System and have heavy-duty diesel vehicles registered with the Alaska
Division of Motor Vehicles and rely on barge deliveries.0 The remaining 65 percent of Alaska's
communities are referred to as "off-highway" or "remote" communities, which are all served by
barge lines. Although barge lines can directly access some off-highway communities, those
communities that are not located on a navigable waterway are served by a two-stage delivery
system: over water by barge line and then over land to reach the community.
Ice formation on the navigable waters during the winter months restricts fuel delivery to
the areas served by barge lines. Therefore, fuel is generally only delivered to these areas between
the months of May and October. For example, Kodiak Oil commented to the 1998 proposal that
during the winter of 1989, Cook Inlet (inlet to Anchorage) froze and "we were down to our last
gallons when a barge arrived from Seattle. We even had to borrow jet fuel from the U.S. Coast
Guard to use as heating oil." This distribution problem also restricts the ability of fuel
distributors in Alaska to supply multiple grades of petroleum products to these communities.
Letter from Alaska Department of Environmental Conservation to EPA dated August 11, 1998.
LX-4
-------
Chapter IX: Sulfur Control in Alaska & Territories
The Alaska Department of Environmental Conservation reported construction costs are
30 percent higher in Alaska than in the lower-48 states.d This increase is due to higher freight
costs for materials, and higher electrical, mechanical and civil costs due mostly to higher labor
costs. Also, because of the State's high latitude, it experiences seasonal extremes in the amount
of daily sunlight and temperature, which in turn affects the period of time during which
construction can occur.
According to Alaska's petition, its extreme northern location places it in a unique
position to fuel transcontinental cargo flights between Europe, Asia, and North America.
Roughly 75 percent of all air transit freight between Europe and Asia lands in Anchorage, as does
that between Asia and the United States. The result is a large market for jet fuel (Jet-A kerosene)
produced by local refiners, which decreases the relative importance of highway diesel fuel to
these refiners. Based on State tax revenue receipts and estimates by Alaska's refiners, diesel fuel
consumption for highway use represents roughly five percent of total Alaska distillate fuel
consumption.6 The Trustees commented in 1998 that, according to 1996 and 1997 information
from the Alaska Department of Revenue, highway diesel actually represents between 11 percent
and 13 percent of the distillate market. We could not confirm the Trustees' figures. Using the
tax data submitted in Appendix Af of the Trustees comments to the 1998 proposal, we calculated
that Alaska used 5.9 percent and 6.3 percent of its total distillate consumption for motor vehicles
for the fiscal years ending June 30, 1996, and June 30, 1997, respectively.8 Using that same data,
the Alaska Department of Environmental Conservation reported to us that it derived
"approximately 5 percent" for these same years.11 In that same submittal, the Alaska Department
d IBID.
e EPA independently verified the distillate consumption estimates based on statistics from the Federal
Highway Administration and the Department of Energy. These statistics show that in 1997 the proportion of jet fuel
consumption compared to total distillate consumption was approximately 66 percent for Alaska, compared to
approximately 27 percent for the United States. The proportion of diesel fuel consumption for highway use
compared to total distillate consumption was approximately 5 percent for Alaska, compared to approximately 30
percent for the United States.
f Attachment A from the comments submitted by the Trustees in response to the 1998 proposal included
state tax revenue data from the Alaska Internet site at http://www.revenue.state.ak.us/iea/. Attachment A includes
Table 9 - Taxable Motor Fuel Gallons Sold in Alaska for the fiscal year ending June 30. 1996. and Table 6 -
Taxable Fuel Gallons Sold in Alaska for the fiscal year ending June 30. 1997.
B These ratios were derived by dividing the adjusted diesel used for highway purposes by the total
distillate. The adjusted diesel usage was calculated by subtracting all non-highway usage from the reported "gross
gallons sold" of highway diesel. The fuel subtracted included "exemptions" for fuel used for heating, exported, off-
highway, utilities, power plants, and drilling well injection. The total distillate was calculated as the sum of the
gross "highway" diesel, Jet fuel, and marine diesel.
h Letter from ADEC to EPA dated August 11, 1998.
LX-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
of Environmental Conservation also reported that industry calculations based on the same data
source show 4.2 percent for 1996. We note that the above calculations ranging from 4.2 percent
to 7 percent agree reasonably well with the original estimate of 5 percent.
Alaska's highway diesel vehicle fleet is relatively small, particularly outside the Federal
Aid Highway System, and these vehicles are predominantly older than the national average.
There are less than 9000 diesel vehicles in Alaska, with less than 600 of these vehicles in all of
rural Alaska. By comparison, there are about 7.2 million diesel vehicles in the entire nation.
Thus, the Alaska diesel fleet represents only about 1/800 of the national diesel fleet. These
vehicles are also older on average. The average age of the diesel fleet in the areas served by the
Federal Aide Highway System is about 14 years, and the average age of the vehicles in the rural
areas is about 18 years. Only about 200 to 400 new model year diesel vehicles are added to the
state's diesel fleet each year, and only about 5 to 15 in the state's rural areas. By comparison, the
average age of the national diesel fleet is about 10 years, and about 300,000 to 500,000 new
model year diesel vehicles are added to the national diesel fleet each year.1
Information provided to us by the State of Alaska indicates that refiners supply and
distribute conventional diesel fuel in the summer which has a sulfur content of approximately
3000 ppm, and supply and distribute Jet-A kerosene in the winter as an Arctic-grade diesel,
which has a sulfur content between 650 and 1100 ppm from Alaskan refiners, and 300 ppm from
one refiner in the lower-48 states. Where the Jet-A kerosene is used as conventional diesel fuel
and not for aviation, it is generally mixed in the same tanks with conventional diesel fuel. The
fuel type supplied and delivered is based strictly on economics, availability, and whether winter
grade fuel is needed. Where some or all of the Jet-A kerosene is used as aviation fuel, it must be
segregated from conventional diesel fuel to avoid contamination. Aviation fuel must meet
stringent specifications, and contamination would disqualify the fuel as aviation fuel. The same
barge and truck tankers and transfer equipment apparently can deliver both types of fuel without
significant contamination of the Jet-A kerosene. However, dyed non-highway diesel fuel is a
serious concern. According to the refiners and distributors in Alaska, if Alaska were required to
dye its non-highway diesel fuel, residual dye in tanks or transfer equipment would be enough to
contaminate and disqualify Jet-A kerosene for use as aviation fuel. Either the tanks and transfer
equipment would have to be thoroughly cleaned prior to handling jet fuel, or separate tankage
and transfer equipment would be needed.
Alaska's climate is colder than that of the other 49 states. The extremely low
temperatures experienced in a large portion of Alaska during the winter imposes a more severe
fuel specification requirement for diesel fuel in those parts of Alaska than in the rest of the
1 We compared diesel vehicle registration data (12,000 pound and greater, unladen weight) as of October
1998 provided by the state of Alaska to national diesel vehicle data as described by the input data to the EPA
MOBILE model for the 1999 calendar year.
LX-6
-------
Chapter IX: Sulfur Control in Alaska & Territories
country. This specification, known as a "cloud point" specification" significantly affects vehicle
start-up and other engine operations. Alaska has the most severe cloud point specification for
diesel fuel in the nation at -56 °F. Because Alaska experiences extremely low temperatures in
comparison to the other 49 states, and the cloud point specifications for diesel fuel in the other 49
states are not as severe, most diesel fuel used in Alaska is produced by refiners located in Alaska.
Jet-A kerosene meets the same cloud point specification as No. 1 diesel fuel (which is marketed
primarily during the winter in Alaska, as opposed to No. 2 diesel fuel which is marketed
primarily in the summer) and is commonly mixed with or used as a substitute for No. 1 diesel
fuel. However, Jet-A kerosene can have a sulfur content as high as 3000 ppm. The current
highway diesel fuel sulfur standard of 500 ppm, and the new sulfur standard of 15 ppm, would
prohibit using Jet-A kerosene from being used as a fuel for motor vehicles, unless the sulfur
content of the Jet-A kerosene were also reduced to meet these standards.
The Trustees, in their 1998 comments, challenged past statements by refiners and the
State regarding the technical feasibility of meeting the 500 ppm sulfur standard in Alaska
compared to the lower-48 states. The Trustees indicated that while Alaska's climate may be
colder on a state-by-state basis, Alaska's climatic conditions are not unique relative to other states
when considered on a city-by-city basis. The Trustees noted that we have not granted cities in
the lower-48 states exemptions from the low sulfur diesel regulations. The Trustees also
commented that in light of Canada's implementation of 500 ppm sulfur diesel fuel regulations,
we should examine the methods used by cities and towns located in northern Canada, including
the Yukon and the Northwest Territories, to overcome cloud point difficulties. We believe that
Alaska is technically capable of complying with the current 500 ppm sulfur standard (as well as
the new 15 ppm sulfur standard), but that cost is the primary issue.
In complying with the 500 ppm (or the new 15 ppm) sulfur standard, refiners have the
option to invest in the process modifications necessary to desulfurize diesel fuel for use in motor
vehicles, or not invest in the process modifications and either import low sulfur diesel fuel from
outside of Alaska or only supply diesel fuel for non-highway purposes (e.g., heating, generation
of electricity, nonroad vehicles). Previously, most of Alaska's refiners indicated that local
J The cloud point defines the temperature at which cloud or haze or wax crystals appears in the fuel. The
purpose of the cloud point specification is to ensure a minimum temperature above which fuel lines and other
engine parts are not plugged by solids that form in the fuel. This specification is designated by the American
Society for Testing and Materials (ASTM) in its "Standard Specification for D975-96 Diesel Fuel Oils", and varies
by area of the country and by month of the year based on historical temperature records. Alaska has the most
stringent cloud point specification in the United States. For example in January, Alaska's cloud point specification
is -56 °F, -26 °F, and -2°F for the northern (above 62° latitude), southern (below 62° latitude), and Aleutian Islands
plus southeastern coast region, respectively. In contrast, the most stringent cloud point specification in January in
the lower-48 states is -29 °F for Minnesota. For the state of Washington, from which some distillate is imported into
Alaska, the January cloud point specification is +19.4°F and 0°F for the western and eastern parts of the state,
respectively.
IX-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
refineries would choose to exit the market for highway diesel fuel if an exemption from the 500
ppm sulfur requirement is not granted. They pointed to their limited refining capabilities, the
small size of the market for highway diesel fuel in Alaska, and the costs that would be incurred to
desulfurize diesel fuel (to meet the 500 ppm standard). Among the reasons for the high cost
include the construction costs in Alaska, which are 25 to 65 percent higher than costs in the
lower-48 states, and the cost of modifying the fuel production process itself.
The 1998 proposal indicated that the Alaskan refineries cannot produce 500 ppm diesel
fuel without significant capital expenditures for desulfurization equipment and operation. In
response to that proposal, MAPCO (now Williams Energy Services) commented the installation
of the necessary equipment at MAPCO's North Pole refinery would cost over $100 million for a
diesel hydrotreater, a hydrogen plant, a sulfur recovery unit, additional tankage and associated
piping, utilities, etc. MAPCO indicated this cost estimate does not include needed new tankage
at MAPI's Anchorage Terminal "where MAPI has found it almost impossible to build additional
tankage to build adequate additional tankage because of difficulties in obtaining the required
permits". Further, MAPCO indicated the refinery would have to dispose of waste sulfur material
in a landfill, likely in an out-of-state disposal site at additional cost for disposal. MAPCO
explained that local borough landfills are nearing capacity and struggling with identifying long
term options, and that other borough landfill sites are not likely to accept waste generated outside
their own borough. MAPCO further commented that all non-taxable non-highway diesel fuel
would have to be dyed (the current temporary exemption also exempts the dye requirement). For
MAPI alone, this would mean dyeing over 100 million gallons of diesel per year at a cost of over
$125,000 per year for blending infrastructure, cost of dye, management of dyeing operation,
maintenance, laboratory, QA, accounting and loading personnel. Similarly, the Alaskan Oil and
Gas Association (AOGA) commented the total cost to retrofit Alaskan refineries (for 500 ppm
sulfur diesel fuel) has been estimated to be "in excess of $150 million." Tesoro indicated that it
is currently producing 500 ppm diesel fuel at a cost of about four to six cents per gallon using
low sulfur crude product, but would likely need desulfurization equipment to reach the new
sulfur limit of 15 ppm.
In its 1998 comments, theTrustees noted that in March of 1998, Petro Star, Inc. signed a
commercial license agreement with Energy Biosystems Corporation for a biocatalytic
desulfurization unit, the costs of which are significantly less than the current hydro-treating
method. They noted that after we granted the "urban" exemption, Tesoro announced plans to
spend $50 million in 1997 to "beef up" its Nikiski refinery (near Anchorage on Kenai Peninsula),
expand its network of Alaska gas stations and renovate existing stores, and to spend an additional
$30 million over 1998-99 for further store expansion and renovations. They noted that the State
of Alaska's Industrial Development and Export Authority later announced it would help finance
$8.1 million of the cost of Tesoro's refinery improvements. They further noted that MAPCO also
was in the midst of a $11 million refinery expansion in 1997. The Trustees commented that EPA
IX-8
-------
Chapter IX: Sulfur Control in Alaska & Territories
should, therefore, ignore complaints by local refineries that the cost of retrofitting are too much
for them to bear in order to supply Alaskans with cleaner fuel.
