Control of Emissions from Marine SI
and Small SI Engines, Vessels, and
Equipment
Final Regulatory Impact Analysis
United States
Environmental Protection
Agency
-------
Control of Emissions from Marine SI
and Small SI Engines, Vessels, and
Equipment
Final Regulatory Impact Analysis
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
v>EPA
United States EPA420-R-08-014
Environmental Protection _ ^ , „„„
Agency September 2008
-------
Table of Contents
Executive Summary ES-1
CHAPTER 1: Industry Characterization
1.1 Manufacturers of Small SI Engines 1-1
1.2 Manufacturers of Marine Spark-Ignition Engines 1-2
1.3 Fuel System Components 1-6
CHAPTER 2: Air Quality, Health, and Welfare Concerns
2.1 Ozone 2-3
2.2 Paniculate Matter 2-14
2.3 Air Quality Modeling Methodology 2-33
2.4 Air Toxics 2-39
2.5 Carbon Monoxide 2-44
2.6 Acute Exposure to Air Pollutants 2-47
CHAPTER 3: Emission Inventory
3.1 Overview of Small Nonroad and Marine SI Engine Emissions Inventory
Development 3-2
3.2 Baseline Emission Inventory Estimates 3-4
3.3 Contribution of Small Nonroad and Marine SI Engines to National Emissions
Inventories 3-19
3.4 Controlled Nonroad Small Spark-Ignition and Marine Engine Emission Inventory
Development 3-26
3.5 Projected Emissions Reductions from the New Standards 3-34
3.6 Emission Inventories Used for Air Quality Modeling 3-43
CHAPTER 4: Feasibility of Exhaust Emission Control
4.1 General Description of Spark-Ignition Engine Technology 4-1
4.2 General Description of Exhaust Emission Control Technologies 4-8
4.3 Feasibility of Small SI Engine Standards 4-23
4.4 Feasibility of Outboard/Personal Watercraft Marine Engine Standards 4-38
4.5 Feasibility of Sterndrive/Inboard Marine Engine Standards 4-43
4.6 Feasibility of Standard for Marine Generator Sets 4-55
4.7 Test Procedures 4-56
4.8 Impacts on Safety, Noise, and Energy 4-72
APPENDIX 4A: Normalized Modal Emissions for a 7.4 L MPI SD/I 4-81
CHAPTER 5: Feasibility of Evaporative Emission Control
5.1 Diurnal Breathing Loss Evaporative Emissions 5-2
5.2 Running Loss Emissions 5-26
5.3 Fuel Tank Permeation 5-29
5.4 Fuel/Vapor Hose Permeation 5-70
5.5 Other Evaporative Emissions 5-89
5.6 Evaporative Emission Test Procedures 5-91
5.7 Impacts on Noise, Energy, and Safety 5-109
-------
APPENDIX 5A: Diurnal Temperature Traces 5-112
APPENDIX 5B: Emission Results for Small SI Equipment Fuel Tanks Showing Effect of
Venting on Diffusion 5-115
APPENDIX 5C: Diurnal Emission Results: Canister and Passive-Purge 5-120
APPENDIX 5D: Material Properties of Common Fuel System Materials 5-121
APPENDIX 5E: Diurnal Test Temperature Traces 5-125
CHAPTER 6: Costs of Control
6.1 Methodology 6-1
6.2 Exhaust Emission Control Costs for Small SI Engines 6-2
6.3 Exhaust Emission Control Costs for Outboard and Personal Watercraft Engines ... 6-44
6.4 Exhaust Emission Control Costs for Sterndrive/Inboard Marine Engines 6-56
6.5 Evaporative Emission Control Costs for Small SI Equipment 6-64
6.6 Costs of Evaporative Emission Controls for Marine Vessels 6-78
6.7 Cost Sensitivity Analysis 6-92
CHAPTER 7: Cost Per Ton
7.1 30-Year Net Present Value Cost effectiveness 7-1
7.2 Results 7-2
CHAPTER 8: Cost-Benefit Analysis
8.1 Overview 8-2
8.2 Air Quality Impacts for Benefits Analysis 8-5
8.3 PM-Related Health Benefits Estimation - Methods and Inputs 8-7
8.4 Health Impact Functions 8-9
8.5 Economic Values for Health Outcomes 8-27
8.6 Benefits Analysis Results for the Final Standards 8-35
8.7 Comparison of Costs and Benefits 8-45
Appendix 8 A: Sensitivity Analyses of Key Parameters in the Benefits Analysis 8-52
Appendix 8B: Health-Based Cost-Effectiveness of Reductions in Ambient O3 and PM25
Associated with the Final Small SI and Recreational Marine Engine Rule 8-61
CHAPTER 9: Economic Impact Analysis
9.1 Overview and Results 9-1
9.2 Economic Methodology 9-16
9.3 EIM Data Inputs and Model Solution 9-36
9.4 Methods for Describing Uncertainty 9-66
Appendix 9A: Impacts on Small SI Markets 9-74
Appendix 9B: Impacts on Marine SI Markets 9-92
Appendix 9C: Time Series of Social Cost 9-104
Appendix 9D: Overview of Model Equations and Calculation 9-108
Appendix 9E: Elasticity Parameters for Economic Impact Modeling 9-113
Appendix 9F: Derivation of Supply Elasticity 9-127
Appendix 9G: Initial Market Equilibrium - Price Forecasts 9-129
Appendix 9H: Sensitivity Analysis 9-131
CHAPTER 10: Small-Business Flexibility Analysis
10.1 Overview of the Regulatory Flexibility Act 10-1
-------
10.2 Need for and Objective of the Rulemaking 10-2
10.3 Summary of Significant Public Comments 10-3
10.4 Definition and Description of Small Entities 10-4
10.5 Type and Numbers of Small Entities Affected 10-5
10.6 Reporting, Recordkeeping, and Compliance Requirements 10-7
10.7 Steps Taken to Minimize the Impact on Small Entities 10-8
10.8 Projected Economic Effects of the Rulemaking 10-17
CHAPTER 11: Regulatory Alternatives
11.1 Identification of Alternative Program Options 11-1
11.2 Cost per Engine 11-7
11.3 Emission Reduction 11-18
11.4 Cost Effectiveness 11-20
11.5 Summary and Analysis of Alternative Program Options 11-21
APPENDIX 11 A: Emission Factors for the Less Stringent Alternative 11-26
APPENDIX 1 IB: Emission Factors for the More Stringent Alternative 11-27
-------
Executive Summary
Executive Summary
The Environmental Protection Agency (EPA) is establishing new requirements to reduce
emissions of hydrocarbon (HC) and oxides of nitrogen (NOx) from nonroad small spark ignited
engines below 19kW ("Small SI engines") and marine spark ignited engines ("Marine SI
engines"). This rule includes exhaust and evaporative emission standards for these engines as
well as related gasoline fuel tanks and fuel lines.
This executive summary describes the relevant air-quality issues, highlights the new exhaust
and evaporative emission standards, and gives an overview of the analyses in the rest of this
document.
Air Quality Background and Environmental Impact of the Rule
Emissions from Small SI engines and equipment and Marine SI engines and vessels
contribute to a number of serious air pollution problems and will continue to do so in the future
absent further reduction measures. Such emissions lead to adverse health and welfare effects
associated with ozone, particulate matter (PM), nitrogen oxides (NOx), volatile organic
compounds (VOC) including toxic compounds, and carbon monoxide (CO). These emissions
also cause significant public welfare harm, such as damage to crops and regional haze.
Millions of Americans continue to live in areas with unhealthy air quality that may endanger
public health and welfare. As of March 2008 approximately 139 million people live in the 72
areas that are designated as nonattainment for the 8-hour ozone National Ambient Air Quality
Standards (NAAQS). In addition, approximately 88 million people live in areas that are
designated as nonattainment for the PM2 5 NAAQS. Federal, state, and local governments are
working to bring ozone and PM levels into attainment with the NAAQS. The reductions
included in this rule will be useful to states in attaining and maintaining the ozone, CO, and PM
NAAQS.
In 2002, emissions from land-based nonroad Small SI engines and Marine SI
engines were estimated to be about 26 percent of the total mobile-source inventory of VOC
emissions and 1 percent of the NOx inventory. As presented in Figures 1 and 2, this rule will
significantly reduce future Small SI and Marine SI emission inventories.
ES-1
-------
Final Regulatory Impact Analysis
Figure 1: Small Spark Ignition VOC+NOx Baseline and Phase 3 Control Emission
Inventory
1 400 000
1 200 000
1 000 000
800 000
V)
1
600 000
400 000
200 000
n
20
\ ^^
\ ^^^
.---'"'
*- s ..--'"
00 2005 2010 2015 2020 2025 2030 2035 20
Year
40
Figure 2: Marine Spark Ignition VOC+NOx Baseline and Phase 3 Control Emission
Inventory
1 Tnnnn
1 nnn nnn
finnnnn
in
O 600000
I-
idnnnnn
Ofw-t (YY1
0 -
Bss^linfi
N^ ___
\
».
*\
*.
2000 2005 2010 2015 2020 2025 2030 2035 2040
Year
ES-2
-------
Executive Summary
Exhaust and Evaporative Emission Standards
Tables 1 through 4 show the exhaust and evaporative emission standards and when they will
apply. For Small SI nonhandheld engines, the standards are expected to result in the use of
engine modificiations, aftertreatment systems, and some use of electronic fuel injection in Class
II engines. As shown in Tables 1 through 4, we are phasing in many of the standards over time
to address considerations of lead time, workload, and overall feasibility. In addition, the rule
includes other provisions designed to address the transition to meeting the standards.
Table 1: Small SI Nonhandheld Engine Exhaust Emission Standards and Schedule
Engine Class
Class I (>80cc to <225cc)b
Class II (>225cc)
Model Year
2012
2011
HC+NOx
[g/kW-hr]
10.0
8.0
coa
[g/kW-hr]
610
610
a 5 g/kW-hr CO for Small SI engines powering marine generators.
b Nonhandheld engines at or below 80cc will be subject to the emission standards for handheld engines.
Table 2: Small SI Equipment Evaporative Emission Standards and Schedule
Standard Level
Handheld
Class I
Class II
Fuel Line Permeation
15 g/nf/day
2012a'b
2009
2009
Tank Permeation
1.5 g/m2/day
2009-2013°
2012
2011
Running Loss
Design Standard
NA
2012
2011
a 2013 for small-volume families.
b A separate set of declining fuel line permeation standards applies for cold-weather equipment from 2012 through
2016. A standard of 225 g/nf/day for cold-weather equipment fuel lines applies for 2016 and later.
0 2009 for families certified in California, 2013 for small-volume families, 2011 for structurally integrated nylon fuel
tanks, and 2010 for remaining families.
ES-3
-------
Final Regulatory Impact Analysis
Table 3a: Outboard/PWC Marine SI Engine Exhaust Standards and Schedule3
Pollutant
HC+NOx
CO
Powe^
P < 4.3 kW
P> 4.3 kW
P < 40 kW
P> 40 kW
Emission Standard15
30.0
2.1+0.09 x (151 + 557/P09)
500 - 5.0 x p
300
Model Year
2010
2010
2010
2010
a These engines are also subject to not-to-exceed standards
b P = maximum engine power in kilowatts (kW).
Table 3b: Sterndrive/Inboard Marine SI Engine Exhaust Standards and Schedule
Power3
P < 373 kWb
High-performance engines < 485 kW d
High-performance engines < 485 kW d
Model Year
2010C
2010
2011
2010
2011
HC+NOx fs/kW-hrl
5.0
20.0
25.0
16.0
22.0
CO Ts/kW-hrl
75
350
350
350
350
a P = maximum engine power in kilowatts (kW).
b These engines are also subject to not-to-exceed standards. This category also includes engines >373 kW that do
not otherwise meet the definition of "high-performance."
0 2011 for small-businesses and for engines built using the 4.3L or 8.1L GM engine blocks.
d For small businesses, the 2010 standards do not apply and the 2011 standards are delayed until 2013.
Table 4: Marine SI Engine Evaporative Emissions Standards and Schedule
Standard Level
Portable Tanks
PWC
Other Installed Tanks
Fuel Line Permeation
15 g/nf/day
2009a
2009
2009a
Tank Permeation
1.5 g/mVday
2011
2011
2012
Diurnal
0.40 g/gal/day
2010b
2010
2011c'd
a 2011 for primer bulbs. Phase-in for under cowl fuel lines, by length, on OB engines: 30% 2010, 60% 2011, 90%
2012, 100% 2015.
b Design standard.
0 Fuel tanks installed in nontrailerable boats (> 26 ft. in length or >8.5 ft. in width) may meet a standard of 0.16
g/gal/day over an alternative test cycle.
d The standard is effective July 31, 2011. For boats with installed fuel tanks, this standard is phased-in 50%/100%
over the first two years. As an alternative, small manufacturers may participate in a diurnal allowance program.
ES-4
-------
Executive Summary
EPA has also taken steps to ensure that engines built to these standards achieve more
accurate emissions reductions and is upgrading the test requirements to those listed in
40CFR1065 as outlined in Preamble Section IX General Test Procedures.
Feasibility of Meeting the Small SI Engine Exhaust Emission Standards
Since 1997, exhaust emission control development for Small SI engines has concentrated on
engine redesign including carburetor design, improved engine combustion and engine cooling.
The primary technical focus of the new emission standards will be engine upgrades as needed,
catalyst application to the majority of Small SI engines and electronic fuel injection on some
Class II engines. Related information is in Chapter 4.
We are finalizing, more stringent exhaust HC+NOx standards for Class I and II Small SI
engines. We are also establishing a new CO standard for Small SI engines used in marine
generator applications. The standards differ by engine size. Class I engines have a total engine
displacement of < 225cc. Class II engines have a total engine displacement of >225cc.
In the 2008 model year, manufacturers certified nearly 235 Class I and II engine families to
the Phase 2 standards using a variety of engine designs and emission control technology. All
Class I engines were produced using carbureted air-fuel induction systems and are air cooled.
An extremely small number of engines used catalyst-based emission control technology.
Similarly, Class II engines were predominantly carbureted and air cooled. A limited number of
these engines used catalyst technology, electronic engine controls and fuel injection, and/or
water
cooling.
The market focus has a large part to play in the engine design and quality. The large number
of residential and commercial applications have led to a wide variety of engine qualities and
designs in the marketplace today. Some of the more durable engine designs already incorporate
the base design requirements needed to incorporate a catalyst to meet the Phase 3 emission
standards. In addition, a number of engine families in both classes are currently certified at
levels that would comply with the Phase 3 standards.
Based on our own testing of advanced technology for these engines, our engineering
assessments, and statements from the affected industry, we believe the requirements will lead
many engine manufacturers to adopt exhaust aftertreatment technology using catalyst-based
systems. Other likely engine changes include improvements in engine designs, cooling system
designs and fuel delivery systems. The addition of electronic controls and/or fuel injection
systems to some Class II engine families may obviate the need for catalytic aftertreatment, with
the most likely candidates being multi-cylinder engine designs.
Information herein on the feasibility assessment of exhaust emissions on Small SI engines
includes the emission evaluation of current product and advanced technology engines. Areas
covered include laboratory and field evaluations, review of patents of existing catalyst/muffler
ES-5
-------
Final Regulatory Impact Analysis
designs for Class I engines, discussions with engine manufacturers and suppliers of emission
control-related engine components regarding recent and expected advances in emissions
performance, and an analysis of catalyst/muffler units that were already in mass production by an
original equipment manufacturer for use on European walk-behind lawn mowers.
EPA used this information to design, build and emission test prototype catalyst-based
emission control systems that were capable of effectively and safely achieving the Phase 3
emission standards on both Class I and Class II engines. Chapter 4 projects that in some cases
manufacturers of Class I and Class II engines may need to improve the durability of their basic
engine designs, cooling system designs, ignition systems, or fuel metering systems for some
engines in order to comply with the Phase 3 emission regulations over the useful life. EPA also
built and tested electronic fuel injection systems on two twin cylinder Class II engines and
emission tested them with and without catalysts. EFI improves the management of air-fuel
mixtures and ignition spark timing and each of the engines achieved the requisite emission limit
for HC+NOx (e.g., 8.0 g/kW-hr). Based on this work and information from one manufacturer of
emission controls, we believe that either a catalyst-based system or electronic engine controls
appear sufficient to meet the standard. Manufacturers adopting the EFI approach will likely
realize other advantages such as easier starting, more stable and reliable engine operation, and
reduced fuel consumption.
We also used the information and the results of our engine testing to assess the potential need
for improvements to engine, cooling and fuel system designs. A great deal of this effort was
conducted in association with our more in-depth study regarding the efficacy and safety of
implementing advanced exhaust emission controls on Small SI and recreational Marine SI
engines, as well as new evaporative requirements for these engines, equipment, and vessels. The
results of that study are also discussed in Chapter 4.
There are a number of Class II engines that use gaseous fuels (i.e., liquid propane gas or
compressed natural gas). Based on our engineering evaluation of current and likely emission
control technology for these engines, we conclude that these engines will use catalysts, or larger
catalysts than current, in order to achieve the Phase 3 HC+NOx standard. Some engines
currently meet the Phase 3 emission standards.
Regarding the marine generator CO standard, two manufacturers that produce the majority of
marine generators have announced that as a result of boat builder demand, they are converting
their marine generator product lines to new designs which can achieve more than a 99 percent
reduction in CO emissions in order to reduce the risk of CO poisoning. These low CO emission
designs used closed-loop electronic fuel injection and catalytic control on engines which are
water cooled using the lake or sea water. Both of these manufacturers have certified some low
CO engines and have expressed their intent to convert their full product lines in the near future.
These manufacturers also make use of electronic controls to monitor catalyst function.
Feasibility of Meeting the Marine SI Exhaust Emission Standards
The technology is available for marine engine manufacturers to use to meet the new
ES-6
-------
Executive Summary
standards. This technology is the same that manufacturers are anticipated to use to meet the
California ARB standards in 2008. For outboards and personal watercraft (OB/PWC) this
largely means extended use of lower-emitting engine technology widely used today. For
sterndrive and inboard (SD/I) marine engines, this means the use of catalytic converters in the
exhaust system. Chapter 4 includes detailed descriptions of low emission technologies for
marine engines, including emissions test data on these technologies.
OB/PWC
Over the past several years, manufacturers have demonstrated their ability to achieve
significant HC+NOx emission reductions from OB/PWC engines. This has largely been
accomplished through the introduction of two-stroke direct injection engines in some
applications and conversion to four-stroke engines. Current certification data for these types of
engines show that these technologies may be used to achieve emission levels significantly below
the existing exhaust emission standards. In fact, California has adopted standards requiring a 65
percent reduction beyond the current federal standards beginning in 2008.
Our own analysis of recent certification data shows that most four-stroke outboard engines
and many two-stroke direct injection outboard engines currently meet the new HC+NOx
standard. Similarly, although PWC engines tend to have higher HC+NOx emissions,
presumably due to their higher power densities, many of these engines also meet the new
HC+NOx standard. Although there is currently not a CO emission standard for OB/PWC
engines, OB/PWC manufacturers are required to report CO emissions from their engines. These
emissions are based on test data from new engines and do not consider deterioration or
compliance margins. Based on this data, all of the two-stroke direct injection engines show
emissions well below the new standards. In addition, the majority of four-stroke engines meet
the CO standards as well.
We therefore believe the HC+NOx and CO emission standards can be achieved by phasing
out conventional carbureted two-stroke engines and replacing them with four-stroke engines or
two-stroke direct injection engines. This has been the market-driven trend over the last five
years. Chapter 4 compares recent certification data to the new standards.
SD/I
Engine manufacturers can adapt readily available technologies to control emissions from
SD/I engines. Electronically controlled fuel injection gives manufacturers more precise control
of the air/fuel ratio in each cylinder, thereby giving them greater flexibility in how they calibrate
their engines. With the addition of an oxygen sensor, electronic controls give manufacturers the
ability to use closed-loop control, which is especially valuable when using a catalyst. In
addition, manufacturers can achieve HC+NOx reductions through the use of exhaust gas
recirculation. However, the most effective technology for controlling emissions is a three-way
catalyst in the exhaust stream.
In SD/I engines, the exhaust manifolds are water-jacketed and the water mixes with the
ES-7
-------
Final Regulatory Impact Analysis
exhaust stream before exiting the vessel. Manufacturers add a water jacket to the exhaust
manifold to meet temperature-safety protocol. They route this cooling water into the exhaust to
protect the exhaust couplings and to reduce engine noise. Catalysts must therefore be placed
upstream of the point where the exhaust and water mix-this ensures the effectiveness and
durability of the catalyst. Because the catalyst must be small enough to fit in the exhaust
manifold, potential emission reductions are not likely to exceed 90 percent, as is common in
land-based applications. However, as discussed in Chapter 4 of the Final RIA, data on
catalyst-equipped SD/I engines show that emissions may be reduced by 70 to 80 percent for
HC+NOx and 30 to 50 percent for CO over the test cycle. Larger reductions, especially for CO,
have been achieved at lower-speed operation.
Chapter 4 discusses issues that have been addressed in catalyst designs for SD/I engines such
as sustained operation at high load, potential saltwater effects on catalyst efficiency, and thermal
shock from cold water contacting a hot catalyst. Test programs have been performed to evaluate
catalysts in the laboratory and on the water. Three SD/I engine manufacturers have certified SD/I
engines to the California ARB standards, and some catalyst-equipped engines are available for
purchase nationwide. Manufacturers have indicated that they have successfully completed
durability testing, including extended in-use testing on saltwater.
Feasibility of Meeting the Evaporative Emission Standards
There are many feasible control technologies that manufacturers can use to meet the
evaporative emission standards. We have collected emission test data on a wide range of
technologies for controlling evaporative emissions. Chapter 5 presents a description of the
evaporative emission sources which include permeation, diurnal, running loss, hot soak, and
refueling emissions. In addition, Chapter 5 presents evaporative emission test data for current
Small SI and marine fuel systems and on a wide range of evaporative emission control
technologies. Below is an overview of technologies that are available for meeting the
evaporative emission standards.
Low-permeation fuel lines are in production today. One fuel line design, already used in
some marine applications, uses a thermoplastic layer between two rubber layers to control
permeation. This thermoplastic barrier may either be nylon or ethyl vinyl acetate (EVOH).
Barrier approaches in automotive applications include fuel lines with fluoroelastomers such as
FKM and fluoroplastics such as Teflon and THV. In addition to presenting data on
low-permeation fuel lines, Chapter 5 lists several fuel-system materials and their permeation
rates. Molded rubber fuel line components, such as primer bulbs and some handheld fuel lines,
could meet the standard by using a fluoroelastomer such as FKM.
Plastic fuel tanks used in Small SI and Marine SI applications can be molded using several
processes. While no fuel tank permeation control strategy will work for all production processes
and materials, there are multiple control strategies available for fuel tanks manufactured with
each of the molding processes. These molding processes include blow-molding, injection-
molding, thermoforming, rotational-molding, and hand built constructions (fiberglass).
ES-8
-------
Executive Summary
Multi-layer fuel tanks can be formed using most of these molding processes. These fuel tank
constructions include a barrier layer of a low permeation material such as ethylene vinyl alcohol
(EVOH) or nylon. This technology has been used in blow-molded fuel tanks for automotive
applications for many years and can achieve emission levels well below the new standard. For
thermoformed fuel tanks, a similar barrier formed into the plastic sheet that is later molded into a
fuel tank. Rotationally-molded fuel tanks can be produced with an inner barrier layer such as
nylon. As an alternative, in the blow-molding process, a low-permeable resin can be blended
with polyethylene and extruded it with a single screw. Although the barrier is not continuous,
this strategy can still be used to meet the permeation standard. A similar strategy may be used
for fiberglass fuel tank where the barrier material is clay nanocomposites. Finally, fuel tanks
may be formed entirely out of a low permeation material such as nylon or an acetal copolymer.
Many fuel tanks used with handheld equipment use nylon fuel tanks.
Another approach to producing fuel tanks that meet the permeation standards would be to
create permeation barrier through a post-processing step. Regardless of the molding process,
another type of low-permeation technology for high-density polyethylene fuel tanks would be to
treat the surfaces with a barrier layer. Two ways of achieving this are known as fluorination and
sulfonation. In these processes, the tanks are exposed to a gas which forms a permeation barrier
on the surfaces of the fuel tank. Either of these processes can be used to reduce gasoline
permeation by more than 95 percent. Additionally, a barrier layer can be put onto a fuel tank
with the use of an epoxy barrier coating.
There are several technologies that can be used to reduce diurnal emissions from marine fuel
tanks. The simplest approach is to seal the fuel tank. Portable fuel tanks currently use manual
valves that can be closed to seal the fuel tank. PWC typically use sealed fuel systems with
pressure relief valves that open at pressures ranging from 0.5 to 4.0 psi. For other vessels with
installed fuel tanks, manufacturers have commented that even 1.0 psi of pressure would be too
high for their applications. Through the use of a carbon canister in the vent line, diurnal
emissions can be controlled from these fuel tanks without creating significant pressure in the fuel
tank. With this technology, vapor generated in the tank is vented to a canister containing
activated carbon. The fuel tank must be sealed such that the only venting that occurs is through
the carbon canister. The activated carbon collects and stores the hydrocarbons. The activated
carbon bed in the canister is refreshed by purging the vapors with air flow. The standard is
based on the air flow being generated by the natural breathing of the fuel tank as it heats and
cools.
Running loss emissions can be controlled from Small SI equipment by sealing the fuel cap
and routing vapors from the fuel tank to the engine intake. In doing so, vapors generated by heat
from the engine will be burned in the engine's combustion chamber. It may be necessary to use
a valve or limited-flow orifice in the purge line to prevent too much fuel vapor from reaching the
engine and to prevent liquid fuel from entering the line if the equipment flips over. Depending
on the configuration of the fuel system and purge line, a one-way valve in the fuel cap may be
desired to prevent a vacuum in the fuel tank during engine operation. We anticipate that a
system like this would eliminate running loss emissions. However, higher temperatures during
operation and the additional length of vapor line would slightly increase permeation.
ES-9
-------
Final Regulatory Impact Analysis
Considering these effects, we still believe that the system described here would reduce running
losses from Small SI equipment by more than 90 percent.
Many manufacturers today use fuel caps that by their design effectively limit the diffusion of
gasoline from fuel tanks. In any case, we expect that the new running loss design standard will
limit any diffusion emissions from this equipment. As discussed in Chapter 5, venting a fuel
tank through a tube (rather than through an open orifice) greatly reduces diffusion.
Estimated Costs and Cost-Effectiveness for Small SI Engines and Equipment
There are approximately 410 nonroad equipment manufacturers using Small SI engines in
over a thousand different equipment models. There are more than 50 engine manufacturers
certifying Small SI engine families for these applications. Fixed costs consider engine research
and development, engine tooling, engine certification, and equipment redesign. Variable costs
include estimates for new emission-control hardware. Near-term and long-term costs for some
example pieces of equipment are shown in Table 5. Also shown in Table 5 are typical prices for
each piece of equipment for reference. See Chapter 6 for detailed information related to our
engine and equipment cost analysis.
Table 5: Estimated Costs for Several Example Pieces of Equipment ($2005)a
Over the Range of Useful Life Categories for Small SI Enginesb
Exhaust Near Term
Long Term
Evaporative Near Term
Long Term
Total (without fuel savings)
Near Term
Long Term
Total (with fuel savings)0
Near Term
Long Term
Estimated Equipment Price Range
Class I
$10 to $26
$10 to $12
$3.05
$2.20
$14 to $26
$11 to $17
$13 to $25
$10 to $16
$100-$2,800
Class II
$17 to $60
$12 to $30
$6.73
$5.16
$46 to $92
$27 to $52
$l-$48/$40-$86
-$18-$6/$21-$46
Engines w/ and w/o EFI
$300-$6800
Handheld
(Class III-V)
$0.28
$0.00
$0.82
$0.69
$1.12
$0.69
$0.72
$0.29
$210 avg
a Near-term costs include both variable costs and fixed costs; long-term costs include only variable costs
and represent those costs that remain following recovery of all fixed costs.
b Class I (125,250, or 500 hours), Class II (250, 500, or 1000 hours)
0 Class I, Class II and handheld have fuel savings from evaporative measures. Class II engines with EFI have fuel
savings of $39 based on the lifetime savings in the use of a residential ride on mower. There are no fuel savings
related to compliance with the exhaust emission standard for Class I, handheld, or Class II engines without EFI.
Chapter 6 presents aggregate costs of compliance for the new exhaust and evaporative
ES-10
-------
Executive Summary
emission standards for Small SI engines. Table 6 presents the annualized aggregate costs and
fuel savings for the period from 2008-2037. The annualized fuel savings for Small SI engines
are due to reduced fuel costs form the sue of electronic fuel injection on Class II engines as well
as fuel savings from evaporative measures on all Small SI engines.
Table 6: Estimated Annualized Cost to manufacturers and Annualized Fuel Savings for
Small SI Engines and Equipment at a 7% Discount Rate (2005$)
Exhaust
Evaporative
Aggregate
Annualized Cost to Manufacturers
(millions/yr)
$182
$65
$247
Annualized fuel savings
(millions/yr)
$24
$53
$77
Chapter 7 describes the cost effectiveness analysis. In this analysis, the aggregate costs of
compliance are determined for the period 2008-2037. The discounted aggregate costs for the
period are divided by the discounted aggregate HC_NOx emission reductions.
Table 7: Aggregate Cost per Ton for Small SI Engines and Equipment
2008-2037 Net Present Values at 7% Discount Rate ($2005)
Pollutant
NOx+HC
7%
Aggregate Discounted
Lifetime Cost per ton
Without Fuel Savings
$978
Aggregate Discounted
Lifetime Cost per ton
With Fuel Savings
$650
Estimated Costs and Cost-Effectiveness for Marine SI Engines
According to the US Coast Guard there are well over a thousand different boat builders
using Marine SI engines. There are about 10 engine manufacturers certifying to the current
OB/PWC exhaust emission standards. We have identified more than 30 companies
manufacturing SD/I marine engines. Fixed costs consider engine research and development,
engine tooling, engine certification, and equipment redesign. Variable costs include estimates
for new emission-control hardware. Near-term and long-term costs for three different Marine SI
applications are shown in Table 8. Also shown in Table 8 are typical prices for these types of
marine vessels. See Chapter 6 for detailed information related to our engine and equipment cost
analysis.
ES-ll
-------
Final Regulatory Impact Analysis
Table 8: Estimated Average Incremental Costs for SI Marine Engines and Vessels ($2005)a
Engine Category
(Fuel Storage System)
Exhaust
Near Term
Long Term
Evaporative
Near Term
Long Term
Total (without fuel savings)*
Near Term
Long Term
Total (with fuel savings)"
Near Term
Long Term
Estimated Vessel Price Range
Outboard
(Portable)
$291
$224
$12
$8
$433
$336
$245
$148
$10,000-50,000
PWC
$359
$272
$17
$11
$376
$283
$165
$72
$6,000-12,000
SD/I
(Installed)
$355
$266
$74
$62
$487
$376
$348
$237
$20,000-200,000
a Near-term costs include both variable costs and fixed costs; long-term costs include only variable costs and
represent those costs that remain following recovery of all fixed costs.
b Total costs are presented as an average per boat and consider that many boats have multiple engines.
Chapter 6 presents aggregate costs of compliance for the new exhaust and evaporative
emission standards for Marine SI engines. Table 9 presents the annualized aggregate costs and
fuel savings for the period from 2008-2037. The annualized fuel savings for Marine SI engines
are due to reduced fuel costs from the use of more fuel efficient engines as well as fuel savings
from evaporative measures.
Table 9: Estimated Annualized Cost to Manufacturers and Annualized Fuel Savings for
Marine SI Engines and Vessels at a 7% Discount Rate (2005$)
Exhaust
Evaporative
Aggregate
Annualized Cost to Manufacturers
(millions/yr)
$123
$22
$144
Annualized Fuel Savings
(millions/year)
$56
$22
$78
Chapter 7 describes the cost effectiveness analysis. In this analysis, the aggregate costs of
compliance are determined for the period 2008-2037. The discounted aggregate costs for the
period are divided by the discounted aggregate HC+NOx emission reductions over that same
period. Table 10 presents the cost per ton estimates with and without fuel savings.
ES-12
-------
Executive Summary
Table 10: Aggregate Cost per Ton for SI Marine Engines and Vessels
2008-2037 Net Present Values at 7% Discount Rate ($2005)
Pollutant
NOx+HC
7%
Aggregate Discounted
Lifetime Cost per ton
Without Fuel Savings
$780
Aggregate Discounted
Lifetime Cost per ton
With Fuel Savings
$360
Economic Impact Analysis
We prepared a final Economic Impact Analysis estimate the market and social welfare
impacts of the new standards. This analysis can be found in Chapter 9. According to this
analysis, the average price of a Marine SI engine in 2030 is projected to increase by less than 2
percent ($213) as a result of the new standards, and the average price of a Marine SI vessel is
projected to increase by between 0.7 percent and 2.4 percent ($218 to $702), depending on the
type of vessel. The average price of a Small SI engine in 2030 is projected to increase by about
7.4 percent ($12), and the average price of Small SI nonhandheld equipment is projected to
increase by between 2.2 percent and 5.6 percent ($15 to $20), depending on equipment class.
Changes in quantity produced are expected to be small, at less than 2 percent. The exceptions
are PWC (4.8 percent) and Class II equipment (2.4 percent).
The net social costs of the program in 2030 are estimated to be $186 million. This includes
$459 million of direct social costs and $273 million on fuel savings for the end users of these
products. Overall, the consumers of Marine SI vessels and Small SI equipment are expected to
bear the majority of the costs of complying with the program: 76 percent of the Marine SI
program social costs in 2030, and 91 percent of the Small SI program social costs. However,
when the fuel savings are considered, the social costs burden for consumers of Marine SI
equipment becomes a net benefit (the fuel savings are greater than the compliance costs of the
program), while the end-user share of the Small SI program drops to 86 percent.
Benefits
We estimate that the requirements in this rulemaking will result in substantial benefits to
public health and welfare and the environment, as described in Chapter 8. The benefits analysis
performed for this rulemaking uses sophisticated air quality and benefit modeling tools and is
based on peer-reviewed studies of air quality and health and welfare effects associated with
improvements in air quality and peer-reviewed studies of the dollar values of those public health
and welfare effects.
The range of benefits associated with this program are estimated based on the risk of several
sources of PM- and ozone-related mortality effect estimates, along with all other PM and ozone
non-mortality related benefits information. These benefits are presented in Table 11. The
benefits reflect two different sources of information about the impact of reductions in PM on
ES-13
-------
Final Regulatory Impact Analysis
reduction in the risk of premature death, including an estimate of mortality derived from the
epidemiological literature (the American Cancer Society (ACS) cohort study - Pope et al., 2002)
and an expert elicitation study conducted by EPA in 2006. In order to provide an indication of
the sensitivity of the benefits estimates to alternative assumptions, in Chapter 8 of the RIA we
present a variety of benefits estimates based on two epidemiological studies (including the ACS
Study and the Six Cities Study) and the expert elicitation. EPA intends to ask the Science
Advisory Board to provide additional advice as to which scientific studies should be used in
future RIAs to estimate the benefits of reductions in PM.
The range of ozone benefits associated with the final standards is also estimated based on
risk reductions estimated using several sources of ozone-related mortality effect estimates.
There is considerable uncertainty in the magnitude of the association between ozone and
premature mortality. This analysis presents four alternative estimates for the association based
upon different functions reported in the scientific literature. We use the National Morbidity,
Mortality and Air Pollution Study (NMMAPS), which was used as the primary basis for the risk
analysis in the ozone Staff Paper and reviewed by the Clean Air Science Advisory Committee
(CASAC). We also use three studies that synthesize ozone mortality data across a large number
of individual studies. Note that there are uncertainties within each study that are not fully
captured by this range of estimates. Chapter 8 of the RIA presents the results of each of the
ozone mortality studies separately.
In a recent report on the estimation of ozone-related premature mortality published by the
National Research Council (NRC), a panel of experts and reviewers concluded that
ozone-related mortality should be included in estimates of the health benefits of reducing ozone
exposure. The report also recommended that the estimation of ozone-related premature mortality
be accompanied by broad uncertainty analyses while giving little or no weight to the assumption
that there is no causal association between ozone exposure and premature mortality. Because
EPA has yet to develop a coordinated response to the NRC report's findings and
recommendations, however, we have retained the approach to estimating ozone-related
premature mortality used in RIA for the final Ozone NAAQS. EPA will specifically address the
report's findings and recommendations in future rulemakings.
The range of total ozone- and PM-related benefits associated with the final standards is
presented in Table 11. We present total benefits based on the PM- and ozone-related premature
mortality function used. The benefits ranges therefore reflect the addition of each estimate of
ozone-related premature mortality (each with its own row in Table 11) to estimates of
PM-related premature mortality, derived from either the epidemiological literature or the expert
elicitation.