It appears there are two related cost issues that need to be addressed. One is whether
refiners can afford the retrofits to produce highway diesel fuel that meets the 500 ppm (and the
new 15 ppm) sulfur standard. Their 1998 comments seem to have focused not on whether the
refiners can afford the retrofits, but on cost benefits: the anticipated high costs of 500 ppm diesel
fuel and the lack of significant environmental benefits for Alaska. According to the Alaska
Department of Environmental Conservation, MAPCO's and Tesoro's expansions "were built
because they are profitable investments. This is unlike the capital cost required for 500 ppm
diesel. One refiner indicated that they do not see this as a profitable investment, but a burden to
consumers, refiners, and fuel distributors in Alaska."k However, we have indications that
Alaska's refiners may now be considering options for sulfur control.
The second related issue is the cost of producing 500 ppm (or 15 ppm) diesel fuel. Our
analysis of the costs to meet the 15 ppm sulfur standard, described in Chapter V, are significantly
lower than that indicated by MAPCO (now Williams) to meet the 500 ppm standard. Further, the
costs of desulfurization may prove to be less than estimates using the standard hydrogenation
process if new technologies as discussed in Chapter IV become prove to be commercially viable.
The 1998 proposal discussed Alaska's potential ability to import diesel fuel from the
lower-48 states. In that proposal, we explained that the fuel currently being imported into
Southeast Alaska either does not meet the federal sulfur standard (500 ppm) for motor vehicles,
or is not arctic grade, or both, thus the cost of requiring 500 ppm diesel fuel would likely
increase. The Trustees subsequently commented that there are many northern cities and towns in
the lower-48 states that need 500 ppm diesel fuel with a low cloud point, and that we failed to
explain where those northern cities and towns obtain their low sulfur, low cloud point diesel fuel,
or the inability of the lower-48 state refineries to provide Alaska with 500 ppm diesel fuel.
Based on available information, we believe lower-48 state refineries are probably capable of
producing and marketing 500 ppm (and 15 ppm) arctic grade diesel fuel for Alaska, but there are
likely cost implications for producing that fuel in the lower-48 states and in transporting it to
Alaska.
The Trustees commented in 1998 that based on information provided in the 1998
proposal, the price of imported fuel is not overly expensive or unreasonable. In response to this
comment, we looked at the price of diesel fuel in the lower-48 states, and the cost to distribute
fuel to Alaska from the lower-48 states. Two sources of distillate prices indicate the average
price difference between high sulfur and 500 ppm diesel fuel in the lower-48 states is about 2.1
IBID.
IX-9
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
to 2.5 cents per gallon.1 MAPCO submitted an invoice for one vessel shipment of 42,779.20
barrels (1,796,726 gallons) of gasoline from Puget Sound to Anchorage at a cost of $300,000, or
16.7 cents per gallon. We assume that the cost to ship diesel fuel depends on the distance the
fuel is shipped, and that it is similar to the cost of shipping gasoline. Adding this shipping cost
(16.7 cents per gallon) to the average price difference between high sulfur and 500 ppm sulfur
diesel fuel in the lower-48 states (2.1 to 2.5 cents per gallon), the estimated direct cost of
replacing locally produced high sulfur highway diesel fuel in Anchorage with imported 500 ppm
diesel fuel from the lower-48 states could be 18.8 to 19.2 cents per gallon.
This cost estimate does not consider any additional cost that may be associated with
producing special batches of 500 ppm diesel fuel having a low cloud point specification in the
lower-48 states. EPA has no estimates of such cost, but at least two factors must be considered.
First, it is likely that low-cloud point specification fuel costs more to produce because of its
special formulation, and the fact that it is produced only to the extent it is specifically requested
by purchasers. Second, the demand for low-cloud point specification fuel for Alaska would not
likely coincide in time with the demand for low cloud point specification fuel in the lower-48
states. Alaska distributes and stocks its winter fuel during the summer and fall months, when
summer fuel is being produced and delivered in the lower-48 states. For example, the Colonial
Pipeline requests winter grade fuel beginning in September, so that it will be available for use in
the Northeast states by about the end of October."1 Thus, refiners in the lower-48 states would
just be gearing up to produce winter fuel for the lower-48 states when they would be gearing
down production of winter fuel for Alaska. Those refiners would have to produce special
batches of fuel destined for Alaska in the spring and summer months, and refiners and
distributors in the lower-48 states would have to store and transport this fuel while keeping it
segregated from other diesel fuel destined for use in the lower-48 states and from any other high-
sulfur diesel fuel destined for use in Alaska.
The Trustees also commented in 1998 that we should provide evidence that transportation
costs would increase. In the 1998 proposal, we concluded that increased transportation costs
would be either zero or minimal for Southeast Alaska because most if not all its fuel is already
imported to that area. However, the shipping costs would increase for other areas which
currently obtain their fuel from in-state refineries, and that increase would likely be "more than
minimal." Comments to that proposal submitted by MAPCO (Docket A-96-26, IV-D-6)
indicated our information was dated, and that much of the diesel fuel in Southeast Alaska now
1 According to the DOE "Monthly & Annual Petroleum Supply Data" for 1997, the average price of 500
ppm #2 diesel fuel in the United States in 1997 was 2.1 cents per gallon more than the average price of high-sulfur
#2 diesel. Also, according to Hart's Diesel Fuel News for January through March of 2000, 500 ppm distillate
averaged 2.5 cents per gallon more than low sulfur distillate for New York, Houston and Chicago.
m Electronic mail to EPA staff by the American Petroleum Institute on October 12, 1998.
IX-10
-------
Chapter IX: Sulfur Control in Alaska & Territories
comes from in-state refiners. Regardless, we consider the cost of transporting fuel produced in
the lower-48 states to Alaska is the sum of the cost of transporting the fuel in the lower-48 states
(e.g., from the refinery to Seattle) and the cost of transporting the fuel to Alaska from the lower-
48 states (e.g., from Seattle to Anchorage). The transportation cost within the lower-48 states,
for which we do not have a cost estimate, is incurred regardless of whether the fuel produced in
the lower-48 states is destined for use in the lower-48 states or Alaska. Thus, the transportation
cost within the lower-48 states should already be incorporated into the cost of fuel being sold in
the lower-48 states. An example of transportation cost from the lower-48 states to Alaska was
provided by MAPCO, which submitted an invoice for one vessel shipment of 42,779.20 barrels
(1,796,726 gallons) of gasoline from Puget Sound to Anchorage at a cost of $300,000, or 16.7
cents per gallon.
We stated in the 1998 proposal that Canadian 500 ppm diesel fuel does not seem to be
available for export to Alaska. The Trustees commented in 1998 that this conclusion was based
on 1995 information, that Canada recently passed regulations requiring all on-road vehicles to
use 500 ppm diesel fuel, and Canadian refineries may now welcome the opportunity to supply
Alaska with 500 ppm diesel. In response to this comment, EPA contacted Transport Canada,
which indicated Western Canada has greater ability to produce 500 ppm products than other
areas of Canada, but that some Western Canadian refiners have indicated that they do not have
enough capacity to supply 500 ppm diesel fuel with a -57° F cloud point to Alaska. The cloud
point limit requires the fuel to be made from the same streams used to make Jet-A kerosene, and
"there is simply not enough of this material available in Western Canada to supply Alaska, in
addition to meeting domestic Jet A-l and northern diesel demands." But even if some Canadian
refiners could supply 500 ppm diesel fuel to Alaska, "[t]he costs for distribution would be
significant."" In addition, Canada's highway diesel fuel would not meet the requirements of
today's rule (15 ppm sulfur), unless Canada subsequently would adopt these new requirements
for Canada. Canada has recently publically announced its intention to adopt a highway diesel
sulfur standard consistent with our final rule.0
Cost is relevant not only for the likely higher cost of replacement of imported fuel (due to
importation costs, decreased competition, etc.), but for total impact on the Alaskan economy.
For example, in response to the 1998 proposal, the Anchorage International Airport (Airport)
commented that increased costs of highway diesel fuel could significantly impact Alaskans
through increased capital and operating costs for airports, tenants, as well as other Alaskan
industries and businesses relying on diesel fuel. It said these increased costs would be passed on
to end-users. The Airport also indicated many small businesses are relied on in Alaska due to the
n Information submitted to EPA on July 20, 1998 by Transport Canada via electronic mail
0 "Process Begins to Develop Long Term Agenda to Reduce Air Pollution from Vehicles and Fuels",
Environment Canada press release, May 26, 2000
IX-11
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
state's expansive size, undeveloped transportation system, and rural markets. It said that these
added capital and operating costs combined with logistical difficulties would place many small
but needed businesses common at airports and throughout Alaska at a great disadvantage. The
Alaska Railroad Corporation commented that although railroads would not be required to use
500 ppm highway diesel, it anticipates a reduction in the quantity of petroleum products it
transports from the MAPCO refinery in Fairbanks for use throughout Alaska. It said some of its
diesel fuel business would be diverted to out-of-state refiners, reducing railroad revenues, thereby
increasing the cost of transporting other goods and services, and thus causing customers to
evaluate other less fuel-efficient alternatives.
Whether 500 ppm (or 15 ppm) diesel fuel is produced in Alaska or imported from the
lower-48 states or Canada, there remains the problem of segregating the two fuels for transport to
communities that are accessible only by navigable waterways and subsequent storage of the fuels
in those communities. Fuel is delivered to these communities only between the months of May
and October due to ice formation which blocks waterways leading to these communities for
much of the remainder of the year. The fuel supplied to these communities during the summer
months must last through the winter and spring months until resupply can occur. Additionally,
the existing fuel storage facilities limit the number of fuel types that can be stored for use in these
communities. The cost of constructing separate storage facilities and providing separate tanks
for transport of 500 ppm diesel fuel for motor vehicles could be significant. This is largely due
to the high cost of construction in Alaska relative to the lower 48 states, and the constraints
inherent in distributing fuel in Alaska.
One alternative to constructing separate storage facilities is to supply only 500 ppm diesel
fuel to these communities. However, this would require use of the higher cost, 500 ppm diesel
fuel for all diesel fuel needs. This would increase the already high cost of living in these
communities, since a large percentage of distillate consumption in these communities is for non-
highway uses, such as operating diesel powered electrical generators. The Trustees commented
in 1998 that we offered little evidence that 500 ppm diesel fuel would cost more, and in fact, 500
ppm diesel fuel may be less expensive. In our 1998 proposal, we indicated that "the distributors
import the more expensive Jet-A kerosene for all uses because limited storage prevents
segregation among the intended uses." Where segregation is not possible, we believe the
distributors have to supply fuel that meets the most stringent requirements among the various
uses of the fuel. Presently, the most stringent requirements are for jet fuel. As long as fuel is
needed for aviation purposes and can not be segregated from the fuel used for other purposes, all
fuel has to meet the more stringent and more costly jet fuel specifications. Similarly, if 500 ppm
diesel fuel were required for highway use and segregation were not possible, all fuel would have
to meet both the more costly jet fuel specifications and the more costly 500 ppm motor vehicle
diesel fuel specification. Our cost estimates for refiners to desulfurize to the new 15 ppm sulfur
standard are discussed in Chapter V.
IX-12
-------
Chapter IX: Sulfur Control in Alaska & Territories
Also, Alaska would not be able to avoid segregating the fuel simply by supplying only
500 ppm diesel fuel for all uses. If the EPA 500 ppm sulfur content exemption were to expire or
be terminated, the automatic Internal Revenue Service (IRS) exemption for Alaska would also
expire, and the IRS would require that the non-highway fuel (except jet fuel) be segregated and
dyed. The IRS tax code has a diesel fuel dye requirement that parallels that of the EPA, but areas
of Alaska covered by an EPA diesel sulfur exemption are also exempt from the IRS dye
requirement (26 CFR 48.4082-5T). Consequently, if the EPA exemption were to expire or be
terminated and fuel producers and distributors decided to supply only 500 ppm diesel fuel for
economic or other reasons, the producers and distributors would still have to dye and segregate
the non-highway fuel (except jet fuel) for federal tax purposes or pay the highway fuel tax. The
jet fuel intended for aviation would have to be segregated from jet fuel intended for general usage
to avoid being contaminated with the dye.
In response to the 1998 proposal, we received significant comments from industry and
businesses in Alaska on the issue of storage and segregation. We listed the most significant of
these comments below. Note that while these issues were raised in response to the exemption
decision for the current 500 ppm standard, they are generally relevant to the new 15 ppm sulfur
standard as well.
The Anchorage International Airport commented that construction of separate storage and
distribution facilities will be at "substantial capital and operating costs."
Kodiak Oil commented that there is not a lot of use of diesel fuel by vehicles. Diesel fuel is used
for other purposes, including fishing boats, ocean transportation (tug boats), heating oil, off-road
construction equipment, off-road logging and fish processing. Currently there is no need to
segregate the fuel by use. The costs associated with 500 ppm diesel would be borne by a small
segment of the market and would require a large expense in new tanks and pipelines for a small
part of sales. Kodiak Oil further stated that fuel distributors can't send tank trucks to a fuel
terminal in another town to load it up and drive it back. They must receive fuel by barge in large
volumes and must have a large investment in tanks and equipment to store it.
Petro Marine Services commented that the vast majority of its product is used by vessels, and
very little of it is sold to motor vehicles. If it were required to segregate 500 ppm diesel for use
by motor vehicles, separate tanks would be required for distribution and storage, which would
represent a huge cost for a very small volume of fuel.
Petro Star commented that segregated storage would be required to distribute 500 ppm diesel in
many areas, and such storage would be unfeasible or prohibitively expensive.
ARCO commented the demand for motor vehicle fuel is small compared to other uses of its
"Arctic Heating Fuel" (similar to No. 1 diesel fuel produced by ARCO Alaska Inc. for use in
IX-13
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
exploration and production in the North Slope) and it would be very expensive to construct
separate storage tanks for a second type of diesel fuel.
The Alaska Oil and Gas Association commented that because 500 ppm diesel must be segregated
from non-road diesel, separate distribution, storage and delivery systems must be installed. It
estimates that there are approximately 80 tank farms located throughout Alaska, many of them in
remote areas. Without the exemption all of these locations would be required to construct and
maintain separate facilities.
Yutana Barge Lines. Inc. commented that over 95 percent of the diesel that Yutana Barge Lines,
Inc. hauls is either for marine or power generation. It calculated the cost to be approximately 18
cents per gallon to properly segregate the dyed and undyed fuel, and build the additional facilities
and equipment to ship and store the dyed and undyed fuel.
MAPCO (now Williams Energy Services) commented that additional tankage and segregation of
piping systems in the Alaska distribution system could easily cost $50 million, assuming that
permits can be obtained. There is a shortage of tankage for storage and distribution of products
throughout the state. As a result, in Anchorage MAPCO ships and stores imported ethanol (for
oxygenated gasoline during the control period for CO) in rail cars at a cost of over 27 cents per
gallon. MAPCO has tried to build new tanks to expand its storage and distribution capabilities,
but has been unable to obtain the associated air permits. It indicated similar permitting and
storage challenges exist throughout the state. There are over 80 tank farms in Alaska that would
require additional tankage. These tank farms are widely dispersed and delivery times are such
that considerable storage volume would have to be dedicated to 500 ppm diesel. The cost of
constructing one 10,000 barrel tank with all required spill containment and supporting
infrastructure is at least $600,000. MAPCO also commented that segregated tanker trucks would
be required to haul the dyed fuels in order to avoid cross contamination with other non-dyed
diesel and jet kerosene. The cost for each additional truck is approximately $250,000.
Contamination of other fuels by the dye (dying non-highway fuel, except jet fuel, would be
required without the exemption) is a serious potential problem. Jet fuel is quite easily stained by
the dye and in Alaska many of the same trucks that haul the heating fuel also haul the highway
diesel and jet fuel.
The Trustees commented in 1998 that we should clarify why distributors are presently
capable of segregating summer and winter fuels, but if the exemption were denied, distributors
would not be capable of segregation. In response, it is likely that distributors are presently
capable of segregating the currently available fuels, but not additional fuels. For example, where
only Jet-A kerosene is currently used for all purposes (under the current sulfur and dye
exemptions), no segregation may presently be needed. Where both Jet-A kerosene and
conventional diesel fuel are currently available (under the current sulfur and dye exemptions),
adequate segregated storage may presently exist for the Jet-A kerosene (for jet use and all other
IX-14
-------
Chapter IX: Sulfur Control in Alaska & Territories
uses in the winter) and conventional diesel fuel (including a mixture of conventional diesel and
Jet-A kerosene) in the summer. In each case, if we do not extend the sulfur and dye exemptions
beyond January 1, 2004, additional tankage would likely be required to segregate the 500 ppm
highway diesel fuel, and additional tankage may be required to segregate jet fuel intended for
aviation purposes from the jet fuel intended for other non-highway uses, which would have to be
dyed in accordance with the EPA and IRS requirements.
The Trustees commented in 1998 that studies by members of the diesel engine
manufacturing and petroleum refining industries prompted by the 1990 diesel rule (55 FR 34121,
August 21, 1990) found that the potential cost benefit to truck owners and operators of using 500
ppm diesel fuel compared to conventional diesel fuel is between 8 and 30 cents per gallon, and
this benefit would more than offset any cost increase incurred in requiring the use of 500 ppm
diesel. Actually, in promulgating the 500 ppm requirement in 1990, we estimated cost savings to
truck owners and operators from sulfur control to range from 0.8 to 30 cents per gallon,
depending on the method used to evaluate the benefits and the vehicle class evaluated. While we
have not gone back to reevaluate the benefits of the 500 ppm sulfur cap, we do believe the use of
the new 15 ppm fuel in upcoming engine technologies is very likely to offer benefits in addition
to those resulting from the 500 ppm cap (see chapters in and V). More importantly, as discussed
in chapter HI, we believe that the use of diesel fuel with sulfur levels greater than the 15 ppm
sulfur limit will not be viable with the use of the engine technologies necessary to meet the new
2007 vehicle emission standards.
b. Environmental and Health Factors
Since Alaska has been operating under diesel sulfur exemptions since 1993, continuing
the diesel sulfur exemption would not cause emissions to significantly increase, but would mean
Alaska would forego the potential benefits to its air quality resulting from the use of 500 ppm
diesel fuel. The only violations of national ambient air quality standards in Alaska have been for
carbon monoxide (CO) and paniculate matter (PM10). CO violations have been recorded in the
State's two largest communities: Anchorage and Fairbanks. PM10 violations have been recorded
in two rural communities, Mendenhall Valley of Juneau and Eagle River in Anchorage. The
most recent PM10 inventories for these two communities, although more than a decade old, show
that these violations are largely the result of fugitive dust from paved and unpaved roads, and that
diesel motor vehicles are responsible for less than one percent of the overall PM10 being emitted
within the borders of each of these areasp. Moreover, Eagle River has not had a violation of the
PM10 standard since 1986. Mendenhall Valley has initiated efforts for road paving to be
implemented to control road dust. The sulfur content of highway diesel fuel will not likely have
a significant impact on ambient PM10 or CO levels in any of these areas because of the minimal
p "PM10 Emission Inventories for the Mendenhall Valley and Eagle River Areas," prepared for the U.S.
Environmental Protection Agency, Region X, by Engineering-Science, February 1988.
IX-15
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
contribution by diesel motor vehicles to PM10 in these areas and the insignificant effect of diesel
fuel sulfur content on CO emissions.
The use of high-sulfur (>500 ppm) diesel fuel may cause plugging or increased particulate
sulfate emissions in diesel vehicles equipped with trap systems or oxidation catalysts, and could
impair the ability of oxidation catalysts to reduce hydrocarbons (HC) and CO exhaust emissions.
However, any increase in sulfate particulate emissions would likely have only a minor effect on
ambient PM10 levels in Alaska since current diesel motor vehicle contributions to PM10 emissions
are minimal.
The Trustees asserted in their comments to the 1998 proposal that given half of Alaska's
highway fuel consumption is diesel, diesel particles can be expected to be significant contributors
to wintertime air pollution. They also claim the State of Alaska bases an assertion that diesel
vehicles are negligible contributors to the ambient PM10 problems on inaccurate data. Alaska
calculated that by using 500 ppm diesel fuel the maximum reduction in PM10 emissions would be
70 tons per year. Alaska based this figure on the assumption that the annual highway diesel fuel
consumption is 50 million gallons. According to the Trustees, recent records show Alaskans
consume closer to 200 million gallons of highway diesel fuel each year. They concluded that if
Alaska used 500 ppm diesel fuel, the State would reduce its PM10 emissions by 280 tons per year
instead of the 70 tons estimated by the State. The Trustees indicated this reduction, whether
significant in proportion to other PM10 sources, is in the best interests of the health and well-
being of Alaskan residents. The Trustees also commented they are concerned about inventory
numbers during wintertime air pollution events in Anchorage and Fairbanks. They claimed
Alaska has evidence that Anchorage and Fairbanks events are mostly vehicular pollution with
measurable amounts of woodsmoke and that fine particle pollution is present during wintertime
inversions that result in high pollution levels.
Any reduction of particulate emissions is likely to be beneficial to health at some level.
However, we could not substantiate that half of Alaska's highway fuel consumption is diesel fuel.
Compared to the total highway consumption of gasoline and diesel fuel combined, the diesel fuel
highway consumption is approximately 24 percent, based on Alaska revenue data for the fiscal
year ending 1997 submitted by the Trustees in Attachment A to their comments to the 1998
proposal. That same revenue data indicate that highway vehicles in Alaska consume about 80
million gallons of diesel fuel annually. Using these statistics, the reduction in PM10 by using 500
ppm diesel fuel for highway vehicles would be about 112 tons per year.
The Trustees noted in their 1998 comments that in 1994, engine manufacturers began
designing their engines with catalysts to meet the 1994 emission standards with the use of 500
ppm diesel fuel. They further concluded that when a diesel engine equipped with a catalyst burns
high-sulfur diesel fuel, sulfate emissions increase. Alaska had argued that because of the small
size of the new (1994 and later) vehicle fleet on Alaskan roads, the increase in sulfate emissions
IX-16
-------
Chapter IX: Sulfur Control in Alaska & Territories
would be insignificant. The Trustees noted that at the time the Petition was submitted in 1995,
Alaska's assertion may have been true, but as each year passes, more and more diesel engines
equipped with catalysts join the Alaska fleet. The Trustees commented that EPA should examine
the effect that increased numbers of diesel vehicles with catalysts have on emissions inventories.
Based on available information and analysis, increased sulfate emissions with the use of
high-sulfur (>500 ppm) diesel fuel and oxidation catalyst technology is a possibility. However,
with few exceptions, engine manufacturers are now complying with our current emission
standards without the use of oxidation catalysts. Therefore, the concern is unlikely to grow.
The Trustees commented in 1998 that we should consider the impact of the new national
ambient air quality standard (NAAQS) for particulate matter (PM2 5) that we recently
promulgated. However, as we stated in the 1998 proposal, it is untimely to consider the impact
of the new NAAQS for PM2 5. We may revisit this issue again in the future if and when actions
on the NAAQS revisions are complete. Following a NAAQS revision, state Governors must
submit recommendations for designations. These are designations of "nonattainment,"
"attainment," and "unclassifiable." Under section 6102(c)(l) and 6102(d) of the Transportation
Equity Act for the 21st Century (TEA-21) concerning the designation process for a PM25
NAAQS, we will not make designations for PM2 5 until after at least three calendar years of air
quality data have been gathered, which have been measured by Federal reference method
monitors, or equivalent monitors. We anticipate that three years of quality assured data will be
available from the first set of PM2.5 monitors by 2002, and from all of the monitors by 2003.
Under section 6102(c)(l) of the TEA-21, states will be required to submit designations referred
to in section 107(d)(l) of the Clean Air Act for each area concerning PM25 within one year after
receipt of three years of quality assured air quality data from Federal reference method monitors
or equivalent monitors. Under section 6102(d) of the TEA-21, we would then have to
promulgate designations referred to in section 106(d)(l) of the Clean Air Act for a PM2 5
NAAQS by the earlier of one year after the date States are required to make their submittal or by
December 31, 2005. Consequently, data indicating whether Alaska is complying with a new
PM2 5 NAAQS would not be available until at least 2002 or 2003, and the compliance
designation for Alaska would not be promulgated until at least 2004 or 2005. Nevertheless,
based on very limited information available at this time, there is no substantive indication that
Alaska would be in noncompliance with a PM25 NAAQS, or that requiring 500 ppm (or 15 ppm)
diesel fuel for motor vehicles in Alaska would significantly impact Alaska's prospects for
attainment with a PM2 5 NAAQS.
The Trustees commented in 1998 that emissions monitoring results submitted in support
of Alaska's petition do not provide a reliable basis for judging the potential adverse health effects
of continuing the Alaska exemption in perpetuity. They explained that in Anchorage, for
example, the Alaska Department of Environmental Conservation has two fixed monitoring sites
for PM10 and one fixed monitoring site for PM2 5. The Trustees noted that neither Alaska's
IX-17
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
petition nor our 1998 proposal indicate whether these sites are located in areas of high diesel use,
and commented that the Alaska Department of Environmental Conservation must accurately
monitor diesel emissions in representative locations if we intend to rely on such data in making
its decision. In response to this comment, the Alaska Department of Environmental
Conservation submitted a letter to us that includes technical detail regarding the monitoring
sites.q
The Trustees commented in 1998 that, while we, and the state of Alaska, may be correct
in our assertion that sulfur content in motor vehicle diesel fuel will not significantly affect
ambient PM10 problems in Alaska, we failed to consider that Congress enacted the 500 ppm
requirement for additional reasons, especially because of sulfur effects on human health. The
Trustees stated that Congress passed Section 21 l(i) of the Clean Air Act because it was
concerned about particulates which cause cancer, genetic mutation and other health related
ailments. There is a discussion of the legal requirements for today's action in the Appendix.