ES-14
-------
Executive Summary
Table 11: Estimated Monetized PM- and Ozone-Related Health Benefits of the Small SI
and Marine SI Engine Standards
2030 Total Ozone and PM Benefits - PM Mortality Derived from Epidemiology Studies3
Premature Ozone Mortality
Function or Assumption
NMMAPS
Meta-analysis
Reference
Bell etal, 2004
Bell etal, 2005
Ito etal., 2005
Levy etal., 2005
Assumption that association is not causal6
Mean Total Benefits
(Billions, 2005$, 3% discount
rate)c'd
$2.4
$3.7
$4.4
$4.4
$1.8
2030 Total Ozone and PM Benefits - PM Mortality Derived from Expert Elicitationb
Premature Ozone Mortality
Function or Assumption
NMMAPS
Meta-analysis
Reference
Bell etal., 2004
Bell etal., 2005
Ito etal., 2005
Levy etal., 2005
Assumption that association is not causal6
Mean Total Benefits
(Billions, 2005$, 3% discount
rate)c'd
$1.7 to $9.7
$3.0 to $11
$3. 7 to $12
$3. 7 to $12
$1.1 to $9.1
a Total includes ozone and PM2.5 benefits. Range was developed by adding the estimate from the ozone premature
mortality function to an estimate of PM2.5-related premature mortality derived from the ACS (Pope et al., 2002)
study.
b Total includes ozone and PM2.5 benefits. Range was developed by adding the estimate from the ozone premature
mortality function to both the lower and upper ends of the range of the PM2.5 premature mortality functions
characterized in the expert elicitation. The effect estimates of five of the twelve experts included in the elicitation
panel fall within the empirically-derived range provided by the ACS and Six-Cities studies. One of the experts fall
below this range and six of the experts are above this range. Although the overall range across experts is
summarized in this table, the full uncertainty in the estimates is reflected by the results for the full set of 12 experts.
The twelve experts'judgments as to the likely mean effect estimate are not evenly distributed across the range
illustrated by arraying the highest and lowest expert means.
0 Note that total benefits presented here do not include a number of unqualified benefits categories. A detailed
listing of unqualified health and welfare effects is provided in Table 8.4-1.
d Results reflect the use of a 3 percent discount rate. Monetary results presented in Table 8.6-2 use both a 3 and 7
percent discount rate, as recommended by EPA's Guidelines for Preparing Economic Analyses and OMB Circular
A-4. Results are rounded to two significant digits for ease of presentation and computation.
eA recent report published by the National Research Council (NRC, 2008) recommended that EPA "give little or no
weight to the assumption that there is no causal association between estimated reductions in premature mortality and
reduced ozone exposure."
ES-15
-------
Final Regulatory Impact Analysis
We estimate that by 2030, the annual emission reductions associated with these more
stringent standards will annually prevent 230 PM-related premature deaths (based on the ACS
cohort study), between 77 and 350 ozone-related premature deaths (assuming a causal
relationship between ozone and mortality), 1,700 hospital admissions and emergency room
visits, 23,000 work days lost, and approximately 590,000 minor restricted-activity days.
Impact on Small Businesses
Chapter 10 discusses our Small Business Flexibility Analysis, which evaluates the impacts of
the emission standards on small entities. As a part of this analysis, we interacted with several
small entities representing the various affected sectors and convened a Small Business Advocacy
Review (SBAR) Panel to gain feedback and advice from these representatives. The small
entities that participated in the process included engine manufacturers, equipment manufacturers,
vessel manufacturers, fuel tank manufacturers, and fuel hose manufacturers. The feedback from
these companies was used to develop regulatory options which could address the impacts of the
rule on small businesses. Small entities raised general concerns related to potential difficulties
and costs of meeting the new standards.
The SBAR Panel consisted of representatives from EPA, the Office of Management and
Budget, and the Small Business Administration. The Panel developed a wide range of regulatory
flexibilities to mitigate the impacts of the standards on small entities, and recommended that we
propose and seek comment on the flexibilities. Chapter 10 discusses the flexibilities
recommended by the Panel, and the flexibilities we are finalizing with today's rule. We are
establishing several provisions that give affected small entities several compliance options aimed
specifically at reducing their compliance burdens. In general the options are similar to small
entity provisions adopted in prior rulemakings where EPA set standards for other types of
nonroad engines. The provisions include extra lead time for complying with the new standards,
reduced testing requirements for demonstrating compliance with the new standards, and hardship
provisions to address significant economic impacts and unusual circumstances related to the new
standards. These provisions are intended to reduce the burden on small entities that will be
required to meet the new emission standards when they are implemented. Given all of the
flexibilities being adopted for small entities, we believe that this action will not have a
significant economic impact on a substantial number of small entities.
Alternative Program Options
In developing the emission standards, we considered several alternatives including less
and/or more stringent options. The paragraphs below summarize the information considered in
Chapter 11 of the Draft RIA.
Small SI Engines
For Small SI engines, we considered what was achievable with catalyst technology. Our
technology assessment work indicated that the emission standards are feasible in the context of
ES-16
-------
Executive Summary
provisions for establishing emission standards prescribed in section 213 of the Clean Air Act.
We also considered what could be achieved with larger, more efficient catalysts and improved
fuel induction systems. In particular, Chapter 4 of the Draft RIA presents data on Class I
engines with more active catalysts and on Class II engines with closed-loop control fuel injection
systems in addition to a catalyst. In both cases larger emission reductions were achieved.
Based on this work we considered HC+NOx standards which would have involved a 50
percent reduction for Class I engines and a 65-70 percent reduction for Class II engines. Chapter
11 of the Draft RIA evaluates these alternatives, including an assessment of the overall
technology and costs of meeting more stringent standards. For Class I engines a 50 percent
reduction standard would require base engine changes not necessarily involved with the new
standards and the use of a more active catalyst. For Class II engines this would require the
widespread use of closed loop control fuel injection systems rather than carburetors, some
additional engine upgrades, and the use three-way catalysts. We believe it is not appropriate at
this time to establish more stringent exhaust emission standards for Small SI engines. Our key
concern is lead time. More stringent standards would require several years (3-5) more lead time
beyond the 2011 model year start date. We believe it would be more effective to implement the
Phase 3 standards we are finalizing today to achieve near-term emission reductions needed to
reduce ozone precursor emissions and to minimize growth in the Small SI exhaust emissions
inventory in the post 2010 time frame. More efficient catalysts, engine improvements, and
closed loop electronic fuel injection could be the basis for more stringent emission standards at
some point in the future.
Marine SI Engines
In developing the final emission standards for SD/I engines, we considered both what was
achievable without catalysts and what could be achieved with larger, more efficient catalysts
than those used in our test programs. Without catalysts, we believe exhaust gas recirculation is a
technologically feasible and cost-effective approach to reducing emissions from SD/I marine
engines. However, we believe greater reductions could be achieved through the use of catalysts.
We considered basing an interim standard on EGR, but were concerned that this will divert
manufacturers' resources away from catalyst development and could have the effect of delaying
emission reductions from this sector.
Several of the marine engines with catalysts that were tested as part of the development of
the standards had HC+NOx emission rates appreciably lower than 5 g/kW-hr, even with
consideration of expected in-use emissions deterioration associated with catalyst aging. We
considered a 2.5 g/kW-hr HC+NOx standard in our analysis of alternatives. However, we
believe a standard of 5 g/kW-hr is still appropriate given the potential variability of in-use
performance and in test data.
For OB/PWC engines, we considered a level of 10 g/kW-hr HC+NOx for OB/PWC engines
greater than 40 kW with an equivalent percent reduction below the new standards for engines
less than 40 kW. This second tier of standards could apply in the 2012 or later time frame. Such
a standard would be consistent with currently certified emission levels from a significant number
ES-17
-------
Final Regulatory Impact Analysis
of four-stroke outboard engines. We have three concerns with adopting this second tier of
OB/PWC standards. First, while some four-stroke engines may be able to meet a 10 g/kW-hr
standard with improved calibrations, it is not clear that all engines could meet this standard
without applying catalyst technology. As described in Section IV.H.3 of the preamble, we
believe it is not appropriate to base standards in this rule on the use of catalysts for OB/PWC
engines. The technology is yet to be adequately demonstrated. Second, certification data for
personal watercraft engines show somewhat higher exhaust emission levels, so setting the
standard at 10 g/kW-hr would likely require catalysts for many models. Third, two-stroke direct
injection engines operate with lean air-fuel ratios, so reducing NOx emissions with any kind of
aftertreatment is challenging.
Therefore, unlike the standards for SD/I engines, we are not pursuing OB/PWC standards
that will require the use of catalysts. Catalyst technology would be necessary for significant
additional control of HC+NOx and CO emissions. While there is good potential for eventual
application of catalyst technology to OB/PWC engines, we believe the technology is not
adequately demonstrated at this point.
Evaporative Emission Controls
We considered both less and more stringent evaporative emission control alternatives for fuel
systems used in Small SI equipment and Marine SI vessels. Chapter 11 of the Draft RIA
presents details on this analysis of regulatory alternatives. The results of this analysis are
summarized below. We believe that the permeation standards are reflective of available
technology and represent a step change in emissions performance. Therefore, we consider the
same permeation control scenario in the less stringent and more stringent regulatory alternatives.
For Small SI equipment, we considered a less stringent alternative without running loss
emission standards for Small SI engines. However, we believe that controlling running loss
emissions from non-handheld equipment is feasible at a relatively low cost. Running loss
emissions can be controlled by changing the fuel tank and cap venting scheme and routing
vapors from the fuel tank to the engine intake. Not requiring these controls would be
inconsistent with section 213 of the Clean Air Act. For a more stringent alternative, we
considered applying a diurnal emission standard for all Small SI equipment. We believe that
passively purging carbon canisters could reduce diurnal emissions by 50 to 60 percent from
Small SI equipment. However, we believe there would be significant costs to add carbon
canisters to all Small SI equipment nationwide, especially when taking packaging and vibration
into account. The cost sensitivity is especially noteworthy given the relatively low emissions
levels (on a per-equipment basis) from such small fuel tanks.
For Marine SI vessels, we considered a less stringent alternative, where there would be no
diurnal emission standard for vessels with installed fuel tanks. However, installed fuel tanks on
marine vessels are much larger in capacity than those used in Small SI applications. Our
analysis indicates that traditional carbon canisters are feasible for boats at relatively low cost.
While packaging and vibration are also issues with marine applications, we believe these issues
have been addressed. Carbon canisters were installed on fourteen boats by industry in a pilot
ES-18
-------
Executive Summary
program. The results demonstrated the feasibility of this technology. The standards would be
achievable through engineering design-based certification with canisters that are very much
smaller than the fuel tanks. In addition, sealed systems, with pressure control strategies would
be accepted under the engineering design-based certification. For a more stringent scenario, we
consider a standard that would require boat builders to use an actively purged carbon canister.
This means that, when the engine is operating, it would draw air through the canister to purge the
canister of stored hydrocarbons. However, we rejected this option because active purge occurs
infrequently due to the low hours of operation per year seen by many boats. The gain in overall
efficiency would be quite small relative to the complexity active purge adds into the system in
that the engine must be integrated into a vessel-based control strategy. The additional benefit of
an actively purged diurnal control system is small in comparison to the cost and complexity of
such a system.
Conclusion
We believe the new emission standards reflect what manufacturers can achieve through the
application of available technology. We believe the lead time is necessary and adequate for
manufacturers to select, design, and produce emission control strategies that will work best for
their product lines. We expect that meeting these requirements will pose a challenge, but one
that is feasible when taking into consideration the availability and cost of technology, lead time,
noise, energy, and safety.
ES-19
-------
-------
Industry Characterization
CHAPTER 1: Industry Characterization
The information contained in this chapter on the Small SI engine and Marine SI engine
industries was assembled by RTI International, a Health, Social and Economics Research firm in
cooperation with EPA. RTI prepared one report each on the Small SI and Marine SI industries,
"Industry Profile for Small Nonroad Spark Ignition Engines and Equipment"1 and "Industry
Profile for Marine SI Industry"2 report. The following sections provide a brief report overview.
The reader is encouraged to refer to the reports for greater detail. In addition, this chapter
includes an overview of production practices for fuel system component manufacturers. Chapter
10 provides more information on businesses that would be affected by new standards.
1.1 Manufacturers of Small SI Engines
The nonroad spark-ignition (SI) industry includes a wide variety of handheld and
nonhandheld equipment. Nonhandheld equipment is powered mainly by four-stroke gasoline
engines; handheld equipment is powered mainly by two-stroke gasoline engines. Comprising
much of what the general public considers "lawn and garden (L&G) equipment," this industry
also produces significant numbers of generators, compressors, and construction and maintenance
equipment. The industry often refers to itself as the "outdoor power equipment" industry.
The industry profile report prepared by RTI for Small SI provides background
information on the engines and equipment that make up the small nonroad SI industry, defined
as those products rated less than or equal to 19 kilowatt (kW) (roughly equivalent to 25
horsepower [hp]). The profile describes markets for engines and equipment, and discusses their
use in both consumer and commercial applications. In each market, producers and consumers
are described, along with product attributes and the effect of those attributes on production cost
and demand. The market analysis emphasizes assessing suppliers' cost of production and
industry structure, along with demanders' price responsiveness and consumption alternatives.
The variety of products in this industry is usefully partitioned by both application
categories and engine type. Figure 1-1 illustrates the links between the market segments of the
Small SI engine supply chain included in the profile, from engine manufacturing and sale to
equipment production, and on to purchase by consumers and commercial customers. Although
more than 98 percent of total unit sales in the L&G equipment sector go to households, other
sectors' sales are dominated by commercial equipment. Because of the significantly higher
prices of commercial units, commercial sales represent a considerable share of the total value of
production.
It should be noted that there is a fair amount of vertical integration in the handheld
industry, with the same parent firm making both engines and the equipment in which those
engines are used. Handheld equipment includes string trimmers, leaf blowers, and chainsaws.
This situation is known as "captive" engine production; data on internal consumption of engines
and transfer prices are typically not available outside the firm. The makers of non-handheld
l-l
-------
Final Regulatory Impact Analysis
engines typically sell their engines to independent equipment manufacturers in a merchant
engine market, where prices and quantities exchanged can be directly observed.
The industry profile report prepared by RTI for Small SI presents information on product
characteristics, supply-side considerations, consumer demand, and market structure for small
nonroad SI engines. The report also includes similar types of information on equipment
markets, broken down by application category. Considerations related to consumer and
commercial markets are included in the report.
Figure 1-1: The Nonroad Small SI Industry
Consumers
C o m m e re ia I U s e rs
EquipmentMarkets
• Lawnmowers
Handheld Lawn and Garden
• Other Lawn and Garden
• Generators and Welders
• Compressors and Pumps
• Recreational Products
•Snow Blowers
• Other S m a II SI
I m p o rts
• Engines
• Equipment
N onintegrated
E q u ip m e n t
M a n u fa c t u re rs
S m a I
I Engine
• Class
•Size
M a rk e ts
M e re h a n t
E n g in e
M a n u fa c tu re rs
In te g ra te d
E q u ip m e n t
M a n u fa c tu re rs
C a ptiv e
E n g in e
M a n u fa c t u re rs
1.2 Manufacturers of Marine Spark-Ignition Engines
The Marine SI industry is dominated by recreational applications with some commercial
use and includes markets for several types of boats, personal watercraft (PWC), and SI engines
that power them. The industry profile presented in the "Industry Profile for Marine SI Industry"
report by RTI describes producers and consumers for each market segment; product attributes
and the effects of these attributes on production costs and demand are described as well. As part
of the market characterization, particular emphasis is placed on assessing suppliers' industrial
organization and cost of production and demanders' price responsiveness and substitution
possibilities. The Marine SI industry is divided into three applications areas: outboard (OB)
boats, sterndrive and inboard (SD/I) boats, and PWC.
1-2
-------
Industry Characterization
1.2.1 OB Boats
An OB boat is a vessel powered by one or more gasoline engines, which are located
outside the hull at the back of the boat. The engine and drive unit are combined in a single
package. An engine can easily be removed from the boat for inspection or repair, and it is quite
common for the boat owner to change engines during the life of the vessel. The OB boat
segment is the largest of the three application areas; in 2002, 213,000 units were sold, which is
more than the combined sales of SD/I and PWC.
The OB application area can be further divided into "recreational" and "luxury"
categories. The luxury category includes more-expensive vessels, for which the engine
constitutes only a small portion of the cost of the entire vessel. The NMMA distinguishes
between 14 types of OB vessels, 10 of which are considered recreational and 4 luxury.
1.2.2 SD/I Boats
SD/I vessels have an engine installed inside the hull of the vessel. An inboard vessel is a
boat in which the engine is located inside the hull at the center of the boat with a propeller shaft
going through the rear of the boat. A sterndrive (or inboard/outboard) vessel is a boat in which
the engine is located inside the hull at the back of the boat with a drive assembly couple directly
to the propeller, propeller shaft going through the rear of the boat. In contrast to OB vessels,
SD/I vessels' engine is an integral part. Removal or replacement is significantly more difficult,
so most repair work is done with the engine in place. Just like OBs, the SD/I application area is
divided into recreational and luxury categories.
1.2.3 PWC
According to the Personal Watercraft Industry Association (PWIA), a PWC is defined as
a "vessel with an inboard motor powering a water jet pump as its primary source of motive
power, and which is designed to be operated by a person sitting, standing, or kneeling on the
vessel."
The PWC application area is divided into the entry level, high end, and performance
categories based on the horsepower ratings of the vessel. These categories correspond to 50 to
100 hp, 100 to 175 hp, and over 175 hp accordingly. Our study considers two categories that
were available in 2002: entry level and high end. The performance category was introduced in
2003.
1.2.4 Marine SI Engines
Some OB engine manufacturers specifically build their engines to be incorporated into
boats produced by another division within the same parent company. Other manufacturers
produce and sell their engines to independent OB boat builders or consumers who need a
replacement engine. SD/I engine manufacturers typically build custom engines for SD/I boats
by marinizing automotive engines. All PWC vessel manufacturers build their own engines for
1-3
-------
Final Regulatory Impact Analysis
their vessels
Marine SI engines sold today are a mix of three primary technologies: crankcase
scavenged two-stroke engines, direct-injection two-stroke engines, and four-stroke engines.
Table 6.2.2-11 in Chapter 6 presents our best estimate of the technology mix for OB and PWC
engines by power class. This technology mix is based on data submitted by manufacturers when
the certify to our existing HC+NOx exhaust emission standards. Prior to the implementation of
the existing standards, the vast majority of outboard and PWC engines were crankcase
scavenged two-stroke engines.
The following Figures show the flow of engines from the engine manufacturer to the consumer
for the different engine types.
Figure 1-2. OB Marine Economic Model Conceptual Flow Chart
Consumers
Loose Engines
Imports
• Engines
• Equipment
Consumers
New Boats
i
k
Equipment Market
Noninteg rated
Equipment
Manufacturers
Marine SI Engine
Markets
Merchant
Engine
Manufacturers
Integrated
Equipment
Manufacturers
Captive
Engine
Manufacturers
1-4
-------
Industry Characterization
Figure 1-3: PWC Economic Model Conceptual Flow Chart
Consumer
>
k
Equipment Market
t
Integrated
Manufa
i
Captive
Manufa
k
Equipment
cturers
k
Engine
cturers
Figure 1-4: Inboard Marine Economic Model Conceptual Flow Chart
Integrated
Equipment
Manufacturer
i
k.
Merchant
Engine
Manufacturer
1-5
-------
Final Regulatory Impact Analysis
1.3 Fuel System Components
The primary fuel system components that would be affected by the rule are the fuel tanks
and fuel lines on affected equipment and vessels. This section gives an overview of the
production practices for these products.
1.3.1 Fuel Tank Production Practices
Plastic fuel tanks are either blow-molded, injection-molded, or rotational-molded.
Generally, portable, PWC, and mid-sized Small SI fuel tanks are blow-molded. Blow-molding
involves forming polyethylene in large molds using air pressure to shape the tank. Because this
has high fixed costs, blow molding is only used where production volumes are high. This works
for portable fuel tanks where the volumes are high and a single shape can be used for most
applications. For portable tanks, the fuel tank manufacturer will generally design the tank, then
send it out to a blow molder for production.
Smaller fuel tanks used in Small SI equipment are often injection-molded. In the
injection molding process, fuel tanks are formed by forcing heated plastic into molds at high
pressure. Generally, two fuel tank halves are formed, which are later fused together. This
process requires high tooling costs, but lower total fixed costs than blow-molding. Injection-
molding is typically used for smaller fuel tanks and has the advantage of giving manufacturers
the ability to work with complex tank designs.
Larger fuel tanks used on Class II equipment and in boats with installed fuel tanks are
typically rotational-molded out of cross-link polyethylene. Rotational-molding is a lower cost
alternative for smaller production volumes. In this method, a mold is filled with a powder form
of polyethylene with a catalyst material. The mold is rotated in an oven; the heat melts the
plastic and activates the catalyst which causes a strong cross-link material structure to form.
This method is used for Class II fuel tanks where the tanks are unshielded on the equipment.
These fuel tanks also used meet specific size and shape requirements for boats and are preferred
because they do not rust like metal tanks, but at the same time are more fire resistant than high-
density polyethylene fuel tanks.
Metal fuel tanks are also used on both Small SI equipment and boats. Typically, metal
tanks on Small SI equipment are made of steel. These tanks are typically stamped out in two
pieces and either welded or formed together with a seal. Aluminum fuel tanks are also used
primarily for installed marine fuel tanks because aluminum is more resistant to oxidation than
steel. In the marine industry, tank manufacturers generally custom make each tank to meet the
boat manufacturers needs. Generally, sheet aluminum is used and is cut, bent, and welded into
the required configuration.
1.3.2 Fuel Hose Production Practices
Marine hose is designed to meet the Coast Guard performance requirements as defined
by the Society of Automotive Engineer's recommended practice SAE J 1527. For fuel supply
1-6
-------
Industry Characterization
lines, this includes a permeation rate of 100 g/m2/day at 23°C (Class 1). For other fuel hose not
normally continuously in contact with fuel (vent and fuel fill neck), the permeation standard is
300 g/m2/day (Class 2). In general, boat builders will use Class 1 hose for both fuel supply and
vent lines for simplicity. Some boat builders use low permeation barrier hose, for which,
specifications are now included in SAE J 1527. For fuel fill necks, boat builders generally use
Class 2 hose. Small SI hose is typically produced to manufacturer specifications. However,
manufacturers may specify hose based on industry standards such as those listed in SAE J30.
Most fuel supply and vent hose is extruded nitrile rubber with a coating for better wear
and flame resistance. Hose may also be reinforced with fabric or wire. (In contrast, plastic
automotive fuel lines are extruded without reinforcement and are generally referred to as
"tubing.") Hose manufacturers offer a wide variety of fuel hoses including those with a barrier
layer of low permeability material, such as nylon, THV, FKM or ethyl vinyl alcohol, either on
the inside surface or sandwiched between layers of nitrile rubber. These technologies are
discussed in more detail in Chapter 5.
Fuel fill hose used on boats is generally manufactured by hand wrapping layers of rubber
and reinforcement materials around a steel mandril. This hose is then heated to cure the rubber.
Fuel fill hose generally has a much larger diameter than fuel supply and vent hose and this
process offers an effective method of producing this larger diameter hose.
Pre-formed fuel lines are made in two ways. The first, and more common method, is to
cut lengths of extruded hose, before it is vulcanized, and slip them over a contoured mandril.
The hose is then vulcanized in the oven on the mandril to give it a preformed shape. The second
way, primarily used on handheld equipment, but also for some outboard engine fuel system
components, is to injection-mold small parts. To make the parts hollow, they are molded with a
mandril inside. To remove the mandril, the part is typically inflated with air for just long enough
to pull it off the mandril. Primer bulbs are also made in this manner.
1-7
-------
Final Regulatory Impact Analysis
Chapter 1 References
1. "Industry Profile for Small Nonroad Spark-Ignition Engines and Equipment," RTI International, October 2006.
2. "Industry Profile for Marine SI Industry," RTI International, October 2006.
1-8
-------
-------
Final Regulatory Impact Analysis
CHAPTER 2: Air Quality, Health, and Welfare Concerns
The standards finalized in this action will reduce emissions of hydrocarbons (HC),
oxides of nitrogen (NOx), carbon monoxide (CO) and air toxics from the engines, vessels and
equipment subject to this rule. Emissions of these pollutants contribute to ozone, PM and CO
nonattainment and to adverse health effects associated with air toxics. The emissions from
these engines, vessels and equipment also contribute to adverse environmental effects.
The health and environmental effects associated with emissions from Small SI engines
and equipment and Marine SI engines and vessels are a classic example of a negative
externality (an activity that imposes uncompensated costs on others). With a negative
externality, an activity's social cost (the cost on society imposed as a result of the activity
taking place) exceeds its private cost (the cost to those directly engaged in the activity). In
this case, as described in this chapter, emissions from Small SI engines and equipment and
Marine SI engines and vessels impose public health and environmental costs on society. The
market system itself cannot correct this externality. The end users of the equipment and
vessels are often unaware of the environmental impacts of their use for lawn care or
recreation. Because of this, consumers fail to send the market a signal to provide cleaner
equipment and vessels. In addition, producers of these engines, equipment, and vessels are
rewarded for emphasizing other aspects of these products (e.g., total power). To correct this
market failure and reduce the negative externality, it is necessary to give producers social cost
signals. The standards EPA is finalizing will accomplish this by mandating that Small SI
engines and equipment and Marine SI engines and vessels reduce their emissions to a
technologically feasible limit. In other words, with this rule the costs of the services provided
by these engines and equipment will account for social costs more fully.
In this Chapter we will discuss the impacts of the pollutants emitted by Small SI
engines and equipment and Marine SI engines and vessels on health and welfare, National
Ambient Air Quality Standard (NAAQS) attainment, and personal exposure. Air quality
modeling and monitoring data presented in this chapter indicate that a large number of people
live in counties that are designated as nonattainment for either or both of the 8-hour ozone or
PM2.5 NAAQS. Figure 2-1 illustrates the widespread nature of the ozone and PM2.5
nonattainment areas and also depicts mandatory class I areas. The emission standards in this
rule will help reduce HC, NOx, PM, air toxic and CO emissions and their associated health
and environmental effects.
2-2
-------
Air Quality, Health and Welfare Concerns
Legend
| Class I Areas
m PM2 5 and Ozone NonAttainment
Ozone NonAttainment
PM2.5 NonAttainment
Map current through March 12. 2008
Figure 2-1: 8-Hour Ozone and PMi.s Nonattainment Areas and Mandatory Class I
Federal Areas
2.1 Ozone
In this section we review the health and welfare effects of ozone exposure. We also
describe the air quality monitoring and modeling data that indicates people in many areas
across the country are exposed to levels of ambient ozone above the 1997 and 2008 ozone
NAAQS. The data also indicates that in the future people will continue to live in counties
with ozone levels above the NAAQS without additional federal, state or local measures.
Emissions of volatile organic compounds (VOCs), of which HC are a subset, and NOx from
the engines, vessels and equipment subject to this rule contribute to these ozone
concentrations. Information on air quality was gathered from a variety of sources, including
monitored ozone concentrations, air quality modeling forecasts conducted for this rulemaking,
and other state and local air quality information.
2-3
-------
Final Regulatory Impact Analysis
2.1.1 Science of Ozone Formation
Ground-level ozone pollution is formed by the reaction of VOCs and NOx in the
atmosphere in the presence of heat and sunlight. These pollutants, often referred to as ozone
precursors, are emitted by many types of pollution sources such as highway vehicles and
nonroad engines (including those subject to this rule), power plants, chemical plants,
refineries, makers of consumer and commercial products, industrial facilities, and smaller area
sources.
The science of ozone formation, transport, and accumulation is complex.1 Ground-
level ozone is produced and destroyed in a cyclical set of chemical reactions, many of which
are sensitive to temperature and sunlight. When ambient temperatures and sunlight levels
remain high for several days and the air is relatively stagnant, ozone and its precursors can
build up and result in more ozone than typically would occur on a single high-temperature
day. Ozone can be transported hundreds of miles downwind of precursor emissions, resulting
in elevated ozone levels even in areas with low VOC or NOx emissions.
The highest levels of ozone are produced when both VOC and NOx emissions are
present in significant quantities on clear summer days. Relatively small amounts of NOx
enable ozone to form rapidly when VOC levels are relatively high, but ozone production is
quickly limited by removal of the NOx. Under these conditions NOx reductions are highly
effective in reducing ozone while VOC reductions have little effect. Such conditions are
called "NOx-limited". Because the contribution of VOC emissions from biogenic (natural)
sources to local ambient ozone concentrations can be significant, even some areas where man-
made VOC emissions are relatively low can be NOx-limited.
Ozone concentrations in an area also can be lowered by the reaction of nitric oxide
(NO) with ozone, forming nitrogen dioxide (NO2); as the air moves downwind and the cycle
continues, the NO2 forms additional ozone. The importance of this reaction depends, in part,
on the relative concentrations of NOx, VOC, and ozone, all of which change with time and
location. When NOx levels are relatively high and VOC levels relatively low, NOx forms
inorganic nitrates (i.e., particles) but relatively little ozone. Such conditions are called "VOC-
limited". Under these conditions, VOC reductions are effective in reducing ozone, but NOx
reductions can actually increase local ozone under certain circumstances. Even in VOC-
limited urban areas, NOx reductions are not expected to increase ozone levels if the NOx
reductions are sufficiently large.
Rural areas are usually NOx-limited, due to the relatively large amounts of biogenic
VOC emissions in such areas. Urban areas can be either VOC- or NOx-limited, or a mixture
of both, in which ozone levels exhibit moderate sensitivity to changes in either pollutant.
2.1.2 Health Effects of Ozone Pollution
Exposure to ambient ozone contributes to a wide range of adverse health effectsA.
A Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notable different ozone concentrations. Also, the amount of ozone
2-4
-------
Air Quality, Health and Welfare Concerns
These health effects are well documented and are critically assessed in the EPA ozone air
91
quality criteria document (ozone AQCD) and EPA staff paper. ' We are relying on the data
and conclusions in the ozone AQCD and staff paper, regarding the health effects associated
with ozone exposure.
Ozone-related health effects include lung function decrements, respiratory symptoms,
aggravation of asthma, increased hospital and emergency room visits, increased asthma
medication usage, and a variety of other respiratory effects. Cell-level effects such as,
inflammation of lungs, have been documented as well. In addition, there is suggestive
evidence of a contribution of ozone to cardiovascular-related morbidity and highly suggestive
evidence that short-term ozone exposure directly or indirectly contributes to non-accidental
and cardiopulmonary-related mortality, but additional research is needed to clarify the
underlying mechanisms causing these effects. In a recent report on the estimation of ozone-
related premature mortality published by the National Research Council (NRC), a panel of
experts and reviewers concluded that short-term exposure to ambient ozone is likely to
contribute to premature deaths and that ozone-related mortality should be included in
estimates of the health benefits of reducing ozone exposure.4 People who appear to be more
susceptible to effects associated with exposure to ozone include children, asthmatics and the
elderly. Those with greater exposures to ozone, for instance due to time spent outdoors (e.g.,
children and outdoor workers), are also of concern.
Based on a large number of scientific studies, EPA has identified several key health
effects associated with exposure to levels of ozone found today in many areas of the country.
Short-term (1 to 3 hours) and prolonged exposures (6 to 8 hours) to higher ambient ozone
concentrations have been linked to lung function decrements, respiratory symptoms, increased
hospital admissions and emergency room visits for respiratory problems.5'6'7'8'9'10 Repeated
exposure to ozone can increase susceptibility to respiratory infection and lung inflammation
and can aggravate preexisting respiratory diseases, such as asthma.11'12'13'14'15 Repeated
exposure to sufficient concentrations of ozone can also cause inflammation of the lung,
impairment of lung defense mechanisms, and possibly irreversible changes in lung structure,
which over time could affect premature aging of the lungs and/or the development of chronic
respiratory illnesses, such as emphysema and chronic bronchitis.16'17'18'19
Children and adults who are outdoors and active during the summer months, such as
construction workers, are among those most at risk of elevated ozone exposures.20 Children
and outdoor workers tend to have higher ozone exposure because they typically are active
outside, working, playing and exercising, during times of day and seasons (e.g., the summer)
when ozone levels are highest.21 For example, summer camp studies in the Eastern United
States and Southeastern Canada have reported statistically significant reductions in lung
function in children who are active outdoors.22'23'24'25'26'27'28'29 Further, children are more at
risk of experiencing health effects from ozone exposure than adults because their respiratory
systems are still developing. 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
delivered to the lung is not only influenced by the ambient concentration but also by the individuals breathing
route and rate.
2-5
-------
Final Regulatory Impact Analysis
ozone levels during prolonged periods of moderate exertion.30'31; 32'33
2.1.3 Current Ozone Levels
The small SI and marine SI engine emission reductions will assist ozone
nonattainment areas in reaching the standard by each area's respective attainment date and/or
assist in maintaining the ozone standard in the future. In this and the following section we
present information on current and model-projected future ozone levels.
A nonattainment area is defined in the CAA as an area that is violating a NAAQS or is
contributing to a nearby area that is violating the NAAQS. EPA designated nonattainment
areas for the 1997 ozone NAAQS in June 2004. The final rule on Air Quality Designations
and Classifications for the 1997 Ozone NAAQS (69 FR 23858, April 30, 2004) identifies the
criteria that EPA considered in making the 1997 8-hour ozone nonattainment designations,
including 2001-2003 measured data, air quality in adjacent areas, and other factors.8
As of March 12, 2008 there are approximately 140 million people living in 72 areas
designated as nonattainment with the 1997 8-hour ozone NAAQS. There are 337 full or
partial counties that make up the 8-hour ozone nonattainment areas. These numbers do not
include the people living in areas where there is a future risk of failing to maintain or attain
the 8-hour ozone NAAQS. The 1997 8-hour ozone nonattainment areas, nonattainment
counties, and populations are listed in Appendix 2A to this RIA.
EPA has recently amended the ozone NAAQS (73 FR 16436, March 27, 2008). The
final ozone NAAQS rule addresses revisions to the primary and secondary NAAQS for ozone
to provide increased protection of public health and welfare, respectively. With regard to the
primary standard for ozone, EPA has revised the level of the 8-hour standard to 0.075 parts
per million (ppm), expressed to three decimal places. With regard to the secondary standard
for ozone, EPA has revised the current 8-hour standard by making it identical to the revised
primary standard.
States with ozone nonattainment areas are required to take action to bring those areas
into compliance in the future. The attainment date assigned to an ozone nonattainment area is
based on the area's classification. Most ozone nonattainment areas will be required to attain
the 1997 8-hour ozone NAAQS in the 2007 to 2013 time frame and then be required to
maintain it thereafter.0 The attainment dates associated with the potential nonattainment areas
B 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. For example, an 8-hour ozone design value is the fourth highest daily maximum 8-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. For a county, the design value is the highest design value from among all the monitors with
valid design values within that county. If a county does not contain an ozone monitor, it does not have a design
value. However, readers should note that ozone design values generally represent air quality across abroad area
and that absence of a design value does not imply that the county is in compliance with the ozone NAAQS.
c The Los Angeles South Coast Air Basin 8-hour ozone nonattainment area is designated as severe and will have
to attain before June 15, 2021. The South Coast Air Basin has recently applied to be redesignated as an extreme
2-6
-------
Air Quality, Health and Welfare Concerns
based on the 2008 8-hour ozone NAAQS will likely be in the 2013 to 2021 timeframe,
depending on the severity of the problem. Table 2-1 provides an estimate, based on 2004-06
air quality data, of the counties with design values greater than the 2008 ozone NAAQS. We
expect many of the ozone nonattainment areas will need to adopt additional emissions
reduction programs to attain and maintain the ozone NAAQS. The expected VOC and NOX
reductions from these standards, which take effect between 2009 and 2013, will be useful to
states as they seek to either attain or maintain the ozone NAAQS.
Table 2-1 Counties with Design Values Greater Than the 2008 Ozone NAAQS Based on
2004-2006 Air Quality Data
1997 Ozone Standard: counties within the 72
areas currently designated as nonattainment
2008 Ozone Standard: additional counties that
would not meet the 2008 NAAQS6
Total
Number of Counties
337
74
411
Population"
139,633,458
15,984,135
155,617,593
Notes:
" Population numbers are from 2000 census data.
* Attainment designations for the 2008 ozone NAAQS have not yet been made. Nonattainment for the 2008
Ozone NAAQS will be based on three years of air quality data from later years. Also, the county numbers in the
table include only the counties with monitors violating the 2008 Ozone NAAQS. The numbers in this table may
be an underestimate of the number of counties and populations that will eventually be included in areas with
multiple counties designated nonattainment.
2.1.4 Projected Ozone Levels
In conjunction with this rulemaking, we performed a series of air quality modeling
simulations for the continental U.S. The model simulations were performed for several
emissions scenarios including the following: 2002 baseline projection, 2020 baseline
projection, 2020 baseline projection with small Si/marine SI engine controls, 2030 baseline
projection, and 2030 baseline projection with small Si/marine SI engine controls. Information
on the air quality modeling methodology is contained in Section 2.3 as well as the air quality
modeling technical support document (AQ TSD). In the following sections we describe our
modeling of 8-hour ozone levels in the future with and without the controls being finalized in
this action.
2.1.4.1 Projected 8-Hour Ozone Levels without this Rulemaking
EPA has already adopted many emission control programs that are expected to reduce
ambient ozone levels. These control programs include the Locomotive and Marine Rule (73
FR 25098, May 6, 2008), Clean Air Interstate Rule (70 FR 25162, May 12, 2005), the Clean
Air Nonroad Diesel rule (69 FR 38957, June 29, 2004), and the Heavy Duty Engine and
Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements (66 FR 5002, Jan.
18, 2001). As a result of these programs, the number of areas that continue to violate the 8-
hour ozone NAAQS in the future is expected to decrease.
nonattainment area which will make their attainment date June 15, 2024.
2-7
-------
Final Regulatory Impact Analysis
The baseline air quality modeling completed for this rule predicts that without
additional local, regional or national controls there will continue to be a need for reductions in
8-hour ozone concentrations in some areas in the future. The determination that an area is at
risk of exceeding the 8-hour ozone standard in the future was made for all areas with current
design values greater than or equal to 85 ppb (or within a 10 percent margin) and with
modeling evidence that concentrations at and above these levels will persist into the future.0
Those interested in greater detail should review the air quality modeling TSD which is
included in the docket for this rule.34
The baseline inventories that underlie the modeling conducted for this rulemaking
include emission reductions from existing federal, state and local controls. There was no
attempt to examine the prospects of areas attaining or maintaining the standard with future
possible controls. We expect many of the areas to adopt additional emission reduction
programs, but we are unable to quantify or rely upon future reductions from additional
programs since they have not yet been promulgated. With reductions from programs already
in place (but excluding the emission reductions from this rule), the number of counties in
2020 with projected 8-hour ozone design values at or above 85 ppb is expected to be 8 with a
population of 22 million people. In addition, in 2020, 37 counties where 27 million people
are projected to live, will be within 10 percent of violating the 1997 8-hour ozone NAAQS.
The results should therefore be interpreted as indicating counties at risk for violating the
ozone NAAQS in the future without additional federal, state or local measures in addition to
this rulemaking.
2.1.4.2 Projected 8-Hour Ozone Levels with this Rulemaking
This section summarizes the results of our modeling of ozone air quality impacts in
the future due to the reductions in small SI and marine SI emissions finalized in this action.
Specifically, we compare baseline scenarios to scenarios with controls. Our modeling
indicates that the reductions from this rule will provide nationwide improvements in ambient
ozone concentrations and minimize the risk of exposures in future years. Since some of the
VOC and NOX emission reductions from this rule go into effect during the period when some
areas are still working to attain the 8-hour ozone NAAQS, the projected emission reductions
will assist state and local agencies in their effort to attain the 8-hour ozone standard and help
others maintain the standard. Emissions reductions from this rule will also help to counter
potential ozone increases due to climate change, which are expected in many urban areas in
the United States, but are not reflected in the modeling shown here.35
On a population-weighted basis, the average modeled future-year 8-hour ozone design
values will decrease by 0.57 ppb in 2020 and 0.76 ppb in 2030. Table 2-2 shows the average
change in future year eight-hour ozone design values for: (1) all counties with 2002 baseline
design values, (2) counties with baseline design values that exceeded the standard in 2000-
2004 ("violating" counties), (3) counties that did not exceed the standard, but were within 10
percent of it in 2000-2004, (4) counties with future year design values that exceeded the
D Ozone design values are reported in parts per million (ppm) as specified in 40 CFR Part 50. Due to the scale of
the design value changes in this action results have been presented in parts per billion (ppb) format.
2-8
-------
Air Quality, Health and Welfare Concerns
standard, and (5) counties with future year design values that did not exceed the standard, but
were within 10 percent of it in 2020 and 2030. Counties within ten percent of the standard are
intended to reflect counties that meet the standard, but will likely benefit from help in
maintaining that status in the face of growth. All of these metrics show a decrease in 2020
and 2030, indicating in five different ways the overall improvement in ozone air quality.
Table 2-2 Average Change in Projected Future Year 8-hour Ozone Design Value as a Result
of the Small SI and Marine SI controls
Average"
All
All, population-weighted
Counties whose base year is violating the 1997
8 -hour ozone standard
Counties whose base year is violating the 1997
8-hour ozone standard, population-weighted
Counties whose base year is within 10 percent of
the 1997 8-hour ozone standard
Counties whose base year is within 10 percent of
the 1997 8-hour ozone standard, population-
weighted
Counties whose future year is violating the 1997
8 -hour ozone standard
Counties whose future year is violating the 1997
8-hour ozone standard, population-weighted
Counties whose future year is within 10 percent
of the 1997 8-hour ozone standard
Counties whose future year is within 10 percent
of the 1997 8-hour ozone standard, population-
weighted
Number
of US
Counties
660
660
261
261
223
223
8 (2020)
6 (2030)
8 (2020)
6 (2030)
37 (2020)
23 (2030)
37 (2020)
23 (2030)
Change in
2020 design
value6 (ppb)
-0.47
-0.57
-0.62
-0.61
-0.42
-0.55
-0.13
-0.17
-0.71
-0.54
Change in
2030 design
value6 (ppb)
-0.66
-0.76
-0.88
-0.80
-0.61
-0.78
-0.10
-0.13
-1.05
-0.79
Notes:
" averages are over counties with 2002 modeled design values
* Ozone design values are reported in parts per million (ppm) as specified in 40 CFR Part 50. Due to the scale of
the design value changes in this action results have been presented in parts per billion (ppb) format.
Table 2-3 lists the counties with projected 8-hour ozone design values that violate or
are within 10 percent of the 1997 8-hour ozone standard in 2020after application of the small
SI and marine SI controls. Counties are marked with a "V" in the table if their projected
design values are greater than or equal to 85 ppb. Counties are marked with an "X" in the
table if their projected annual design values are greater than or equal to 76.5 ppb, but less than
85 ppb. The counties marked "X" are not projected to violate the standard, but to be close to
it, so the rule will help assure that these counties continue to meet the standard. The current
design values are also presented in Table 2-3. Recall that we project future design values only
for counties that have current design values, so this list is limited to those counties with
ambient monitoring data sufficient to calculate current 3-year design values.
2-9
-------
Final Regulatory Impact Analysis
Figure 2-2 illustrates the geographic impact of the small SI and marine SI engine
controls on 8-hour ozone design values in 2020. Some of the most significant decreases will
occur in the great lakes region, the gulf coast region, the northeast corridor and in the Seattle
region. The maximum decreases in a 2020 design values is 2.0 ppb in Cape Cod,
Massachusetts.
2-10
-------
Air Quality, Health and Welfare Concerns
Table 2-3 Counties with 2020 8-hour Ozone Design Values in Violation or Within 10 percent
of the 1997 Ozone Standard as a Result of the Small SI and Marine SI Controls
State
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CT
CT
IN
LA
MD
NJ
NJ
NJ
NJ
NY
OH
OH
PA
PA
TX
TX
TX
WI
WI
WI
County
El Dorado
Fresno
Kern
Kings
Los Angeles
Madera
Merced
Nevada
Orange
Placer
Riverside
Sacramento
San Bernardino
San Diego
Stanislaus
Tulare
Tuolumne
Fairfield
New Haven
Lake
East Baton Rouge
Harford
Camden
Gloucester
Mercer
Ocean
Suffolk
Ashtabula
Geauga
Bucks
Philadelphia
Brazoria
Harris
Jefferson
Kenosha
Racine
Sheboygan
2000-2004
Average 8-Hour
Ozone DV (ppb)"
105.0
110.0
114.3
95.7
121.3
91.0
101.7
97.7
85.3
98.3
115.0
99.0
128.7
92.3
95.0
105.7
91.0
98.3
98.3
88.3
87.0
100.3
99.7
98.0
97.7
105.7
97.0
95.7
99.0
99.0
96.7
94.0
102.0
91.0
98.3
91.7
97.0
2020 modeling
projections of
8-Hour Ozone
DV
X
X
X
V
X
V
V
V
V
V
X
V
X
V
V
X
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
X
V
V
V
V
2020
Population
236,310
1,066,878
876,131
173,390
10,376,013
173,940
277,863
131,831
3,900,599
451,620
2,252,510
1,640,590
2,424,764
3,863,460
607,766
477,296
70,570
962,824
898,415
509,293
522,399
317,847
547,817
304,105
392,236
644,323
1,598,742
108,355
114,438
711,275
1,394,176
322,385
4,588,812
272,075
184,825
212,351
128,777
Notes:
a Ozone design values are reported in parts per million (ppm) as specified in 40 CFR Part 50. Due to the scale of
the design value changes in this action results have been presented in parts per billion (ppb) format.
2-11
-------
Final Regulatory Impact Analysis
Legend Number of Counties
| | <= -2.0 1
J >-2.0 to -1.0 69
| >-1.0 to -0.5 186
| >-0.5 to -0.1 360
^^ = 0.0 44
I |>0.0 0
2020cc_bond impact
Figure 2-2 Impact of Small SI and Marine SI controls on 8-hour Ozone Design Values in 2020 (units are ppb)
2-12
-------
2.1.5 Environmental Effects of Ozone Pollution
There are a number of public welfare effects associated with the presence of ozone in
the ambient air.36 In this section we discuss the impact of ozone on plants, including trees,
agronomic crops and urban ornamentals.
2.1.5.1 Impacts on Vegetation
The Air Quality Criteria Document for Ozone and related Photochemical Oxidants
notes that "ozone affects vegetation throughout the United States, impairing crops, native
vegetation, and ecosystems more than any other air pollutant. Like carbon dioxide (CO2) and
other gaseous substances, ozone enters plant tissues primarily through apertures (stomata) in
leaves in a process called "uptake".37 Once sufficient levels of ozone, a highly reactive
substance, (or its reaction products) reaches the interior of plant cells, it can inhibit or damage
essential cellular components and functions, including enzyme activities, lipids, and cellular
membranes, disrupting the plant's osmotic (i.e., water) balance and energy utilization
patterns.38'39 This damage is commonly manifested as visible foliar injury such as chlorotic or
necrotic spots, increased leaf senescence (accelerated leaf aging) and/or reduced
photosynthesis. All these effects reduce a plant's capacity to form carbohydrates, which are
the primary form of energy used by plants.40 With fewer resources available, the plant
reallocates existing resources away from root growth and storage, above ground growth or
yield, and reproductive processes, toward leaf repair and maintenance. Studies have shown
that plants stressed in these ways may exhibit a general loss of vigor, which can lead to
secondary impacts that modify plants' responses to other environmental factors. Specifically,
plants may become more sensitive to other air pollutants, more susceptible to disease, insect
attack, harsh weather (e.g., drought, frost) and other environmental stresses. Furthermore,
there is evidence that ozone can interfere with the formation of mycorrhiza, essential
symbiotic fungi associated with the roots of most terrestrial plants, by reducing the amount of
carbon available for transfer from the host to the symbiont.41'42
Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure. Ozone effects also tend to accumulate
over the growing season of the plant, so that even lower concentrations experienced for a
longer duration have the potential to create chronic stress on sensitive vegetation. Not all
plants, however, are equally sensitive to ozone. Much of the variation in sensitivity between
individual plants or whole species is related to the plant's ability to regulate the extent of gas
exchange via leaf stomata (e.g., avoidance of Os uptake through closure of stomata).43'44'45
Other resistance mechanisms may involve the intercellular production of detoxifying
substances. Several biochemical substances capable of detoxifying ozone have been reported
to occur in plants including the antioxidants ascorbate and glutathione. After injuries have
occurred, plants may be capable of repairing the damage to a limited extent.46
Because of the differing sensitivities among plants to ozone, ozone pollution can also
exert a selective pressure that leads to changes in plant community composition. Given the
range of plant sensitivities and the fact that numerous other environmental factors modify
plant uptake and response to ozone, it is not possible to identify threshold values above which
2-13
-------
Regulatory Impact Analysis
ozone is consistently toxic for all plants. The next few paragraphs present additional
information on ozone damage to trees, ecosystems, agronomic crops and urban ornamentals.
Ozone also has been conclusively shown to cause discernible injury to forest trees.47'48
In terms of forest productivity and ecosystem diversity, ozone may be the pollutant with the
greatest potential for regional-scale forest impacts. Studies have demonstrated repeatedly that
ozone concentrations commonly observed in polluted areas can have substantial impacts on
plant function.49'50
Because plants are at the center of the food web in many ecosystems, changes to the
plant community can affect associated organisms and ecosystems (including the suitability of
habitats that support threatened or endangered species and below ground organisms living in
the root zone). Ozone impacts at the community and ecosystem level vary widely depending
upon numerous factors, including concentration and temporal variation of tropospheric ozone,
species composition, soil properties and climatic factors.51 In most instances, responses to
chronic or recurrent exposure in forested ecosystems are subtle and not observable for many
years. These injuries can cause stand-level forest decline in sensitive ecosystems.52'53'54 It is
not yet possible to predict ecosystem responses to ozone with much certainty; however,
considerable knowledge of potential ecosystem responses has been acquired through long-
term observations in highly damaged forests in the United States.
Laboratory and field experiments have also 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 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 Unites States."55 In addition, economic studies have shown
reduced economic benefits as a result of predicted reductions in crop yields associated with
observed ozone levels.56'57'58
Urban ornamentals represent an additional vegetation category likely to experience
some degree of negative effects associated with exposure to ambient ozone levels. It is
estimated that more than $20 billion (1990 dollars) are spent annually on landscaping using
ornamentals, both by private property owners/tenants and by governmental units responsible
for public areas.59 This is therefore a potentially costly environmental effect. However, 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 analysis
has been conducted.
2.2 Participate Matter
In this section we review the health and welfare effects of PM. We also describe air
quality monitoring and modeling data that indicate many areas across the country continue to
be exposed to levels of ambient PM above the NAAQS. Emissions of PM, HCs and NOx
from the engines, vessels and equipment subject to this rule contribute to these PM
concentrations. Information on air quality was gathered from a variety of sources, including
2-14
-------
monitored PM concentrations, air quality modeling forecasts conducted for this rulemaking,
and other state and local air quality information.
2.2.1 Science of PM Formation
Particulate matter (PM) represents a broad class of chemically and physically diverse
substances. It can be principally characterized as discrete particles that exist in the condensed
(liquid or solid) phase spanning several orders of magnitude in size. PMio refers to particles
generally less than or equal to 10 micrometers (//m) in aerodynamic diameter. PM2.5 refers to
fine particles, generally less than or equal to 2.5 //m in aerodynamic diameter. Inhalable (or
"thoracic") coarse particles refer to those particles generally greater than 2.5 //m but less than
or equal to 10 //m in aerodynamic diameter. Ultrafine PM refers to particles generally less
than 100 nanometers (0.1 //m) in aerodynamic diameter. Larger particles (>10 //m) tend to be
removed by the respiratory clearance mechanisms, whereas smaller particles are deposited
deeper in the lungs.
Fine particles are produced primarily by combustion processes and by transformations
of gaseous emissions (e.g., SOx, NOx and VOCs) in the atmosphere. The chemical and
physical properties of PM2.5 may vary greatly with time, region, meteorology and source
category. Thus, PM2.s, may include a complex mixture of different pollutants including
sulfates, nitrates, organic compounds, elemental carbon and metal compounds. These
particles can remain in the atmosphere for days to weeks and travel through the atmosphere
hundreds to thousands of kilometers.
Particles span many sizes and shapes and consist of hundreds of different chemicals.
Particles are emitted directly from sources and are also formed through atmospheric chemical
reactions; the former are often referred to as "primary" particles, and the latter as "secondary"
particles. In addition, there are also physical, non-chemical reaction mechanisms that
contribute to secondary particles. Particle pollution also varies by time of year and location
and is affected by several weather-related factors, such as temperature, clouds, humidity, and
wind. A further layer of complexity comes from a particle's ability to shift between
solid/liquid and gaseous phases, which is influenced by concentration, meteorology, and
temperature.
2.2.2 Health Effects of PM
As stated in EPA's Paniculate Matter Air Quality Criteria Document (PM AQCD),
available scientific findings "demonstrate well that human health outcomes are associated
with ambient PM."E We are relying on the data and conclusions in the PM AQCD and PM
Staff Paper, which reflects EPA's analysis of policy-relevant science from the PM AQCD,
regarding the health effects associated with particulate matter.60'61 We also present additional
E Personal exposure includes contributions from many different types of particles, from many sources, and in
many different environments. Total personal exposure to PM includes both ambient and nonambient
components; and both components may contribute to adverse health effects.
2-15
-------
Regulatory Impact Analysis
recent studies published after the cut-off date for the PM AQCD.F62 Taken together this
information supports the conclusion that PM-related emissions such as those controlled in this
action are associated with adverse health effects. Information on PM-related mortality and
morbidity is presented first, followed by information on near-roadway exposure studies,
marine ports and rail yard exposure studies.
2.2.2.1 Short-term Exposure Mortality and Morbidity Studies
As discussed in the PM AQCD, short-term exposure to PM2.5 is associated with
mortality from cardiopulmonary diseases (PM AQCD, p. 8-305), hospitalization and
emergency department visits for cardiopulmonary diseases (PM AQCD, p. 9-93), increased
respiratory symptoms (PM AQCD, p. 9-46), decreased lung function (PM AQCD Table 8-34)
and physiological changes or biomarkers for cardiac changes (PM AQCD, Section 8.3.1.3.4).
In addition, the PM AQCD describes a limited body of new evidence from epidemiologic
studies for potential relationships between short term exposure to PM and health endpoints
such as low birth weight, preterm birth, and neonatal and infant mortality. (PM AQCD,
Section 8.3.4).
Among the studies of effects from short-term exposure to PM2.5, several specifically
address the contribution of mobile sources to short-term PM2 5 effects on daily mortality.
These studies indicate that there are statistically significant associations between mortality
and PM related to mobile source emissions (PM AQCD, p.8-85). The analyses incorporate
source apportionment tools into daily mortality studies and are briefly mentioned here.
Analyses incorporating source apportionment by factor analysis with daily time-series studies
of daily death indicated a relationship between mobile source PM2.5 and mortality.63'64
Another recent study in 14 U.S. cities examined the effect of PMi0 exposures on daily hospital
admissions for cardiovascular disease. This study found that the effect of PMio was
significantly greater in areas with a larger proportion of PMio coming from motor vehicles,
indicating that PMio from these sources may have a greater effect on the toxicity of ambient
PMio when compared with other sources.65 These studies provide evidence that PM-related
emissions, specifically from mobile sources, are associated with adverse health effects
Long-term Exposure Mortality and Morbidity Studies
Long-term exposure to elevated ambient PM2.5 is associated with mortality from
cardiopulmonary diseases and lung cancer (PM AQCD, p. 8-307), and effects on the
respiratory system such as decreased lung function or the development of chronic respiratory
disease (PM AQCD, pp. 8-313, 8-314). Of specific importance to this rulemaking, the PM
AQCD also notes that the PM components of gasoline and diesel engine exhaust represent
F These additional studies are included in the 2006 Provisional Assessment of Recent Studies on Health Effects
of Paniculate Matter Exposure. The provisional assessment did not and could not (given a very short timeframe)
undergo the extensive critical review by EPA, CASAC, and the public, as did the PM AQCD. The provisional
assessment found that the "new" studies expand the scientific information and provide important insights on the
relationship between PM exposure and health effects of PM. The provisional assessment also found that "new"
studies generally strengthen the evidence that acute and chronic exposure to fine particles and acute exposure to
thoracic coarse particles are associated with health effects.
2-16
-------
one class of hypothesized likely important contributors to the observed ambient PM-related
increases in lung cancer incidence and mortality (PM AQCD, p. 8-318).
The PM AQCD and PM Staff Paper emphasize the results of two long-term studies,
the Six Cities and American Cancer Society (ACS) prospective cohort studies, based on
several factors - the inclusion of measured PM data, the fact that the study populations were
similar to the general population, and the fact that these studies have undergone extensive
reanalysis (PM AQCD, p. 8-306, Staff Paper, p.3-18).66'67'68 These studies indicate that there
are significant associations for all-cause, cardiopulmonary, and lung cancer mortality with
long-term exposure to PM2 5. One analysis of a subset of the ACS cohort data, which was
published after the PM AQCD was finalized but in time for the 2006 Provisional Assessment,
found a larger association than had previously been reported between long-term PM2.s
exposure and mortality in the Los Angeles area using a new exposure estimation method that
accounted for variations in concentration within the city.69
As discussed in the PM AQCD, the morbidity studies that combine the features of
cross-sectional and cohort studies provide the best evidence for chronic exposure effects.
Long-term studies evaluating the effect of ambient PM on children's development have
shown some evidence indicating effects of PM2.5 and/or PMio on reduced lung function
growth (PM AQCD, Section 8.3.3.2.3). In another recent publication included in the 2006
Provisional Assessment, investigators in southern California reported the results of a cross-
sectional study of outdoor PM2 5 and measures of atherosclerosis in the Los Angeles basin.70
The study found significant associations between ambient residential PM2.5 and carotid
intima-media thickness (CIMT), an indicator of subclinical atherosclerosis, an underlying
factor in cardiovascular disease.
2.2.2.3 Roadway-Related PM Exposure and Health Studies
A recent body of studies examines traffic-related PM exposures and adverse health
effects. These studies are relevant to this rule because highway SI vehicles and nonroad SI
engines, vessels and equipment have similar chemical and physical exhaust properties.
However, this comparison is qualitative in nature since the near-road environment is
influenced by both gasoline (SI) and diesel vehicles, as well as re-entrained road dust and
brake and tire wear. One study was done in North Carolina looking at concentrations of PM2.5
inside police cars and corresponding physiological changes in the police personnel driving the
cars. The authors report significant elevations in markers of cardiac risk associated with
concentrations of PM2.5 inside police cars on North Carolina state highways.71 Other studies
have found associations between traffic-generated particle concentrations at residences and
adverse effects, including all-cause mortality, infant respiratory symptoms, and reduced
cognitive functional development.72'73'74'75 There are other pollutants present in the near
roadway environment, including air toxics which are discussed in Section 2.4. Additional
information on near-roadway health effects can be found in the recent Mobile Source Air
Toxics rule (72 FR 8428, February 26, 2007).
2-17
-------
Regulatory Impact Analysis
2.2.3 Current and Projected PM Levels
The emission reductions from this rule will assist PM nonattainment areas in reaching
the standard by each area's respective attainment date and assist PM maintenance areas in
maintaining the PM standards in the future. In this and the following section we present
information on current and model-projected future PM levels.
2.2.3.1 Current PM2.5 Levels
The small SI and marine SI engine emission reductions will assist PM nonattainment
areas in reaching the standard by each area's respective attainment date and/or assist in
maintaining the PM standard in the future. In this and the following section we present
information on current and model-projected future PM levels.
A nonattainment area is defined in the Clean Air Act (CAA) as an area that is
violating an ambient standard or is contributing to a nearby area that is violating the standard.
In 2005, EPA designated 39 nonattainment areas for the 1997 PM2.5 NAAQS based on air
quality design values and a number of other factors (70 FR 943, January 5, 2005; 70 FR
19844, April 14, 2005).° These areas are comprised of 208 full or partial counties with a total
population exceeding 88 million. The 1997 PM2.5 nonattainment counties, areas and
populations, as of March 2008, are listed in Appendix 2B to this RIA.
EPA has recently amended the NAAQS for PM25 (71 FR 61144, October 17, 2006).
The final PM NAAQS rule addressed revisions to the primary and secondary NAAQS for
PM25 to provide increased protection of public health and welfare, respectively. The primary
PM25NAAQS includes a short-term (24-hour) and a long-term (annual) standard. The level
of the 24-hour PM25 NAAQS has been revised from 65 ug/m3to 35 ug/m3 to provide increased
protection against health effects associated with short-term exposures to fine particles. The
current form of the 24-hour PM25 standard was retained (e.g., based on the 98th percentile
concentration averaged over three years). The level of the annual PM25NAAQS was retained
at 15 ug/m3, continuing protection against health effects associated with long-term exposures.
The current form of the annual PM25 standard was retained as an annual arithmetic mean
averaged over three years, however, the following two aspects of the spatial averaging criteria
were narrowed: (1) the annual mean concentration at each site will now be within 10 percent
of the spatially averaged annual mean, and (2) the daily values for each monitoring site pair
will now yield a correlation coefficient of at least 0.9 for each calendar quarter.
With regard to the secondary standards for PM25, EPA has revised these standards to
be identical in all respects to the revised primary standards. Specifically, EPA has revised the
current 24-hour PM25 secondary standard by making it identical to the revised 24-hour PM25
primary standard and retained the annual PM25 secondary standard. This suite of secondary
PM25 standards is intended to provide protection against PM-related public welfare effects,
including visibility impairment, effects on vegetation and ecosystems, and material damage
and soiling.
' The full details involved in calculating a PM2 5 design value are given in Appendix N of 40 CFR Part 50.
2-18
-------
States with PM2.5 nonattainment areas will be required to take action to bring those
areas into compliance in the future. Most PM2 5 nonattainment areas will be required to attain
the 1997 PM2.5 NAAQS in the 2010 to 2015 time frame and then be required to maintain the
1997 PM2.s NAAQS thereafter.11 Nonattainment areas will be designated with respect to the
2006 PM2.5 NAAQS in early 2010. The attainment dates associated with the potential
nonattainment areas based on the 2006 PM2.5 NAAQS will likely be in the 2014 to 2019
timeframe. Table 2-4 provides an estimate, based on 2003-05 air quality data, of the counties
with design values greater than the 2006 PM25 NAAQS. The emission standards being
finalized in this action will become effective between 2009 and 2013. The expected PM2 5
inventory reductions will be useful to states in attaining or maintaining the PM2 5 NAAQS.
Table 2-4 Counties with Design Values Greater Than the 2006 PM25NAAQS Based on 2003-
2005 Air Quality Data
1997 PM25 Standards: counties within the 39
areas currently designated as nonattainment
2006 PM25 Standards: additional counties that
would not meet the 2006 NAAQS6
Total
Number of Counties
208
49
257
Population"
88,394,000
18,198,676
106,592,676
Notes:
a Population numbers are from 2000 census data.
* Attainment designations for the 2006 PM2 5 NAAQS have not yet been made. Nonattainment for the 2006
PM2 5 NAAQS will be based on three years of air quality data from later years. Also, the county numbers in the
table include only the counties with monitors violating the 2006 PM2 5 NAAQS. The numbers in this table may
be an underestimate of the number of counties and populations that will eventually be included in areas with
multiple counties designated nonattainment.
H The EPA finalized PM25 attainment and nonattainment areas in April 2005. The EPA finalized the PM
Implementation rule in March 2007.
2-19
-------
Regulatory Impact Analysis
2.2.3.2 Current PMio Levels
EPA designated PMio nonattainment areas in 1990.1 As of March 2008,
approximately 28 million people live in the 47 areas that are designated as PMio
nonattainment, for either failing to meet the PMio NAAQS or for contributing to poor air
quality in a nearby area. There are 46 full or partial counties that make up the PMio
nonattainment areas/
2.2.3.3 Projected PM2.5 Levels
In conjunction with this rulemaking, we performed a series of air quality modeling
simulations for the continental U.S. The model simulations were performed for several
emissions scenarios including the following: 2002 baseline projection, 2020 baseline
projection, 2020 baseline projection with small Si/marine SI engine controls, 2030 baseline
projection, and 2030 baseline projection with small Si/marine SI engine controls. Information
on the air quality modeling methodology is contained in Section 2.3 as well as the air quality
modeling technical support document (AQ TSD). In the following sections we describe
projected PM2 5 levels in the future with and without the controls being finalized in this action.
2.2.3.2.1 ProjectedPM2.s Levels without this Rulemaking
Even with the implementation of all current state and federal regulations, including the
Locomotive and Marine Rule, CAIR Rule, the NOX SIP call, nonroad and on-road diesel rules
and the Tier 2 rule, there are projected to be U.S. counties violating the PM2 5 NAAQS well
into the future. The model outputs from the 2002, 2020 and 2030 baselines, combined with
current air quality data, were used to identify areas expected to exceed the PM2 5 NAAQS in
the future.
The baseline air quality modeling conducted for this final rule projects that in 2020,
with all current controls in effect, up to 11 counties, with a population of 25 million people,
may not attain the annual standard of 15 |ig/m3 This does not account for additional areas
that have air quality measurements within 10 percent of the PM25 standard. These areas,
although not violating the standard, will also benefit from the emissions reductions, ensuring
long term maintenance of the PM NAAQS. For example, in 2020, an additional 16 million
people are projected to live in 13 counties that have air quality measurements within 10
percent of the 2006 PM NAAQS. This modeling supports the conclusion that there are a
substantial number of counties across the US projected to experience PM2.5 concentrations at
or above the PM2 5 NAAQS into the future. Emission reductions from small SI and marine SI
engines will be helpful for these counties in attaining and maintaining the PM2 5 NAAQS.
1 A PM10 design value is the concentration that determines whether a monitoring site meets the NAAQS for
PM10. The full details involved in calculating a PM10 design value are given in Appendices H and I of 40 CFR
Part 50.
1 The PM10 nonattainment areas are listed in Appendix 2C to this RIA.
2-20
-------
2.2.3.2.2 ProjectedPM25Levels With this Rulemaking
The impacts of the small SI and marine SI engine controls were determined by
comparing the model results in the future year control runs against the baseline simulations of
the same year. On a population-weighted basis, the average modeled future-year annual
PM2 5 design value (DV) for all counties is expected to decrease by 0.02 |ig/m3 in 2020 and
2030. There are areas with larger decreases in their future-year annual PM2.5 DV, for instance
the Chicago region will experience a 0.08 |ig/m3 reduction by 2030. Figure 2-3 illustrates the
geographic impact of the small SI and marine SI engine controls on annual PM25 design
values in 2020.
2-21
-------
Regulatory Impact Analysis
Legend
Number of Counties
<= -0.50 0
>= -0.49 to -0.25 0
>=-0.24 to-0.10 0
>= -0.09 to -0.05 3
>= -0.04 to 0.0 553
>0.0 0
2020cc_bond impact
Figure 2-3 Impact of Small SI and Marine SI controls on annual PM2 s Design Values (DV) in 2020 (units are ug/m3)
2-22
-------
Table 2-5 lists the counties with projected annual PM2.5 design values that violate
or are within 10 percent of the annual PM2 5 standard in 2020. Counties are marked with
a "V" in the table if their projected design values are greater than or equal to 15.05
|ig/m3. Counties are marked with an "X" in the table if their projected annual design
values are greater than or equal to 13.55 |ig/m3, but less than 15.05 |ig/m3. The counties
marked "X" are not projected to violate the standard, but to be close to it, so the rule will
help assure that these counties continue to meet the standard. The current design values
are also presented in Table 2-5. Recall that we project future design values only for
counties that have current design values, so this list is limited to those counties with
ambient monitoring data sufficient to calculate current 3-year design values.
Table 2-5 Counties with 2020 Projected Annual PM2 5 Design Values in Violation or
Within 10 percent of the Annual PM2.5 Standard as a Result of the Small SI and Marine
SI Controls
State
Alabama
California
California
California
California
California
California
California
California
California
California
California
California
California
Georgia
Illinois
Illinois
Kentucky
Michigan
Montana
New York
Ohio
Pennsylvania
West Virginia
County
Jefferson Co
Fresno Co
Imperial Co
Kern Co
Kings Co
Los Angeles Co
Merced Co
Orange Co
Riverside Co
San Bernardino Co
San Diego Co
San Joaquin Co
Stanislaus Co
Tulare Co
Fulton Co
Cook Co
Madison Co
Jefferson Co
Wayne Co
Lincoln Co
New York Co
Cuyahoga Co
Allegheny Co
Hancock Co
2000-
2004
Average
annual
PM25
DV
(ug/m3)
18.36
20.02
14.44
21.77
18.77
23.16
16.47
18.27
27.15
24.63
15.65
14.84
16.49
21.33
18.29
17.06
17.27
16.58
19.32
15.85
17.16
18.36
20.99
17.30
2020
modeling
projections
of
annual
PM2 5 DV
V
X
V
X
X
X
X
X
X
X
V
V
V
X
V
V
V
V
X
V
V
V
X
V
2020
Population
681,549
1,066,878
161,555
876,131
173,390
10,376,013
277,863
3,900,599
2,252,510
2,424,764
3,863,460
743,469
607,766
477,296
929,278
5,669,479
278,167
726,257
1,908,196
20,147
1,700,384
1,326,680
1,242,587
30,539
2-23
-------
Regulatory Impact Analysis
2.2.4 Environmental Effects of PM Pollution
In this section we discuss some of the public welfare effects of PM and its
precursors, including NOx, such as visibility impairment, atmospheric deposition, and
materials damage and soiling.
2.2.4.1 Visibility Impairment
Visibility can be defined as the degree to which the atmosphere is transparent to
visible light.76 Visibility impairment manifests in two principal ways: as local visibility
impairment and as regional haze77 Local visibility impairment may take the form of a
localized plume, a band or layer of discoloration appearing well above the terrain as a
result of complex local meteorological conditions. Alternatively, local visibility
impairment may manifest as an urban haze, sometimes referred to as a "brown cloud."
This urban haze is largely caused by emissions from multiple sources in the urban area
and is not typically attributable to only one nearby source or to long-range transport. The
second type of visibility impairment, regional haze, usually results from multiple
pollution sources spread over a large geographic region. Regional haze can impair
visibility over large regions and across states.
Visibility is important because it has direct significance to people's enjoyment of
daily activities in all parts of the country. Individuals value good visibility for the well-
being it provides them directly, where they live and work and in places where they enjoy
recreational opportunities. Visibility is also highly valued in significant natural areas
such as national parks and wilderness areas, and special emphasis is given to protecting
visibility in these areas.
Fine particles are the major cause of reduced visibility in parts of the United
States. To address the welfare effects of PM on visibility, EPA sets secondary PM2.5
standards which work in conjunction with the regional haze program. The secondary
(welfare-based) PM2.5 NAAQS is equal to the suite of primary (health-based) PM2.5
NAAQS. The regional haze rule (64 FR 35714, July 1999) was put in place to protect the
visibility in mandatory class I federal areas. These areas are defined in Section 162 of the
Act as those national parks exceeding 6,000 acres, wilderness areas and memorial parks
exceeding 5,000 acres, and all international parks which were in existence on August 7,
1977. A list of the mandatory class I federal areas is included in Appendix 2D. Visibility
is impaired in both PM2.5 nonattainment areas and mandatory class I federal areas.
Control of small SI and marine SI emissions will improve visibility. The small SI
and marine SI engines subject to this rule emit PM and PM precursors and thus contribute
to visibility impairment. In the next sections we present current information and
projected estimates about visibility impairment related to ambient PM2.5 levels across the
country and visibility impairment in mandatory class I federal areas. We conclude that
visibility will continue to be impaired in the future and the emission reductions from this
rule will help improve visibility conditions across the country and in mandatory class I
2-24
-------
federal areas. For more information on visibility see the PM AQCD as well as the 2005
PM Staff Paper.78'79
2.2.4.1.1 Current Visibility Impairment in PM2.sNonattainment Areas
As mentioned above, the secondary PM2.5 standards were set as equal to the suite
of primary PM2 5 standards. Almost 90 million people live in the 208 counties that are in
nonattainment for the 1997 PM2.5 NAAQS, (see Appendix 2A for the complete list of
current nonattainment areas). These populations, as well as large numbers of individuals
who travel to these areas can experience visibility impairment.
2.2.4.1.2 Current Visibility Impairment at Mandatory Class I Federal Areas
Detailed information about current and historical visibility conditions in
mandatory class I federal areas is summarized in the EPA Report to Congress and the
2002 EPA Trends Report.80'81 The conclusions draw upon the Interagency Monitoring of
Protected Visual Environments (IMPROVE) network data. One of the objectives of the
IMPROVE monitoring network program is to provide regional haze monitoring
representing all mandatory class I federal areas where practical. The National Park
Service report also describes the state of national park visibility conditions and discusses
the need for improvement.82
The regional haze rule requires states to establish goals for each affected
mandatory class I federal area that 1) improves visibility on the haziest days (20% most
impaired days), 2) ensures no degradation occurs on the cleanest days (20% least
impaired days), and 3) achieves natural background visibility levels by 2064. Although
there have been general trends toward improved visibility, progress is still needed on the
haziest days. Specifically, as discussed in the 2002 EPA Trends Report, without the
effects of pollution a natural visual range in the United States is approximately 75 to 150
km in the East and 200 to 300 km in the West. In 2001, the mean visual range for the
QO
worst days was 29 km in the East and 98 km in the West.