The Trustees commented in 1998 that Alaska provides a brief and wholly inadequate
analysis of problems caused by high-sulfur content (> 500 ppm) diesel fuel and then summarily
concludes that Alaskan residents will not experience any health problems should we chose to
grant the exemption to the 500 ppm standard. The Trustees asserted that a complete analysis of
health effects caused by exposure to diesel exhaust must be included in our final decision to grant
or deny Alaska's petition. The Trustees further asserted that we must recognize and discuss the
fact that sulfur causes health problems in two primary ways from Alaskan residents. First,
inhalation of sulfur or sulfur dioxide (SO2) causes health problems. Second, high sulfur content
diesel fuels increase diesel exhaust which in turn causes an increase in emission of all major
criteria air pollutants in exhaust, including carbon monoxide, hydrocarbons, and nitrogen oxides.
We are concerned about the emissions and health impact of sulfur in diesel fuel, as
discussed at length in Chapter n. This discussion is prefaced by the presumption that
implementing the new 15 ppm fuel standard in Alaska, along with the new diesel motor vehicle
engine and diesel motor vehicle emission standards, is expected to eventually reduce emissions
from diesel motor vehicles in Alaska from current levels (assuming diesel vehicle miles traveled
in Alaska do not significantly increase). If the new fuel standard is not implemented in Alaska,
those emission reduction benefits will not be achieved, and some increased emissions would
result from the use of high sulfur (> 15 ppm) fuel in upcoming engine and vehicle technologies.
In fact, as discussed in chapter IE, the upcoming engine and vehicle technologies would not be
feasible on diesel fuel having a sulfur content greater than 15 ppm.
q Letter of August 11, 1998 from the Alaska Department of Environmental Conservation. Copies of that
letter are in the EPA Public Docket A-96-26 located in Washington D.C. and in Anchorage, Alaska.
IX-18
-------
Chapter IX: Sulfur Control in Alaska & Territories
The Trustees commented in 1998 that granting a fuel exemption for Alaska's mobile
sources prevents the development of 500 ppm diesel supply sources in Alaska that can be
economically tapped by stationary sources and off-highway engines. The Trustees indicated that
according to the Alaska Department of Environmental Conservation, almost all villages in
Alaska depend on diesel generators to produce power for utility purposes. Most of these diesel
generators are located in close proximity to residences and community centers, such as schools,
where the exposure to diesel fumes is most problematic. The Trustees commented that these
communities would benefit by the availability of 500 ppm diesel fuel.
We agree that there could be secondary air quality benefits to requiring 500 ppm (or 15
ppm) diesel fuel for highway vehicles, but producers and distributors would not be required to
provide 500 ppm (or 15 ppm) diesel fuel for non-highway usage. The higher cost of producing
500 ppm fuel might discourage an oversupply. The extent of any 500 ppm fuel produced for
non-highway uses depends on the refiners' decisions whether to produce 500 ppm fuel for the
highway market, as well as for the non-highway market. If the refiners in Alaska decide to not
produce excess 500 ppm diesel fuel, a significant amount of diesel fuel in Alaska would still
have a high sulfur content. Also, regardless of the sulfur content, producers and distributors
would have to segregate and dye the fuel not destined for motor vehicle use, or pay the Federal
fuel tax in accordance with the Internal Revenue Service requirements. As previously noted, the
State of Alaska indicates that a large portion of the estimated high cost associated with
distribution of 500 ppm diesel fuel is due to the EPA and Internal Revenue Service dye
requirements for the non-highway fuel. However, if Alaska were to be exempt from those dye
requirements, much of the high costs of segregating the fuels could be avoided and excess 500
ppm fuel for the non-highway market would be more economical. Our final rule grants Alaska a
permanent exemption from the EPA dye requirements, but an exemption from the IRS dye
requirement would require legislation by Congress. We understand that the Alaska congressional
delegation is working on such legislation.
For reasons discussed in more detail in the following section, cross-border traffic is
expected to raise an air quality issue for the lower-48 states and Canada. We are concerned about
the impact on emissions in Canada and the lower-48 states of diesel vehicles and trucks from
Alaska being driven to locations in Canada and the lower-48 states. We are also concerned about
the impact on emissions in the lower-48 states, and of damage to the engines and emission
control systems, of diesel vehicles and trucks from the lower-48 states driven to Alaska and
refueled with high-sulfur fuel before returning to the lower-48 states.
c. Engine and emission control system factors
The impacts of using high-sulfur (>500 ppm) diesel fuel on current engine and emission
control system technologies, and those anticipated that will be used to meet the 2004 emission
standards, are discussed in our 1998 proposal (63 FR 23241, April 28, 1998 ) to grant Alaska a
IX-19
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
permanent exemption from the diesel sulfur standard (500 ppm), and are not repeated here. The
impacts of using diesel fuel with higher than 15 ppm sulfur on upcoming engine and emission
control systems designed to meet the new 2007 emission standards are discussed in chapter HI of
this document.
In response to our 1998 proposal to grant Alaska a permanent exemption from the diesel
fuel sulfur standard of 500 ppm, but prior to our announced action to propose more stringent
emission standards for heavy-duty diesel engines and sulfur standard for highway diesel fuel of
today's action, the Engine Manufacturers Association commented that, if we grant Alaska a
permanent exemption for highway diesel fuel, the engine manufacturers should be exempted
from the recall liability requirements and warranty liability requirements of Section 207 of the
Clean Air Act for any engine affected by a fuel exemption in Alaska. The Engine Manufacturers
Association stated that the engine manufacturers' ability to meet current and future emission
standards, and the proper operation and durability of highway engines and after treatment
technology, is dependent on the availability of 500 ppm (or the new low sulfur diesel fuel,
depending on the applicable emission standards) diesel fuel.
In response to our preamble language in the August 1998 final rule to extend Alaska's
sulfur exemption for nine months, the Engine Manufacturers Association commented that this is
not a local issue, confined solely within Alaska's borders. It stated the 2004 engine technologies
that will require 500 ppm diesel fuel will be used on all heavy-duty diesel engines, including
those used in line-haul operations, and that vehicles from the lower-48 states and Canada with
technologies requiring 500 ppm diesel fuel surely will deliver goods in Alaska. The Engine
Manufacturers Association indicated that any exposure to Alaska high-sulfur diesel fuel may
permanently reduce the effectiveness of the emission control technologies employed on those
engines and substantially reduce their overall durability and performance. It further indicated
that not only will the owners of those vehicles suffer damage (for which the engine manufacturer
should not be responsible), but as a result, the lower-48 states and Canada will also suffer
adverse and excessive emission contributions.
As discussed in Chapter IE, we expect that these concerns of the Engine Manufacturers
Association will be even greater for the upcoming engine technologies designed to meet the new
2007 emission standards for diesel motor vehicles and diesel motor vehicle engines, if they
would be operated using diesel fuel with sulfur levels higher than 15 ppm. We are denying
Alaska's petition for permanent sulfur exemption for the areas served by the Federal Aid
Highway System. While we are not taking action in the final rule on the existing permanent
sulfur exemption for the rural areas, both the rural areas and the areas served by the Federal Aid
Highway System will be required to meet the new 15 ppm sulfur standard. However, we are
allowing Alaska's 500 ppm exemptions to continue until the new 15 ppm sulfur standard
becomes effective in 2006. Thus, today's action will fully address the concerns of the Engine
Manufacturers Association in the long term, but only partially for the transition period.
IX-20
-------
Chapter IX: Sulfur Control in Alaska & Territories
Consequently, the following comments pertain to the period when Alaska will still be covered by
a diesel fuel sulfur exemption, and 500 ppm diesel fuel may not be available as in the rest of the
nation.
Engine manufacturers previously expressed concern that, during the period of exemption,
2004 technology emission control equipment will likely be damaged, and those engine
technologies using cooled-EGR will likely have increased wear, if they are operated using high-
sulfur (>500 ppm) fuel in Alaska. The Alaska Trucking Association, which is the other
stakeholder most directly impacted by the use of high sulfur fuel, commented in a 1998 letter that
it is most concerned about the near-term cost of 500 ppm fuel. That association indicated that the
Alaska market will provide 500 ppm fuel for those engines as needed, without a mandate from
the state or federal government. It noted that Alaska's truck fleet is somewhat older than that of
the lower-48 states, and the new technology engines (those designed to meet the 2004 emission
standards) will likely not be "common" in Alaska's fleet until sometime after 2010. The Alaska
Trucking Association recommended a five-year extension to the sulfur exemption and a
subsequent "reasonable" transition period, after which 500 ppm diesel fuel would be required.
We are encouraging, but not requiring, Alaska to make 500 ppm diesel fuel available to
the 2004 technology vehicles and trucks that need it during this period of exemption. We have
learned that Tesoro is currently producing and supplying 500 ppm diesel fuel in Alaska.
Therefore, for example, for those areas supplied by Tesoro (or other Alaska refiners that choose
to produce 500 ppm sulfur diesel fuel, or distributors that choose to import 500 ppm sulfur diesel
fuel), Alaska would need only to segregate and label enough of Tesoro's product for the new
technology vehicles and trucks. In areas not served by Tesoro (or other Alaska refiners
producing 500 ppm fuel, or distributors importing 500 ppm fuel), Alaska would need to transport
some of the Tesoro fuel, or import some 500 ppm fuel from the lower-48 states as well.
In our June 2, 2000 proposal to establish a 15 ppm sulfur standard in 2006, as in previous
actions to grant Alaska sulfur exemptions, we proposed to not base any recall on emissions
exceedances caused by the use of high-sulfur (>500 ppm) fuel in Alaska during this period of
exemption. Our in-use testing goals are to establish whether representative engines, when
properly maintained and used, will meet emission standards for their useful lives. These goals
are consistent with the requirements for recall outlined in Section 207(c)(l) of the CAA. Further,
we believe that manufacturers may have a reasonable basis for denying emission related
warranties where damage or failures are caused by the use of high-sulfur (>500 ppm) fuel in
Alaska. These issues are addressed in the previous rules to grant Alaska extensions to its
temporary exemption for Federal Aid Highway areas. In response to the proposal, the Engine
Manufacturers Association commented that the level of protection provided to engine
manufacturers under the current exemption for Alaska and the proposal, as described above, falls
short of what is reasonable and necessary. It asserted that the use of high sulfur diesel fuel by an
engine should raise a "rebuttable presumption" that the fuel has caused the engine failure, and
IX-21
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
that EPA should have the burden of rebutting that presumption. It also asserted that the
emissions warranty is a regulatory requirement under Section 207, that only EPA has the
authority to exclude claims based on the use of high sulfur diesel fuel.
We understand and concur with the manufacturers' concerns about in-use testing of
engines operated in an area exempt from fuel sulfur requirements. Consequently, we affirm that,
for recall purposes, we will not seek to conduct or cause the in-use testing of engines we know
have been exposed to high sulfur fuels. We will likely screen any engines used in our testing
program to see if they have been operated in the exempt area. We believe we can readily obtain
sufficient samples of engines without testing engines from exempt areas. Also, in any recall that
we order, manufacturers have the option to request a public hearing under §85.1807(b)(ii). The
use of engines that have seen high sulfur fuel would increase the likelihood that EPA would grant
a request for a recall hearing. We would expect manufacturers to scrutinize any test engines for
sulfur usage that were used to justify an ordered recall. In reviewing the warranty concerns of the
Engine Manufacturers Association, we have determined that our position regarding warranties, as
previously stated and described above, is consistent with section 207(a) and (b) of the CAA and
does not require any new or amended regulatory language to implement.
Subsequent to the 1995 petition for a permanent exemption from the diesel fuel sulfur
(500 ppm) requirements, the Engine Manufacturers Association requested enforcement
discretion regarding the removal of catalytic converters because of an indicated plugging problem
caused by the high-sulfur (>500 ppm) diesel fuel in Alaska. However, information subsequently
collected by us from several heavy-duty engine manufacturers demonstrates that catalyst
plugging is mainly a cold weather problem and not a high-sulfur fuel issue. We are also aware
that the majority of the plugged catalysts have been eliminated. In a September 19, 1997 letter to
us, the Engine Manufacturers Association indicated that the immediate problems that led to the
Engine Manufacturers Association's earlier request have been resolved, although the
manufacturers indicate they continue to have concerns.