2.2.4.1.3 Future Visibility Impairment
Additional emission reductions will be needed from a broad set of sources,
including those in this action, as part of the overall strategy to achieve the visibility goals
of the Act and the regional haze program.
Modeling was used to project visibility conditions in 133 mandatory class I
federal areas across the US in 2020 and 2030 as a result of the small SI and marine SI
engine standards. The AQ modeling TSD and Section 2.3 of this RIA provide
information on the modeling methodology. Table 2-6 below indicates the current
monitored deciview values, the natural background levels each area is attempting to
reach, and also the projected deciview values in 2020 and 2030 with and without the
standards. In 2030, the greatest visibility improvement due to this rule (0.14 deciview)
will occur at Brigantine, New Jersey.
2-25
-------
Regulatory Impact Analysis
Table 2-6 Current (2002) and Future (2020 and 2030) Projected Visibility Conditions With and
Without Small SI and Marine SI Rule in Mandatory Class I Federal Areas (20% Worst Days)
Class 1 Area
Sipsey Wilderness
Caney Creek Wilderness
Upper Buffalo Wilderness
Chiricahua NM
Chiricahua Wilderness
Galiuro Wilderness
Grand Canyon NP
Mazatzal Wilderness
Petrified Forest NP
c Baseline
btate Visibility
AL
AR
AR
AZ
AZ
AZ
AZ
AZ
AZ
Pine Mountain Wilderness AZ
Saguaro NM
Sierra Ancha Wilderness
Sycamore Canyon
Wilderness
Agua Tibia Wilderness
Caribou Wilderness
Cucamonga Wilderness
Desolation Wilderness
Dome Land Wilderness
Emigrant Wilderness
Hoover Wilderness
Joshua Tree NM
Lassen Volcanic NP
Lava Beds NM
Mokelumne Wilderness
Pinnacles NM
Point Reyes NS
Redwood NP
San Gabriel Wilderness
San Gorgonio Wilderness
San Jacinto Wilderness
South Warner Wilderness
Thousand Lakes
Wilderness
Ventana Wilderness
Yosemite NP
Black Canyon of the
Gunnison NM
Eagles Nest Wilderness
Flat Tops Wilderness
Great Sand Dunes NM
AZ
AZ
AZ
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
CO
CO
29.03
26.36
26.27
13.43
13.43
13.43
11.66
13.35
13.21
13.35
14.83
13.67
15.25
23.50
14.15
19.94
12.63
19.43
17.63
12.87
19.62
14.15
15.05
12.63
18.46
22.81
18.45
19.94
22.17
22.17
15.05
14.15
18.46
17.63
10.33
9.61
9.61
12.78
2020 2020 Bond
Base Rule
23
22
22
13
13
13
11
12
12
12
14
13
14
21
13
17
12
18
17
12
17
13
14
12
17
21
17
17
20
19
14
13
17
17
9
9
9
12
.73
.05
.35
.09
.09
.07
.09
.72
.83
.58
.47
.20
.94
.14
.60
.36
.13
.34
.21
.72
.93
.54
.42
.30
.36
.99
.86
.25
.22
.87
.59
.52
.64
.14
.79
.03
.25
.35
23
22
22
13
13
13
11
12
12
12
14
13
14
21
13
17
12
18
17
12
17
13
14
12
.72
.03
.33
.09
.09
.06
.09
.71
.82
.56
.48
.20
.93
.13
.60
.38
.13
.34
.20
.72
.97
.54
.42
.30
17.34
21
17
17
20
19
14
13
17
17
9
.98
.86
.25
.24
.90
.59
.52
.63
.14
.79
9.03
9.25
12
.35
2030 2030 Bond Natural
Base Rule Background
23.66
21.92
22.19
13.09
13.09
13.09
11.08
12.73
12.75
12.54
14.44
13.15
14.93
20.94
13.51
17.10
12.12
18.11
17.19
12.74
17.71
13.43
14.32
12.31
17.09
21.79
17.79
16.93
19.70
19.55
14.52
13.41
17.62
17.11
9.77
8.96
9.24
12.34
23
21
22
13
13
13
11
12
12
12
14
13
14
20
13
17
12
18
17
12
17
13
14
12
17
21
17
16
19
19
14
13
17
17
9
8
9
12
.64
.89
.17
.09
.09
.09
.08
.71
.75
.53
.45
.14
.93
.94
.51
.10
.12
.11
.19
.74
.72
.43
.32
.30
.09
.79
.78
.93
.71
.52
.52
.40
.62
.11
.77
.95
.24
.34
10.99
11.58
11.57
7.21
7.21
7.21
7.14
6.68
6.49
6.68
6.46
6.59
6.69
7.64
7.31
7.06
6.12
7.46
7.64
7.91
7.19
7.31
7.86
6.12
7.99
15.77
13.91
7.06
7.30
7.30
7.86
7.31
7.99
7.64
6.24
6.54
6.54
6.66
2-26
-------
Class 1 Area
La Garita Wilderness
Maroon Bells-Snowmass
Wilderness
Mesa Verde NP
Mount Zirkel Wilderness
Rawah Wilderness
Rocky Mountain NP
Weminuche Wilderness
West Elk Wilderness
Chassahowitzka
Everglades NP
St. Marks
Cohutta Wilderness
Okefenokee
Wolf Island
Craters of the Moon NM
Sawtooth Wilderness
Mammoth Cave NP
Acadia NP
Moosehorn
Roosevelt Campobello
International Park
Isle Royale NP
Seney
Voyageurs NP
Hercules-Glades
Wilderness
Anaconda-Pintler
Wilderness
Bob Marshall Wilderness
Cabinet Mountains
Wilderness
Gates of the Mountains
Wilderness
Medicine Lake
Mission Mountains
Wilderness
Scapegoat Wilderness
Selway-Bitterroot
Wilderness
UL Bend
Linville Gorge Wilderness
Swanquarter
Lostwood
Theodore Roosevelt NP
0. . Baseline
State ... ......
Visibility
CO
CO
CO
CO
CO
CO
CO
CO
FL
FL
FL
GA
GA
GA
ID
ID
KY
ME
ME
ME
Ml
Ml
MN
MO
MT
MT
MT
MT
MT
MT
MT
MT
MT
NC
NC
ND
ND
10.33
9.61
13.03
10.52
10.52
13.83
10.33
9.61
26.09
22.30
26.03
30.30
27.13
27.13
14.00
13.78
31.37
22.89
21.72
21.72
20.74
24.16
19.27
26.75
13.41
14.48
14.09
11.29
17.72
14.48
14.48
13.41
15.14
28.77
25.49
19.57
17.74
2020 2020 Bond
Base Rule
9.89
9.21
12.39
10.05
10.04
13.08
9.85
9.15
21.94
19.77
21.82
23.33
23.42
23.37
12.97
13.63
25.48
19.77
18.63
18.45
19.10
21.72
17.58
22.93
13.14
14.13
13.54
10.91
16.19
14.04
14.16
13.04
14.64
22.45
21.15
17.70
16.49
9.89
9.21
12.39
10.05
10.03
13.06
9.85
9.15
21.92
19.76
21.81
23.32
23.41
23.35
12.96
13.63
25.47
19.75
18.62
18.44
19.08
21.70
17.56
22.92
13.13
14.12
13.53
10.91
16.19
14.04
14.15
13.04
14.63
22.44
21.11
17.70
16.48
2030 2030 Bond Natural
Base Rule Background
9.88
9.20
12.37
10.04
10.04
13.01
9.85
9.14
21.91
19.94
21.83
23.28
23.40
23.32
12.82
13.63
25.44
19.81
18.64
18.47
19.04
21.66
17.43
22.81
13.11
14.08
13.46
10.87
16.09
13.99
14.12
12.99
14.58
22.41
21.15
17.60
16.34
9.87
9.20
12.37
10.03
10.02
12.99
9.84
9.14
21.88
19.91
21.81
23.26
23.39
23.29
12.80
13.63
25.42
19.78
18.62
18.45
19.01
21.63
17.41
22.78
13.11
14.07
13.44
10.86
16.09
13.99
14.11
12.99
14.57
22.39
21.10
17.60
16.34
6.24
6.54
6.83
6.44
6.44
7.24
6.24
6.54
11.21
12.15
11.53
11.14
11.44
11.44
7.53
6.43
11.08
12.43
12.01
12.01
12.37
12.65
12.06
11.30
7.43
7.74
7.53
6.45
7.90
7.74
7.74
7.43
8.16
11.22
11.94
8.00
7.79
2-27
-------
Regulatory Impact Analysis
Class 1 Area
Great Gulf Wilderness
Presidential Range-Dry
River Wilderness
Brigantine
Bandelier NM
Bosque del Apache
Gila Wilderness
Pecos Wilderness
Salt Creek
San Pedro Parks
Wilderness
Wheeler Peak Wilderness
White Mountain
Wilderness
Jarbidge Wilderness
Wichita Mountains
Crater Lake NP
Diamond Peak
Wilderness
Eagle Cap Wilderness
Gearhart Mountain
Wilderness
Hells Canyon Wilderness
Kalmiopsis Wilderness
Mount Hood Wilderness
Mount Jefferson
Wilderness
Mount Washington
Wilderness
Mountain Lakes
Wilderness
Strawberry Mountain
Wilderness
Three Sisters Wilderness
Cape Remain
Badlands NP
Wind Cave NP
Great Smoky Mountains
NP
Joyce-Kilmer-Slickrock
Wilderness
Big Bend NP
Carlsbad Caverns NP
Guadalupe Mountains NP
Arches NP
Bryce Canyon NP
0. . Baseline
State ... ......
Visibility
NH
NH
NJ
NM
NM
NM
NM
NM
NM
NM
NM
NV
OK
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
SC
SD
SD
TN
TN
TX
TX
TX
UT
UT
22.82
22.82
29.01
12.22
13.80
13.11
10.41
18.03
10.17
10.41
13.70
12.07
23.81
13.74
13.74
18.57
13.74
18.55
15.51
14.86
15.33
15.33
13.74
18.57
15.33
26.48
17.14
15.84
30.28
30.28
17.30
17.19
17.19
11.24
11.65
2020 2020 Bond
Base Rule
19
19
24
11
12
12
9
16
9
9
12
11
20
13
13
17
13
17
14
14
14
14
13
17
14
22
15
14
23
23
16
15
15
11
11
.45
.45
.85
.35
.85
.54
.97
.59
.43
.88
.88
.86
.62
.27
.20
.83
.37
.20
.98
.13
.77
.75
.24
.73
.82
.74
.84
.91
.93
.43
.13
.89
.87
.11
.34
19
19
24
11
12
12
9
16
9
9
12
11
20
13
13
17
13
17
14
14
14
14
13
17
14
22
15
14
23
23
16
15
15
11
11
.43
.43
.75
.35
.85
.54
.97
.58
.43
.88
.89
.85
.60
.25
.19
.82
.37
.19
.97
.12
.76
.74
.23
.72
.81
.72
.83
.91
.92
.42
.13
.89
.86
.11
.34
2030 2030 Bond Natural
Base Rule Background
19.46
19.46
24.91
11.29
12.73
12.54
9.97
16.52
9.40
9.87
12.87
11.85
20.55
13.20
13.12
17.71
13.33
17.04
14.93
14.14
14.76
14.72
13.17
17.60
14.79
22.71
15.74
14.87
23.86
23.37
16.15
15.87
15.84
11.03
11.31
19
19
24
11
12
.43
.43
.77
.28
.73
12.53
9
16
9
9
12
11
20
13
13
17
13
17
14
14
14
14
13
17
14
22
15
14
23
23
16
15
15
11
11
.97
.52
.40
.87
.86
.85
.53
.18
.11
.70
.33
.01
.92
.12
.75
.71
.16
.59
.78
.68
.74
.86
.85
.35
.14
.87
.84
.01
.31
11.99
11.99
12.24
6.26
6.73
6.69
6.44
6.81
6.08
6.44
6.86
7.87
7.53
7.84
7.84
8.92
7.84
8.32
9.44
8.44
8.79
8.79
7.84
8.92
8.79
12.12
8.06
7.71
11.24
11.24
7.16
6.68
6.68
6.43
6.86
2-28
-------
Class 1 Area
Canyonlands NP
Zion NP
James River Face
Wilderness
Shenandoah NP
Lye Brook Wilderness
Alpine Lake Wilderness
Glacier Peak Wilderness
Goat Rocks Wilderness
Mount Adams Wilderness
Mount Rainier NP
North Cascades NP
Olympic NP
Pasayten Wilderness
Dolly Sods Wilderness
Otter Creek Wilderness
Bridger Wilderness
Fitzpatrick Wilderness
Grand Teton NP
North Absaroka
Wilderness
Red Rock Lakes
Teton Wilderness
Washakie Wilderness
Yellowstone NP
0. . Baseline
State ... ......
Visibility
UT
UT
VA
VA
VT
WA
WA
WA
WA
WA
WA
WA
WA
WV
WV
WY
WY
WY
WY
WY
WY
WY
WY
11.24
13.24
29.12
29.31
24.45
17.84
13.96
12.76
12.76
18.24
13.96
16.74
15.23
29.04
29.04
11.12
11.12
11.76
11.45
11.76
11.76
11.45
11.76
2020 2020 Bond
Base Rule
10.81
12.92
23.34
22.80
21.08
16.71
13.60
12.05
12.01
17.24
13.57
15.82
14.84
22.35
22.29
10.80
10.85
11.35
11.16
11.43
11.40
11.17
11.38
10.81
12.95
23.31
22.78
21.06
16.69
13.60
12.03
12.00
17.23
13.57
15.82
14.84
22.34
22.28
10.80
10.85
11.35
11.16
11.42
11.39
11.16
11.37
2030 2030 Bond Natural
Base Rule Background
10.82
12.81
23.26
22.76
21.11
16.60
13.67
12.03
11.97
17.21
13.67
15.89
14.81
22.33
22.27
10.79
10.84
11.31
11.13
11.39
11.36
11.14
11.34
10.85
12.83
23.23
22.73
21.08
16.57
13.66
12.02
11.96
17.18
13.66
15.86
14.81
22.31
22.25
10.78
10.84
11.31
11.12
11.39
11.36
11.14
11.33
6.43
6.99
11.13
11.35
11.73
8.43
8.01
8.36
8.36
8.55
8.01
8.44
8.26
10.39
10.39
6.58
6.58
6.51
6.86
6.51
6.51
6.86
6.51
The level of visibility impairment in an area is based on the light-extinction coefficient and a unitless
visibility index, called a "deciview", which is used in the valuation of visibility. The deciview metric
provides a scale for perceived visual changes over the entire range of conditions, from clear to hazy. Under
many scenic conditions, the average person can generally perceive a change of one deciview. The higher
the deciview value, the worse the visibility. Thus, an improvement invisibility is a decrease in deciview
value.
2.2.4.2 Particulate Matter Deposition
Paniculate matter contributes to adverse effects on vegetation and ecosystems,
and to soiling and materials damage. These welfare effects result predominately from
exposure to excess amounts of specific chemical species, regardless of their source or
predominant form (particle, gas or liquid). Reflecting this fact, the PM AQCD concludes
that regardless of size fractions, particles containing nitrates and sulfates have the greatest
potential for widespread environmental significance, while effects are also related to
other chemical constituents found in ambient PM, such as trace metals and organics. The
following characterizations of the nature of these welfare effects are based on the
information contained in the PM AQCD and PM Staff Paper.
2-29
-------
Regulatory Impact Analysis
2.2.4.2.1 Deposition of Nitrates and Sulfates
At current ambient levels, risks to vegetation from short-term exposures to dry
deposited particulate nitrate or sulfate are low. However, when found in acid or
acidifying deposition, such particles do have the potential to cause direct leaf injury.
Specifically, the responses of forest trees to acid precipitation (rain, snow) include
accelerated weathering of leaf cuticular surfaces, increased permeability of leaf surfaces
to toxic materials, water, and disease agents; increased leaching of nutrients from foliage;
and altered reproductive processes—all which serve to weaken trees so that they are more
susceptible to other stresses (e.g., extreme weather, pests, pathogens). Acid deposition
with levels of acidity associated with the leaf effects described above are currently found
in some locations in the eastern U.S.84 Even higher concentrations of acidity can be
present in occult depositions (e.g., fog, mist or clouds) which more frequently impacts
higher elevations. Thus, the risk of leaf injury occurring from acid deposition in some
areas of the eastern U.S. is high. Nitrogen deposition has also been shown to impact
ecosystems in the western U.S. A study conducted in the Columbia River Gorge
National Scenic Area (CRGNSA), located along a portion of the Oregon/Washington
border, indicates that lichen communities in the CRGNSA have shifted to a higher
oc
proportion of nitrophilous species and the nitrogen content of lichen tissue is elevated.
Lichens are sensitive indicators of nitrogen deposition effects to terrestrial ecosystems
and the lichen studies in the Columbia River Gorge clearly show that ecological effects
from air pollution are occurring.
Some of the most significant detrimental effects associated with excess reactive
nitrogen deposition are those associated with a syndrome known as nitrogen saturation.
These effects include: (1) decreased productivity, increased mortality, and/or shifts in
plant community composition, often leading to decreased biodiversity in many natural
habitats wherever atmospheric reactive nitrogen deposition increases significantly and
critical thresholds are exceeded; (2) leaching of excess nitrate and associated base cations
from soils into streams, lakes, and rivers, and mobilization of soil aluminum; and (3)
fluctuation of ecosystem processes such as nutrient and energy cycles through changes in
the functioning and species composition of beneficial soil organisms.86
In the U.S. numerous forests now show severe symptoms of nitrogen saturation.
These forests include: the northern hardwoods and mixed conifer forests in the
Adirondack and Catskill Mountains of New York; the red spruce forests at Whitetop
Mountain, Virginia, and Great Smoky Mountains National Park, North Carolina; mixed
hardwood watersheds at Fernow Experimental Forest in West Virginia; American beech
forests in Great Smoky Mountains National Park, Tennessee; mixed conifer forests and
chaparral watersheds in southern California and the southwestern Sierra Nevada in
Central California; the alpine tundra/subalpine conifer forests of the Colorado Front
Range; and red alder forests in the Cascade Mountains in Washington.
Excess nutrient inputs into aquatic ecosystems (i.e. streams, rivers, lakes,
estuaries or oceans) either from direct atmospheric deposition, surface runoff, or leaching
from nitrogen saturated soils into ground or surface waters can contribute to conditions of
2-30
-------
severe water oxygen depletion; eutrophication and algae blooms; altered fish
distributions, catches, and physiological states; loss of biodiversity; habitat degradation;
and increases in the incidence of disease.
Severe and persistent eutrophication often directly impacts human activities. For
example, losses in the nation's fishery resources may be directly caused by fish kills
associated with low dissolved oxygen and toxic blooms. Declines in tourism occur when
low dissolved oxygen causes noxious smells and floating mats of algal blooms create
unfavorable aesthetic conditions. Risks to human health increase when the toxins from
algal blooms accumulate in edible fish and shellfish, and when toxins become airborne,
causing respiratory problems due to inhalation. According to a NOAA report, more than
half of the nation's estuaries have moderate to high expressions of at least one of these
symptoms - an indication that eutrophication is well developed in more than half of U.S.
estuaries.87
2.2.4.2.2 Deposition of Heavy Metals
Heavy metals, including cadmium, copper, lead, chromium, mercury, nickel and
zinc, have the greatest potential for influencing forest growth (PM AQCD, p. 4-87).88
Investigation of trace metals near roadways and industrial facilities indicate that a
substantial load of heavy metals can accumulate on vegetative surfaces. Copper, zinc,
and nickel have been documented to cause direct toxicity to vegetation under field
conditions (PM AQCD, p. 4-75). Little research has been conducted on the effects
associated with mixtures of contaminants found in ambient PM. While metals typically
exhibit low solubility, limiting their bioavailability and direct toxicity, chemical
transformations of metal compounds occur in the environment, particularly in the
presence of acidic or other oxidizing species. These chemical changes influence the
mobility and toxicity of metals in the environment. Once taken up into plant tissue, a
metal compound can undergo chemical changes, accumulate and be passed along to
herbivores or can re-enter the soil and further cycle in the environment. Although there
has been no direct evidence of a physiological association between tree injury and heavy
metal exposures, heavy metals have been implicated because of similarities between
metal deposition patterns and forest decline (PM AQCD, p. 4-76). This hypothesized
relationship/correlation was further explored in high elevation forests in the northeastern
U.S. These studies measured levels of a group of intracellular compounds found in plants
that bind with metals and are produced by plants as a response to sublethal concentrations
of heavy metals. These studies indicated a systematic and significant increase in
concentrations of these compounds associated with the extent of tree injury. These data
strongly imply that metal stress causes tree injury and contributes to forest decline in the
northeastern United States (PM AQCD 4-76,77).89 Contamination of plant leaves by
heavy metals can lead to elevated soil levels. Trace metals absorbed into the plant
frequently bind to the leaf tissue, and then are lost when the leaf drops (PM AQCD, p. 4-
75). As the fallen leaves decompose, the heavy metals are transferred into the soil.90'91
The environmental sources and cycling of mercury are currently of particular
concern due to the bioaccumulation and biomagnification of this metal in aquatic
2-31
-------
Regulatory Impact Analysis
ecosystems and the potent toxic nature of mercury in the forms in which is it ingested by
people and other animals. Mercury is unusual compared with other metals in that it
largely partitions into the gas phase (in elemental form), and therefore has a longer
residence time in the atmosphere than a metal found predominantly in the particle phase.
This property enables mercury to travel far from the primary source before being
deposited and accumulating in the aquatic ecosystem. The major source of mercury in the
Great Lakes is from atmospheric deposition, accounting for approximately eighty percent
of the mercury in Lake Michigan.92'93 Over fifty percent of the mercury in the
Chesapeake Bay has been attributed to atmospheric deposition.94 Overall, the National
Science and Technology Council identifies atmospheric deposition as the primary source
of mercury to aquatic systems.95 Forty-four states have issued health advisories for the
consumption offish contaminated by mercury; however, most of these advisories are
issued in areas without a mercury point source.
Elevated levels of zinc and lead have been identified in streambed sediments, and
these elevated levels have been correlated with population density and motor vehicle
use.96'97 Zinc and nickel have also been identified in urban water and soils. In addition,
platinum, palladium, and rhodium, metals found in the catalysts of modern motor
vehicles, have been measured at elevated levels along roadsides.98 Plant uptake of
platinum has been observed at these locations.
2.2.4.2.3 Deposition of Poly cyclic Organic Matter
Poly cyclic organic matter (POM) is a byproduct of incomplete combustion and
consists of organic compounds with more than one benzene ring and a boiling point
greater than or equal to 100 degrees centigrade.99 Poly cyclic aromatic hydrocarbons
(PAHs) are a class of POM that contains compounds which are known or suspected
carcinogens.
Major sources of PAHs include mobile sources. PAHs in the environment may be
present as a gas or adsorbed onto airborne parti culate matter. Since the majority of PAHs
are adsorbed onto particles less than 1.0 jim in diameter, long range transport is possible.
However, studies have shown that PAH compounds adsorbed onto diesel exhaust
particulate and exposed to ozone have half lives of 0.5 to 1.0 hours.100
Since PAHs are insoluble, the compounds generally are particle reactive and
accumulate in sediments. Atmospheric deposition of particles is believed to be the major
source of PAHs to the sediments of Lake Michigan.101'102 Analyses of PAH deposition in
Chesapeake and Galveston Bay indicate that dry deposition and gas exchange from the
atmosphere to the surface water predominate.103'104 Sediment concentrations of PAHs are
high enough in some segments of Tampa Bay to pose an environmental health threat.
EPA funded a study to better characterize the sources and loading rates for PAHs into
Tampa Bay.105 PAHs that enter a water body through gas exchange likely partition into
organic rich particles and can be biologically recycled, while dry deposition of aerosols
containing PAHs tend to be more resistant to biological recycling.106 Thus, dry
deposition is likely the main pathway for PAH concentrations in sediments while
gas/water exchange at the surface may lead to PAH distribution into the food web,
2-32
-------
leading to increased health risk concerns.
Trends in PAH deposition levels are difficult to discern because of highly variable
ambient air concentrations, lack of consistency in monitoring methods, and the
significant influence of local sources on deposition levels.107 Van Metre et al. noted PAH
concentrations in urban reservoir sediments have increased by 200-300% over the last
forty years and correlate with increases in automobile use.108
Cousins et al. estimate that more than ninety percent of semi-volatile organic
compound (SVOC) emissions in the United Kingdom deposit on soil.109 An analysis of
PAH concentrations near a Czechoslovakian roadway indicated that concentrations were
thirty times greater than background.110
2.2.4.2.4 Materials Damage and Soiling
The effects of the deposition of atmospheric pollution, including ambient PM, on
materials are related to both physical damage and impaired aesthetic qualities. The
deposition of PM (especially sulfates and nitrates) can physically affect materials, adding
to the effects of natural weathering processes, by potentially promoting or accelerating
the corrosion of metals, by degrading paints, and by deteriorating building materials such
as concrete and limestone. Only chemically active fine particles or hygroscopic coarse
particles contribute to these physical effects. In addition, the deposition of ambient PM
can reduce the aesthetic appeal of buildings and culturally important articles through
soiling. Particles consisting primarily of carbonaceous compounds cause soiling of
commonly used building materials and culturally important items such as statues and
works of art.
2.3 Air Quality Modeling Methodology
In this section we present information on the air quality modeling, including the
model domain and modeling inputs. Further discussion of the modeling methodology,
including evaluations of model performance, is included in the Air Quality Modeling
Technical Support Document (AQM TSD).111
2.3.1 Air Quality Modeling Overview
A national scale air quality modeling analysis was performed to estimate future
year 8-hour ozone concentrations, annual PM2.5 concentrations, and visibility levels.
These projections were used as inputs to the calculation of expected benefits from the
small SI and marine SI emissions controls considered in this assessment. The 2002-based
CMAQ modeling platform was used as the tool for the air quality modeling of future
baseline emissions and control scenarios. It should be noted that the 2002-based
modeling platform has recently been finalized and the 2001-based modeling platform was
used as the tool for the air quality modeling performed for the proposal. In the next
paragraph we discuss some of the differences between the 2001-based platform used for
the proposal and the 2002-based platform used for this final rule.
2-33
-------
Regulatory Impact Analysis
The 2002-based modeling platform includes a number of updates and
improvements to data and tools compared to the 2001-based platform that was used for
the proposal modeling. For the final rule modeling we used the new 2002 National
Emissions Inventory along with updated versions of the models used to project future
emissions from electric generating units (EGUs) and onroad and nonroad vehicles. The
proposal modeling was based on the 2001 National Emissions Inventory. The new
platform also includes 2002 meteorology and more recent ambient design values which
were used as the starting point for projecting future air quality. For proposal, we used
meteorology for 2001 for modeling the East and 2002 for modeling the West. The
updates to CMAQ between proposal and final include (1) an in-cloud sulfate chemistry
module that accounts for the nonlinear sensitivity of sulfate formation to varying pH; (2)
improved vertical convective mixing; (3) heterogeneous reaction involving nitrate
formation; (4) an updated gas-phase chemistry mechanism, Carbon Bond 2005 (CB05);
and (5) an aqueous chemistry mechanism that provides a comprehensive simulation of
aerosol precursor oxidants.
The CMAQ model is a three-dimensional grid-based Eulerian air quality model
designed to estimate the formation and fate of oxidant precursors, primary and secondary
particulate matter concentrations and deposition over regional and urban spatial scales
(e.g., over the contiguous U.S.).112'113'114 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 pollution control measures that affect PM, ozone and deposition of pollutants to
the surface. In addition to the CMAQ model, the modeling platform includes the
emissions, meteorology, and initial/boundary condition data which are inputs to this
model.
The CMAQ model was peer-reviewed in 2003 for EPA as reported in "Peer
Review of CMAQ Model".115 The latest version of CMAQ (Version 4.6.1) was
employed for this modeling analysis. This version reflects updates, as mentioned above,
in a number of areas to improve the underlying science which include (1) use of a state-
of-the science inorganic and organic aerosol module, (2) an in-cloud sulfate chemistry
module that accounts for the nonlinear sensitivity of sulfate formation to varying pH, (3)
improved vertical convective mixing, (4) heterogeneous reaction involving nitrate
formation and (5) an updated Carbon Bond 05 (CB05) gas-phase chemistry mechanism
and aqueous chemistry mechanism that provides a comprehensive simulation of aerosol
precursor oxidants.
2.3.2 Model Domain and Configuration
The CMAQ modeling domain encompasses all of the lower 48 States and portions
of Canada and Mexico. The modeling domain is made up of a large continental U.S. 36
km grid and two 12 km grids (an Eastern US and a Western US domain), as shown in
Figure 2-4. The modeling domain contains 14 vertical layers with the top of the
modeling domain at about 16,200 meters, or 100 millibars (mb).
2-34
-------
12km Western Domain {WRAP}
Ofigin: -2412000. -972000
col: 213 row: 192
12km Eastern Domain
Figure 2-4. Map of the CMAQ modeling domain
2.3.3 Model Inputs
The key inputs to the CMAQ model include emissions from anthropogenic and
biogenic sources, meteorological data, and initial and boundary conditions. The CMAQ
meteorological input files were derived from a simulation of the Pennsylvania State
University / National Center for Atmospheric Research Mesoscale Model116 for the entire
year of 2002. This model, commonly referred to as MM5, is a limited-area,
nonhydrostatic, terrain-following system that solves for the full set of physical and
thermodynamic equations which govern atmospheric motions. The meteorology for the
national 36 km grid and the 12 km Eastern U.S. grid were developed by EPA and are
described in more detail within the AQM TSD. The meteorology for the 12 km Western
U.S. grid was developed by the Western Regional Air Partnership (WRAP) Regional
Planning Organization. The meteorological outputs from MM5 were processed to create
model-ready inputs for CMAQ using the Meteorology-Chemistry Interface Processor
(MCIP) version 3.1 to derive the specific inputs to CMAQ, for example: horizontal wind
components (i.e., speed and direction), temperature, moisture, vertical diffusion rates, and
rainfall rates for each grid cell in each vertical layer.117
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, the GEOS-CHEM model.118 The
global GEOS-CHEM model simulates atmospheric chemical and physical processes
driven by assimilated meteorological observations from the NASA's Goddard Earth
2-35
-------
Regulatory Impact Analysis
Observing System (GEOS). This model was run for 2002 with a grid resolution of 2
degree x 2.5 degree (latitude-longitude) and 20 vertical layers. The predictions were used
to provide one-way dynamic boundary conditions at three-hour intervals and an initial
concentration field for the 36 km CMAQ simulations. The future base conditions from
the 36 km coarse grid modeling were used as the initial/boundary state for all subsequent
12 km finer grid modeling.
The emissions inputs used for the 2002 base year and each of the future year base
cases and control scenarios are summarized in Chapter 3 of this RIA.
2.3.4 CMAQ Evaluation
An operational model performance evaluation for PM2.5 and its related speciated
components (e.g., sulfate, nitrate, elemental carbon, organic carbon, etc.) was conducted
using the 2002 data in order to estimate the ability of the CMAQ modeling system to
replicate base year concentrations. In summary, model performance statistics were
calculated for observed/predicted pairs of daily/monthly/seasonal/annual concentrations.
Statistics were generated for the following geographic groupings: domain wide, Eastern
vs. Western (divided along the 100th meridian), and each Regional Planning
Organization (RPO) region.K The "acceptability" of model performance was judged by
comparing our results to those found in recent regional PM2.5 model applications for
other, non-EPA studies.L Overall, the performance for the 2002 modeling platform is
within the range of these other applications. A detailed summary of the 2002 CMAQ
model performance evaluation is available within the AQM TSD.
2.3.5 Model Simulation Scenarios
As part of our analysis for this rulemaking the CMAQ modeling system was used
to calculate 8-hour ozone concentrations, annual PM2 5 concentrations, and visibility
estimates for each of the following emissions scenarios:
2002 base year
2020 base line projection
2020 base line projection with small SI and marine SI controls
2030 base line projection
2030 base line projection with small SI and marine SI controls
K Regional Planning Organization regions include: Mid-Atlantic/Northeast Visibility
Union (MANE-VU), Midwest Regional Planning Organization - Lake Michigan Air
Directors Consortium (MWRPO-LADCO), Visibility Improvement State and Tribal
Association of the Southeast (VISTAS), Central States Regional Air Partnership
(CENRAP), and Western Regional Air Partnership (WRAP).
L These other modeling studies represent a wide range of modeling analyses which cover
various models, model configurations, domains, years and/or episodes, chemical
mechanisms, and aerosol modules.
2-36
-------
It should be noted that the emission control scenarios used in the air quality and
benefits modeling are slightly different than the emission control program being
finalized. The differences reflect further refinements of the regulatory program since we
performed the air quality modeling for this rule. Chapter 3 of this RIA describes the
changes in the inputs and resulting emission inventories between the preliminary
assumptions used for the air quality modeling and the final regulatory scenario. These
refinements to the program would not significantly change the results summarized here or
our conclusions drawn from this analysis.
We use the predictions from the model in a relative sense by combining the 2002
base-year predictions with predictions from each future-year scenario and applying these
modeled ratios to ambient air quality observations to estimate annual PM2.5
concentrations, 8-hour ozone concentrations, and visibility levels for each of the 2020
and 2030 scenarios. The ambient air quality observations are average conditions, on a
site by site basis, for a period centered around the model base year (i.e., 2000-2004).
After completing this process, we then calculated the effect of changes in PM, ozone and
visibility air quality metrics resulting from this rulemaking on the health and welfare
impact functions of the benefits analysis.
The projected annual PM2.5 design values were calculated using the Speciated
Modeled Attainment Test (SMAT) approach. The SMAT uses an Federal Reference
Method FRM mass construction methodology that results in reduced nitrates (relative to
the amount measured by routine speciation networks), higher mass associated with
sulfates (reflecting water included in FRM measurements), and a measure of organic
carbonaceous mass that is derived from the difference between measured PM2 5 and its
non-carbon components. This characterization of PM2 5 mass also reflects crustal
material and other minor constituents. The resulting characterization provides a complete
mass balance. It does not have any unknown mass that is sometimes presented as the
difference between measured PM2 5 mass and the characterized chemical components
derived from routine speciation measurements. However, the assumption that all mass
difference is organic carbon has not been validated in many areas of the US. The SMAT
methodology uses the following PM2.5 species components: sulfates, nitrates, ammonium,
organic carbon mass, elemental carbon, crustal, water, and blank mass (a fixed value of
0.5 |ig/m3). More complete details of the SMAT procedures can be found in the report
"Procedures for Estimating Future PM2.5 Values for the CAIR Final Rule by Application
of the (Revised) Speciated Modeled Attainment Test (SMAT)".119 For this latest
analysis, several datasets and techniques were updated. These changes are fully
described within the AQM TSD. The projected 8-hour ozone design values were
calculated using the approach identified in EPA's guidance on air quality modeling
attainment demonstrations.
2.3.6 Visibility Modeling Methodology
The modeling platform described in this section was also used to project changes
in visibility. The estimate of visibility benefits was based on the projected improvement
in annual average visibility at mandatory class I federal areas. There are 156 Federally
2-37
-------
Regulatory Impact Analysis
mandated Class I areas which, under the Regional Haze Rule, are required to achieve
natural background visibility levels by 2064. These mandatory class I federal areas are
mostly national parks, national monuments, and wilderness areas. There are currently
116 Interagency Monitoring of Protected Visual Environments (IMPROVE) monitoring
sites (representing all 156 mandatory class I federal areas) collecting ambient PM25 data
at mandatory class I federal areas, but not all of these sites have complete data for 2002.
For this analysis, we quantified visibility improvement at the 133 mandatory class I
federal areas which have complete IMPROVE ambient data for 2002 or are represented
by IMPROVE monitors with complete data.M
Visibility impairment is quantified in extinction units. Visibility degradation is
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 (bext) and visual range.
The light extinction coefficient is based on the work of Sisler, 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, and soil.120
Visual range is a measure of visibility that is inversely related to the extinction
coefficient. Visual range can be defined as the maximum distance at which one can
identify a black object against the horizon sky. Visual range (in units of kilometers) can
be calculated from bext using the formula: Visual Range (km) = 3912/bext (bext units are
inverse megameters [Mm"1])
The future year visibility impairment was calculated using a methodology which applies
modeling results in a relative sense similar to the Speciated Modeled Attainment Test
(SMAT).