The Engine Manufacturers Association commented in response to the 1998 proposal that,
as a result of the 2004 emission standards, heavy-duty engine manufacturers will likely introduce
EGR systems on their engines and may also, if fully developed, use NOx catalysts. In the EGR
system, exhaust gas is recirculated back into the cylinder to reduce the amount of fresh charge air
or oxygen available for combustion during certain operating conditions. When the engine is
operated on high-sulfur (>500 ppm) fuel, sulfur in the exhaust gas stream is condensed by the
EGR cooler and forms sulfuric acid deposits in the cooler. The Engine Manufacturers
Association claims these sulfuric acid deposits will significantly contribute to the deterioration of
the EGR system and cause decreased engine durability. Similar to oxidation catalysts, the NOx
catalyst will be adversely affected if operated on high-sulfur (>500 ppm) diesel fuel. The Engine
Manufacturers Association claims that a special concern regarding NOx catalysts is the fact that a
IX-22
-------
Chapter IX: Sulfur Control in Alaska & Territories
single exposure to high-sulfur (>500 ppm) diesel fuel will likely poison the catalyst, causing it to
lose its emissions reduction effectiveness permanently.
The Engine Manufacturers Association further commented that, under an exemption from
the sulfur content (500 ppm) for highway diesel fuel, manufacturers should be allowed to sell
engines without catalysts or exhaust gas recirculation (EGR) systems, and to continue to sell
older technology heavy-duty engines (i.e., without catalysts or EGR systems). It also commented
that users of vehicles in which 1994 and later model year heavy-duty engines are placed should
be exempted from tampering liability and be allowed to remove plugged catalysts. The Engine
Manufacturers Association is concerned that using a high-sulfur (>500 ppm) content fuel over a
long period of time may have a tendency to cause plugging of catalysts, which could lead to more
serous engine malfunction and warranty claims. The Engine Manufacturers Association asserts
that vehicle owners are already experiencing engine failures directly resulting from catalyst
plugging, and that the plugging is substantially aggravated when the vehicle is operated in
extremely cold temperatures.
We also are concerned about these problems, but do not believe it is necessary to offer an
exemption that allows the removal of catalysts in the field, or that permits manufacturers to
introduce into commerce catalyzed-engines without catalysts. In response to the June 2000
proposal for today's action, we received no additional information about in-use catalyst plugging
and engine failures being caused by the use of high-sulfur (>500 ppm) diesel fuel, or any data and
descriptions of remedies. Also, as previously discussed, we have indicated we do not intend to
use vehicles from Alaska to show noncompliance by those engines for the purpose of recalling an
engine class, and that the engine manufacturers have a reasonable basis for denying warranty
coverage where the problems are due to the use of high-sulfur (>500 ppm) diesel fuel.
d. Are there alternatives to granting or denying Alaska's Petition for
permanent exemption?
The Alaska Center for the Environment suggested in an April 23, 1996 letter, and the
Trustees supported, three fall back positions. While they strongly advocated that we deny
Alaska's request for a permanent sulfur exemption, they preferred one or a combination of the
three fallback positions rather than a permanent exemption. They believed that the fall back
alternatives would achieve some air quality benefits at reduced cost compared to full compliance
with the fuel sulfur requirement (500 ppm). The first alternative would permanently exempt
most of Alaska, but not Southeast Alaska. Southeast Alaska does not need the low cloud point
diesel fuel in the winter, and it was assumed at that time (but no longer true) that Southeast
Alaska imports most of its diesel fuel from the lower-48 states. The Alaska Center for the
Environment claimed that 500 ppm diesel fuel from the lower-48 states could easily be
substituted for high sulfur (> 500 ppm) diesel fuel currently being imported before shipping it to
Southeast Alaska. The second alternative would permanently exempt Alaska in the summer
IX-23
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
months, when conventional diesel fuel is generally used, but not in the winter months when Jet-A
kerosene is generally used. The Alaska Center for the Environment claimed the problem of
supplying 500 ppm fuel in the winter, when low-cloud point fuel is needed, could be avoided.
The third alternative would require lower sulfur (but generally greater than 500 ppm) Jet-A
kerosene year round. The Alaska Center for the Environment claimed some air quality benefits
might be achieved by using fuel that is already available in Alaska.
In general, we believe some assumptions upon which each of these alternatives were
based are erroneous or no longer applicable because of recent information or changes to the
emission standards and fuel requirements in the final rule. We now believe these alternatives
would not likely achieve significant cost reductions, or significant air quality benefits, or protect
the upcoming engine technologies from reduced performance and damage. In response to the
1998 proposal, no other stakeholder indicated they were in favor of these alternatives, and
refiners and the Alaska Department of Environmental Conservation submitted comments
indicating they are opposed to these alternatives because of cost or lack of air quality benefits.
Since no other stakeholder supported these alternatives, and since the situation has changed with
the new standards, and since we are not granting Alaska's request for a permanent exemption for
the areas covered by the Federal Aid Highway System, we believe that these alternatives need not
be addressed further.
The Trustees also noted in its 1998 comments, that when we granted Alaska's rural areas
a permanent exemption from the 500 ppm diesel fuel requirements, we reserved the right to
withdraw the exemption in the future if circumstances change such that the exemption is no
longer appropriate. The Trustees commented that in light of Canada's promulgation of 500 ppm
diesel fuel requirements, Petro Star's commitment to produce 500 ppm diesel fuel, recent health
studies that concretely link SO2 and diesel exhaust with respiratory problems and risks of cancer,
and worldwide acceptance of the 500 ppm sulfur content limit, we should not only deny Alaska's
current petition for the areas served by the Federal Aid Highway System, we should also
reconsider Alaska's exemption for the rural areas.
e. What Flexibility Are We Offering Alaska?
The Federal Register notice for the final rule fully discusses our flexibility option for
Alaska. In summary, as mentioned above, Alaska has submitted a petition for a permanent
exemption from the 500 ppm standard for areas not served by the Federal Aid Highway System.
We are not taking action on that petition in the final rule. Our goal is to take action on that
petition in a way that minimizes costs through Alaska's transition to the low sulfur (15 ppm)
program. The cost of compliance could be reduced if Alaska refiners were given the flexibility
to meet the low sulfur standard in one step (i.e., going straight from uncontrolled levels to the 15
ppm standard), rather than in two steps (i.e., once for the current 500 ppm sulfur standard in 2004
when the temporary exemption expires, and again for the 15 ppm standard in 2006). Therefore,
IX-24
-------
Chapter IX: Sulfur Control in Alaska & Territories
we are extending the existing temporary exemption from the current sulfur standard of 500 ppm
for the areas of Alaska served by the Federal Aid Highway System from January 1, 2004 (the
current expiration date) to the effective date for the new 15 ppm sulfur standard (i.e., June 1,
2006 at the refinery level; July 15, 2006 at the terminal level; and September 1, 2006 at all
downstream locations).
Also, because of the unique circumstances in Alaska, we are offering an alternative
option for implementing the low sulfur (15 ppm) fuel program in Alaska. We are providing the
State an opportunity to develop an alternative low sulfur transition plan for Alaska. This plan
will need to ensure that sufficient supplies of low sulfur (15 ppm) diesel fuel are available in
Alaska to meet the demand of any new 2007 and later model year diesel vehicles. Given that
Alaska's demand for highway diesel fuel is very low and only a small number of new diesel
vehicles are introduced each year, it may be possible to develop an alternative implementation
plan for Alaska in the early years of the program that provides low sulfur (15 ppm) diesel only in
sufficient quantities to meet the demand from the small number of new diesel vehicles. This
would give Alaska refiners more flexibility during the transition period because they would not
have to desulfurize the entire highway diesel volume, or even the levels required under the
temporary compliance option described in the Federal Register notice for the final rule. Our goal
in offering this additional flexibility is to transition Alaska into the low sulfur (15 ppm) fuel
program in a manner that does not compromise the environmental benefits of the program by
ensuring that the new vehicles receive the low sulfur fuel they need, while minimizing costs. We
expect that the transition plan would begin to be implemented at the same time as the national
program, but the State will have an opportunity to determine what volumes of 15 ppm fuel will
need to be supplied, and in what timeframes, in different areas of the State.
At a minimum, this transition plan will need to: 1) ensure an adequate supply (either
through production or imports of 15 ppm fuel to meet the demand of any 2007 or later model
year vehicles), 2) ensure sufficient retail availability of low sulfur fuel for new vehicles in
Alaska, 3) address the growth of supply and availability over time as more new vehicles enter the
fleet, 4) include measures to ensure segregation of the 15 ppm fuel, prevent contamination, and
prevent misfueling, and 5) ensure enforceability. We anticipate that, to develop a workable
transition plan, the State will likely work in close cooperation with refiners and other key
stakeholders, including retailers, distributors, truckers, engine manufacturers, environmental
groups, and other interested groups. In the Federal Register Notice forme final rule, we discuss
this option in more detail, including the timeframe for the State to submit the plan and for EPA to
approve it.
If the State anticipates that the primary demand for 15 ppm fuel will be along the highway
system (e.g., to address truck traffic from the lower-48 states) in the early years of the program,
then the initial stages of the transition plan could be focused in these areas. We believe it would
be appropriate for the State to consider an extended transition schedule for implementing the low
IX-25
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
sulfur program in rural Alaska, as part of the state's overall plan, based on when they anticipate
the introduction of a significant number of 2007 and later model year vehicles in the remote
areas. While we are not taking action in the final rule on the existing permanent exemption from
the 500 ppm standard for the rural areas, both the rural areas and the areas served by the Federal
Aid Highway System will be required to meet the new 15 ppm sulfur standard, with the
opportunity to develop a transition plan for how the new standard will be implemented as
discussed above.
During this transition period, it is possible that both 15 ppm (for 2007 and later model
year vehicles) and higher sulfur (for older vehicles) highway fuels might be available in Alaska.
To avoid the two-step sulfur program described above during an alternative transition period, we
would consider additional extensions to the temporary exemption of the 500 ppm standard
beyond 2006 (e.g., for that portion of the highway diesel pool that is available for the pre-2007
vehicles during Alaska's transition period). We would make a decision on any additional
temporary extensions, if appropriate, in the context of the separate rulemaking taking action on
the alternative transition plan submitted by Alaska (as described in the Federal Register Notice
for the final rule).
As in previous actions to grant Alaska sulfur exemptions, we will not base any vehicle or
engine recall on emissions exceedences caused by the use of high-sulfur (>500 ppm) fuel in
Alaska during the period of the temporary sulfur exemption. In addition, manufacturers may
have a reasonable basis for denying emission related warranties where damage or failures are
caused by the use of high-sulfur (>500 ppm) fuel in Alaska.
Finally, the costs of complying with both the current 500 ppm sulfur and new 15 ppm
sulfur diesel fuel standards could be reduced significantly if Alaska were not required to dye the
non-highway fuel. Dye contamination of other fuels, particularly jet fuel, is a serious potential
problem. This is a serious issue in Alaska since the same transport and storage tanks used for jet
fuel (which are more than half of Alaska's distillate market) are generally also used for other
diesel products, including off-highway diesel products which are required to be dyed under the
current national program. This issue is discussed further above. Therefore, we also are granting
Alaska's request for a permanent exemption from the dye requirement of 40 CFR 80.29 and 40
CFR 80.446 for the entire State.
IX-26
-------
Chapter IX: Sulfur Control in Alaska & Territories
C. American Samoa, Guam, and Commonwealth of
Northern Mariana Islands (CNMI)
1. Why Are We Considering an Exemption for American Samoa,
Guam, and CNMI?
Prior to the effective date of the current highway diesel sulfur standard of 500 ppm, the
territories of American Samoa, Guam and CNMI petitioned EPA under section 325 of the CAA
for an exemption from the sulfur requirement under section 21 l(i) of the CAA and associated
regulations at 40 CFR 80.29. The petitions were based on geographical, meteorological, air
quality, and economic factors unique to those territories. We subsequently granted the petitions."
We are now revisiting the issue of diesel sulfur exemptions for these territories, because this rule
impacts these territories, either directly by requiring stringent sulfur requirements or indirectly by
the future impact of using high-sulfur diesel fuel in upcoming technologies for diesel vehicles
and engines.
2. What are the Relevant Factors?
a. American Samoa
American Samoa is a group of five volcanic islands and two coral atolls located in
Polynesia, approximately 2,300 miles southwest of Hawaii and 1,600 miles northeast of New
Zealand. It has a total land area of approximately 76 square miles, about two-thirds of which is
mountainous with steep slopes that make it virtually inaccessible. The population was about
52,400 in 1993. Over 96 percent of the population live on the largest island, Tutuila, which is
approximately 53 square miles. The air quality is generally pristine, due to the combined
prevailing winds, climate, remoteness, and low population. American Samoa is classified EPA
"Priority in" for all pollutants. Continuing the existing diesel sulfur exemption or expanding it to
the new diesel sulfur standard would not cause an increase in emissions, but would forego a
small emissions benefit if American Samoa used low sulfur diesel fuel for its highway vehicles.
There are no cross border issues.
Total diesel fuel imports and consumption is about 38,600,000 gallons per year. The 60
diesel fueled vehicles licensed for highway use (in 1991) use approximately 0.12 percent, 46,000
gallons per year. The fishing fleet uses approximately 70 percent, or 27,000,000 gallons per year.