In calculating visibility impairment, the extinction coefficient is made up of
individual component species (sulfate, nitrate, organics, etc). The predicted change in
visibility is calculated as the percent change in the extinction coefficient for each of the
PM species (on a daily average basis). The individual daily species extinction
coefficients are summed to get a daily total extinction value. The daily extinction
coefficients are converted to visual range and then averaged across all days. In this way,
we can calculate annual average extinction and visual range at each IMPROVE site.
Subtracting the annual average control case visual range from the base case visual range
gives a projected improvement in visual range (in km) at each mandatory class I federal
area. This serves as the visibility input for the benefits analysis (See Chapter X).
M There are 100 IMPROVE sites with complete data for 2002. Many of these sites collect data that is
"representative" of other nearby unmonitored mandatory class I federal areas. There are a total of 133
mandatory class I federal areas that are represented by the 100 sites. The matching of sites to monitors is
taken from "Guidance for Tracking Progress Under the Regional Haze Rule".
2-38
-------
For visibility calculations, we are continuing to use the IMPROVE program
species definitions and visibility formulas which are recommended in the modeling
guidance.121 Each IMPROVE site has measurements of PM25 species and therefore we
do not need to estimate the species fractions in the same way that we did for FRM sites
(using interpolation techniques and other assumptions concerning volatilization of
species).
2.4 Air Toxics
Small SI and Marine SI emissions contribute to ambient levels of air toxics
known or suspected as human or animal carcinogens, or that have noncancer health
effects. The population experiences an elevated risk of cancer and other noncancer health
effects from exposure to air toxics.122 These compounds include, but are not limited to,
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, acrolein, poly cyclic organic matter
(POM), and naphthalene. These compounds, except acetaldehyde, were identified as
national or regional risk drivers in the 1999 National-Scale Air Toxics Assessment
(NATA) and have significant inventory contributions from mobile sources.
Table2-7 Mobile Source Inventory Contribution to 1999 Emissions of NATA Risk Drivers"
1999 NATA Risk Driver
Benzene
1,3 -Butadiene
Formaldehyde
Acrolein
Polycyclic organic matter (POM)*
Naphthalene
Diesel PM and Diesel exhaust organic
gases
Percent of Emissions
Attributable to All Mobile
Sources
68%
58%
47%
25%
5%
27%
100%
Percent of Emissions
Attributable to Non-road
Sources
19%
17%
20%
11%
2%
6%
62%
" This table is generated from data contained in the pollutant specific Microsoft Access database files found
in the County-Level Emission Summaries section of the 1999 NATA webpage
(http://www.epa.gov/ttn/atw/natal999/tables.html).
*This POM inventory includes the 15 POM compounds: benzo[b]fluoranthene, benz[a]anthracene,
indeno(l,2,3-c,d)pyrene, benzo[k]fluoranthene, chrysene, benzo[a]pyrene, dibenz(a,h)anthracene, anthracene,
pyrene, benzo(g,h,i)perylene, fluoranthene, acenaphthylene, phenanthrene, fluorine, and acenaphthene.
According to NATA for 1999, mobile sources were responsible for 44 percent of
outdoor toxic emissions and almost 50 percent of the cancer risk. Benzene is the largest
contributor to cancer risk of all 133 pollutants quantitatively assessed in the 1999 NATA
and mobile sources were responsible for 68 percent of benzene emissions in 1999. In
2-39
-------
Regulatory Impact Analysis
response, EPA has recently finalized vehicle and fuel controls that address this public
health risk.N
People are exposed to toxics from spark-ignition engines as a result of operating
these engines and from intrusion into the home of emissions that occur in residential
attached garages. A study of aldehyde exposures among lawn and garden equipment
operators found formaldehyde and acetaldehyde exposure concentrations, during
approximately 30 to 120 minutes of engine use, that were one to two orders of magnitude
greater than those measured at an upwind monitor.123 The study also reported measurable
concentrations of transition metals emitted from most test engines, in addition to high
organic carbon concentrations in PM2.5 samples. Analyses of organic material emitted
from hand-held engines have detected PAHs and other compounds, suggesting that
exposures to hand-held engine emissions are similar in composition to those found in
motor vehicle-affected environments, such as near major roadways.124 Numerous studies
have reported elevated benzene concentrations in residential attached garages. 125>126>127
These studies indicate the potential for elevated exposures as a result of the use and
storage of small spark-ignition engines.
Noncancer health effects can result from chronic,0 subchronic,P or acuteQ
inhalation exposures to air toxics, and include neurological, cardiovascular, liver, kidney,
and respiratory effects as well as effects on the immune and reproductive systems.
According to the 1999 NAT A, nearly the entire U.S. population was exposed to an
average concentration of air toxics that has the potential for adverse noncancer
respiratory health effects. This will continue to be the case in 2030, even though toxics
concentrations will be lower. Mobile sources were responsible for 74 percent of the
noncancer (respiratory) risk from outdoor air toxics in 1999. The majority of this risk
was from exposure to acrolein. The confidence in the RfC for acrolein is medium and
confidence in NATA estimates of population noncancer hazard from ambient exposure to
this pollutant is low.128'129
The NATA modeling framework has a number of limitations which prevent its
use as the sole basis for setting regulatory standards. These limitations and uncertainties
are discussed on the 1999 NATA website.130 Even so, this modeling framework is very
N U.S. EPA (2007) Control of Hazardous Air Pollutants from Mobile Sources. 72 FR 8428; February 26,
2007.
0 Chronic exposure is defined in the glossary of the Integrated Risk Information (IRIS) database
(http://www.epa.gov/iris) as repeated exposure by the oral, dermal, or inhalation route for more than
approximately 10% of the life span in humans (more than approximately 90 days to 2 years in typically
used laboratory animal species).
p Defined in the IRIS database as repeated exposure by the oral, dermal, or inhalation route for more than
30 days, up to approximately 10% of the life span in humans (more than 30 days up to approximately 90
days in typically used laboratory animal species)..
Q Defined in the IRIS database as exposure by the oral, dermal, or inhalation route for 24 hours or less.
2-40
-------
useful in identifying air toxic pollutants and sources of greatest concern, setting
regulatory priorities, and informing the decision making process.
Benzene: The EPA's IRIS database lists benzene as a known human carcinogen
(causing leukemia) by all routes of exposure, and concludes that exposure is associated
with additional health effects, including genetic changes in both humans and animals and
increased proliferation of bone marrow cells in mice.131'132'133 EPA states in its IRIS
database that data indicate a causal relationship between benzene exposure and acute
lymphocytic leukemia and suggest a relationship between benzene exposure and chronic
non-lymphocytic leukemia and chronic lymphocytic leukemia. The International Agency
for Research on Carcinogens (IARC) has determined that benzene is a human carcinogen
and the U.S. Department of Health and Human Services (DHHS) has characterized
benzene as a known human carcinogen.134'135
A number of adverse noncancer health effects including blood disorders, such as
preleukemia and aplastic anemia, have also been associated with long-term exposure to
benzene.136'137 The most sensitive noncancer effect observed in humans, based on current
data, is the depression of the absolute lymphocyte count in blood.138'139 In addition,
recent work, including studies sponsored by the Health Effects Institute (HEI), provides
evidence that biochemical responses are occurring at lower levels of benzene exposure
than previously known.140'141'142'143 EPA's IRIS program has not yet evaluated these new
data.
1,3-Butadiene: EPA has characterized 1,3-butadiene as carcinogenic to humans
by inhalation.144'145 The IARC has determined that 1,3-butadiene is a human carcinogen
and the U.S. DHHS has characterized 1,3-butadiene as a known human carcinogen.146'147
There are numerous studies consistently demonstrating that 1,3-butadiene is metabolized
into genotoxic metabolites by experimental animals and humans. The specific
mechanisms of 1,3-butadiene-induced carcinogenesis are unknown; however, the
scientific evidence strongly suggests that the carcinogenic effects are mediated by
genotoxic metabolites. Animal data suggest that females may be more sensitive than
males for cancer effects associated with 1,3-butadiene exposure; there are insufficient
data in humans from which to draw conclusions about sensitive subpopulations. 1,3-
butadiene also causes a variety of reproductive and developmental effects in mice; no
human data on these effects are available. The most sensitive effect was ovarian atrophy
observed in a lifetime bioassay of female mice.148
Formaldehyde: Since 1987, EPA has classified formaldehyde as a probable
human carcinogen based on evidence in humans and in rats, mice, hamsters, and
monkeys.149 EPA is currently reviewing recently published epidemiological data. For
instance, research conducted by the National Cancer Institute (NCI) found an increased
risk of nasopharyngeal cancer and lymphohematopoietic malignancies such as leukemia
among workers exposed to formaldehyde.150'151 NCI is currently performing an update of
these studies. A recent National Institute of Occupational Safety and Health (NIOSH)
study of garment workers also found increased risk of death due to leukemia among
2-41
-------
Regulatory Impact Analysis
workers exposed to formaldehyde.152 Extended follow-up of a cohort of British chemical
workers did not find evidence of an increase in nasopharyngeal or lymphohematopoietic
cancers, but a continuing statistically significant excess in lung cancers was reported.153
In the past 15 years there has been substantial research on the inhalation
dosimetry for formaldehyde in rodents and primates by the CUT Centers for Health
Research (formerly the Chemical Industry Institute of Toxicology), with a focus on use of
rodent data for refinement of the quantitative cancer dose-response assessment.154'155'156
CIIT's risk assessment of formaldehyde incorporated mechanistic and dosimetric
information on formaldehyde.
Based on the developments of the last decade, in 2004, the working group of the
International Agency for Research on Cancer (IARC) concluded that formaldehyde is
carcinogenic to humans (Group 1), on the basis of sufficient evidence in humans and
sufficient evidence in experimental animals - a higher classification than previous IARC
evaluations. After reviewing the currently available epidemiological evidence, the IARC
(2006) characterized the human evidence for formaldehyde carcinogenicity as
"sufficient," based upon the data on nasopharyngeal cancers; the epidemiologic evidence
on leukemia was characterized as "strong."157 EPA is reviewing the recent work cited
above from the NCI and NIOSH, as well as the analysis by the CUT Centers for Health
Research and other studies, as part of a reassessment of the human hazard and dose-
response associated with formaldehyde.
Formaldehyde exposure also causes a range of noncancer health effects, including
irritation of the eyes (burning and watering of the eyes), nose and throat. Effects from
repeated exposure in humans include respiratory tract irritation, chronic bronchitis and
nasal epithelial lesions such as metaplasia and loss of cilia. Animal studies suggest that
formaldehyde may also cause airway inflammation - including eosinophil infiltration into
the airways. There are several studies that suggest that formaldehyde may increase the
risk of asthma - particularly in the young.158'159
Acetaldehyde: Acetaldehyde is classified in EPA's IRIS database as a probable
human carcinogen, based on nasal tumors in rats, and is considered toxic by the
inhalation, oral, and intravenous routes.160 Acetaldehyde is reasonably anticipated to be a
human carcinogen by the U.S. DHHS in the 11th Report on Carcinogens and is classified
as possibly carcinogenic to humans (Group 2B) by the IARC.161'162 EPA is currently
conducting a reassessment of cancer risk from inhalation exposure to acetaldehyde.
The primary noncancer effects of exposure to acetaldehyde vapors include
irritation of the eyes, skin, and respiratory tract.163 In short-term (4 week) rat studies,
degeneration of olfactory epithelium was observed at various concentration levels of
acetaldehyde exposure.164'165 Data from these studies were used by EPA to develop an
inhalation reference concentration. Some asthmatics have been shown to be a sensitive
subpopulation to decrements in functional expiratory volume (FEV1 test) and
bronchoconstriction upon acetaldehyde inhalation.166 The agency is currently conducting
a reassessment of the health hazards from inhalation exposure to acetaldehyde.
2-42
-------
Acrolein: EPA determined in 2003 that the human carcinogenic potential of
acrolein could not be determined because the available data were inadequate. No
information was available on the carcinogenic effects of acrolein in humans and the
animal data provided inadequate evidence of carcinogen!city.167 The IARC determined
in 1995 that acrolein was not classifiable as to its carcinogenicity in humans.168
Acrolein is extremely acrid and irritating to humans when inhaled, with acute
exposure resulting in upper respiratory tract irritation, mucus hypersecretion and
congestion. Levels considerably lower than 1 ppm (2.3 mg/m3) elicit subjective
complaints of eye and nasal irritation and a decrease in the respiratory rate.169'170 Lesions
to the lungs and upper respiratory tract of rats, rabbits, and hamsters have been observed
after subchronic exposure to acrolein. Based on animal data, individuals with
compromised respiratory function (e.g., emphysema, asthma) are expected to be at
increased risk of developing adverse responses to strong respiratory irritants such as
acrolein. This was demonstrated in mice with allergic airway-disease by comparison to
non-diseased mice in a study of the acute respiratory irritant effects of acrolein.171
EPA is currently in the process of conducting an assessment of acute exposure
effects for acrolein. The intense irritancy of this carbonyl has been demonstrated during
controlled tests in human subjects, who suffer intolerable eye and nasal mucosal sensory
reactions within minutes of exposure.172
Polycyclic Organic Matter (POM): POM is generally defined as a large class of
organic compounds which have multiple benzene rings and a boiling point greater than
100 degrees Celsius. 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.
One of these compounds, naphthalene, is discussed separately below. Polycyclic
aromatic hydrocarbons (PAHs) are a subset of POM that contain only hydrogen and
carbon atoms. A number of PAHs are known or suspected carcinogens. Recent studies
have found that maternal exposures to PAHs (a subclass of POM) in a population of
pregnant women were associated with several adverse birth outcomes, including low
birth weight and reduced length at birth, as well as impaired cognitive development at
age three.173'174 EPA has not yet evaluated these recent studies.
Naphthalene: Naphthalene is found in small quantities in gasoline and diesel
fuels. Naphthalene emissions have been measured in larger quantities in both gasoline
and diesel exhaust compared with evaporative emissions from mobile sources, indicating
it is primarily a product of combustion. EPA recently released an external review draft of
a reassessment of the inhalation carcinogenicity of naphthalene based on a number of
recent animal carcinogenicity studies.175 The draft reassessment recently completed
external peer review.176 Based on external peer review comments received to date,
additional analyses are being undertaken. This external review draft does not represent
official agency opinion and was released solely for the purposes of external peer review
and public comment. Once EPA evaluates public and peer reviewer comments, the
document will be revised. The National Toxicology Program listed naphthalene as
2-43
-------
Regulatory Impact Analysis
"reasonably anticipated to be a human carcinogen" in 2004 on the basis of bioassays
reporting clear evidence of carcinogenicity in rats and some evidence of carcinogenicity
in mice.177 California EPA has released a new risk assessment for naphthalene, and the
IARC has reevaluated naphthalene and re-classified it as Group 2B: possibly
carcinogenic to humans.178 Naphthalene also causes a number of chronic non-cancer
effects in animals, including abnormal cell changes and growth in respiratory and nasal
179
tissues.
The small SI and marine SI standards will reduce air toxics emitted from these
engines, vessels and equipment, thereby helping to mitigate some of the adverse health
effects associated with their operation. The assumption that toxic reductions track
reductions in HC are supported by results from numerous test programs, including recent
testing on small nonroad gasoline engines with and without controls.180
2.5 Carbon Monoxide
Unlike many gases, CO is odorless, colorless, tasteless, and nonirritating. Carbon
monoxide results from incomplete combustion of fuel and is emitted directly from
vehicle tailpipes. Incomplete combustion is most likely to occur at low air-to-fuel ratios
in the engine. These conditions are common during vehicle starting when air supply is
restricted ("choked"), when vehicles are not tuned properly, and at high altitude, where
"thin" air effectively reduces the amount of oxygen available for combustion (except in
engines that are designed or adjusted to compensate for altitude). High concentrations of
CO generally occur in areas with elevated mobile-source emissions. Carbon monoxide
emissions increase dramatically in cold weather. This is because engines need more fuel
to start at cold temperatures and because some emission control devices (such as oxygen
sensors and catalytic converters) operate less efficiently when they are cold. Also,
nighttime inversion conditions are more frequent in the colder months of the year. This is
due to the enhanced stability in the atmospheric boundary layer, which inhibits vertical
mixing of emissions from the surface.
2.5.1 Health Effects of CO Pollution
We are relying on the data and conclusions in the EPA Air Quality Criteria
Document for CO (CO Criteria Document), which was published in 2000, regarding the
health effects associated with CO exposure.R181 Carbon monoxide enters the bloodstream
through the lungs and forms carboxyhemoglobin (COHb), a compound that inhibits the
blood's capacity to carry oxygen to organs and tissues.182'183 Carbon monoxide has long
been known to have substantial adverse effects on human health, including toxic effects
on blood and tissues, and effects on organ functions. Although there are effective
compensatory increases in blood flow to the brain, at some concentrations of COHb,
somewhere above 20 percent, these compensations fail to maintain sufficient oxygen
delivery, and metabolism declines.184 The subsequent hypoxia in brain tissue then
R The NAAQS review process is underway for CO and the CO Integrated Science
Assessment is scheduled to be completed in 2010.
2-44
-------
produces behavioral effects, including decrements in continuous performance and
IOC
reaction time.
Carbon monoxide has been linked to increased risk for people with heart disease,
reduced visual perception, cognitive functions and aerobic capacity, and possible fetal
effects.186 Persons with heart disease are especially sensitive to carbon monoxide
poisoning and may experience chest pain if they breathe the gas while exercising.187
Infants, elderly persons, and individuals with respiratory diseases are also particularly
sensitive. Carbon monoxide can affect healthy individuals, impairing exercise capacity,
visual perception, manual dexterity, learning functions, and ability to perform complex
tasks.188
Several epidemiological studies have shown a link between CO and premature
morbidity (including angina, congestive heart failure, and other cardiovascular diseases).
Several studies in the United States and Canada have also reported an association
between ambient CO exposures and frequency of cardiovascular hospital admissions,
especially for congestive heart failure (CHF). An association between ambient CO
exposure and mortality has also been reported in epidemiological studies, though not as
consistently or specifically as with CHF admissions. EPA reviewed these studies as part
of the CO Criteria Document review process and noted the possibility that the average
ambient CO levels used as exposure indices in the epidemiology studies may be
surrogates for ambient air mixes impacted by combustion sources and/or other
constituent toxic components of such mixes. More research will be needed to better
clarify CO's role.189
As noted above, CO has been linked to numerous health effects. In addition to
health effects from chronic exposure to ambient CO levels, acute exposures to higher
levels are also a problem. Acute exposures to CO are discussed further in Section 2.6.
2.5.2 Attainment and Maintenance of the CO NAAQS
On July 3, 1995 EPA made a finding that small land-based spark-ignition engines
cause or contribute to CO nonattainment (60 FR 34581, July 3, 1995). Marine spark-
ignition engines, which have relatively high per engine CO emissions, can also be a
source of CO emissions in CO nonattainment areas. In the preamble for this proposed
rule EPA makes a finding that recreational marine engines and vessels cause or
contribute to CO nonattainment and we provide information showing CO emissions from
spark-ignition marine engines and vessels in the CO nonattainment areas in 2005. Spark-
ignition marine engines and vessels contribute to CO nonattainment in more than one of
the CO nonattainment areas.
A nonattainment area is defined in the Clean Air Act (CAA) as an area that is
violating an ambient standard or is contributing to a nearby area that is violating the
standard. EPA has designated nonattainment areas for the CO NAAQS by calculating air
quality design values and considering other factors.8
s The full details involved in calculating a CO design value are given in 40 CFR Part 50.8.
2-45
-------
Regulatory Impact Analysis
There are two CO NAAQS. The 8-hour average CO NAAQS is 9 ppm, not to be
exceeded more than once per year, and the 1-hour average CO NAAQS is 35 ppm, not to
be exceeded more than once per year. As of March 12, 2008, there are approximately
850 thousand people living in 4 areas (which include 5 counties) that are designated as
nonattainment for CO, see Table 2-8. The emission reductions in this rule will help areas
to attain and maintain the CO NAAQS.
Table 2-8: Classified Carbon Monoxide Nonattainment Areas as of March 2008a
Area
Las Vegas, NV
El Paso, TX
Reno, NV
Total
Classification
serious
moderate <= 12.7 ppm
moderate <= 12.7 ppm
Population (1000s)
479
62
179
719.5
a This table does not include Salem, OR which is an unclassified CO nonattainment area.
In addition to the CO nonattainment areas, there are areas that have not been
designated as nonattainment where air quality monitoring may indicate a need for CO
control. For example, areas like Birmingham, AL and Calexico, CA have not been
designated as nonattainment although monitors in these areas have recorded multiple
exceedances since 1995.190
There are also almost 69 million people living in CO maintenance areas, see
Table 2-9.T Carbon monoxide maintenance areas may remain at risk for high CO
episodes especially in geographic areas with unusually challenging meteorological and
topographical conditions and in areas with high population growth and increasing vehicle
miles traveled.
Table 2-9: Carbon Monoxide Maintenance Areas as of March 2008
[Number of Areas [Number of Counties [Population (1000s)
Serious
Moderate > 12.7ppm
Moderate <= 12.7ppm
Unclassified
Total
6
4
30
33
73
15
19
62
38
127
20,496,077
17,575,606
23,371,653
7,480,907
68,924,243
A 2003 NAS report found that in geographical areas that have achieved
attainment of the NAAQS, it might still be possible for ambient concentrations of CO to
sporadically exceed the standard under unfavorable conditions such as strong winter
The CO nonattainment and maintenance areas are listed in Appendix 2E to this RIA.
2-46
-------
inversions. Areas like Alaska are prone to winter inversions due to their topographic and
meteorological conditions. The report further suggests that additional reductions in CO
are prudent to further reduce the risk of violations in regions with problematic
topography and temporal variability in meteorology.191 The reductions in CO emissions
from this rule will assist areas in maintaining the CO standard.
As discussed in the preamble, Small SI engines and equipment and Marine SI
engines and vessels do contribute to CO nonattainment. The CO emission benefits from
this rule will help states in their strategy to attain the CO NAAQS. Maintenance of the
CO NAAQS is also challenging and many areas will be able to use the emissions
reductions from this rule to assist in maintaining the CO NAAQS into the future.
2.6 Acute Exposure to Air Pollutants
Emissions from Small SI engines and equipment and Marine SI engines and
vessels contribute to ambient concentrations of ozone, CO, air toxics and PM and acute
exposures to air toxics, CO and PM. The standards being finalized in this action can help
reduce acute exposures to emissions from Marine SI engines and vessels and Small SI
engines and equipment.
2.6.1 Exposure to CO from Marine SI Engines and Vessels
In recent years, a substantial number of CO poisonings and deaths have occurred
on and around recreational boats across the nation. The actual number of deaths
attributable to CO poisoning while boating is difficult to estimate because CO-related
deaths in the water may be labeled as drowning. An interagency team consisting of the
National Park Service, the U.S. Department of Interior, and the National Institute for
Occupational Safety and Health maintains a record of published CO-related fatal and
nonfatal poisonings.192 Between 1984 and 2004, 113 CO-related deaths and 458 non-
fatal CO poisonings have been identified based on hospital records, press accounts, and
other information. Deaths have been attributed to exhaust from both onboard generators
and propulsion engines. Houseboats, cabin cruisers, and ski boats are the most common
types of boats associated with CO poisoning cases. These incidents have prompted other
federal agencies, including the United States Coast Guard and National Park Service, to
issue advisory statements and other interventions to boaters to avoid activities that could
lead to excessive CO exposure.193
CO concentrations can be extremely elevated within several meters of the exhaust
port. Engineers and industrial hygienists from CDC/NIOSH and other state and federal
agencies have conducted field studies of CO concentrations on and around houseboats.
In one study of houseboat concentrations, CO concentrations immediately at the point of
generator exhaust discharge on one houseboat averaged 0.5% (5,000 ppm), and ranged
from 0.0% to 1.28% (12,800 ppm).194 With both propulsion and generators running,
time-averaged concentrations on the swim deck were 0.2 - 169 ppm at different locations
on one boat's swim platform, 17-570 ppm on another's, and 0-108 on another. Other
studies also show the potential for high concentrations with extreme peaks in CO
2-47
-------
Regulatory Impact Analysis
concentrations in locations where boaters and swimmers can be exposed during typical
boating activities, such as standing on a swim deck or swimming near a boat.
2.6.2 Exposure to CO and PM from Small SI Engines and Equipment
A large segment of the population uses small, gasoline-powered SI lawn and
garden equipment on a regular basis. Emissions from many of the Small SI engines
powering this equipment may lead to elevated air pollution exposures for a number of
gaseous and paniculate compounds, especially for individuals such as landscapers, whose
occupations require the daily use of these engines and equipment.
Emission studies with lawn and garden equipment suggest a potential for high
exposures during the Small SI engine operation.195'196 Studies investigating air pollutant
exposures during small engine use did report elevated personal exposure measurements
related to lawn and garden equipment use.197'198 Bunger et al. reported elevated CO
personal measurements related to chainsaw use, with short-term concentrations exceeding
400 ppm for certain cutting activities. This study evaluated personal exposures during
the use of uncontrolled chainsaws. Baldauf at al. evaluated the use of lawnmowers,
chainsaws and string trimmers meeting US EPA Phase 2 standards. In this study, short-
term exposures during lawnmower and chainsaw use exceeded 120 ppm of CO, while
string trimmer use resulted in some short-term exposures approaching 100 ppm of CO.
This study also indicated that short-term PM2.5 exposures could exceed 100 //g/m3.
Pollutant exposures were highly dependent on the operator's orientation to the engine and
wind direction, as well as the activities being conducted.
These studies indicate that emissions from some lawn and garden equipment
meeting EPA's current Phase 2 standards may contribute to elevated exposures to certain
pollutants. The potential for elevated exposure to CO and PM2.5 for operators of Small SI
engines and equipment will be reduced by this rule.
1 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. Error! Main Document Only.This document is
contained in Docket Identification EPA-HQ-OAR-2004-0008-0455 to 0457.
2 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. Error! Main Document Only.This document is
contained in Docket Identification EPA-HQ-OAR-2004-0008-0455 to 0457.
3 U.S. EPA, Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper, Washington, DC, EPA-452/R-07-003, January
2-48
-------
2007. Error! Main Document Only.This document is available in Docket EPA-HQ-OAR-2004-
0008.
4
National Research Council (NRC), 2008. Estimating Mortality Risk Reduction and Economic Benefits
from Controlling Ozone Air Pollution. The National Academies Press: Washington, D.C.
5 Bates, D.V.; Baker-Anderson, M; Sizto, R. (1990) Asthma attack periodicity: a study of hospital
emergency visits in Vancouver. Environ. Res. 51: 51-70.
6 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.
7 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.
8 Lipfert, F.W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital admissions.
Environ. Res. 59: 374-399.
9 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.
10 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
11 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
12 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.
13 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.
14 Koren, H. S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.; Kozumbo, W. J.;
Becker, S.; House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989a) Ozone-induced inflammation in the
lower airways of human subjects. Am. Rev. Respir. Dis. 139: 407-415.
15 Schelegle, E.S.; Siefkin, A.D.; McDonald, R.J. (1991) Time course of ozone-induced neutrophilia in
normal humans. Am. Rev. Respir. Dis. 143:1353-1358.
16 U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants, EPA600-P-93-
004aF; (See page 7-171). This document is contained in Docket Identification EPA-HQ-OAR-2004-0008-
0455 to 0457.
17 Hodgkin, J.E.; Abbey, D.E.; Euler, G.L.; Magie, A.R. (1984) COPD prevalence in nonsmokers in high
and low photochemical air pollution areas. Chest 86: 830-838.
2-49
-------
Regulatory Impact Analysis
18 Euler, G.L.; Abbey, D.E.; Hodgkin, J.E.; Magie, A.R. (1988) Chronic obstructive pulmonary disease
symptom effects of long-term cumulative exposure to ambient levels of total oxidants and nitrogen dioxide
in California Seventh-day Adventist residents. Arch. Environ. Health 43: 279-285.
19 Abbey, D.E.; Petersen, F.; Mills, P.K.; Beeson, W.L. (1993) Long-term ambient concentrations of total
suspended particulates, ozone, and sulfur dioxide and respiratory symptoms in a nonsmoking population.
Arch. Environ. Health 48: 33-46.
20 U.S. EPA, Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper, Washington, DC, EPA-452/R-07-003, January
2007. This document is available in Docket EPA-HQ-OAR-2004-0008.
21 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
22 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
23 Avol, E. L.; Trim, S. C; Little, D. E.; Spier, C. E.; Smith, M. N.; Peng, R.-C; Linn, W. S.; Hackney, J.
D.; Gross, K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in
children attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of
the Air & Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
Management Association; paper no. 90-150.3.
24 Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I.; Avol, E. L.; Gross, K. B. (1990) Effect of exposures to
ambient ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141: 1136-1146.
25 Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung function
responses to ambient acid aerosol exposures in children. Environ. Health Perspect. 79: 179-185.
26 Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air
pollutants: observational studies. Presented at: 80th annual meeting of the Air Pollution Control
Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
27 Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer
camp. In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment:
papers from an international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste
Management Association; pp. 83-89. (A&WMA transaction series no. TR-19).
28 Spektor, D. M.; Thurston, G. D.; Mao, J.; He, D.; Hayes, C.; Lippmann, M. (1991) Effects of single- and
multiday ozone exposures on respiratory function in active normal children. Environ. Res. 55: 107-122.
29 Spektor, D. M.; Lippman, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F.
E.; Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active, normal children. Am.
Rev. Respir. Dis. 137: 313-320.
30 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
31 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.
2-50
-------
32 Horstman, D.H.; Ball, B.A.; Folinsbee, L.J.; Brown, I; Gerrity, T. (1995) Comparison of pulmonary
responses of asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of
ozone. Toxicol. Ind. Health.
33 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.
34 AQ modeling TSD (EPA Doc #)
35 Intergovernmental Panel on Climate Change (2007). Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, NY.
36 U.S. EPA. 1999. The Benefits and Costs of the Clean Air Act, 1990-2010. Prepared for U.S. Congress
by U.S. EPA, Office of Air and Radiation, Office of Policy Analysis and Review, Washington, DC,
November; EPA report no. EPA410-R-99-001. Error! Main Document Only.This document is
contained in Docket Identification EPA-HQ-OAR-2004-0008-0485.
37 Winner, W.E., and CJ. Atkinson. 1986. "Absorption of air pollution by plants, and consequences for
growth." Trends in Ecology and Evolution 1:15-18.
38 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
39 Tingey, D.T., and Taylor, G.E. 1982. "Variation in plant response to ozone: a conceptual model of
physiological events." In: Effects of Gaseous Air Pollution in Agriculture and Horticulture (Unsworth,
M.H., Omrod, D.P., eds.) London, UK: Butterworth Scientific, pp. 113-138.
40 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
41 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
42 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
43 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
44 Ollinger, S.V., J.D. Aber and P.B. Reich. 1997. "Simulating ozone effects on forest productivity:
interactions between leaf canopy and stand level processes." Ecological Applications 7:1237-1251.
45
Winner, W.E., 1994. "Mechanistic analysis of plant responses to air pollution." Ecological Applications,
4(4) :651-661.
46 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
2-51
-------
Regulatory Impact Analysis
47 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
48 Fox, S., and R. A. Mickler, eds. 1996. Impact of Air Pollutants on Southern Pine Forests. Springer-
Verlag, NY, Ecol. Studies, Vol. 118, 513 pp.
49 De Steiguer, J., J. Pye, C. Love. 1990. "Air Pollution Damage to U.S. Forests." Journal of Forestry, Vol
88 (8) pp. 17-22.
50 Pye, J.M. 1988. "Impact of ozone on the growth and yield of trees: A review." Journal of Environmental
Quality 17 pp.347-360.
51 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
52 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
53 McBride, J.R., P.R. Miller, andR.D. Laven. 1985. "Effects of oxidant air pollutants on forest succession
in the mixed conifer forest type of southern California." In: Air Pollutants Effects On Forest Ecosystems,
Symposium Proceedings, St. P, 1985, p. 157-167.
54 Miller, P.R., O.C. Taylor, R.G. Wilhour. 1982. Oxidant air pollution effects on a western coniferous
forest ecosystem. Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research
Laboratory (EPA600-D-82-276).
55 U.S. EPA. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006. This document is contained in Docket Identification
EPA-HQ-OAR-2004-0008-0455 to 0457.
56 Kopp, R. J.; Vaughn, W. J.; Hazilla, M.; Carson, R. 1985. "Implications of environmental policy for
U.S. agriculture: the case of ambient ozone standards." J. Environ. Manage. 20:321-331.
57 Adams, R. M.; Hamilton, S. A.; McCarl, B. A. 1986. "The benefits of pollution control: the case of
ozone and U.S. agriculture." Am. J. Agric. Econ. 34:3-19.
58 Adams, R. M.; Glyer, J. D.; Johnson, S. L.; McCarl, B. A. 1989. "A reassessment of the economic
effects of ozone on U.S. agriculture." JAPCA 39:960-968.
59 Abt Associates, Inc. 1995. Urban ornamental plants: sensitivity to ozone and potential economic losses.
U.S. EPA, Office of Air Quality Planning and Standards, Research Triangle Park. Under contract to
RADIAN Corporation, contract no. 68-D3-0033, WA no. 6. pp. 9-10.
60 U.S. EPA (2004) Air Quality Criteria for Paniculate Matter (Oct 2004), Volume I Document No.
EPA600/P-99/002aF and Volume II Document No. EPA600/P-99/002bF. Error! Main Document
Only.This document is available in Docket EPA-HQ-OAR-2004-0008-0042 and EPA-HQ-OAR-2004-
0008-0043.
2-52
-------
61 U.S. EPA (2005) Review of the National Ambient Air Quality Standard for Paniculate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-05-005. Error!
Main Document Only.This document is available in Docket EPA-HQ-OAR-2004-0008-0454.
62 U.S. EPA 2006. Provisional Assessment of Recent Studies on Health Effects of Paniculate Matter
Exposure. EPA/600/R-06/063. Error! Main Document Only.This document is available in Docket
EP A-HQ-OAR-2004-0008-0461.
63 Laden F; Neas LM; Dockery DW; et al. 2000. "Association of fine paniculate matter from different
sources with daily mortality in six U.S. cities." Environ Health Perspectives 108(10):941-947.
64 Schwartz J; Laden F; Zanobetti A. 2002. "The concentration-response relation between PM(2.5) and
daily deaths." Environ Health Perspect 110(10): 1025-1029.
65 Janssen NA; Schwartz J; Zanobetti A.; et al. 2002. "Air conditioning and source-specific particles as
modifiers of the effect of PM10 on hospital admissions for heart and lung disease." Environ Health
Perspect 110(l):43-49.
66 Dockery, DW; Pope, CA, III; Xu, X; et al. 1993. "An association between air pollution and mortality in
six U.S. cities." N Engl J Med 329:1753-1759.
67 Pope, CA, III; Burnett, RT; Thun, MJ; Calle, EE; et al. 2002. "Lung cancer, cardiopulmonary mortality,
and long-term exposure to fine paniculate air pollution." J Am Med Assoc 287: 1132-1141.
68 Krewski, D; Burnett, RT; Goldberg, M S; et al. 2000. "Reanalysis of the Harvard Six Cities study and the
American Cancer Society study of paniculate air pollution and mortality. A special report of the Institute's
Particle Epidemiology Reanalysis Project." Cambridge, MA: Health Effects Institute.
69
Jerrett, M; Burnett, RT; Ma, R; et al. 2005. "Spatial Analysis of Air Pollution and Mortality in Los
Angeles." Epidemiology. 16(6):727-736.
70
Ktinzli, N.; Jerrett, M.; Mack, W.J.; et al.(2004) Ambient air pollution and atherosclerosis in Los
Angeles. Environ Health Perspect doi: 10. 1289/ehp.7523 [Available at http://dx.doi.org/] .
71 Riediker, M.; Cascio, W.E.; Griggs, T.R.; et al. 2004. "Paniculate matter exposure in cars is associated
with cardiovascular effects in healthy young men." Am JRespir Crit Care Med 169: 934-940.
72 Maynard, D.; Coull, B.A.; Gryparis, A.; Schwartz, J. (2007) Mortality risk associated with short-term
exposure to traffic particles and sulfates. Environmental Health Perspectives 115: 751-755.
73 Ryan, P.H.; LeMasters, O.K.; Biswas, P.; Levin, L.; Hu, S.; Lindsey, M.; Bernstein, D.I.; Lockey, J.;
Villareal, M.; Khurana Hershey, O.K.; Grinshpun, S.A. (2007) A comparison of proximity and land use
regression traffic exposure models and wheezing in infants. Environ Health Perspect 115: 278-84.
74 Morgenstern, V.; Zutavern, A.; Cyrys, J.; Brockow, I.; Gehring, U.; Koletzko, S.; Bauer, C.P.; Reinhart,
D; Wichmann, H-E.; Heinrich, J. (2007) Respiratory health and individual estimated expousure to traffic-
related air pollutant in a cohort of young children. Occupational and Environmental Medicine 64: 8-16.
75 Franco Suglia, S.; Gryparis, A.; Wright, R.O.; Schwartz, J.; Wright, R.J. (2007) Association of black
carbon with cognition among children in a prospective birth cohort study. American Journal of
Epidemiology 167: 280-286.
2-53
-------
Regulatory Impact Analysis
76 National Research Council, 1993. Protecting Visibility in National Parks and Wilderness Areas.
National Academy of Sciences Committee on Haze in National Parks and Wilderness Areas. National
Academy Press, Washington, DC. This book is available electronically at
http://www.nap.edu/books/0309048443/html/ and is available in Docket EPA-HQ-OAR-2004-0008.
77 U.S. EPA, National Ambient Air Quality Standards for Paniculate Matter; Proposed Rule; January 17,
2006, Vol71 p 2676. Error! Main Document Only.This document is available in Docket EPA-
HQ-OAR-2004-0008-0466. This information is available electronically at: http://epa.gov/fedrgstr/EPA-
AIR/2006/January/Day-17/al77.pdf
78 U.S. EPA (2004) Air Quality Criteria for Paniculate Matter (Oct 2004), Volume I Document No.
EPA600/P-99/002aF and Volume II Document No. EPA600/P-99/002bF. This document is available in
Docket EPA-HQ-OAR-2004-0008.
79 U.S. EPA (2005) Review of the National Ambient Air Quality Standard for Paniculate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-05-005. This
document is available in Docket EPA-HQ-OAR-2004-0008.
80 U.S. EPA. 1993. Effects of the 1990 Clean Air Act Amendments on Visibility in Class I Areas: An EPA
Report to Congress. EPA452-R-93-014. Error! Main Document Only.This document is available
in Docket EPA-HQ-OAR-2004-0008-0483.
81 U.S. EPA (2002) Latest Findings on National Air Quality - 2002 Status and Trends. EPA 454/K-03-001.
Error! Main Document Only.This document is available in Docket EPA-HQ-OAR-2004-0008-
0482.
82 National Park Service. Air Quality in the National Parks, Second edition. NFS, Air Resources Division.
D 2266. September 2002. Error! Main Document Only.This document is available in Docket
EPA-HQ-OAR-2004-0008-0481.
83 U.S. EPA (2002) Latest Findings on National Air Quality - 2002 Status and Trends. EPA 454/K-03-001.
This document is available in Docket EPA-HQ-OAR-2004-0008-0482.
84 Environmental Protection Agency (2003). Response of Surface Water Chemistry to the Clean Air Act
Amendments of 1990. National Health and Environmental Effects Research Laboratory, Office of Research
and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 620/R-
03/001.
85 Fenn, M.E. and Blubaugh, TJ. (2005) Winter Deposition of Nitrogen and Sulfur in the Eastern Columbia
River Gorge National Scenic Area, USDA Forest Service.
86 Galloway, J. N.; Cowling, E. B. (2002). Reactive nitrogen and the world: 200 years of change. Ambio
31:64-71.
87 Bricker, Suzanne B., et al., National Estuarine Eutrophication Assessment, Effects of Nutrient
Enrichment in the Nation's Estuaries, National Ocean Service, National Oceanic and Atmospheric
Administration, September, 1999.
88 Smith, W.H. 1991. "Air pollution and Forest Damage." Chemical Engineering News, 69(45): 30-43.
89 Gawel, J.E.; Ahner, B.A.; Friedland, A.J.; and Morel, F.M.M. 1996. "Role for heavy metals in forest
decline indicated by phytochelatin measurements." Nature, 381: 64-65.
2-54
-------
90 Cotrufo, M.F.; DeSanto, A.V.; Alfani, A.; et al. 1995. "Effects of urban heavy metal pollution on organic
matter decomposition in Quercus ilix L. woods." Environmental Pollution, 89: 81-87.
91 Niklinska, M.; Laskowski, R.; Maryanski, M. 1998. "Effect of heavy metals and storage time on two
types of forest litter: basal respiration rate and exchangeable metals." Ecotoxicological Environmental
Safety, 41: 8-18.
92 Mason, R.P. and Sullivan, K.A. 1997. "Mercury in Lake Michigan." Environmental Science &
Technology, 31: 942-947. (from Delta Report "Atmospheric deposition of toxics to the Great Lakes").
93 Landis, M.S. and Keeler, G.J. 2002. "Atmospheric mercury deposition to Lake Michigan during the Lake
Michigan Mass Balance Study." Environmental Science & Technology, 21: 4518-24.
94 U.S. EPA. 2000. EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
Error! Main Document Only.This document is available in Docket EPA-HQ-OAR-2004-0008.
95 NSTC 1999
96 Callender, E. and Rice, K.C. 2000. "The Urban Environmental Gradient: Anthropogenic Influences on
the Spatial and Temporal Distributions of Lead and Zinc in Sediments." Environmental Science &
Technology, 34: 232-238.
97 Rice, K.C. 1999. "Trace Element Concentrations in Streambed Sediment Across the Conterminous
United States." Environmental Science & Technology, 33: 2499-2504.
98 Ely, JC; Neal, CR; Kulpa, CF; et al. 2001. "Implications of Platinum-Group Element Accumulation
along U.S. Roads from Catalytic-Converter Attrition." Environ. Sci. Technol. 35: 3816-3822.
99 U.S. EPA. 1998. EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of
Poly cyclic Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North
Carolina. Error! Main Document Only.This document is available in Docket EPA-HQ-OAR-
2004-0008.
100U.S. EPA. 1998. EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of
Poly cyclic Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North
Carolina. Error! Main Document Only.This document is available in Docket EPA-HQ-OAR-
2004-0008.
101 Simcik, M.F.; Eisenreich, S.J.; Golden, K.A.; et al. 1996. "Atmospheric Loading of Polycyclic Aromatic
Hydrocarbons to Lake Michigan as Recorded in the Sediments." Environmental Science and Technology,
30: 3039-3046.
102 Simcik, M.F.; Eisenreich, S.J.; and Lioy, P.J. 1999. "Source apportionment and source/sink relationship
of PAHs in the coastal atmosphere of Chicago and Lake Michigan." Atmospheric Environment, 33: 5071-
5079.
103 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. 2001. "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-
2183.
104 Park, J.S.; Wade, T.L.; and Sweet, S. 2001. "Atmospheric distribution of polycyclic aromatic
hydrocarbons and deposition to Galveston Bay, Texas, USA." Atmospheric Environment, 35: 3241-3249.
2-55
-------
Regulatory Impact Analysis
105 Poor, N.; Tremblay, R.; Kay, H.; et al. 2002. "Atmospheric concentrations and dry deposition rates of
poly cyclic aromatic hydrocarbons (PAHs) for Tampa Bay, Florida, USA." Atmospheric Environment 38:
6005-6015.
106 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. 2001. "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-
2183.
107 U.S. EPA. 2000. EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina. This
document is available in Docket EPA-HQ-OAR-2004-0008.
108 Van Metre, P.C.; Mahler, B.J.; and Furlong, E.T. 2000. "Urban Sprawl Leaves its PAH Signature."
Environmental Science & Technology, 34: 4064-4070.
109 Cousins, I.T.; Beck, A.J.; and Jones, K.C. 1999. "A review of the processes involved in the exchange of
semi-volatile organic compounds across the air-soil interface." The Science of the Total Environment, 228:
5-24.
110 Tuhackova, J. et al. (2001) Hydrocarbon deposition and soil microflora as affected by highway traffic.
Environmental Pollution, 113: 255-262.
111 U.S. EPA, Technical Support Document for the Final Small Si/Marine SI Rule: Air Quality Modeling,
Research Triangle Park, NC
112 U.S. EPA, Byun, D.W., and Ching, J.K.S., Eds, 1999. Science algorithms of EPA Models-3 Community
Multiscale Air Quality (CMAQ modeling system, EPA/600/R-99/030, Office of Research and
Development).
113 Byun, D.W., and Schere, K.L., 2006. Review of the Governing Equations, Computational Algorithms,
and Other Components of the Models-3 Community Multiscale Air Quality (CMAQ) Modeling System, J.
Applied Mechanics Reviews, 59 (2), 51-77.
114 Dennis, R.L., Byun, D.W., Novak, J.H., Galluppi, K.J., Coats, C.J., and Vouk, M.A., 1996. The next
generation of integrated air quality modeling: EPA's Models-3, Atmospheric Environment, 30, 1925-1938.
115 Amar, P., Bornstein, R., Feldman, H., Jeffries, H., Steyn, D., Yamartino, R., Zhang, Y., 2004. Final
Report Summary: December 2003 Peer Review of the CMAQ Model, p. 7.
116 Grell, G., J. Dudhia, andD. Stauffer, 1994: A Description of the Fifth-Generation Perm State/NCAR
Mesoscale Model (MM5), NCAR/TN-398+STR., 138 pp, National Center for Atmospheric Research,
Boulder CO.
117 U.S. Environmental Protection Agency, Byun, D.W., and Ching, J.K.S., Eds, 1999. Science algorithms
of EPA Models-3 Community Multiscale Air Quality (CMAQ) modeling system, EPA/600/R-99/030,
Office of Research and Development). Please also see: http://www.cmascenter.org/
118
Yantosca, B., 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling Group,
Harvard University, Cambridge, MA, October 15, 2004.
119
U.S. EPA, (2004), "Procedures for Estimating Future PM2.5 Values for the CAIR Final Rule by
Application of the (Revised) Speciated Modeled Attainment Test (SMAT)- Updated 11/8/04".
120 Sisler 1996
2-56
-------
121 U.S. EPA, Guidance on the Use of Models and Other Analyses For Demonstrating Attainment of Air
Quality Goals for Ozone, PM2.5, and Regional Haze; EPA-454/B-07-002; Research Triangle Park, NC;
April 2007.
122
U. S. EPA. 1999 National-Scale Air Toxics Assessment.
http://www.epa.gov/ttn/atw/natal999/risksum.html
123 Baldauf, R.; Fortune, C.; Weinstein, J.; Wheeler, M; Blanchard, F. (2006) Air contaminant exposures
during the operation of lawn and garden equipment. J Exposure Sci & Environ Epidemiol 16: 362-370.
124
Volckens, J.; Olson, D.A.; Hays, M.D. (2008) Carbonaceous species emitted from handheld two-stroke
engines. Atmos Environ 42: 1239-1248.
125 Isbell, M.; Ricker, J.; Gordian, M.E.; Duff, L.K. (1999) Use of biomarkers in an indoor air study: lack
of correlation between aromatic VOCs with respective urinary biomarkers. Sci Total Environ 241: 151-
159.
126 Batterman, S.; Jia, C.; Hatzivasilis, G. (2007) Migration of volatile organic compounds from attached
garages to residences: a major exposure source. Environ Res 104: 224-240.
127 Philips, M.L.; Esmen, N.A.; Hall, T.A.; Lynch, R. (2005) Determinants of exposure to volatile organic
compounds in four Oklahoma cities. J Exposure Anal Environ Epidemiol 15: 35-46.
128 U.S. EPA (2003) Integrated Risk Information System File of Acrolein. National Center for
Environmental Assessment, Office of Research and Development, Washington, D.C. 2003. This material
is available electronically at http://www.epa.gov/iris/subst/0364.htm.
129 U.S. EPA (2006) National-Scale Air Toxics Assessment for 1999. This material is available
electronically at http://www.epa.gov/ttn/atw/natal999/risksum.html.
130 U.S. EPA (2006) National-Scale Air Toxics Assessment for 1999.
http://www.epa.gov/ttn/atw/natal999.
131 U.S. EPA. 2000. Integrated Risk Information System File for Benzene. This material is available
electronically at: http://www.epa.gov/iris/subst/0276.htm.
132 International Agency for Research on Cancer, IARC monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 29, Some industrial chemicals and dyestuffs, International Agency for
Research on Cancer, World Health Organization, Lyon, France, p. 345-389, 1982.
133 Irons, R.D.; Stillman, W.S.; Colagiovanni, D.B.; Henry, V.A. (1992) Synergistic action of the benzene
metabolite hydroquinone on myelopoietic stimulating activity of granulocyte/macrophage colony-
stimulating factor in vitro, Proc. Natl. Acad. Sci. 89:3691-3695.
134 International Agency for Research on Cancer (IARC). 1987. Monographs on the evaluation of
carcinogenic risk of chemicals to humans, Volume 29, Supplement 7, Some industrial chemicals and
dyestuffs, World Health Organization, Lyon, France.
135 U.S. Department of Health and Human Services National Toxicology Program 11th Report on
Carcinogens available at: http://ntp.niehs.nih.gov/go/16183.
136 Aksoy, M. (1989). Hemato to xicity and carcinogenicity of benzene. Environ. Health Perspect. 82:
193-197.
2-57
-------
Regulatory Impact Analysis
137 Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews. 3:541-
554.
138 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, andR.B. Hayes (1996) Hematotoxicity
among Chinese workers heavily exposed to benzene. Am. J. Ind. Med. 29: 236-246.
139 U.S. EPA 2002 Toxicological Review of Benzene (Noncancer Effects). Environmental Protection
Agency, Integrated Risk Information System (IRIS), Research and Development, National Center for
Environmental Assessment, Washington DC. This material is available electronically at
http://www.epa.gov/iris/subst/0276.htm.
140 Qu, O.; Shore, R.; Li, G.; Jin, X.; Chen, C.L.; Cohen, B.; Melikian, A.; Eastmond, D.; Rappaport, S.; Li,
H.; Rupa, D.; Suramaya, R.; Songnian, W.; Huifant, Y.; Meng, M.; Winnik, M.; Kwok, E.; Li, Y.; Mu,
R.;Xu, B.; Zhang, X.; Li, K. (2003). HEI Report 115, Validation & Evaluation of Biomarkers in Workers
Exposed to Benzene in China.
141 Qu, Q., R. Shore, G. Li, X. Jin, L.C. Chen, B. Cohen, et al. (2002). Hematological changes among
Chinese workers with a broad range of benzene exposures. Am. J. Industr. Med. 42: 275-285.
142
Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in Workers Exposed to
Low Levels of Benzene. Science 306: 1774-1776.
143 Turtletaub, K.W. and Mani, C. (2003). Benzene metabolism in rodents at doses relevant to human
exposure from Urban Air. Research Reports Health Effect Inst. Report No.113.
144 U.S. EPA. 2002. Health Assessment of 1,3-Butadiene. Office of Research and Development, National
Center for Environmental Assessment, Washington Office, Washington, DC. Report No. EPA600-P-98-
00IF. This document is available electronically at http://www.epa.gov/iris/supdocs/buta-sup.pdf.
145 U.S. EPA. 2002 "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection
Agency, Integrated Risk Information System (IRIS), Research and Development, National Center for
Environmental Assessment, Washington, DC http://www.epa. gov/iris/subst/013 9.htm.
146 International Agency for Research on Cancer (IARC) (1999) Monographs on the evaluation of
carcinogenic risk of chemicals to humans, Volume 71, Re-evaluation of some organic chemicals, hydrazine
and hydrogen peroxide and Volume 97 (in preparation), World Health Organization, Lyon, France.
147 U.S. Department of Health and Human Services National Toxicology Program 11th Report on
Carcinogens available at: http://ntp.niehs.nih.gov/go/16183.
148 Bevan, C.; Stadler, J.C.; Elliot, G.S.; et al. (1996) Subchronic toxicity of 4-vinylcyclohexene in rats and
mice by inhalation. Fundam. Appl. Toxicol. 32:1-10.
149 U.S. EPA. 1987. Assessment of Health Risks to Garment Workers and Certain Home Residents from
Exposure to Formaldehyde, Office of Pesticides and Toxic Substances, April 1987.
150 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2003. Mortality from
lymphohematopoetic malignancies among workers in formaldehyde industries. Journal of the National
Cancer Institute 95: 1615-1623.
151 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2004. Mortality from solid
cancers among workers in formaldehyde industries. American Journal of Epidemiology 159: 1117-1130.
2-58
-------
152 Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to formaldehyde: an
update. Occup. Environ. Med. 61: 193-200.
153 Coggon, D, EC Harris, J Poole, KT Palmer. 2003. Extended follow-up of a cohort of British chemical
workers exposed to formaldehyde. J National Cancer Inst. 95:1608-1615.
154 Conolly, RB, JS Kimbell, D Janszen, PM Schlosser, D Kalisak, J Preston, and FJ Miller. 2003.
Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Tox Sci
75: 432-447.
155 Conolly, RB, JS Kimbell, D Janszen, PM Schlosser, D Kalisak, J Preston, and FJ Miller. 2004. Human
respiratory tract cancer risks of inhaled formaldehyde: Dose-response predictions derived from
biologically-motivated computational modeling of a combined rodent and human dataset. Tox Sci 82: 279-
296.
156 Chemical Industry Institute of Toxicology (CUT). 1999. Formaldehyde: Hazard characterization and
dose-response assessment for carcinogenicity by the route of inhalation. CUT, September 28, 1999.
Research Triangle Park, NC.
157 International Agency for Research on Cancer (2006) Formaldehyde, 2-Butoxyethanol and 1-tert-
Butoxypropan-2-ol. Monographs Volume 88. World Health Organization, Lyon, France.
ICO
Agency for Toxic Substances and Disease Registry (ATSDR). 1999. Toxicological profile for
Formaldehyde. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc.gov/toxprofiles/tpl 11.html
159
WHO (2002) Concise International Chemical Assessment Document 40: Formaldehyde. Published
under the joint sponsorship of the United Nations Environment Programme, the International Labour
Organization, and the World Health Organization, and produced within the framework of the Inter-
Organization Programme for the Sound Management of Chemicals. Geneva.
160 U.S. EPA (1988). Integrated Risk Information System File of Acetaldehyde. Research and
Development, National Center for Environmental Assessment, Washington, DC. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
161 U.S. Department of Health and Human Services National Toxicology Program 11th Report on
Carcinogens available at: http://ntp.niehs.nih.gov/go/16183.
162 International Agency for Research on Cancer (IARC). 1999. Re-evaluation of some organic chemicals,
hydrazine, and hydrogen peroxide. IARC Monographs on the Evaluation of Carcinogenic Risk of
Chemical to Humans, Vol 71. Lyon, France.
163 U.S. EPA (1988). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
164 U.S. EPA. 2003. Integrated Risk Information System File of Acrolein. Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available electronically
at http://www.epa.gov/iris/subst/0364.htm.
165 Appleman, L.M., R.A. Woutersen, and V.J. Feron. (1982). Inhalation toxicity of acetaldehyde in rats. I.
Acute and subacute studies. Toxicology. 23: 293-297.
2-59
-------
Regulatory Impact Analysis
166 Myou, S.; Fujimura, M; Nishi K.; Ohka, T.; and Matsuda, T. (1993) Aerosolized acetaldehyde induces
histamine-mediated bronchoconstriction in asthmatics. Am. Rev. Respir.Dis. 148(4 Pt 1): 940-943.
167 Integrated Risk Information System File of Acrolein. Research and Development, National Center for
Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm
168 International Agency for Research on Cancer (IARC). 1995. Monographs on the evaluation of
carcinogenic risk of chemicals to humans, Volume 63, Dry cleaning, some chlorinated solvents and other
industrial chemicals, World Health Organization, Lyon, France.
169 Weber-Tschopp, A; Fischer, T; Gierer, R; et al. (1977) Experimentelle reizwirkungen von Acrolein auf
den Menschen. Int Arch Occup Environ Hlth 40(2): 117-130. In German
170 Sim, VM; Pattle, RE. (1957) Effect of possible smog irritants on human subjects. J Am Med Assoc
165(15):1908-1913.
171 Morris JB, Symanowicz PT, Olsen JE, et al. 2003. Immediate sensory nerve-mediated respiratory
responses to irritants in healthy and allergic airway-diseased mice. J Appl Physiol 94(4): 1563-1571.
172 Sim VM, Pattle RE. Effect of possible smog irritants on human subjects JAMA165: 1980-2010, 1957.
173 Perera, P.P.; Rauh, V.; Tsai, W-Y.; et al. (2002) Effect of transplacental exposure to environmental
pollutants on birth outcomes in a multiethnic population. Environ Health Perspect. Ill: 201-205.
174 Perera, P.P.; Rauh, V; Whyatt, R.M.; Tsai, W.Y.; Tang, D.; Diaz, D.; Hoepner, L.; Barr, D.; Tu, Y.H.;
Camann, D.; Kinney, P. (2006) Effect of prenatal exposure to airborne polycyclic aromatic hydrocarbons
on neurodevelopment in the first 3 years of life among inner-city children. Environ Health Perspect 114:
1287-1292.
175 U. S. EPA. 2004. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available electronically
at http://www.epa.gov/iris/subst/0436.htm.
176 Oak Ridge Institute for Science and Education. (2004). External Peer Review for the IRIS
Reassessment of the Inhalation Carcinogenicity of Naphthalene. August 2004.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=84403
177 National Toxicology Program (NTP). (2004). 11th Report on Carcinogens. Public Health Service, U.S.
Department of Health and Human Services, Research Triangle Park, NC. Available from: http://ntp-
server.niehs.nih.gov.
178 International Agency for Research on Cancer (IARC). (2002). Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals for Humans. Vol.82. Lyon, France.
179 U. S. EPA. 1998. Toxicological Review of Naphthalene, Environmental Protection Agency, Integrated
Risk Information System, Research and Development, National Center for Environmental Assessment,
Washington, DC. This material is available electronically at http://www.epa.gov/iris/subst/0436.htm
1 80
US EPA (2004) Characterization of Emissions from Small, Hand-Held,In-Use 2-Cycle Engines Internal
Report, EPA/600/X-04/191.
2-60
-------
181 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
182 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
183 Coburn, R.F. (1979) Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322.
184 Helfaer, M.A., and Traystman, R.J. (1996) Cerebrovascular effects of carbon monoxide. In: Carbon
Monoxide (Penney, D.G., ed). Boca Raton, CRC Press, 69-86.
185
Benignus, V. A. (1994) Behavioral effects of carbon monoxide: meta analyses and extrapolations. J.
Appl. Physiol. 76:1310-1316.
186 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
187 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
188 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
189 U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1, 2000. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, D.C. This document is available online at
http://www.epa.gov/ncea/pdfs/coaqcd.pdf. A copy of this document is available in Docket Identification
EPA-HQ-OAR-2004-0008-0050.
190 National Academy of Sciences (2003). Managing Carbon Monoxide Pollution in Meteorological and
Topographical Problem Areas. The National Academies Press, Washington, D.C. This document is
available on the internet at http://www.nap.edu.
191 National Academy of Sciences (2003). Managing Carbon Monoxide Pollution in Meteorological and
Topographical Problem Areas. The National Academies Press, Washington, D.C. This document is
available on the internet at http://www.nap.edu.
192 National Park Service: Department of Interior; National Institute for Occupational Safety and Health.
(2004) Boat-related carbon monoxide (CO) poisonings. Updated: October 2004 [Online at
2-61
-------
Regulatory Impact Analysis
http://safetynet.smis.doi.gov/thelistbystatelO-19-04.pdf] A copy of this document is available in Docket
Identification EPA-HQ-OAR-2004-0008-0106.
193 U.S. Department of Interior. (2004) Carbon monoxide dangers from generators and propulsion engines.
On-board boats - compilation of materials. [Online at http://safetynet.smis.doi.gov/COhouseboats.htm] A
copy of this document is available in Docket Identification EPA-HQ-OAR-2004-0008-0110.
194 Letter dated April 5, 2001 to Mr. John Stenseth, Fun Country Marine Industries, Inc., from Ronald M.
Hall, NIOSH.
195 Gabele, P. (1997) Exhaust emissions from four-stroke lawn mower engines. Air Waste Manage Assoc.
47: 945-952.
196 Priest, M.W.; Williams, D.J.; and Bridgman, H.A. (2000) Emissions from in-use lawn-mowers in
Australia. Atmos Environ. 34:657-664.
197 Hunger, J.; Bombosch, F.; Mesecke, U.; and Hairier, E. (1997) Monitoring and analysis of occupational
exposure to chain saw exhausts. Am IndHyg Assoc J. 58:747-751.
198 Baldauf, R.; Fortune, C.; Weinstein, J.; Wheeler, M.; and Blanchard, F. (2006) Air contaminant
exposures during the operation of lawn and garden equipment. Journal of Exposure Science and
Environmental Epidemiology. 16(4):362-370.
2-62
-------
Emission Inventory
CHAPTER 3: Emission Inventory
This chapter presents our analysis of the emission impact of the final Phase 3 standards
for spark ignition (SI) small nonroad engines (<25 horsepower (hp) or < 19 kilowatts (kW) used
in land-based or auxiliary marine applications (hereafter collectively termed small nonroad SI
engines) and marine SI engines. The control requirements include exhaust and evaporative
emission standards for small non-handheld SI engines (Class I <225 cubic centimeters (cc) and
Class II >225 cc), an evaporative emission standards for small handheld SI engines (Classes III-
V), and exhaust and evaporative emission standards for all marine SI engines.
Section 3.1 presents an overview of methodology used to develop the emission
inventories for the small nonroad and marine engines that are subject to the final rulemaking.
Section 3.2 identifies the specific modeling inputs that were used to develop the baseline
scenario emission inventories. The resulting baseline emission inventories are also presented in
that section. Section 3.3 then describes the contribution of the small nonroad and marine SI
engines to national baseline inventories. Section 3.4 describes the development of the controlled
inventories, specifically the changes made to the baseline modeling inputs to incorporate the new
standards. The control inventories are also presented in this section. Section 3.5 follows with
the projected emission reductions resulting from the final rule. Section 3.6 describes the
emission inventories used in the air quality modeling described in Chapter 2. This discussion
includes a description of the changes in the inputs and resulting emission inventories between the
preliminary baseline and control scenarios used for the air quality modeling and the slightly
refined baseline and control scenarios reflected in the actual final rule.
The emission inventory estimates contained in Sections 3.2, 3.4, 3.5, and 3.6, for small
nonroad and marine SI engines are reported for the 50-state geographic area that comprises the
United States (including the District of Columbia). These inventories reflect the emissions from
the engines subject to the final Phase 3 standards, i.e., federal engines. As such, they exclude the
emissions from engines that are regulated by the State of California as provided for by section
209 of the Clean Air Act.
More specifically, California has been granted a waiver under the Clean Air Act to
regulate the emissions from all nonroad SI engines, except for engines with less than 175
horsepower that are used in farm and construction equipment. Therefore, these latter engines are
subject to federal regulation and are included in our 50-state inventories. By contrast, we do not
include any of the emissions from California marine SI engines in these inventories. As with
certain nonroad engine classes, the State has been granted a waiver to regulate the exhaust
emissions from all marine SI engines and evaporative emissions from outboard and personal
watercraft SI engines. That State also has indicated its intent to adopt the final Phase 3 standards
for evaporative emissions from sterndrive engines. Therefore, the 50-state inventories presented
in Sections 3.2, 3.4, 3.5, and 3.6 only reflect the emissions from small nonroad and marine SI
engines that are subject to federal regulation.
3-1
-------
Final Regulatory Impact Analysis
Section 3.3 presents a nationwide comparison of the emissions from small nonroad and
marine SI engines to those from other source categories, i.e., stationary, area, and other mobile
sources. Unlike the 50-state inventories described earlier, these inventories reflect the emissions
from all sources, whether they are separately regulated by a state government or the federal
government, e.g., all of California's small SI and marine SI engines are included.
Inventories are generally presented for the following pollutants: exhaust and evaporative
hydrocarbons reported as total hydrocarbons (THC) and volatile organic compounds (VOC),
oxides of nitrogen (NOX), particulate matter (PM25 and PM10), and carbon monoxide (CO). The
VOC category is a broader class of hydrocarbon compounds than THC that is primarily
important for air quality modeling purposes. The additional compounds that comprise this
category are reactive oxygenated species represented by aldehydes (RCHO) and alcohols
(RCOH), and less reactive species represented by methane (CH4) and ethane (CH3CH3). The PM
inventories for particle sizes of <2.5 microns or <10 microns in diameter include directly emitted
PM only, although secondary sulfates are taken into account in the air quality modeling as noted
below. Toxic pollutant inventories are also presented because the final Phase 3 requirements
will reduce hazardous air pollutants such as benzene, formaldeyde, acetaldehyde, 1,3-butadiene,
acrolein, napthalene, and 15 other compounds grouped together as poly cyclic organic matter
(POM).
Finally, none of the controlled inventory estimates include the potential uses of the
averaging, banking, and trading (ABT) program for engine manufacturers, since these are
flexibilities that would be difficult to predict and model. More information regarding these
provisions can be found in the preamble for this final rule that is published in the Federal
Register.
3.1 Overview of Small Nonroad and Marine SI Engine Emissions Inventory
Development
This section describes how the final emission inventories were modeled for the small
nonroad and marine SI engines that are affected by the Phase 3 standards. Generally, the
inventories were generated using a modified version of our NONROAD2005 model. More
specifically, we started with the most recent public version of the model, i.e., NONROAD2005a,
which was released in February 2006. A copy of that model and the accompanying technical
reports that detail of the modeling inputs (e.g., populations, activity, etc.) are available in the
docket for this final rule.1 They can also be accessed on our website at:
http ://www. epa. gov/otaq/nonrdmdl .htm.
The NONROAD2005a model was modified to incorporate new emission test data and
other improvements for this rulemaking. This special version is named NONROAD2005d. A
copy of the model and the accompanying documentation are available in the docket.2'3'4 The
inputs we used to model the effects of the Phase 3 standards are also described in Chapter 4 for
exhaust emissions and Chapter 5 for evaporative emissions. Finally, the modifications we made
to NONROAD2005a to reflect the baseline and control scenarios related to the final rule are
3-2
-------
Emission Inventory
summarized in Sections 3.2 and 3.4, respectively.
The nonroad model estimates emission inventories of important air pollutant species from
a diverse universe of nonroad equipment. The model's scope includes all off-highway sources
with the exception of locomotives, aircraft and commercial marine vessels. The model can
distinguish emissions on the basis of equipment type, horsepower, and technology group. For
the engines subject to the final rule, the nonroad model evaluates numerous equipment types
with each type containing multiple horsepower categories and technology groups. A central
feature of the model is the projection of past, present, or future emissions between 1970 and
2050.
The chemical species NOx, PM, and CO are exhaust emissions, i.e., pollutants emitted
directly as exhaust from the combustion of fuel (both liquid and gaseous fuels) in the engine.
Hydrocarbon species, e.g., THC and VOC, consist of both exhaust and evaporative emissions.
The exhaust component represents hydrocarbons emitted as products of combustion, which can
also include emissions vented from the crankcase. The evaporative hydrocarbon component
includes compounds from unburned fuel that are emitted either while the engine is being
operated or when the equipment is not in use. The various categories of evaporative emissions
that are included in the nonroad model are:
Diurnal. These emissions result from changes in temperature during the day. As the day
gets warmer there is a concomitant rise in the temperature of the liquid fuel in the fuel tank. This
causes the vapor pressure inside the tank to increase, forcing vaporized fuel to escape into the
atmosphere. For modeling purposes, this category also includes diffusion losses that come from
fuel vapor exiting the orifice of a vented fuel tank cap regardless of temperature.
Permeation. These emissions occur when fuel molecules transfuse through plastic or
rubber fuel-related components (fuel lines and fuel tanks) into the atmosphere.
Hot Soak. These emissions occur after the engine is shut off and the engine's residual
heat causes fuel vapors from the fuel tank or fuel metering device to be released into the
atmosphere.
Running Loss. Similar in form to diurnal losses, these emissions are caused from the
engine's heat during equipment operation.
Vapor Displacement or Refueling Loss. These are vapors displaced from the fuel tank
when liquid fuel is being added during a refueling event.
Liquid Spillage. This refers to the liquid fuel that is spilled when equipment is refueled
either from a portable fuel container or fuel pump, which subsequently evaporates into the
atmosphere.
Equipment fueled by compressed natural gas, liquified petroleum gas, or diesel fuel are
assumed to have zero evaporative emissions. Consequently, all evaporative emissions are from
-------
Final Regulatory Impact Analysis
gasoline or gasoline blends, i.e., ethanol and gasoline.
The control scenario analyzed in Section 3.4 reflects the final Phase 3 standards for
exhaust hydrocarbons, CO, and NOx from small nonhandheld nonroad and marine SI engines.a
New standards to control evaporative emissions from hose permeation and tank permeation from
these engine classes and handheld equipment are also included. Further, the final requirements
also establish new standards for running loss and diffusion emissions from small nonhandheld
nonroad SI engines and diurnal emissions from marine SI engines. Finally, we expect that the
technology necessary to achieve the final exhaust emission standards will indirectly lower
exhaust PM. All of these effects are reflected in the controlled emission inventories presented in
this chapter.
3.2 Baseline Emission Inventory Estimates
This section describes more specifically how we developed the baseline exhaust and
evaporative inventories for small nonroad and marine SI engines. The resulting baseline
inventories are also presented. Section 3.2.1 provides this information for exhaust and
evaporative emissions.
The inventory estimates presented throughout this section include only equipment that
would be subject to the final standards. For small nonroad SI equipment, California's Air
Resources Board (ARE) has promulgated standards that are roughly equivalent in stringency
overall to final Phase 3 federal standards, although some of the specific requirements and test
procedures are different. However, the Clean Air Act prohibits California from regulating
engines used in farm and construction equipment with maximum power levels below 175 hp or
130 kW. Therefore, the requirements contained in this final rule for small nonroad SI engines
will apply in California to the above farm and construction equipment power levels. As a result,
these engines are included in the inventories presented in this chapter. However, the majority of
the small nonroad SI equipment in California is subject to ARE regulations, so the effect of the
federal Phase 3 standards for these engines in that State is rather limited.
For marine SI engines, ARB also has its own exhaust emission standards that are roughly
equivalent overall to the final Phase 3 federal standards. In addition, ARB has stated its intend
to develop evaporative emissions standards for marine SI equipment in California. Therefore,
the exhaust and evaporative inventory estimates for marine SI engines/equipment completely
exclude California.
3.2.1 Baseline Exhaust and Evaporative Emissions Estimates for THC, VOC, NOx, PM2 5,
PM10, and CO
The baseline exhaust and evaporative emission inventories for small nonroad and marine
SI engines include the effects of all existing applicable federal emission standards. We
The CO standard applies to small nonhandheld SI engines used in auxiliary marine applications.
3-4
-------
Emission Inventory
generated these inventories by starting with the NONROAD2005a emissions model, which was
released to the public in February 2006. That model was then modified to incorporate new
emission test data and other improvements for this rulemaking. This special version of the
model is named NONROAD2005d. The modifications to the base model are described below.
3.2.1.1 Changes from NONROAD2005a to NONROAD2005d
As already mentioned, a number of improvements to the most publically available
nonroad emissions inventory model were made to develop the NONROAD2005d, which is used
in this final rulemaking. These revisions were based on recent testing programs, other
information, and model enhancements. The changes are summarized below for small nonroad
and marine SI engines. Many of the most important revisions are discussed in greater detail in
the following sections.
3.2.1.1.1 Revisions for Small SI Engines
The modifications that we made to the NONROAD2005a model for Small SI engines
that are most relevant to the final rule are summarized below:
1. Revised fuel tank and hose permeation emission factors;
2. Added new fuel tank diffusion losses to the diurnal emission estimates;
3. Updated or corrected exhaust emission factors and deterioration rates, and
technology-type sales fractions for Phase 2 engines;
4. Adjusted equipment populations to properly account for the application of federal
emission requirements to engines in California;
5. Added the ability to specifically model the effects of ethanol blends on exhaust
emissions and on fuel tank and hose permeation losses;
6. Added hot soak and running losses for handheld equipment;
7. Corrected snowblower technology types to include 4-stroke engines; and
8. Corrected running loss emission factors for Class 1 snowblowers to account for
cold weather applications.
3.2.1.1.2 Revisions for Recreation Marine SI Engines
The modifications that we made to the NONROAD2005a model for marine SI engines
that are most relevant to the final rule are summarized below:
1. Revised brake-specific fuel consumption factors;
2. Revised PM emission factors for 2-stroke technology engines;
3. Revised fuel tank and hose permeation emission factors and temperature effects;
4. Updated modeling inputs for high performance sterndrive and inboard (SD/I)
engines; and
5. Added the ability to specifically model the effects of ethanol blends on exhaust
emissions and fuel tank and hose permeation losses.