Power generation uses approximately 18 percent, or 7,000,000 gallons per year. Tuna Canneries
use approximately 10 percent, or 4,000,000 gallons per year in boilers for steam generation.
1 See 57 FR 32010, July 20, 1992 for American Samoa;: 57 FR 32010, July 20, 1992 for Guam; and 59 FR
26129, May 19, 1994 for CNMI
IX-27
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
Other uses (stand-by generators and small engines) consume a small quantity. American Samoa
has two providers of diesel fuel. Mobil Australia provides about 70 percent and Tesoro Hawaii
provides about 30 percent. All petroleum is shipped to American Samoa by medium range
tankers.
The economy of American Samoa is ill suited to handling the additional expense of low
sulfur diesel fuel and upgrades to its storage and distribution system. It lacks internal petroleum
supplies and refining capabilities. Diesel fuel must be imported over long distances and in small
cargo parcels, making the cost about 10 cents per gallon higher than in the mainland United
States (exclusive of the effects of taxes). It was estimated in 1991 that shipping a segregated
quantity of low sulfur diesel fuel would cost an additional eight to ten cents per gallon. The new
15 ppm sulfur standard would add additional cost to the low sulfur fuel. Compliance with low
sulfur requirements for highway fuel would require construction of separate storage and handling
facilities. American Samoa, which owns the petroleum storage facilities, would have to
construct a new storage tank and fuel lines at an estimated 1991 cost of $550,000.
If American Samoa alternatively decided to use low sulfur fuel for all purposes to avoid
segregation, compliance with low sulfur requirements would, in effect, bar importation of diesel
fuel by the only suppliers in the area outside of Tesoro. The effect of a monopoly from prior
years' experience was to see an increase in the price of fuels by four to five cents per gallon.
While Australia, Philippines and certain other Asian countries have or will soon require low
sulfur diesel fuel, they require a 500 ppm sulfur limit, not the new 15 ppm sulfur limit
established by today's action for the United States.
The fishing fleet buys its fuel in Samoa, but has the alternative to fuel in Fiji or from
tankers at sea. If the cost of fuel in American Samoa increased, the fishing fleet would probably
significantly reduce its fueling in American Samoa resulting in a "disastrous decline in the local
economy and lost revenues to the government." In 1997, the price of fuel in Samoa rose to over
ten cents a gallon higher than the price in Fiji, and Samoa lost about one-half of its sales to the
fleet. Currently, the price of diesel fuel is about six to seven cents per gallon higher than in Fiji.
Any additional fuel switching by the fleet due to a low sulfur requirement would damage the
economy without providing any significant improvement in air quality.
Any increased fuel costs of producing steam by the canneries would be another
competitive disadvantage for them to remain in American Samoa. Cheap labor and fuel in the
Philippines and Thailand have attracted canneries to those countries that compete directly with
the canneries in American Samoa. Over the past eight years, Samoa has granted larger and larger
tax exemptions to retain the canneries in Samoa.
IX-28
-------
Chapter IX: Sulfur Control in Alaska & Territories
b. Guam
Guam is the southern-most island in the Marianas Archipelago, located in Polynesia,
approximately 3,700 miles west-southwest of Honolulu and 1,550 miles south of Tokyo. The
island is about 28 miles long and between 4 and 8.5 miles wide, with a total land area of about
209 square miles. In the early 1990's, the population was 133,152.
Guam is in attainment with the primary NAAQS with the exception of sulfur dioxide in
two areas. One area is a radius of 3.5 km around the Piti Power Plant. The other area is a radius
of 3.5 km around Tanguisson. Both areas are designated nonattainment for sulfur dioxide as a
result of monitored and modeled exceedences in the 1970's prior to implementing changes to
power generation facilities. Guam believes the area around Piti is now in attainment. The
Tanuisson area includes only two small villages and a U.S. Air Force Annex, none of which
attract significant vehicle traffic.
In the early 1990's, there were 735 registered diesel-fueled motor vehicles, approximately
one percent of the total vehicle population on Guam. On an annual basis, the diesel-fueled
vehicles on Guam were estimated in 1993 to emit less than 0.1 percent of the maximum potential
sulfur dioxide emissions from other sources, given the 6000 ppm (maximum) level of sulfur in
diesel fuel at that time. Therefore, Guam's continued use of 6000 ppm (maximum) diesel fuel is
not expected to have any significant impact on the ambient air quality status of Guam, including
the status of the two areas designated as nonattainment for sulfur dioxide. Continuing the
existing diesel sulfur exemption or expanding it to the new diesel sulfur standard would not
cause an increase in emissions, but would forego a small emissions benefit if Guam used low
sulfur diesel fuel for its highway vehicles. There are no cross border issues.
The economy of Guam is ill suited to handling the additional expense of low sulfur diesel
fuel and upgrades to its storage and distribution system. It lacks internal petroleum supplies and
refining capabilities and relies on long distance imports. Essentially all of the island's petroleum
products were refined in Singapore in the early 1990's, but there were four sources: Singapore,
Indonesia, Australia and the Philippines. While Australia, Philippines and certain other Asian
countries have or will soon require low sulfur diesel fuel, they require a 500 ppm sulfur limit, not
the new 15 ppm sulfur limit required by today's action for the United States. In 1992, the No. 2
diesel fuel imports had sulfur content ranging between 3,900 to 5,000 ppm and cetane ranging
from 48 to 55.
If stationary sources continue to use high-sulfur diesel fuel, importing low sulfur diesel
fuel would require the costly construction of separate storage facilities. Even if Guam were to
import low sulfur diesel fuel for all its diesel needs from the United States mainland to avoid the
need to segregate the highway fuel, new storage facilities would be necessary to store larger
quantities of fuel, since shipments would be less frequent and possibly less reliable, due to the
IX-29
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
increased round-trip distances. It was estimated that Guam's fuel suppliers could conceivably be
subjected to $14,500,000 and $22,300,000 annually to comply with the low sulfur requirement,
and the cost per gallon of diesel fuel could increase by 30-46 cents per gallon. This high cost of
compliance is due to additional transportation costs associated with importing fuel from the
mainland, construction of new storage facilities needed to segregate low sulfur and high sulfur
fuel, and to store larger quantities of fuel, and the higher purchase price of the low sulfur [500
ppm] fuel. The new 15 ppm sulfur standard would add additional cost to the low sulfur fuel.
c. Commonwealth of Northern Mariana Islands (CNMI)
CNMI consists of 14 islands of volcanic origin that extend in a general north-south
direction for 388 nautical miles, with a land area of 176.5 square miles. It lies in the western part
of the Pacific Ocean about 1150 miles south of Tokyo, 108 miles north of Guam, and 5280 miles
from the United States mainland. Development and the population of Saipan, the most populated
island, are predominately on the western side. The population centers exist on Saipan (38,896),
Tinian (2,118), and Rota (2,295). The northern islands of the commonwealth have a population
of about 36 people.
The development being concentrated on the west side, meteorology (westward trade
winds) and lack of heavy industry have a beneficial impact on CNMI's air quality. As of 1994,
CNMI was in attainment with all primary NAAQS. The islands have approximately 200 miles of
roads, of which approximately 50 percent are paved, and about 500 diesel vehicles, of which
about 60 are owned by the CNMI government. Continuing the existing diesel sulfur exemption
or expanding it to the new sulfur standard would not cause an increase in emissions, but would
forego a small emissions benefit if CNMI used low sulfur diesel fuel for its highway vehicles.
There are no cross border issues.
Saipan has two bulk storage facilities for diesel fuel, and Rota and Tinian each have one.
The main use of diesel fuel is for electrical power generation. CNMI relies exclusively on diesel
fuel to provide electrical power. An insignificant amount of the diesel fuel is used for motor
vehicles. CNMI lacks internal petroleum supplies and refining capabilities and relies on long
distance imports. At the time of the request for exemption, CNMI relied on the importation of
fuels exclusively from refineries in Singapore. While Australia, Philippines and certain other
Asian countries have or will soon require low sulfur diesel fuel, they require a 500 ppm sulfur
limit, not the new 15 ppm sulfur limit required by today's action for the United States.
The economy of CNMI is ill suited to handling the additional expense of low sulfur diesel
fuel and upgrades to its storage and distribution system. The cost of diesel fuel on Saipan is
approximately 20 cents per gallon higher than that on the United States mainland [1994]. The
cost of diesel fuel on Rota and Tinian are higher than on Saipan. In order to meet the sulfur
requirement, it was estimated that low sulfur (500 ppm) diesel fuel would have to be imported
IX-30
-------
Chapter IX: Sulfur Control in Alaska & Territories
from the United States at a cost increase of 10 to 20 cents per gallon. The new 15 ppm sulfur
standard would add additional cost. A fuel supplier on CNMI estimated that in order to comply
with the low sulfur diesel fuel standard, four new storage tanks would need to be built for the
three [populated] islands at a cost of $2,000,000. Because of the added costs of shipping
materials to CNMI for construction projects, construction costs on CNMI are generally 20 to 40
percent higher than those of the United States mainland.
3. What Are the Options for the Territories?
We could include or exclude the territories in the geographical areas for which the new
diesel fuel sulfur standard applies. As in the early 1990's with the 500 ppm sulfur standard, we
believe that compliance with the new 15 ppm sulfur standard would result in small
environmental benefit, but major economic burden. We are also concerned about the impact to
vehicle owners and operators if new engine and emission control technologies were run using
high-sulfur fuel. We believe that for the sulfur exemption to be viable to vehicle owners and
operators, they would need access to either low sulfur fuel or pre-2007 technology vehicles that
could be run on high-sulfur fuel without significant engine damage or performance degradation.
Consequently, the territories will either need to be exempted from both the new fuel and vehicle
emission standards, or neither. Exempting them from only the fuel requirement would virtually
preclude the sale of new vehicles in the territories due to their sulfur sensitivity.
4. What Flexibility are we Offering the Territories?
As we proposed, today's action excludes American Samoa, Guam and the
Commonwealth of Northern Mariana Islands from the new diesel fuel sulfur requirement of 15
ppm and the 2007 heavy-duty diesel vehicle and engine emissions standards, and other
requirements associated with those emission standards. The territories will continue to have
access to 2006 heavy-duty diesel vehicle and engine technologies, at least as long as
manufacturers choose to market those technologies. We will not, however, allow the emissions
control technology in the territories to backslide from those available in 2006. If, in the future,
manufacturers choose to market only heavy-duty diesel vehicles and engines with 2007 and later
emission control technologies, we believe the market will determine when and if the territories
will make the investment needed to obtain and distribute the low sulfur diesel fuel necessary to
support these technologies.
This exclusion from emission standards does not apply to the new heavy-duty gasoline
engine and vehicle emission standards because low sulfur gasoline that complies with our
regulations will be available, and so concerns about damage to engines and emissions control
systems will not exist. This exclusion from emission standards also does not apply to light-duty
diesel vehicles and trucks because gasoline vehicles and trucks meeting the emission standards
and capable of fulfilling the same functions will be available. We believe that the market will
IX-31
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
determine when and if having access to new light-duty diesel technologies, in place of or in
addition to gasoline technologies, is important enough to obtain and distribute the low sulfur
diesel fuel needed to support those technologies.
As we also proposed, we are requiring all heavy-duty diesel motor vehicles and engines
for these territories to be certified and labeled to the applicable requirements (either to the 2006
model year standards and associated requirements under the exclusion, or to the standards and
associated requirements applicable for the model year of production under the nationwide
requirements) and warranted, as otherwise required under the Clean Air Act and EPA
regulations. Special recall and warranty considerations due to the use of excluded high sulfur
fuel are the same as those for Alaska during its exemption and transition periods (see the
discussion in previous section). To protect against this exclusion being used to circumvent the
emission requirements applicable to the rest of the United States (i.e., continental United States,
Alaska, Hawaii, Puerto Rico and the U.S. Virgin Islands) after 2006 by routing exempted (pre-
2007 technology) vehicles and engines through one of these territories, we are restricting the
importation of vehicles and engines from these territories into the rest of the United States. After
the 2006 model year, diesel vehicles and engines certified under this exclusion to meet the 2006
model year emission standards for sale in American Samoa, Guam and the Commonwealth of the
Northern Mariana Islands will not be permitted entry into the rest of the United States.
IX-32
-------
Appendix A: Legal Authority for Diesel Fuel Sulfur Control
Appendix A: Legal Authority for Diesel Fuel Sulfur
Control
We are promulgating diesel sulfur controls pursuant to our authority under Section
21 l(c)(l) of the Clean Air Act. This section gives us the authority to "control or prohibit the
manufacture, introduction into commerce, offering for sale, or sale" of any fuel or fuel additive
(A) whose emission products, in the judgment of the Administrator, cause or contribute to air
pollution "which may be reasonably anticipated to endanger the public health or welfare" or (B)
whose emission products "will impair to a significant degree the performance of any emission
control device or system which is in general use, or which the Administrator finds has been
developed to a point where in a reasonable time it would be in general use" were the fuel control
or prohibition adopted. The following sections describe current regulatory requirements that
affect diesel sulfur content, and explain our bases for controlling diesel sulfur under Section
21 l(c)(l). This section contains a summary of the issues. Refer to the Preamble and RIA for
more details.