3-5
-------
Final Regulatory Impact Analysis
3.2.1.2 Baseline Exhaust Emission Calculations
3.2.1.2.1 Small SI Exhaust Calculations
We revised the Phase 2 exhaust emission factors in the NONROAD2005d inventory
model to reflect new information and our better understanding of the in-use emissions of these
engines, as discussed further below.
The nonroad model estimates exhaust emissions in a given year by applying an
appropriate emission factor based on the engines age or hours of use. This reflects the fact that
an engine's exhaust emissions performance degrades over its lifetime due to normal use or
misuse (i.e., tampering or neglect). More specifically, the emission factor is a combination of a
"zero-hour" emission level (ZHL) and a deterioration factor (DF). The ZHL represents the
emission rate for recently manufactured engines, i.e., engines with few operating hours. The DF
to the degree of emissions degradation per unit of activity. Nonroad engine activity is expressed
in terms of hours of use or fraction of its median life. This later term refers to the age at which
50 percent of the engines sold in a given year ceased to function and have been scrapped. The
following formula describes the basic form of the calculation:
EFaged = ZHLxDF
where: EFaged is the emission factor for an aged engine
ZML is the zero hour emission factor for a new engine
DF is the deterioration factor
The form of the DF for nonroad SI engines is as follows:
DF = 1 + A x (Age Factor) for Age Factor < 1
DF = 1 + A for Age Factor > 1
where: Age Factor = [Cumulative Hours x Load Factor]
Median Life at Full Load, in Hours
A, = constants for a given technology type; b < 1.
The constants A and b can be varied to approximate a wide range of deterioration
patterns. "A" can be varied to reflect differences in maximum deterioration. For example,
setting A equal to 2.0 would result in emissions at the engine's median life being three times the
emissions when new. The shape of the deterioration function is determined by the second
constant, b. This constant can be set at any level between zero and 1.0; currently, the
NONROAD model sets b equal to either 0.5 or 1.0. The first case results in a curvilinear
deterioration rate in which most of the deterioration occurs in the early part of an engine's life.
The second case results in a linear deterioration pattern in which the rate of deterioration is
constant throughout the median life of an engine. In both cases, we previously decided to cap
deterioration at the end of an engine's median life, under the assumption that an engine can only
3-6
-------
Emission Inventory
deteriorate to a certain point beyond which it becomes inoperable. For spark ignition engines at
or below 25 horsepower, which are the subject of this final rule, the nonroad model sets the
constant b equal to 0.5. The emission factor inputs for Phase 2 small nonroad SI engines used in
this analysis are shown in Table 3.2-1.
Table 3.2-1: Phase 2 Modeling Emission Factors for Small SI Engines(g/kW-hr)b
Class/
Tprhnnlnpy
Class I - SV
Class I - OHV
Class II
THCZML
10.30
8.73
5.58
THC "A"
1.753
1.753
1.095
NOxZML
2.57
3.28
3.71
NOx "A"
0.180
0.180
0.000
COZML
386.53
392.93
472.80
CO "A"
0.070
0.070
0.080
PM10
7MT,
0.35
0.05
0.08
PM10
"A"
1.753
1.753
1.095
Some of the values shown in Table 3.2-1 have been updated from the NONROAD2005a
inventory model based on data collected by EPA on in-use engines as well as
manufacturer-supplied certification data. The ZHL emission factors for Class I engines were
updated based on testing performed by EPA on 16 in-use walk-behind lawnmowers. The Class I
side-valve engine A values were revised to be the same as the Class I overhead engine A values
based on the same in-use testing of lawnmowers which showed similar in-use deterioration
characteristics between overhead valve and sidevalve Class I engines. The Class I and Class II
engine A values for CO emissions were revised to better reflect the level of deterioration seen in
both the in-use lawnmower testing noted above as well as certification data provided by
manufacturers to EPA. Finally, based on data collected from another test program of in-use
lawnmowers, the assumption that there was no deterioration of Class I and II emissions after the
median life was reached was revised to reflect further continued emissions deterioration after
that point.
Also, the model was modified to acknowledge the continued use of side-valve engine
designs in Class I nonhandheld engines meeting Phase 2 standards. In the rulemaking that
established those regulatory requirements, side-valve technology was assumed to be superceded
by overhead valve designs and was modeled accordingly. In reality, side-valve technology has
continued to be used in small nonroad SI engines. The resulting technology mixture is shown in
Table 3.2-2. The estimated sales fractions by engine class and technology are based on sales
information provided by engine manufacturers to EPA for the 2005 model year. A full
description of the emission modeling information for Phase 2 engines and the basis for the
estimates can be found in the docket for this rule.
b The nonroad model calculates VOC by multiplying THC by an adjustment factor depending on engine
and fuel type for exhaust emissions as follows: 2-stroke gasoline = 1.034; 4-stroke gasoline = 0.933; liquified
petroleum (LPG) = 0.995; and compressed natural gas (CNG) = 0.004. Crankcase and evaporative VOC for all fuels
other than CNG is assumed to be equivalent to THC. CNG fueled equipment do not emit crankcase and evaporative
VOC emissions because CNG is comprised almost exclusively of methane. PM2 5 is calculated by multiplying PM10
0.92.
3-7
-------
Final Regulatory Impact Analysis
Table 3.2-2: Phase 3 Small Nonroad SI Engine Technology Classes
Engine Class
Class I
Class I
Class II
Technology Class
Side Valve
Overhead Valve
Overhead Valve
Percent Sales (%)
60
40
100
3.2.1.2.2 Marine SI Exhaust Calculations
The NONROAD2005a model included a number of updates to the emission rates and
technology mix of marine SI engines.5 These updates were largely based on data submitted to
EPA by marine engine manufacturers as part of the certification process and on new test data
collected by EPA.6 However, NONROAD2005a did not include high-performance SD/I marine
engines. High-performance marine engines are niche product and were not included in the data
set used to develop the engine populations for the NONROAD2005a model.
Manufacturers have more recently commented that approximately 1,500 high-
performance engines are produced in the U.S. per year. These engines range from 500 to 1500
horsepower and are used in both racing and non-racing applications. Based on conversations
with individual high-performance engine manufacturers, we estimate that about two thirds of
these engines are sold for use in the U.S. with an average power of about 650 horsepower. These
engines are designed to sacrifice service life for power, but with rebuilds, generally are used for
7-8 years (we use 8 years for our modeling). Based on these estimates and the growth rate in the
NONROAD2005a model, we estimate a 1998 population of SD/I engines >600 horsepower of
7500 units. One manufacturer stated that they performed a survey on the annual use of these
engines for warranty purposes and the result was an average annual use of about 30 hours per
year. We also updated the baseline emission factors for high performance marine engines based
on the emission data presented in Chapter 4. Note that no changes were made to the PM
emission factors because no new data was available. Table 3.2-3 presents the updated emission
factors for high-performance SD/I marine engines.
Table 3.2-3: Emission Factors for High-Performance Marine Engines [g/kW-hr]
Pollutant
HC
CO
NOx
PM
BSFC
Carbureted Engines
(MS4C, Bin 12)
13.8
253
8.4
0.08
400
Fuel-Injected Engines
(MS4D, Bin 12)
13.8
207
6.8
0.08
362
-------
Emission Inventory
3.2.1.3 Baseline Evaporative Emission Calculations
Chapter 5 presents a great deal of information on evaporative emission rates from fuel
systems used in nonroad equipment. Much of this information was incorporated into the
NONROAD2005a model.7 However, we have continued to collect evaporative emission data
and incorporate the new information into our evaporative emission inventory calculations.
These updates are described below. A technical memorandum that documents the methodology
and input values for modeling the effects of ethanol blends on nonroad engine fuel hose and tank
permeation is also available in the docket for this final rulemaking.8
3.2.1.3.1 Fuel Ethanol Content
Currently, about 55 percent of fuel sold in the U.S. contains ethanol. With the recent
establishment of the Energy Policy Act of 2005,9 this percentage is expected to increase. The
significance of the use of ethanol in fuel, for the inventory calculations, is that ethanol in fuel can
affect both exhaust and evaporative emissions from nonroad equipment. The oxygen content of
the ethanol tends to make combustion mixtures leaner, which can decrease exhaust HC and CO
emissions while increasing NOx. Also, fuel blends containing ethanol typically increase the
permeation rate for most materials used in gasoline fuel systems. This is discussed in more
detail below.
Title XV, section 1501, of the Energy Policy Act requires that the total volume of
renewable fuel increase from 4.0 to 7.5 billion gallons per year from 2006 to 2012, and the
Energy Information Administration (EIA) predicts that production will actually surpass 11
billion gallons per year by then. Based on these figures and projected gasoline sales from the
Energy Information Administration,10'11 we estimate that about three-fourths of gasoline sold in
2009 and later will contain ethanol. Table 3.2-4 presents our estimates for ethanol blended fuels
into the future. The blend market shares shown in the last column of this table assume 10
percent for ethanol content of blended gasoline in all areas except California, where it is 5.7
volume percent until switching to 10 percent in 2010.
3-9
-------
Final Regulatory Impact Analysis
Table 3.2-4: Estimated Fraction of Gasoline Containing Ethanol
Calendar Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
U.S. Gasoline Sales
[109gal.]
129.9
132.0
135.6
137.0
140.1
140.2
142.2
141.7
142.5
144.1
145.4
147.1
149.0
U.S. Ethanol Sales
[109gal.]
1.6
1.8
2.1
2.8
3.5
4.0
5.5
6.1
7.6
10.0
10.9
11.0
11.2
Fraction of Gas with
Ethanol
12.5%
13.4%
15.7%
22.2%
27.3%
31.0%
41.5%
46.3%
57.5%
75.1%
78.8%
78.8%*
78.8%*
ethanol fraction projected to be constant after 2010
3.2.1.3.2 Hose Permeation
We developed hose permeation emission factors based on the permeation data and hose
requirements presented in Chapter 5. Because permeation is a function of surface area and
because hose lengths and inner diameters are defining parameters, hose permeation rates are
based on g/m2/day. These emission factors incorporate a more complete set of data than those in
the NONROAD2005a model. In addition, distinctions are now made between permeation rates
for liquid fuel versus fuel vapor exposure and between permeation rates for gasoline versus
ethanol-blend fuels. The updated hose emission factors are discussed below and presented in
Table 3.2-5.
Fuel hoses in small nonroad SI applications vary greatly in construction depending on the
individual specifications of the engine and equipment manufacturers. However most fuel hose
used on non-handheld equipment meets the SAE J30 R7 hose requirements which includes a
permeation requirement of 550 g/m2/day on Fuel C at 23°C.12 Chapter 5 presents data on several
hose constructions that range from 190 to 450 g/m2/day on Fuel C. As discussed in Chapter 5,
permeation is typically lower on gasoline than on Fuel C. At the same time, blending ethanol
into the fuel increases permeation. Based on data presented in Chapter 5, we estimate that non-
handheld fuel hose permeation rates range from 27 to 180 g/m2/day on gasoline and 80-309
g/m2/day on gasoline blended with 10 percent ethanol (E10). Of the data presented in Chapter 5,
the lowest two permeation rates for SAE J30 R7 hose were from an unknown fuel hose
construction and from a hose (used in some small nonroadSI applications) that was specially
constructed of fuel resistant materials to facilitate painting. Dropping the unknown hose
construction (which is not known to be used in Small SI applications), we get average
permeation rates of 122 g/m2/day on gasoline and 222 g/m2/day on E10 at 23°C.
3-10
-------
Emission Inventory
Chapter 5 also presents permeation data on fuel lines used in handheld equipment tested
using E10 fuel. Based on this data, we estimate an average permeation rate at 23°C, on fuel
containing 10 percent ethanol of 255 g/m2/day. To determine an emission factor for handheld
fuel lines on gasoline, we used the ratio of permeation rates for NBR rubber samples on E10
versus gasoline. The resulting permeation rate for handheld hose on gasoline was estimated to
bel40g/m2/dayat23°C.
Fuel hose for portable marine fuel tanks is not subject to any established recommended
practice. For this reason, we consider fuel hose used on portable marine fuel tanks to be
equivalent to the hose used in Small SI applications. The supply hose for each portable marine
fuel tank is modeled to include a primer bulb with the same permeation rate as the hose.
Recommended practices for marine hose on SD/I vessels include a permeation rate of
100 g/m2/day on Fuel C and 300 g/m2/day on fuel CM15 (15 percent methanol).13'14
Accordingly, these vessels have fuel hose with lower permeation. Rather than using the
recommended permeation rate limits for this hose, we base the permeation emission factors for
this hose on the data presented in Chapter 5 on gasoline with ethanol which is more
representative of in-use fuels. Chapter 5 also includes data on commercially available low
permeation fuel hose which is used by some manufacturers. However, we do not include this in
the baseline emission factor calculation because its use is primarily in anticipation of upcoming
permeation standards and would therefore not be expected to remain in the baseline without
enactment of this final rule.
For other vessels with installed fuel tanks (OB and PWC), we based the permeation
emission factors on the test data in Chapter 5 on marine hose not certified to Coast Guard Class I
requirements.
The Coast Guard specifications for fill neck hose call for a permeation limit of 300
g/m2/day on Fuel C and 600 on Fuel CM15. However, fill neck hose are not usually exposed to
liquid fuel. Therefore, we used the vapor line data presented in Chapter 5 for both fill neck and
vent line permeation rates. Hose permeation rates for both gasoline and E10 are presented in
Table3.2.-5.
Table 3.2-5: Hose Permeation Emission Factors at 23"C [g/m2/day]
Hose Type
Handheld equipment fuel hose
Non-handheld equipment fuel hose
Portable fuel tank supply hose*
Installed system OB/PWC fuel lines
Installed system SD/I fuel lines
Fill necks and vent lines (vapor exposure)
Gasoline
140
122
122
42
22
2.5
E10
255
222
222
125
40
4.9
* this permeation rate is used for primer bulbs as well
3-11
-------
Final Regulatory Impact Analysis
The above permeation rates do not include any effects of deterioration. Over time, the
fuel can draw some of the plasticizers out of the rubber in the hose, making it more brittle and
subject to cracking. This is especially true for higher permeation fuel hoses which are generally
less fuel resistant. Exposure to ozone over time can also deteriorate the hose. This deterioration
would presumably increase the permeation rate over time. However, we do not have any data to
quantify this effect and are not including deterioration in this analysis at this time. Lower
permeation fuel hose, such as that designed to meet the final standard would likely have much
lower deterioration due to the use of more fuel resistant materials. Therefore, this analysis may
underestimate the inventory and benefits associated with the final fuel permeation standards.
3.2.1.3.3 Hose Lengths
The hose lengths used in NONROAD2005a are based primarily on confidential
information supplied by equipment manufacturers. Hose lengths for handheld equipment are
based on survey data provided by the Outdoor Power Equipment Institute.15 Recently, we
received comment from a boatbuilder using outboard motors that the hose lengths in our
calculations were too short.16 Because our existing data set did not include outboard boats with
installed fuel tanks, we updated the hose lengths for these vessels based on the data supplied by
this boat builder. In addition, the vent line lengths in the NONROAD2005a were divided by two
to account for a vapor gradient throughout the fuel line caused by diurnal breathing and
diffusion. This factor has been removed in lieu of the new emission factors for vent lines based
on vapor exposure. Table 3.2-6 presents the updated hose lengths for outboard boats with
installed fuel tanks.
Table 3.2-6: Updated Hose Lengths for Outboard Boats with Installed Fuel Systems
Engine Power
Category
18.7-29.8 kW
29.9-37.3 kW
37.4.74.6 kW
74.7-130.5 kW
130.6+ kW
Fill Neck
Length [m]
1.8
2.4
3.1
3.7
4.3
Fuel Supply Hose
Length [m]
1.8
2.4
3.1
3.7
4.3
Vent Hose
Length [m]
1.5
1.8
2.1
2.4
2.7
3.2.1.3.4 Tank Permeation
For fuel tanks, the NONROAD2005a model does not include a fuel ethanol effect on
permeation. Data in Chapter 5 suggest that even polyethylene fuel tanks see a small increase in
permeation on E10 compared to gasoline. This increase is much larger for nylon fuel tanks like
those used in handheld equipment with structurally-integrated fuel tanks. Table 3.2-7 presents
the updated emission factors on E10 fuel and compares them to the emission factors based on
gasoline permeation rates. The primary difference between the permeation rates for installed
marine tanks, compared to smaller HDPE fuel tanks, is largely due to the wall thickness of the
different constructions rather than material permeation properties. Permeation rate is a function
3-12
-------
Emission Inventory
of wall thickness, so as tank thickness doubles, permeation rate halves. The model considers
permeation from metal fuel tanks to be zero.
Table 3.2-7: Tank Permeation Emission Factors at 29°C [g/mVday]
Tank Type
Nylon handheld fuel tanks
Small SI HOPE <0.25 gallons
Small SI HOPE >0.25 gallons
Portable and PWC HOPE fuel tanks
Installed non-metal marine fuel tanks
Metal tanks
Gasoline
1.25
6.5
9.7
9.9
8.0
0
E10
2.5
7.2
10.7
10.9
8.8
0
3.2.1.3.5 Diffusion
The NONROAD2005a model includes an adjustment factor to diurnal emissions to
account for diffusion. The data used to create this adjustment factor is included in Chapter 5.
This adjustment factor is applied to all small nonroad SI equipment in the NONROAD2005a
model. However, we believe that handheld equipment are all produced with either sealed fuel
tanks or slosh/spill resistant fuel caps. Therefore, we do not include diffusion emissions for
handheld equipment in this analysis.
3.2.1.3.6 Modeling of Nonlinear Ethanol Blend Permeation Effects
Based on the limited available test data it appears that the effect of alcohol-gasoline
blends on permeation is nonlinear, tending to increase permeation at lower alcohol
concentrations up to about 20 percent ethanol, but then decreasing permeation at higher alcohol
concentrations.
17
Starting with the zero and 10 percent ethanol points described above, a simple
exponential curve was selected to connect the zero and 10 percent points continuing up to the 20
percent ethanol level. Then to get a nonlinear decreasing curve above 20 percent a simple
decreasing exponential curve was used. Since effects above 85 percent are especially uncertain,
and no such fuels are foreseen for use in nonroad equipment, the effect above 85 percent was set
equal to the E85 effect. The equations used are shown here, and an example curve based on
these equations is shown in Figure 3.2-1.
3-13
-------
Final Regulatory Impact Analysis
Hose and Tank Permeation for zero to 20 percent ethanol volume percent:
Permeation EF = GasEF + GasEF x (ElOfac - 1) x [ (EthVfrac / 0.10) A 0.4 ]
Hose and Tank Permeation for ethanol volume percent greater than 20 percent:
Permeation EF = GasEF x { 1 + (ElOfac - 1) x [ (20 / 10) A 0.4 ] }
x { 1 - [ (MIN(EthVfrac, 0.85) - 0.20 ) / 0.80 ] A (1 / 0.4 ) }
where:
Permeation EF = Permeation emission factor for modeled fuel (grams per meter2 per day)
GasEF = Gasoline hose permeation emission factor from input EF data files (grams per
meter2 per day)
ElOfac = permeation emission adjustment factor for E10 relative to gasoline. This is the
ratio of the E10 to gasoline permeation emission factors (unitless)
EthVfrac = Volume fraction ethanol in the fuel being modeled. El0 = 0.10
0.4 = exponent chosen to yield a reasonable shape of curve.
Figure 3.2-1: Ethanol Blend Hose Permeation Example Curve
300.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Ethanol Volume Fraction
It should be noted that all ethanol blends currently modeled with the NONROAD model
are less than or equal to E10, so no parts of this curve above E10 are used. Also, the value of
ElOfac used in the modeling of the control case is 2.0 for all the tank and hose permeation
sources listed above in Tables 3.2-6 and 3.2-7.
3-14
-------
Emission Inventory
3.2.1.3.7 Modeling Effect ofEthanol Blend Market Share on Permeation
The effect of ethanol blend market share is modeled linearly. In most areas the ethanol
blend market share is either zero or 100 percent, but in areas where it is between those two
market shares, or when doing a nationwide model run, the effect is calculated as a simple
proportion. For instance a 30 percent market share of E10 would be modeled using a permeation
rate 30 percent of the way between the EO permeation rate and the E10 permeation rate.
3.2.1.4 Baseline Exhaust and Evaporative Inventory Results for THC, VOC, NOx,
PM2 5, PM10, and CO
Table 3.2-8 presents the 50-state baseline emission inventories, respectively, for small
nonroad SI engines. Table 3.2-9 provides the same information for marine SI engines.
3-15
-------
Final Regulatory Impact Analysis
Table 3.2-8: Baseline 50-State Annual Exhaust and Evaporative Emissions for
Small Nonroad Spark-Ignition Engines (short tons)
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
2036
2037
2038
2039
2040
THC
1,064,625
1,026,922
963,709
886,524
825,413
768,239
718,564
682,088
665,762
663,620
665,666
671,328
679,634
689,045
699,225
709,899
720,944
732,284
743,755
755,254
766,782
778,333
789,900
801,493
813,217
824,971
836,736
848,508
860,287
872,069
883,856
895,645
907,437
919,232
931,029
942,828
954,629
966,431
978,235
voc
1,047,374
1,009,822
945,601
867,081
804,926
747,552
700,010
665,890
650,954
649,358
651,743
657,539
665,819
675,131
685,164
695,658
706,498
717,615
728,853
740,118
751,412
762,727
774,056
785,411
796,890
808,397
819,915
831,439
842,970
854,504
866,042
877,583
889,126
900,672
912,219
923,769
935,321
946,874
958,429
NOx
106,804
106,852
106,610
106,847
109,233
109,439
111,235
116,329
118,376
119,424
120,820
122,506
124,400
126,395
128,468
130,573
132,701
134,846
137,002
139,160
141,317
143,475
145,636
147,806
150,003
152,209
154,418
156,628
158,840
161,053
163,266
165,479
167,692
169,906
172,119
174,332
176,546
178,759
180,973
PM2.5
23,382
23,480
23,483
23,417
23,498
23,804
24,335
24,882
25,402
25,888
26,364
26,832
27,291
27,747
28,202
28,655
29,107
29,558
30,009
30,460
30,911
31,362
31,813
32,265
32,718
33,173
33,627
34,081
34,535
34,990
35,444
35,898
36,353
36,807
37,261
37,716
38,170
38,625
39,079
PM10
25,416
25,522
25,525
25,453
25,541
25,874
26,451
27,045
27,611
28,139
28,657
29,165
29,664
30,160
30,654
31,146
31,638
32,128
32,618
33,109
33,599
34,089
34,579
35,070
35,563
36,057
36,551
37,045
37,538
38,032
38,526
39,020
39,514
40,008
40,502
40,995
41,489
41,983
42,477
CO
15,091,835
14,351,829
13,690,337
12,923,819
12,252,479
11,711,607
10,861,441
9,992,801
9,623,727
9,568,610
9,579,040
9,644,512
9,751,728
9,879,027
10,020,040
10,169,185
10,324,079
10,483,706
10,645,870
10,808,929
10,972,659
11,136,954
11,301,731
11,467,292
11,634,934
11,803,402
11,972,207
12,141,251
12,310,505
12,479,899
12,649,385
12,818,940
12,988,554
13,158,228
13,327,954
13,497,732
13,667,556
13,837,424
14,007,335
3-16
-------
Emission Inventory
Table 3.2-9: Baseline 50-State Annual Exhaust and Evaporative Emissions for
Marine Spark-Ignition Engines (Short Tons)
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
2036
2037
2038
2039
2040
THC
906,318
873,287
836,493
796,279
756,781
717,924
680,702
645,730
612,180
580,750
553,441
530,682
511,166
495,178
481,650
470,667
462,602
455,864
451,176
447,624
445,371
443,951
443,203
443,196
443,770
444,806
446,152
447,768
449,626
451,666
453,836
456,159
458,592
461,115
463,691
466,323
468,999
471,706
474,437
voc
931,132
896,969
858,916
817,340
776,480
736,303
697,795
661,588
626,901
594,415
566,191
542,671
522,503
505,981
492,000
480,648
472,308
465,336
460,481
456,799
454,457
452,972
452,179
452,151
452,720
453,764
455,126
456,766
458,656
460,733
462,943
465,310
467,790
470,363
472,990
475,674
478,403
481,164
483,949
NOx
46,311
49,694
53,397
57,862
63,366
67,730
73,894
82,123
87,140
90,516
93,662
96,528
99,197
101,703
104,022
106,158
108,084
109,885
111,525
113,063
114,479
115,802
117,052
118,228
119,344
120,401
121,411
122,387
123,335
124,260
125,166
126,052
126,922
127,777
128,621
129,454
130,278
131,095
131,907
PM2.5
15,092
14,417
13,679
12,886
12,090
11,311
10,553
9,824
9,149
8,525
7,983
7,534
7,144
6,823
6,549
6,324
6,156
6,012
5,908
5,826
5,768
5,726
5,696
5,680
5,675
5,678
5,687
5,701
5,719
5,741
5,765
5,792
5,821
5,851
5,883
5,915
5,948
5,982
6,016
PM10
16,404
15,670
14,869
14,007
13,142
12,295
11,470
10,678
9,945
9,266
8,678
8,189
7,766
7,416
7,118
6,874
6,691
6,535
6,422
6,333
6,270
6,224
6,191
6,174
6,168
6,172
6,182
6,197
6,217
6,240
6,266
6,296
6,327
6,360
6,394
6,429
6,465
6,502
6,539
CO
2,472,251
2,407,992
2,346,538
2,266,733
2,170,374
2,103,059
2,007,804
1,885,970
1,823,844
1,788,830
1,758,115
1,732,653
1,710,005
1,690,755
1,673,978
1,660,415
1,650,631
1,642,841
1,638,114
1,635,047
1,634,065
1,634,672
1,636,603
1,639,914
1,644,492
1,650,159
1,656,748
1,663,933
1,671,627
1,679,753
1,688,228
1,697,074
1,706,229
1,715,675
1,725,343
1,735,210
1,745,240
1,755,400
1,765,651
3-17
-------
Final Regulatory Impact Analysis
3.2.2 Baseline Hazardous Air Pollutant Estimates
The analysis of toxic air pollutants from small nonroad and marine SI engines focuses on
seven major pollutants: benzene, formaldehyde, acetaldehyde, 1,3-butadiene, acrolein,
naphthalene, and 15 other compounds grouped together as poly cyclic organic matter (POM) for
this analysis.0 All of these compounds, except acetaldehyde, were identified as national or
regional cancer or noncancer "risk" drivers in the 1999 National Scale Air Toxics Assessment
(NATA)18 and have significant inventory contributions from mobile sources. That is, for a
significant portion of the population, these compounds pose a significant portion of the total
cancer or noncancer risk from breathing outdoor air toxics. The health effects of these hazardous
pollutants are specifically discussed in Chapter 2. Many of these compounds are also part of the
THC and VOC inventories. An exception is formaldehyde, which is not measured by the
analytic technique used to measure THC, and part of the mass of other aldehydes as well.
However, all are included in the VOC inventories presented in this chapter.
The baseline inventories for each of the toxic air pollutants described above were
developed using EPA's National Mobile Inventory Model (NMIM). This model is an analytical
framework that links a county-level database to our NONROAD model and collates the output
into a single database table. The resulting estimates for small nonroad and marine SI engines
account for local differences in fuel characteristics and temperatures on a much finer scale than
is possible when running the standalone NONROAD model. Emissions were modeled for all of
the small nonroad and marine SI equipment categories in the continental United States, including
California. Hence, the hazardous emission inventories presented here include the emissions
from engines that are regulated by that State. As a result, the emission inventories are slightly
overstated relative to the those from the small nonroad and marine SI engines subject to the
federal Phase 3 requirements.
Table 3.2-10 presents the 50-state baseline inventories for toxic air emissions from small
nonroad SI engines in 2002, 2020, 2030. Table 3.2-11 provides the same information for marine
SI engines.
Table 3.2-10: Baseline 50-State Air Toxic Emissions for
Small Nonroad Spark-Ignition Engines (short tons)
Year
2002
2020
2030
Benzene
35,086
23,216
26,776
1,3
Butadiene
5,561
3,468
3,999
Formalde-
hyde
8,664
5,802
6,691
Acetalde-
hyde
2,900
2,792
3,217
Acrolein
505
291
336
Naphthalene
447
638
739
POM
97
131
152
0 The 15 POMs summarized in this chapter are acenaphthylene, anthracene, benz(a)anthracene,
benzo(a)pyrene, benzo(b)fluoranthene, benzo(g,h,i)perylene, beno(k)fluoranthene, chrysene, dibenzo(a,h)anthracene,
fluoranthene, fluorene, ideno(l,2,3-c,d)pyrene, phenanthrene, andpyrene.
3-18
-------
Emission Inventory
Table 3.2-11: Baseline 50-State Air Toxic Emissions for
Marine Spark-Ignition Engines (short tons)
Year
2002
2020
2030
Benzene
23,110
8,417
8,476
1,3
Butadiene
2,053
763
751
Formalde-
hyde
2,153
535
525
Acetalde-
hyde
1,543
818
811
Acrolein
211
27
27
Naphthalene
37
42
46
POM
31
16
16
3.3 Contribution of Small Nonroad and Marine SI Engines to National
Emissions Inventories
This section describes the nationwide contribution of small nonroad and marine SI
engines to the emissions of other source categories. Information is presented for the pollutants
that are directly controlled by the final standards, i.e., VOC, NOX, and CO, and those that are
indirectly reduced by some of the requisite control technology, i.e., PM25 and PM10. The VOC
inventories includes both exhaust and evaporative hydrocarbon emissions.
The national inventories are presented for 2002, 2020, and 2030 for all 50-states and the
District of Columbia.19 The stationary, highway, and aircraft inventories were taken directly
from EPA's National Emissions Inventory (NEI) modeling platform for 2002.20 The emission
inventories for locomotives, and Classes 1 and 2 commercial marine engines were taken from
our recent final rule for these mobile source categories.21 The inventories for Class 3
commercial marine engines was take from our recent advance notice of final rulemaking for that
category of engines.22 The emission estimates for portable fuel containers was taken from the
recent final MS AT rule.23 All of the land-based nonroad engine emission inventories was
developed using the NONROAD2005d model, as previously described.d Finally, these
inventories account for the future use of renewable fuels as required by the Energy Policy Act of
2005.
3.3.1 VOC Emissions Contribution
Table 3.3-1 provides the contribution of small nonroad SI engines, marine SI engines and
other source categories to nationwide VOC emissions. The emissions from small nonroad
(<19kW) and marine SI engines are 26 percent of the mobile source inventory and 12 percent of
the total manmade VOC emissions in 2002. These percentages decrease slightly to 23 percent
and 10 percent, respectively, by 2030.
d The modeling inputs for diesel nonroad engines that were used in NONROAD2005d for this final rule
are the same as those contained in the NONROAD2005a public.
3-19
-------
Final Regulatory Impact Analysis
3.3.2 NOx Emissions Contribution
Table 3.3-2 provides the contribution of nonroad small nonroad SI engines, marine SI
engines and other source categories to nationwide NOx emissions. The emissions from small
nonroad and marine SI engines are about 1 percent of the mobile source inventory and about 1
percent of the total manmade NOx emissions in 2002. These percentages increase to 6 percent
and 3 percent, respectively, by 2030.
3.3.3 PM Emissions Contribution
Table 3.3-3 and 3.3-4 provide the contribution of small nonroad SI engines, marine SI
engines and other source categories to nationwide PM2 5 and PM10 emissions, respectively. Both
particle size categories from small nonroad and marine SI engines are about 8 percent of the
mobile source inventory and approximately 1 percent of the total manmade PM25 emissions in
2002. The mobile source percentage increases to about 12 percent and the total manmade
percentage stay about the same at 1 percent by 2030.
3.3.4 CO Emissions Contribution
Table 3.3-5 provides the contribution of small nonroad SI engines, marine SI engines,
and other source categories to nationwide CO emissions. The emissions from small nonroad and
marine SI engines are 23 percent of the mobile source inventory and 10 percent of the total
manmade CO emissions in 2002. These percentages decrease to 30 percent and increase to 25
percent, respectively, by 2030.