A. EPA's Current Regulatory Requirements for Diesel
We currently have regulatory requirements for diesel fuel adopted under Sections 21 l(c)
and 21 l(i) of the Act. Section 21 l(i)(l) prohibits the manufacture, sale, supply, offering for sale
or supply, dispensing, transport, or introduction into commerce of motor vehicle diesel fuel
which contains a concentration of sulfur in excess of 0.05 percent by weight, and which fails to
meet a cetane index minimum of 40, or aromatics maximum of 35 percent beginning October 1,
1993. Section 21 l(i)(2) requires the Administrator to promulgate regulations to implement and
enforce the requirements of section 21 l(i)(l), and authorizes the Administrator to require that
diesel fuel not intended for motor vehicles be dyed in order to segregate that fuel from motor
vehicle diesel fuel. See 40 CFR §80.29.
B. How the Proposed Diesel 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 diesel fuel based on both of these criteria. Under the first criterion, we believe that emissions
A-l
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
products of sulfur in diesel fuel used in heavy-duty engines contribute to PM pollution. Under
the second criterion, we believe that sulfur in diesel fuel will significantly impair the emissions
control systems expected to be used in heavy-duty engines designed to meet proposed emissions
standards. The following sections summarize our analysis of each criterion.
1. Health and Welfare Concerns of Air Pollution Caused by Sulfur
in Diesel Fuel
We believe that the emission products of diesel sulfur contribute to air pollution that can
reasonably be anticipated to endanger public health and welfare. Sulfur in diesel fuel leads
directly to emissions of SO2 and sulfate PM from the exhaust of diesel vehicles. SO2 emissions
from diesel engines are directly proportional to the amount of sulfur in the fuel. SO2 is oxidized
in the atmosphere to SOS which then combines with water to form sulfuric acid (H2SO4) and
further combines with ammonium in the atmosphere to form ammonium sulfate aerosols. These
aerosols are what is often referred to as sulfate PM. This sulfate PM comprises a significant
portion of the "secondary" PM that does not come directly from the tailpipe, but is nevertheless
formed in the atmosphere from exhaust pollutants. Exposure to secondary PM may be different
from that of PM emitted directly from the exhaust, but the health concerns of secondary PM are
just as severe as for directly emitted particulate matter, with the possible exception of the
carcinogenicity concerns with diesel PM.
Approximately 1-2 percent of the sulfur in diesel fuel is not converted into SO2, but is
instead further oxidized into SOS which then forms sulfuric acid aerosols (sulfate PM) as it
leaves the tailpipe. While only a small fraction of the overall sulfur is converted into sulfate
emissions in the exhaust of diesel vehicles today, it nevertheless accounts for approximately 10
percent of the total PM emissions from diesel engines today. Furthermore, with the application
of the exhaust emission control devices that will be necessary to meet either the PM or NOx
emission standards in 2007, the conversion rate of sulfur in the fuel to sulfate PM in the exhaust
increases dramatically. This sulfate PM is directly proportional to the sulfur concentration in the
fuel. The health and welfare implications of the emissions of these compounds are discussed in
greater detail in Section II of the Preamble and RIA.
Section 21 l(c)(2)(A) requires that, prior to adopting fuel controls 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 emissions from diesel vehicles which are
affected by sulfur in diesel fuel is described in more detail in Sections II of the Preamble and
RIA.
A-2
-------
Appendix A: Legal Authority for Diesel Fuel Sulfur Control
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)(B) 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 is finalizing standards for fuels and vehicles together. Thus, it is
first important to consider that the sulfur standards are not being promulgated as an alternative to
emissions standards, but in addition to such standards, and as a necessary prerequisite to ensuring
that heavy-duty engines can meet the engine standards. In addition, the heavy-duty emission
standards being finalized today will begin with the 2007 model year, and even at that time, many
older technology heavy-duty 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 begin achieving some emissions benefits immediately through
reducing emissions from the existing fleet of motor vehicles.
EPA has also considered more stringent emissions standards under § 202 as an alternative
to regulating diesel sulfur. However, for the reasons described in Section in of the Preamble and
RIA, the Agency concludes that the heavy-duty emission standards represent the levels of
emission control that are economically and technologically feasible from heavy-duty engines and
vehicles beginning in 2007. Moreover, EPA considered heavy-duty standards alone as an
alternative to regulating diesel sulfur. However, as described in Preamble Section HI, the Agency
concludes that the heavy-duty standards would not be feasible without control of diesel sulfur.
For these reasons, EPA finds that the alternatives of either more stringent engine and vehicle
standards, or engine and vehicle standards without sulfur control, are not technologically or
economically feasible options to regulating diesel sulfur.
EPA's consideration of other technologically and economically feasible means of
achieving emissions standards under § 202 of the Act supports the conclusion that the diesel
sulfur standards finalized today represent an appropriate exercise of the Agency's discretion
under § 21 l(c)(l)(A), even when the heavy-duty engine and vehicle standards are considered.
2. Impact of Diesel Sulfur Emission Products on Emission Control
Systems
EPA believes that sulfur in diesel fuel can significantly impair the emissions control
technology of engines designed to meet the final emissions standards. We know that diesel
sulfur has a negative impact on vehicle emission controls. This is not a new development. As
A-3
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
discussed in Section in of the Preamble to the final rule, aftertreatment technologies exist that we
believe will be capable of achieving dramatic reductions in NOx and PM emissions from diesel
engines for the 2007 model year. The aftertreatment technology for PM is already in an
advanced state of development and being tested in diesel vehicle fleet demonstrations in the U.S.
and Europe. The NOx adsorber aftertreatment technology, while already in commercial use in
other applications is in a comparatively earlier state of development for use on diesel vehicles,
but as discussed in Section HI of the preamble and RIA, tremendous progress is already being
made and EPA believes the lead time between now and 2007 will provide ample opportunity to
bring this technology into the diesel vehicle marketplace. EPA believes that these PM and NOx
aftertreatment technologies will be in general use on diesel vehicles by MY 2007, with the diesel
sulfur controls adopted in this rule.
These aftertreatment technologies are ineffective in reducing NOx and PM emissions and
incapable of being introduced widely into the marketplace at the diesel fuel sulfur concentrations
typical today. Not only does their efficiency at reducing NOx and PM emissions fall off
dramatically at elevated fuel sulfur concentrations, but vehicle driveability impacts and
permanent damage to the aftertreatment systems are also possible.
In order to ensure passive regeneration of the diesel particulate filter as described in
Section HI of the Preamble and RIA, we are expecting that significant amounts of precious group
metals (primarily platinum) will be used in their washcoat formulations. There are two primary
mechanisms by which sulfur in diesel fuel can limit the effectiveness or robustness of diesel
particulate filters which rely on an oxidizing catalyst function from platinum. The first is
inhibition of the oxidation of NO to NO2 and the second is the preferential oxidation of SO2 to
SOS, resulting in production of sulfate particulate matter.
All of the NOx aftertreatment technologies discussed in Section HI of the Preamble and
RIA are expected to utilize platinum to oxidize NO to NO2 to improve the NOx reduction
efficiency of the catalysts at low temperatures . In the case of the NOx absorber, conversion of
NO to NO2 is also an essential part of the process of NOx storage. This reliance of NO2 as an
integral part of the reduction process means that the functioning of the NOx aftertreatment
technologies, like the PM aftertreatment technologies, will be significantly impaired by sulfur in
diesel fuel.
3. Sulfur Levels that Exhaust Aftertreatment for Heavy-Duty
Vehicles Can Tolerate
As discussed in Section in.F. of the Preamble, there are three key factors which when
taken together lead us to believe that a diesel fuel sulfur cap of 15 ppm is necessary so that the
NOx and PM aftertreatment technology on diesel engines will function properly and be able to
meet the emission standards . These factors, as discussed in more detail in Section HI of the
A-4
-------
Appendix A: Legal Authority for Diesel Fuel Sulfur Control
Preamble and RIA, are the implications of sulfur levels in excess of 15 ppm on the efficiency and
reliability of the systems and their impact on the fuel economy of the vehicle.
The efficiency of emission control technologies at reducing harmful pollutants is directly
impacted by sulfur in diesel fuel. Initial and long term conversion efficiencies for NOx, HC, CO
and diesel PM emissions are significantly reduced by catalyst poisoning and catalyst inhibition
due to sulfur. NOx conversion efficiencies with the NOx adsorber technology in particular are
dramatically reduced in a very short time due to sulfur poisoning of the NOx storage bed. In
addition total PM control efficiency is negatively impacted by the formation of sulfate PM. The
formation of sulfate PM is likely to be in excess of the total PM standard proposed today, unless
diesel fuel sulfur levels are no higher than 15 ppm. When sulfur is kept at these low levels, both
PM and NOx aftertreatment devices are expected to operate at high levels of conversion
efficiency, allowing compliance with the PM and NOx emissions standards.
The reliability of the emission control technologies to continue to function as required
under all operating conditions for the life of the vehicle is also directly impacted by sulfur in
diesel fuel. As discussed in Section in of the Preamble and RIA, sulfur in diesel fuel can prevent
proper operation and regeneration of both NOx and PM control technologies leading to
permanent loss in emission control effectiveness and even catastrophic failure of the systems.
For example, if regeneration of a PM filter does not occur, catastrophic failure of the filter will
occur. We believe, based on information available to us, that diesel fuel sulfur levels of 15 ppm
are needed and would allow these technologies to operate properly throughout the life of the
vehicle, including proper periodic or continuous regeneration .
The sulfur content of diesel fuel will also impact the fuel economy of vehicles equipped
with NOx and PM aftertreatment technologies. As discussed in detail in Section HI of the
Preamble and RIA, NOx adsorbers are expected to consume diesel fuel in order to cleanse
themselves of stored sulfates and maintain efficiency. The larger the amount of sulfur in diesel
fuel, the greater this impact on fuel economy. Likewise PM trap regeneration is inhibited by
sulfur in diesel fuel. This leads to increased PM loading in the diesel particulate filter, increased
exhaust backpressure, and poorer fuel economy. Thus for both NOx and PM technologies the
lower the fuel sulfur level the better the fuel economy of the vehicle.
As a result of these factors, we believe that it is appropriate to ensure that 15 ppm sulfur
diesel fuel is available and are therefore capping in-use sulfur levels there.
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
A-5
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
which are or will be in general use that require the proposed fuel control with such devices or
systems which are or will be in general use that do not require the proposed fuel control. As
described below, we conclude that the aftertreatment technology expected to be used to meet the
final heavy-duty standards would be significantly impaired by operation on high sulfur diesel
fuel. Our analysis of the available scientific and economic data can be found in the Preamble and
RIA, 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 finalized in the rule, and
a cost effectiveness analysis of the final sulfur control and heavy-duty emissions standards.
Under Section 21 l(c)(2)(B), EPA is also required to compare the costs and benefits of achieving
the adopted vehicle emissions standards through emissions control systems that would not
require the proposed control of sulfur , if any such systems are or 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 on heavy-duty engines and vehicles that can meet the
final heavy-duty emission standards and would not be significantly impaired by diesel fuel with
high sulfur levels. NOx and PM emissions can not be reduced anywhere near the magnitude
contemplated by the standards promulgated today without the application of aftertreatment
technology. While some improvement may yet be possible in engine out emissions, these
improvements will not allow the engines to meet the set of emission standards promulgated
today. As discussed in Sections HI and IV of the Preamble and RIA, there are a number of
aftertreatment technologies that are currently being developed for both NOx and PM control with
varying levels of effectiveness, sulfur sensitivity, and potential application to heavy-duty diesel
vehicles.
As discussed in Sections HI of the Preamble and RIA, all of the aftertreatmrent
technologies that could be used to meet the PM or NOx standards are significantly impaired by
the sulfur in diesel fuel. For PM control, EPA is not aware of a PM aftertreatment technology
that is capable of meeting the PM standard adopted today and that would not need the level of
sulfur control adopted in this rule. In addition, the NOx aftertreatment technologies evaluated by
EPA all rely on the use of catalytic processes to increase the effectiveness of the device in
reducing NOx emissions. For example both NOx adsorbers and compact SCR would rely on
noble metals to oxidize NO to NO2, to increase NOx conversion efficiency at the lower exhaust
temperatures found in diesel motor vehicle operation. This catalytic process, however, produces
sulfate PM from the sulfur in the diesel fuel, and these NOx aftertreatment devices need the level
of sulfur control adopted in this rule in order for the vehicle to comply with the PM standard.
For NOx control, both NOx adsorbers and compact SCR are significantly impaired by
sulfur in diesel fuel, and both technologies would need very large reductions in sulfur from
current levels to meet the NOx standard adopted today.
A-6
-------
Appendix A: Legal Authority for Diesel Fuel Sulfur Control
In addition, compact SCR is not a technology that would be generally available by the
model year 2007 time frame. Significant and widespread changes to the fuel distribution system
infrastructure would have to be made and in place by then, and there is no practical expectation
that this would occur, with or without the low sulfur standard adopted today. While it is feasible
and practical to expect that compact SCR may have a role in specific controlled circumstances,
such as certain centrally fueled fleets, it is not realistic at this time to expect that the fuel
distribution system infrastructure changes needed for widespread and general use of compact
SCR on heavy-duty diesel vehicles will be in place by the model year 2007 time frame. In
addition, even if SCR were used to obtain the emission performance required by today's
standards, it is not clear that the vehicles would continue to maintain that level of performance
in-use. SCR technology requires continued replacement of the urea supply on the vehicle by the
vehicle operator. Failure to do so would make the SCR system completely ineffective. While
various options to encourage vehicle operators to maintain their urea supply have been suggested
(e.g., electronically monitoring urea injection and reducing engine power if not), none provide
reasonable assurance, and often raise other serious concerns such as the safety of vehicle
operation. Finally, EPA believes that the requirement of a cost benefit analysis under section
21 l(c)(2)(B) is not aimed at evaluating emissions control technologies that would require
significant additional or different EPA fuel control regulations before the technology could be
considered generally available.
In sum, EPA believes that both PM and NOx aftertreatment technologies require the level
of sulfur control adopted today to meet the PM standards. There is no PM or NOx emissions
control device or system that would be in general use that does not need this level of sulfur
control for purposes of controlling PM. EPA also believes that NOx aftertreatment technologies
either need the level of sulfur adopted today to be considered generally available for use to meet
the NOx standard, or need sulfur controls approximating those adopted and even with such sulfur
control would not be considered generally available for use to meet the NOx standard.
As described in Section in of the Preamble, EPA anticipates that all the diesel heavy-
duty engine and vehicle technologies expected to be used to meet the final heavy-duty standards
will require the use of low sulfur diesel fuel. If we do not control diesel sulfur to the finalized
levels, we would not be able to set heavy-duty standards as stringent as those we are finalizing
today. Moreover, vehicles already on the road would continue to emit at slightly higher levels
than they would if operated on low sulfur fuel. Consequently, EPA concludes that the benefits
that would be achieved through implementation of the vehicle and sulfur control programs
cannot be achieved through the use of emission control technology that does not need the sulfur
control adopted in this rule, and would be generally available to meet the emissions standards
adopted in this rule .
A-7
-------
Heavy-Duty Standards / Diesel Fuel RIA - December 2000 EPA420-R-00-026
This also means that if EPA were to adopt emissions control standards without
controlling diesel sulfur content, the standards would be significantly less stringent than those
finalized today based on what would be technologically feasible with current sulfur levels.
5. Effect of Diesel 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 diesel fuel, but rather is controlling the levels of sulfur in diesel, this finding is not required
prior to regulation. However, EPA does not believe that the finalized sulfur control will result in
the use of any other fuel or additive that will produce emissions that will endanger public health
or welfare to the same or greater degree as the emissions produced by diesel with uncontrolled
sulfur levels.
Unlike in the case of unleaded gasoline in the past where lead served a primary function
of providing the necessary octane for the vehicles to function properly, sulfur does not serve any
useful function in diesel fuel. It is not added to diesel fuel, but comes naturally in the crude oil
into which diesel fuel is processed. If it were not for the fact that it costs money to remove sulfur
from diesel fuel, it would have been removed years ago to improve the maintenance and
durability characteristics of diesel engines. EPA is unaware of any function of sulfur in diesel
fuel that might have to be replaced once sulfur is removed, with the possible exception of
lubricity characteristics of the fuel. As discussed in the Preamble, there is some evidence to
suggest that as sulfur is removed from diesel fuel the natural lubricity characteristics of diesel
fuel may be reduced. Depending on the crude oil and the manner in which desulfurization occurs
some low sulfur diesel fuels can exhibit poor lubricity characteristics. To offset this concern
lubricity additives are sometimes added to the diesel fuel. These additives, however, are already
in common use today and EPA is unaware of any health hazards associated with the use of these
additives in diesel fuel and would merely be used in larger fractions of the diesel fuel pool. We
do not anticipate that their use would produce emissions which would reduce the large public
health and welfare benefits that this rule would achieve.
EPA is unaware of any other additives that might be necessary to add to diesel fuel to
offset the existence of sulfur in the fuel. EPA is also unaware of any additives that might need to
be added to diesel fuel to offset any other changes to diesel fuel which might occur during the
process of removing sulfur from diesel fuel.
A-8
-------
EPA420-R-00-026
Appendix B: VMT Distribution
APPENDIX B: Vehicle Miles Traveled by HDDE Class for Split by Pre-2007 and
2007+ Model Years (MY)
Calendar
Year
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
2035
2036
LHDDE VMT [million miles]
<2007 2007+ VMT°A
MY MY Total 2^+
47,678 0 47,678 0%
49,089 0 49,089 0%
45,046 5,454 50,500 10.8%
39,542 12,485 52,027 24.0%
34,682 18,872 53,554 35.2%
30,394 24,687 55,081 44.8%
26,612 29,996 56,608 53.0%
23,280 34,855 58,135 60.0%
20,345 39,317 59,662 65.9%
17,761 43,428 61,189 71.0%
15,487 47,229 62,716 75.3%
13,487 50,756 64,243 79.0%
11,728 54,043 65,770 82.2%
10,181 57,116 67,297 84.9%
8,821 60,003 68,824 87.2%
7,626 62,725 70,351 89.2%
6,575 65,303 71,878 90.9%
5,652 67,753 73,405 92.3%
4,840 70,092 74,932 93.5%
4,127 72,332 76,459 94.6%
3,499 74,487 77,986 95.5%
2,947 76,566 79,513 96.3%
2,462 78,578 81,040 97.0%
2,034 80,533 82,567 97.5%
1,658 82,436 84,094 98.0%
1,326 84,295 85,621 98.5%
1,023 86,125 87,148 98.8%
755 87,920 88,675 99.1%
526 89,677 90,202 99.4%
319 91,411 91,729 99.7%
138 93,118 93,256 99.9%
0 94,783 94,783 100%
MHDDE VMT [million miles]
<2007 2007+ VMT°A
MY MY Total 2^+
42,877 0 42,877 0%
44,145 0 44,145 0%
41,914 3,500 45,414 7.7%
37,025 9,762 46,787 20.9%
32,698 15,462 48,160 32.1%
28,870 20,664 49,534 41.7%
25,484 25,423 50,907 49.9%
22,489 29,791 52,280 57.0%
19,842 33,811 53,653 63.0%
17,500 37,526 55,026 68.2%
15,428 40,971 56,400 72.6%
13,594 44,178 57,773 76.5%
11,970 47,176 59,146 79.8%
10,531 49,989 60,519 82.6%
9,254 52,639 61,893 85.0%
8,119 55,147 63,266 87.2%
7,110 57,529 64,639 89.0%
6,212 59,801 66,012 90.6%
5,410 61,975 67,385 92.0%
4,694 64,065 68,759 93.2%
4,052 66,080 70,132 94.2%
3,477 68,028 71,505 95.1%
2,959 69,919 72,878 95.9%
2,493 71,759 74,252 96.6%
2,071 73,554 75,625 97.3%
1,689 75,309 76,998 97.8%
1,340 77,031 78,371 98.3%
1,022 78,722 79,745 98.7%
745 80,373 81,118 99.1%
472 82,019 82,491 99.4%
224 83,640 83,864 99.7%
0 85,237 85,237 100%
HHDDE VMT [million miles]
<2007 2007+ VMT°A
MY MY Total 2^+
166,573 0 166,573 0%
171,502 0 171,502 0%
161,793 14,638 176,431 8.3%
141,114 40,651 181,766 22.4%
122,939 64,162 187,101 34.3%
106,983 85,453 192,436 44.4%
92,990 104,780 197,770 53.0%
80,733 122,372 203,105 60.3%
70,006 138,435 208,440 66.4%
60,625 153,150 213,775 71.6%
52,429 166,681 219,110 76.1%
45,274 179,171 224,445 79.8%
39,031 190,749 229,780 83.0%
33,587 201,528 235,115 85.7%
28,844 211,606 240,450 88.0%
24,712 221,073 245,785 89.9%
21,115 230,005 251,120 91.6%
17,984 238,470 256,455 93.0%
15,261 246,529 261,790 94.2%
12,892 254,233 267,124 95.2%
10,832 261,628 272,459 96.0%
9,040 268,754 277,794 96.7%
7,483 275,647 283,129 97.4%
6,128 282,336 288,464 97.9%
4,950 288,849 293,799 98.3%
3,925 295,209 299,134 98.7%
3,009 301,460 304,469 99.0%
2,208 307,596 309,804 99.3%
1,565 313,574 315,139 99.5%
963 319,511 320,474 99.7%
444 325,364 325,809 99.9%
0 331,144 331,144 100%
SUBTOTAL [million miles]
VMT%
<2007 2007+ Total 2°°7+
MY MY ^
257,128 0 257,128 0%
264,736 0 264,736 0%
248,753 23,591 272,344 8.7%
217,682 62,898 280,580 22.4%
190,319 98,496 288,815 34.1%
166,246 130,804 297,050 44.0%
145,086 160,199 305,285 52.5%
126,502 187,018 313,520 59.7%
110,192 211,564 321,756 65.8%
95,886 234,105 329,991 70.9%
83,344 254,881 338,226 75.4%
72,355 274,106 346,461 79.1%
62,729 291,968 354,696 82.3%
54,299 308,633 362,931 85.0%
46,918 324,248 371,167 87.4%
40,457 338,945 379,402 89.3%
34,800 352,836 387,637 91.0%
29,848 366,024 395,872 92.5%
25,511 378,596 404,107 93.7%
21,712 390,630 412,342 94.7%
18,383 402,194 420,578 95.6%
15,465 413,348 428,813 96.4%
12,904 424,144 437,048 97.0%
10,655 434,628 445,283 97.6%
8,679 444,839 453,518 98.1%
6,940 454,813 461,754 98.5%
5,372 464,617 469,989 98.9%
3,986 474,238 478,224 99.2%
2,835 483,624 486,459 99.4%
1,753 492,941 494,694 99.6%
806 502,123 502,929 99.8%
0 511,165 511,165 100%
Urban Bus VMT [million miles]
< 2007 MY 2007+ VMT
MY Total °A
2007+
3,189 0 3,189 0*%"
3,283 0 3,283 0%
3,243 134 3,378 4.0%
3,073 407 3,480 11.7%
2,896 686 3,582 19.1%
2,713 971 3,684 26.4%
2,523 1,263 3,786 33.4%
2,328 1,560 3,888 40.1%
2,128 1,863 3,990 46.7%
1,923 2,170 4,092 53.0%
1,713 2,482 4,195 59.2%
1,499 2,797 4,297 65.1%
1,283 3,116 4,399 70.8%
1,065 3,436 4,501 76.3%
848 3,755 4,603 81.6%
639 4,066 4,705 86.4%
451 4,356 4,807 90.6%
302 4,608 4,909 93.9%
210 4,802 5,012 95.8%
168 4,946 5,114 96.7%
134 5,082 5,216 97.4%
106 5,212 5,318 98.0%
84 5,337 5,420 98.5%
65 5,457 5,522 98.8%
51 5,574 5,624 99.1%
39 5,688 5,726 99.3%
26 5,803 5,829 99.6%
19 5,912 5,931 99.7%
13 6,020 6,033 99.8%
6 6,129 6,135 99.9%
3 6,234 6,237 99.9%
0 6,339 6,339 100%
Total VMT [million miles]
<2007 2007+ VMT°A
MY MY Total 2^+
260,317 0 260,317 0%
268,019 0 268,019 0.0%
251,996 23,726 275,722 8.6%
220,755 63,305 284,059 22.3%
193,215 99,182 292,397 33.9%
168,958 131,776 300,734 43.8%
147,609 161,462 309,071 52.2%
128,830 188,578 317,409 59.4%
112,320 213,426 325,746 65.5%
97,808 236,275 334,083 70.7%
85,057 257,363 342,420 75.2%
73,854 276,903 350,758 78.9%
64,011 295,084 359,095 82.2%
55,363 312,069 367,432 84.9%
47,766 328,003 375,770 87.3%
41,096 343,011 384,107 89.3%
35,251 357,193 392,444 91.0%
30,149 370,632 400,782 92.5%
25,721 383,398 409,119 93.7%
21,880 395,576 417,456 94.8%
18,517 407,277 425,793 95.7%
15,570 418,560 434,131 96.4%
12,987 429,481 442,468 97.1%
10,721 440,085 450,805 97.6%
8,730 450,413 459,143 98.1%
6,979 460,501 467,480 98.5%
5,397 470,420 475,817 98.9%
4,005 480,150 484,155 99.2%
2,848 489,644 492,492 99.4%
1,760 499,069 500,829 99.6%
810 508,357 509,167 99.8%
0 517,504 517,504 100%
B-l
------- |