3-20
-------
Table 3.3-1: 50-State Annual VOC Baseline Emission Levels for
Mobile and Other Source Categories
Category
Small Nonroad SI total
A Tsnuan -Hi "\-i Small SI
Handheld S"iall SI
Recreational Marine SI
Locomotive
Recreational Marine Diesel
Commercial Marine (Cl & C2)
Land-Based Nonroad Diesel
Commercial Marine (C3)
SI Recreational Vehicles
Large Nonroad SI (>25hp)
Portable Fuel Containers
Aircraft
Total Off Highway
Highway Diesel
Highway Non-Diesel
Total Highway
Total Diesel (distillate) Mobile
Total Mobile Sources
Stationary Point and Area Sources
Total Man-Made Sources
2002
short tons
1,165,257
675,311
489,946
1,003,325
50,665
1,538
17,229
184,868
26,175
551,285
134,950
243,994
52,651
3,431,938
191,514
4,684,391
4,875,904
445,814
8,307,843
9,433,356
17,741,198
%of
mobile
14.0%
8.1%
5.9%
12.1%
0.6%
0.0%
0.2%
2.2%
0.3%
6.6%
1.6%
2.9%
0.6%
41.3%
2.3%
56.4%
58.7%
5.4%
100%
-
-
% of total
6.6%
3.8%
2.8%
5.7%
0.3%
0.0%
0.1%
1.0%
0.1%
3.1%
0.8%
1.4%
0.3%
19.3%
1.1%
26.4%
27.5%
2.5%
46.8%
53.2%
100%
2020
short tons
808, 8 73
590,186
218,688
496,183
27,974
2,485
11,478
76,817
53,204
442,121
11,957
38,185
63,251
2,032,530
129,321
1,973,180
2,102,501
248,075
4,135,030
8,740,057
12,875,088
%of
mobile
19.6%
14.3%
5.3%
12.0%
0.7%
0.1%
0.3%
1.9%
1.3%
10.7%
0.3%
0.9%
1.5%
49.2%
3.1%
47.7%
50.8%
6.0%
100%
-
-
% of total
6.3%
4.6%
1.7%
3.9%
0.2%
0.0%
0.1%
0.6%
0.4%
3.4%
0.1%
0.3%
0.5%
15.8%
1.0%
15.3%
16.3%
1.9%
32.1%
67.9%
100%
2030
short tons
935,619
684,057
251,562
494,217
17,722
2,912
6,911
63,342
79,697
382,468
9,953
43,375
67,730
2,103,947
140,959
1,800,856
1,941,815
231,847
4,045,762
8,740,057
12,785,819
%of
mobile
23.;%
16.9%
6.2%
12.2%
0.4%
0.1%
0.2%
1.6%
2.0%
9.5%
0.2%
1.1%
1.7%
52.0%
3.5%
44.5%
48.0%
5.7%
100%
-
-
% of total
7.3%
5.4%
2.0%
3.9%
0.1%
0.0%
0.1%
0.5%
0.6%
3.0%
0.1%
0.3%
0.5%
16.5%
1.1%
14.1%
15.2%
1.8%
31.6%
68.4%
100%
-------
Table 3.3-2: 50-State Annual NOx Baseline Emission Levels
for Mobile and Other Source Categories
Category
Small Nonroad SI
AT,™ U,™ ,«,„;,; f™,,;; or
Hctn^h"!"1 ^mall e/
Recreational Marine SI
Locomotive
Recreational Marine Diesel
Commercial Marine (Cl & C2)
Land-Based Nonroad Diesel
Commercial Marine (C3)
SI Recreational Vehicles
Large Nonroad SI (>25hp)
Aircraft
Total Off Highway
Highway Diesel
Highway Non-Diesel
Total Highway
Total Diesel (distillate) Mobile
Total Mobile Sources
Stationary Point and Area Sources
Total Man-Made Sources
short tons
119,833
116,743
3,090
49,902
1,118,786
40,437
834,025
1,555,812
745,224
10,614
336,292
103,591
4,914,515
3,529,046
4,293,733
7,822,779
7,078,105
12,737,294
8,613,718
21,351,012
2002
%of
mobile
source
0.9%
0.9%
0.0%
0.4%
8.8%
0.3%
6.5%
12.2%
5.9%
0.1%
2.6%
0.8%
38.6%
27.7%
33.7%
61.4%
55.6%
100%
-
-
% of total
0.6%
0.5%
0.0%
0.2%
5.2%
0.2%
3.9%
7.3%
3.5%
0.0%
1.6%
0.5%
23.0%
16.5%
20.1%
36.6%
33.2%
59.7%
40.3%
100%
short tons
753,872
148,004
5,867
120,172
669,405
43,579
499,798
683,481
1,368,420
30,108
48,270
132,278
3,749,382
681,142
1,270,269
1,951,411
2,577,404
5,700,793
5,773,927
11,474,721
2020
%of
mobile
source
2.7%
2.6%
0.1%
2.1%
11.7%
0.76%
8.77%
12.0%
24.00%
0.5%
0.8%
2.32%
65.8%
11.9%
22.3%
34.2%
45.2%
100%
-
-
% of total
1.3%
1.3%
0.1%
1.0%
5.8%
0.38%
4.36%
5.96%
11.93%
0.3%
0.42%
1.15%
32.7%
5.9%
11.1%
17.0%
22.5%
49.7%
50.3%
100%
short tons
178,406
171,650
6,756
132,898
437,245
43,665
308,614
435,774
2,023,974
34,318
47,766
143,986
3,786,645
355,817
1,144,199
1,500,016
1,581,115
5,286,661
5,773,927
11,060,589
2030
%of
mobile
source
3.4%
3.2%
0.1%
2.5%
8.3%
0.8%
5.8%
8.2%
38.3%
0.6%
0.9%
2.7%
71.6%
6.7%
21.6%
28.4%
29.9%
100%
-
-
% of total
1.6%
1.6%
0.1%
1.2%
4.0%
0.4%
2.8%
3.9%
18.3%
0.3%
0.4%
1.3%
34.2%
3.2%
10.3%
13.6%
14.3%
47.8%
52.2%
100%
-------
Table 3.3-3: 50-State Annual PM2 5 Baseline Emission Levels
for Mobile and Other Source Categories
Category
Small Nonfood SI
^onHan^h"!^ ^mcdl e/
Hctn^h"!"1 ^mall e/
Recreational Marine SI
Locomotive
Recreational Marine Diesel
Commercial Marine (Cl & C2)
Land-Based Nonroad Diesel
Commercial Marine (C3)
SI Recreational Vehicles
Large Nonroad SI (>25hp)
Aircraft
Total Off Highway
Highway Diesel
Highway non-diesel
Total Highway
Total Diesel (distillate) Mobile
Total Mobile Sources
Stationary Point and Area Sources
Total Man-Made Sources
short tons
25, 700
4,841
20,859
16,262
29,660
1,096
28,730
159,111
54,667
13,710
1,652
17,979
348,568
94,982
51,694
146,676
313,581
495,245
3,025,244
3,520,488
2002
%of
mobile
source
5.2%
1.0%
4.2%
3.3%
5.99%
0.22%
5.80%
32.1%
11.04%
2.8%
0.3%
3.63%
70.4%
19.2%
10.4%
29.6%
63.3%
100%
-
-
% of total
0.7%
0.1%
0.6%
0.5%
0.84%
0.03%
0.82%
4.52%
1.55%
0.4%
0.05%
0.51%
9.9%
2.7%
1.5%
4.2%
8.9%
14.1%
85.9%
100%
short tons
32,905
6,957
25,947
6,367
15,145
973
15,787
46,056
110,993
11,901
2,421
22,176
264,722
20,145
45,329
65,474
98,106
330,196
3,047,714
3,377,911
2020
%of
mobile
source
10.0%
2.1%
7.9%
1.9%
4.59%
0.29%
4.78%
13.9%
33.61%
3.6%
0.7%
6.72%
80.2%
6.1%
13.7%
19.8%
29.7%
100%
-
-
% of total
7.0%
0.2%
0.8%
0.2%
0.45%
0.03%
0.47%
1.36%
3.29%
0.4%
0.07%
0.66%
7.8%
0.6%
1.3%
1.9%
2.9%
9.8%
90.2%
100%
short tons
37,878
8,065
29,813
6,163
8,584
1,053
10,017
17,902
166,161
10,090
2,844
24,058
284,749
18,802
51,621
70,423
56,358
355,172
3,047,714
3,402,887
2030
%of
mobile
source
10.7%
2.3%
8.4%
1.7%
2.42%
0.30%
2.82%
5.0%
46.78%
2.8%
0.8%
6.77%
80.2%
5.3%
14.5%
19.8%
15.9%
100%
-
-
% of total
1.1%
0.2%
0.9%
0.2%
0.25%
0.03%
0.29%
0.53%
4.88%
0.3%
0.08%
0.71%
8.4%
0.6%
1.5%
2.1%
1.7%
10.4%
89.6%
100%
-------
Table 3.3-4: 50-State Annual PM10 Baseline Emission Levels
for Mobile and Other Source Categories
Category
Small Nonfood SI
^onHan^h"!^ ^mcdl e/
Hctn^h"!"1 ^mall e/
Recreational Marine SI
Locomotive
Recreational Marine Diesel
Commercial Marine (Cl & C2)
Land-Based Nonroad Diesel
Commercial Marine (C3)
SI Recreational Vehicles
Large Nonroad SI (>25hp)
Aircraft
Total Off Highway
Highway Diesel
Highway non-diesel
Total Highway
Total Diesel (distillate) Mobile
Total Mobile Sources
Stationary Point and Area Sources
Total Man-Made Sources
short tons
27,935
5,262
22,673
17,676
30,578
1,130
29,619
164,032
59,409
14,902
1,672
24,622
371,575
109,097
92,531
201,628
334,456
573,203
3,158,011
3,731,215
2002
%of
mobile
source
4.9%
0.9%
4.0%
3.1%
5.33%
0.20%
5.17%
28.6%
10.36%
2.6%
0.3%
4.30%
64.8%
19.0%
16.1%
35.2%
58.3%
100%
-
-
% of total
0.7%
0.1%
0.6%
0.5%
0.82%
0.03%
0.79%
4.40%
1.59%
0.4%
0.04%
0.66%
10.0%
2.9%
2.5%
5.4%
9.0%
15.4%
84.6%
100%
short tons
35, 766
7,562
28,204
6,920
15,613
1,003
16,275
47,480
120,617
12,936
2,441
30,211
289,263
32,733
96,380
129,113
113,105
418,376
3,194,610
3,612,986
2020
%of
mobile
source
8.5%
1.8%
6.7%
1.7%
3.73%
0.24%
3.89%
11.3%
28.83%
3.1%
0.6%
7.22%
69.1%
7.8%
23.0%
30.9%
27.0%
100%
-
-
% of total
1.0%
0.2%
0.8%
0.2%
0.43%
0.03%
0.45%
1.31%
3.34%
0.4%
0.07%
0.84%
8.0%
0.9%
2.7%
3.6%
3.1%
11.6%
88.4%
100%
short tons
41,172
8,767
32,406
6,699
8,849
1,086
10,327
18,455
180,566
10,967
2,866
32,714
313,702
34,746
110,796
145,542
73,464
459,244
3,194,610
3,653,854
2030
% of mobile
source
9.0%
1.9%
7.1%
1.5%
1.93%
0.24%
2.25%
4.0%
39.32%
2.4%
0.6%
7.12%
68.3%
7.6%
24.1%
31.7%
16.0%
100%
-
-
% of total
1.1%
0.2%
0.9%
0.2%
0.24%
0.03%
0.28%
0.51%
4.94%
0.3%
0.08%
0.90%
8.6%
1.0%
3.0%
4.0%
2.0%
12.6%
87.4%
100%
-------
Table 3.3-5: 50-State Annual CO Baseline Emission Levels
for Mobile and Other Source Categories
Category
Small Nonfood SI
Ar,™zj,™,«,,,;,7 e™,,;; er
fjan^h"^ ^mall e/
Recreational Marine SI
Locomotive
Recreational Marine Diesel
Commercial Marine (Cl & C2)
Land-Based Nonroad Diesel
Commercial Marine (C3)
SI Recreational Vehicles
Large Nonroad SI (>25hp)
Aircraft
Total Off Highway
Highway Diesel
Highway non-diesel
Total Highway
Total Diesel (distillate) Mobile
Total Mobile Sources
Stationary Point and Area Sources
Total Man-Made Sources
short tons
16,943,267
15,884,854
1,058,413
2,663,932
123,210
6,467
151,331
860,257
59,515
1,741,702
1,801,104
557,820
24,908,605
940,898
59,178,847
60,119,745
2,082,164
85,028,351
11,354,201
96,382,552
2002
%of
mobile
source
19.9%
18.7%
1.2%
3.1%
0.1%
0.0%
0.2%
1.0%
0.1%
2.0%
2.1%
0.7%
29.3%
1.1%
69.6%
70.7%
2.4%
100%
-
-
% of total
17.6%
16.5%
1.1%
2.8%
0.1%
0.0%
0.2%
0.9%
0.1%
1.8%
1.9%
0.6%
25.8%
1.0%
61.4%
62.4%
2.2%
88.2%
11.8%
100%
short tons
11,934,654
11,084,472
850,181
1,765,122
167,488
9,374
139,712
310,250
120,889
1,910,030
289,382
677,930
17,324,829
260,238
29,211,716
29,471,955
887,061
46,796,783
11,049,239
57,846,022
2020
%of
mobile
source
25.5%
23.7%
1.8%
3.8%
0.4%
0.0%
0.3%
0.7%
0.3%
4.1%
0.6%
1.4%
37.0%
0.6%
62.4%
63.0%
1.9%
100%
-
-
% of total
20.6%
19.2%
1.5%
3.1%
0.3%
0.0%
0.2%
0.5%
0.2%
3.3%
0.5%
1.2%
29.9%
0.4%
50.5%
50.9%
1.5%
80.9%
19.1%
100%
short tons
13,803,668
12,826,965
976,703
1,801,234
195,882
10,930
143,791
155,576
181,032
1,916,102
270,827
723,842
19,202,883
219,594
32,038,635
32,258,229
725,772
51,461,112
11,049,239
62,510,350
2030
% of mobile
source
26. 8%
24.9%
1.9%
3.5%
0.4%
0.0%
0.3%
0.3%
0.4%
3.7%
0.5%
1.4%
37.3%
0.4%
62.3%
62.7%
1.4%
100%
-
-
% of total
22.1%
20.5%
1.6%
2.9%
0.3%
0.0%
0.2%
0.2%
0.3%
3.1%
0.4%
1.2%
30.7%
0.4%
51.3%
51.6%
1.2%
82.3%
17.7%
100%
-------
Final Regulatory Impact Analysis
3.4 Controlled Nonroad Small Spark-Ignition and Marine Engine Emission
Inventory Development
This section describes how the controlled emission inventories were developed for the
small nonroad and marine SI engines that are subject to the Phase 3 standards. The resulting
controlled emission inventories are also presented. Section 3.4.1 provides this information for
exhaust and evaporative emissions.
Once again, the inventory estimates presented throughout this section only include
equipment that would be subject to the final standards. Specifically for California, this includes
small nonroad SI engines used in farm and construction equipment with maximum power levels
below 175 hp or 130 kW. For marine SI engines, our analysis assumes that the final standards
have no effect because that state already has equivalent exhaust emission standards and is
expected to adopt equivalent evaporative hydrocarbon requirements.
3.4.1 Controlled Exhaust and Evaporative Emissions Estimates for THC, VOC, NOx,
PM2 5, PM10, and CO
The controlled exhaust and evaporative emission inventories for small nonroad and
marine SI engines were generated by modifying the input files for NONROAD2005d to account
for the engine and equipment controls associated with the Phase 3 standards. (See the baseline
emission inventory discussion in Section 3.2 for the changes we made to the publically available
NONROAD2005a model to develop NONROAD2005d.) The modifications that were made to
estimate the controlled emissions inventories are described below.
3.4.1.1 Controlled Exhaust Emission Standards, Zero-Hour Emission Factors and
Deterioration Rates
3.4.1.1.1 Small SI Exhaust Emission Calculations
The final Phase 3 emission standards and implementation schedule are shown in Table
3.4-1. While the new standards take effect in 2011 for Class II engines and 2012 for Class I
engines, we providing a number of flexibilities for engine and equipment manufacturers that will
allow the continued production and use of Class II engines meeting the Phase 2 standards in
limited numbers over the first four years of the Phase 3 program. The implementation schedule
shown in the table is used for modeling purposes only. It is based on our assumption that engine
and equipment manufacturers take full advantage of these flexibilities.
3-26
-------
Emission Inventory
Table 3.4-1: Phase 3 Emission Standards and Estimated Implementation Schedule
for Class I and II Small SI Engines3 (g/kW-hr or Percent)
Engine
Class
Class I
Class II
Requirement
HC+NOx
CO (marine generator
sets only)
Estimated Sales
Percentage
HC+NOx
CO (marine generator
sets only)
Estimated Sales
Percentage
2011
~
~
~
8
5
83
2012
10
5
95
8
5
83
2013
10
5
95
8
5
93
2014
10
5
100
8
5
93
2015+
10
5
100
8
5
100
a Reflects maximum use of compliance flexibilities by engine and equipment manufacturers. Used for modeling
purposes only.
The modeled emission factors corresponding to the final Phase 3 standards are shown in
Table 3.4-2. (See Section 3.2.1.2.1 for a discussion of how the model uses zero hour emission
levels (ZML) and deterioration rates (A values.) We developed these new emission factors
based on testing of catalyst-equipped engines both in the laboratory and in-use. A full
description of the emission factor information for Phase 3 engines and the basis for the estimates
can be found in the docket for this rule.
Table 3.2-2: Phase 3 Modeling Emission Factors for Small SI Engines (g/kW-hr)
Class/
Technology
Class f - SV
Class f - OHV
Class ff
HCZML
5.60
5.09
4.25
HC "A"
0.797
0.797
0.797
NOxZML
1.47
1.91
1.35
NOx "A"
0.302
0.302
0.302
CO ZML
319.76
325.06
431.72
CO "A"
0.070
0.070
0.080
PM10
ZML
0.24
0.05
0.08
PM10 "A"
1.753
1.753
1.095
We left the proportion of sales in each technology classification unchanged from those
used for Phase 2 engines. The technology mix was previously shown Table 3.2-2.
Finally, as discussed in more detail in Chapter 6, were developed a new brake-specific
fuel consumption (BSFC) estimate for Class II engines to reflect the expected fuel consumption
benefit associated with the use of additional electronic fuel injection technology on Phase 3
compliant engines. The resulting BFSC for Phase 3 Class II engines is 0.735 pounds per
horsepower-hour (Ib/hp-hr).
3-27
-------
Final Regulatory Impact Analysis
3.4.1.1.2 Marine SI Exhaust Emission Calculations
For the control case, we developed new technology classifications for engines meeting
the final standards. For outboards and personal watercraft, we no longer will attempt to
determine the technology mix between low emitting technology options (such as DI 2-stroke
versus 4 stroke). The new technology classifications for these engines are simply tied to the
standard. These new technology classifications are titled MO09 and MP09 for outboards and
personal watercraft, respectively. In determining the combined HC+NOx emission factor, we
used the final emission standards with a 10 percent compliance margin (with deterioration factor
applied). To determine the NOx emission factors, we used certification data to determine the
sales weighted average NOx for low emission technologies in each power bin. HC was then
determined as the difference between the HC+NOx and the NOx emission factors. Because we
are establishing the same standards for OB and PWC and because they use similar engines, we
use the same HC+NOx emission factors and deterioration factors for both engine types.
Because the final CO standard primarily acts as a cap on CO, the CO emission factors
were determined based on the emission factors for existing low emission engines in each power
bin. Fuel consumption factors were calculated in the same manner. Therefore, some differences
are seen between the projected CO and BSFC factors for OB and PWC. No changes were made
to the PM emission factors. Also, the existing deterioration factors for 4-stroke carbureted
engines were applied to the control case (1.05 for HC, NOx, and CO). Table 3.4-3 presents the
zero-hour OB/PWC emission factors for the control case.
Table 3.4-3: Control Case Emission Factors for OB/PWC (g/kW-hr)
Power Bin
0-2.2 kW
2.3-4.5 kW
4.6-8.2 kW
8.3-11.9 kW
12.0-18. 6 kW
18. 7-29.8 kW
29.9-37.3 kW
37.4-55. 9 kW
55. 9-74.6 kW
74.7-130.5 kW
130.6+ kW
HC
20.9
22.1
15.5
11.6
12.5
10.2
9.3
9.2
9.2
9.2
10.2
NOx
4.8
3.6
5.6
6.8
4.3
5.7
5.9
5.4
5.4
5.0
3.7
CO
OB PWC
542
357
292
248
205
189
167
169
169
173
137
640
538
243
231
218
206
206
206
206
202
178
BSFC
OB PWC
563
560
555
552
543
528
507
471
471
415
387
563
560
555
552
543
528
507
486
486
394
380
For sterndrive and inboards, we developed a new engine classification similar to the
OB/PWC discussion above. SD/I engines at or below 373 kW are modeled to meet the final
standard through the use of aftertreatment. HC and NOx emission factors are based on test data
presented in Chapter 4 for SD/I engines equipped with catalysts. High performance engines
have two tiers of standards that can be achieved through the use of engine-based technology.
Although the standards distinguish between two power ranges for high-performance, a single
3-28
-------
Emission Inventory
weighted EF is used here. CO emission factors are based on meeting the final standard at the
end of useful life (with the deterioration factor applied). No emission reductions are modeled for
PM. The fuel consumption factor for fuel-injected 4-stroke SD/I engines is applied to the control
case. Deterioration factors for catalyst-equipped engines are the same as those used in the
NONROAD2005a model for catalyst-equipped large SI engines. Table 3.4-4 presents the zero-
hour emission factors and the accompanying deterioration factors for the control case.
Table 3.4-4: Control Case EFs (g/kW-hr) and DFs for SD/I
Engine Category
kW <373 kW
> 373 kW, Tier 1
>373kW, Tier 2
HC
EF
1.80
11.80
8.58
DF
1.64
1.69
1.69
NOx
EF
1.60
6.70
6.80
DF
1.15
1.38
1.38
CO
EF
55.0
207
207
DF
1.36
1.81
1.81
BSFC
345
362
362
3.4.1.2 Controlled Evaporative Emission Rates
Below, we present the effect of the final Phase 3 evaporative emission standards on hose
permeation, tank permeation, diurnal, and running loss emission inventories.
3.4.1.2.1 Hose Permeation
Similar to the baseline case, hose permeation rates are based on g/m2/day and are
modeled as a function of temperature. The fuel hose test procedures are based on Fuel CE10 as a
test fuel. Based on data presented in Chapter 5, we would expect in-use emissions on gasoline-
based E10 to be about half of the measured level on Fuel CE10. In addition, we believe that
hose designed to meet the final 15 g/m2/day standard on 10 percent ethanol fuel will permeate at
least 50 percent less when gasoline is used. Therefore, we model permeation from hoses
designed to meet 15 g/m2/day on Fuel CE10 to be 3.75 g/m2/day on gasoline at 23°C. Consistent
with the baseline emission case, we weight the gasoline and E10 emission factors by our
estimates of gasoline sales with and without ethanol added. The same correction factors used to
account for the effect of ethanol on permeation are used in the control and baseline cases.
Fill neck and vent hose containing vapor rather than liquid fuel are not subject to the final
standards. No emission reductions are modeled for these hose types. In addition, no emission
reductions are modeled for hose on handheld equipment used in cold weather applications (e.g.
Class V chainsaws).
3.4.1.2.2 Tank Permeation
Similar to the baseline case, fuel tank permeation rates are based on units of g/m2/day and
are modeled as a function of temperature. We believe that fuel tanks using alternative materials
3-29
-------
Final Regulatory Impact Analysis
to meet the final 1.5 g/m2/day standard on 10 percent ethanol fuel will typically permeate at least
50 percent less when gasoline is used. Therefore, we model permeation from fuel tanks to be
0.75 g/gal/day at 29°C on gasoline. Consistent with the baseline emission case, we weight the
gasoline and E10 emission factors by our estimates of gasoline sales with and without ethanol
added. The same correction factors used to account for the effect of ethanol on permeation are
used in the control and baseline cases.
One exception to the above discussion is metal tanks. For these fuel tanks, we do not
include any emissions reductions from baseline.
3.4.1.2.3 Diurnal
We are not establishing a diurnal emission requirement for small nonroad SI equipment.
Therefore, we do not model direct reductions in diurnal emissions. However, we are placing a
limit on diffusion emissions. As a result, we set the diffusion multiplier to 1.0 for all non-
handheld Small SI equipment for the control case. Note that this multiplier was already set to
1.0 for handheld equipment in the baseline case. This is equivalent to applying a 32 percent
reduction to the diurnal emission factors.
In the control case for marine SI engines, we model portable fuel tanks as having 90
percent lower diurnal emissions than an open vent system. Also, we set the diffusion multiplier
to 1.0 because the tanks would be sealed. Presumably, the diurnal temperature cycles would
build some pressure in the fuel tank causing hydrocarbons to be released when the tank is
opened. Therefore, we do not model these tanks as having zero diurnal emissions. For PWC, we
use the baseline scenario of sealed systems with a 1.0 psi pressure relief valve. For installed fuel
tanks, we model a 60 percent reduction due to a carbon canister in the fuel line with passive
purge. This reduction is based on data presented in Chapter 5. As in the baseline case, no
diffusion is modeled for PWC and installed fuel tanks.
3.4.1.2.4 Running Loss
For Class I engines, we believe that the final running loss control requirement will be met
by routing vapor from the fuel take to the engine air intake system. Therefore, all vapor
generated in the fuel tank should be consumed by the engine, thereby eliminating running loss
emissions. However, there may be some inefficiencies in the system such as vapor escaping out
the intake at idle. Therefore, we model the running loss emission reduction as only 90 percent.
For Class II equipment, we believe that some equipment will inherently meet the final standard
because they will have low enough temperature fluctuation in the fuel tanks during operation to
certify by design. Based on the data presented in Chapter 5 on fuel tank temperatures during
operation, we estimate an 80 percent reduction in running loss for Class II equipment.
3.4.1.3 Controlled Exhaust and Evaporative Inventory Results for THC, VOC,
NOx, PM2 5, PM10, CO and SO2
Tables 3.4-5 presents the 50-state controlled emission inventories for small nonroad SI
3-30
-------
Emission Inventory
engines. Tables 3.4-6 provides the same information for marine SI engines.
Table 3.4-5: Controlled 50-State Annual Exhaust and Evaporative Emissions for
Small Nonroad Spark-Ignition Engines (short tons)
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
2036
2037
2038
2039
2040
THC
1,064,625
1,026,922
963,709
886,524
825,413
768,239
713,266
669,782
646,659
618,463
573,611
535,035
509,895
495,565
486,951
484,102
485,460
488,969
493,763
499,618
506,071
512,988
520,130
527,386
534,763
542,202
549,676
557,179
564,711
572,261
579,823
587,396
594,978
602,567
610,167
617,776
625,391
633,010
640,633
voc
1,047,374
1,009,822
945,601
867,081
804,926
747,552
694,712
653,584
631,851
604,977
562,317
525,722
501,985
488,517
480,492
477,989
479,510
483,092
487,905
493,736
500,141
506,990
514,054
521,227
528,516
535,864
543,246
550,657
558,094
565,549
573,015
580,492
587,978
595,471
602,974
610,486
618,003
625,525
633,050
NOx
106,804
106,852
106,610
106,847
109,233
109,439
111,235
116,329
118,376
107,135
96,222
86,623
81,011
76,412
73,517
72,202
71,768
71,822
72,175
72,848
73,667
74,592
75,564
76,578
77,629
78,700
79,782
80,875
81,977
83,086
84,197
85,312
86,429
87,547
88,669
89,795
90,922
92,051
93,181
PM2.5
23,382
23,480
23,483
23,417
23,498
23,804
24,335
24,882
25,402
25,888
26,364
26,832
27,291
27,747
28,202
28,655
29,107
29,558
30,009
30,460
30,911
31,362
31,813
32,265
32,718
33,173
33,627
34,081
34,535
34,990
35,444
35,898
36,353
36,807
37,261
37,716
38,170
38,625
39,079
PM10
25,416
25,522
25,525
25,453
25,541
25,874
26,451
27,045
27,611
28,139
28,445
28,747
29,071
29,473
29,896
30,336
30,795
31,259
31,727
32,198
32,671
33,146
33,622
34,098
34,576
35,055
35,534
36,013
36,492
36,971
37,450
37,929
38,408
38,887
39,366
39,845
40,324
40,803
41,282
CO
15,091,835
14,351,829
13,690,337
12,923,819
12,252,479
11,711,607
10,861,441
9,992,801
9,623,727
9,427,359
9,240,448
9,125,886
9,107,104
9,135,515
9,201,411
9,298,167
9,415,010
9,543,762
9,679,462
9,820,562
9,964,572
10,110,794
10,258,216
10,406,862
10,557,631
10,709,373
10,861,555
11,014,081
11,166,921
11,319,980
11,473,166
11,626,453
11,779,824
11,933,276
12,086,826
12,240,473
12,394,188
12,547,962
12,701,792
3-31
-------
Final Regulatory Impact Analysis
Table 3.4-6: Controlled 50-State Annual Exhaust and Evaporative Emissions for
Marine Spark-Ignition Engines (short tons)
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
2036
2037
2038
2039
2040
THC
906,318
873,287
836,493
796,279
756,781
717,924
680,702
644,330
593,432
543,080
495,189
452,028
412,280
376,203
342,807
312,281
285,067
259,742
238,704
219,826
203,027
188,065
175,954
166,147
157,943
151,063
145,397
140,670
136,990
134,079
131,797
130,067
128,766
127,853
127,275
126,952
126,816
126,828
126,959
voc
931,132
896,969
858,916
817,340
776,480
736,303
697,795
660,187
607,656
555,760
506,468
462,066
421,190
384,108
349,793
318,441
290,507
264,519
242,957
223,621
206,427
191,121
178,752
168,751
160,391
153,384
147,624
142,822
139,083
136,124
133,803
132,045
130,723
129,798
129,212
128,888
128,753
128,769
128,906
NOx
46,311
49,694
53,397
57,862
63,366
67,730
73,894
82,123
84,822
85,353
85,673
85,732
85,609
85,334
84,890
84,279
83,468
82,546
81,398
80,081
78,657
77,197
75,802
74,424
73,057
71,713
70,421
69,236
68,639
68,339
68,148
68,038
67,985
67,975
68,009
68,077
68,174
68,302
68,461
PM2.5
15,092
14,417
13,679
12,886
12,090
11,311
10,553
9,824
8,832
7,891
7,035
6,275
5,577
4,951
4,376
3,856
3,399
2,978
2,640
2,341
2,081
,851
,673
,534
,419
,323
,247
,185
,137
,099
,067
,043
,024
,010
,001
995
991
989
989
PM10
16,404
15,670
14,869
14,007
13,142
12,295
11,470
10,678
9,600
8,577
7,647
6,820
6,062
5,381
4,756
4,191
3,695
3,237
2,869
2,545
2,262
2,012
1,818
1,668
1,542
1,438
1,355
1,288
1,236
1,194
1,160
1,134
1,113
1,098
1,088
1,081
1,077
1,075
1,075
CO
2,472,251
2,407,992
2,346,538
2,266,733
2,170,374
2,103,059
2,007,804
1,885,970
1,808,304
1,755,638
1,707,370
1,664,442
1,624,423
1,587,889
1,553,983
1,523,443
1,496,863
1,472,528
1,452,196
1,433,655
1,417,440
1,403,195
1,391,146
1,380,739
1,371,913
1,364,592
1,358,936
1,354,638
1,353,989
1,355,439
1,357,905
1,361,273
1,365,343
1,370,010
1,375,199
1,380,822
1,386,805
1,393,116
1,399,715
3-32
-------
Emission Inventory
3.4.2 Controlled Hazardous Air Pollutant Estimates
The final hydrocarbon emission standards for small nonroad and marine SI engines will
also reduce toxic air pollutants. To calculate the controlled toxic air emission inventories, we
multiplied the baseline hazardous air pollutant estimates (Section 3.2.2) by the ratio of control
and baseline emission inventories (Section 3.2.1.4 and 3.4.1.3, respectively) for VOC or PM, as
appropriate. More specifically, we used the VOC ratio for all toxic pollutant species that are
found in the gas phase. The gas phase pollutants are all the species described below, except for
naphthalene and the polycyclic organic matter (POM) compounds that are found in both the gas
and particulate phase. In these cases, we used the PM ratio to estimate the controlled
inventories.
Controlled inventories were calculated for the seven major types of air toxic emissions:
benzene, formaldeyde, acetaldehyde, 1,3-butadiene, acrolein, naphthalene, and 15 other
compounds grouped together as POM for this analysis.6 Table 3.4-7 presents the 50-state
controlled inventories, respectively, small nonroad SI engines. Table 3.4-8 provides the same
information for marine SI engines.
Table 3.4-7: Controlled 50-State Air Toxic Emissions for
Small Nonroad Spark-Ignition Engines (short tons)
Year
2002
2020
2030
Benzene
35,086
15,413
17,577
1,3
Butadiene
5,561
2,504
2,859
Formalde-
hyde
8,664
4,189
4,784
Acetalde-
hyde
2,900
2,015
2,300
Acrolein
505
210
240
Naphthalene
447
620
718
POM
97
128
147
Table 3.4-8: Controlled 50-State Air Toxic Emissions for
Marine Spark-Ignition Engines (short tons)
Year
2002
2020
2030
Benzene
23,110
4,453
2,582
1,3
Butadiene
2,053
390
216
Formalde-
hyde
2,153
273
151
Acetalde-
hyde
1,543
418
234
Acrolein
211
14
8
Naphthalene
37
19
9
POM
31
7
3
e The 15 POMs summarized in this chapter are acenapthylene, anthracene, benz(a)anthracene,
benzo(a)pyrene, benzo(b)fluoranthene, benzo(g,h,i)perylene, beno(k)fluoranthene, chrysene, dibenzo(a,h)anthracene,
fluoranthene, fluorene, indeno(l,2,3-c,d)-pyrene, phenanthrene, andpyrene.
3-33
-------
Final Regulatory Impact Analysis
3.5 Projected Emissions Reductions from the Final Rule
This section presents the projected total emission reductions associated with the final
Phase 3 standards. We calculated the reductions by subtracting the baseline inventories from
Section 3.2 by the controlled inventories from Section 3.4.
3.5.1 Results for THC, VOC, NOx, PM25, PM10, and CO
Tables 3.5-1 presents the 50-state exhaust and evaporative emission inventories and
percent reductions, respectively, for small nonroad SI engines. Tables 3.5-2 provides the same
information for marine SI engines. Tables 3.5-3 summarizes the combined emission reductions
for the final rule. The earliest Phase 3 evaporative standards for small nonroad SI engines begin
in 2008. Similar final evaporative standards affect Marine SI engines one year later. Therefore
the emission reductions are shown beginning in 2008 for small nonroad SI engines and 2009 for
Marine SI engines. Figures 3.5-1 though 3.5-5 show the combined baseline, controlled, and by
contrast the reduction emission inventories over time for small nonroad and Marine SI engines.
3-34
-------
Table 3.5-1: Total 50-State Annual Exhaust
for Small Spark-Ignition
and Evaporative Emission Reductions
Engines (short tons)
Year
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
2037
2038
2039
2040
THC
Tons
5,298
12,306
19,103
45,157
92,055
136,293
169,739
193,479
212,274
225,797
235,483
243,315
249,991
255,636
260,711
265,345
269,770
274,107
278,455
282,769
287,060
291,328
295,576
299,808
304,033
308,250
312,460
316,665
320,862
325,052
329,238
333,421
337,602
%
1
2
3
7
14
20
25
28
30
32
33
33
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
35
35
voc
Tons
5,298
12,306
19,103
44,381
89,426
131,817
163,834
186,614
204,672
217,669
226,988
234,523
240,948
246,382
251,271
255,737
260,002
264,185
268,375
272,533
276,668
280,782
284,876
288,955
293,027
297,091
301,148
305,201
309,246
313,284
317,318
321,349
325,379
%
1
2
3
7
14
20
25
28
30
32
33
33
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
34
35
35
NOx
Tons
0
0
0
12,289
24,598
35,882
43,389
49,983
54,951
58,371
60,933
63,024
64,827
66,312
67,650
68,883
70,072
71,228
72,374
73,509
74,635
75,753
76,863
77,967
79,068
80,167
81,264
82,359
83,450
84,537
85,623
86,708
87,792
%
0
0
0
10
20
29
35
40
43
45
46
47
47
48
48
48
48
48
48
48
48
48
48
48
48
48
48
48
48
48
48
49
49
PM2.5
Tons
0
0
0
0
195
385
546
632
697
745
776
800
820
838
854
868
881
894
908
922
936
949
963
977
990
1,004
1,018
1,031
1,045
1,059
1,072
1,086
1,100
%
0
0
0
0
1
1
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
PM10
Tons
0
0
0
0
212
419
594
687
758
810
843
870
892
911
928
943
958
972
987
1,002
1,017
1,032
1,047
1,062
1,076
1,091
1,106
1,121
1,136
1,151
1,166
1,181
1,195
%
0
0
0
0
1
1
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
CO
Tons
0
0
0
141,251
338,592
518,625
644,624
743,512
818,628
871,019
909,068
939,944
966,407
988,367
1,008,088
1,026,161
1,043,515
1,060,430
1,077,303
1,094,028
1,110,652
1,127,170
1,143,584
1,159,920
1,176,219
1,192,487
1,208,730
1,224,953
1,241,128
1,257,259
1,273,368
1,289,462
1,305,543
%
0
0
0
1
4
5
7
8
8
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
-------
Table 3.5-2: Total 50-
for
State Annual Exhaust and Evaporative Emission Reductions
Marine Spark-Ignition Engines (short tons)
Year
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
2037
2038
2039
2040
THC
Tons
1,400
18,748
37,670
58,252
78,654
98,886
118,974
138,843
158,386
177,535
196,122
212,471
227,798
242,345
255,886
267,248
277,049
285,828
293,743
300,755
307,097
312,636
317,588
322,039
326,092
329,826
333,261
336,417
339,371
342,183
344,878
347,477
%
0
3
6
11
15
19
24
29
34
38
43
47
51
54
58
60
63
64
66
67
69
70
70
71
71
72
72
73
73
73
73
73
voc
Tons
1,400
19,245
38,656
59,723
80,605
101,313
121,873
142,207
162,207
181,801
200,818
217,524
233,178
248,030
261,851
273,428
283,400
292,329
300,379
307,503
313,944
319,573
324,609
329,140
333,265
337,067
340,565
343,778
346,786
349,649
352,395
355,043
%
0
3
7
11
15
19
24
29
34
38
43
47
51
55
58
60
63
65
66
68
69
70
70
71
72
72
72
73
73
73
73
73
NOx
Tons
0
2,318
5,163
7,989
10,796
13,588
16,369
19,131
21,879
24,617
27,340
30,128
32,981
35,822
38,605
41,250
43,803
46,286
48,689
50,990
53,151
54,696
55,921
57,018
58,015
58,937
59,802
60,611
61,377
62,104
62,793
63,445
%
0
3
6
9
11
14
16
18
21
23
25
27
29
31
33
35
37
39
40
42
43
44
45
46
46
46
47
47
47
48
48
48
PM2.5
Tons
0
317
634
948
1,259
1,567
1,872
2,173
2,468
2,756
3,034
3,269
3,485
3,688
3,875
4,023
4,146
4,256
4,355
4,440
4,516
4,582
4,642
4,698
4,749
4,797
4,841
4,882
4,920
4,957
4,993
5,027
%
0
3
7
12
17
22
27
33
39
45
50
55
60
64
68
71
73
75
77
78
79
80
81
81
82
82
83
83
83
83
83
84
PM10
Tons
0
344
689
1,031
1,369
1,703
2,035
2,362
2,683
2,996
3,298
3,553
3,788
4,008
4,212
4,373
4,506
4,626
4,734
4,826
4,909
4,981
5,046
5,107
5,162
5,214
5,262
5,306
5,348
5,388
5,427
5,464
%
0
3
7
12
17
22
27
33
39
45
50
55
60
64
68
71
73
75
77
78
79
80
81
81
82
82
83
83
83
83
83
84
CO
Tons
0
15,540
33,192
50,745
68,211
85,582
102,867
119,995
136,972
153,767
170,313
185,917
201,392
216,625
231,477
245,457
259,174
272,579
285,568
297,811
309,295
317,638
324,313
330,323
335,801
340,886
345,665
350,145
354,388
358,435
362,283
365,936
%
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
17
18
19
19
19
20
20
20
20
20
20
21
21
21
-------
Table 3.5-3: Total 50-State Annual Exhaust and Evaporative Emission Reductions
for Small Nonroad and Marine Spark-Ignition Engines (short tons)
Year
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
2037
2038
2039
2040
THC
Tons
5,298
13,706
37,851
82,827
150,307
214,947
268,625
312,454
351,117
384,183
413,019
439,437
462,463
483,433
503,056
521,231
537,018
551,156
564,282
576,512
587,814
598,426
608,211
617,396
626,072
634,342
642,286
649,926
657,279
664,423
671,421
678,299
685,079
%
0
1
3
7
12
18
23
26
30
33
35
37
39
40
42
43
44
44
45
45
46
46
46
47
47
47
47
47
47
47
47
47
47
voc
Tons
5,298
13,706
38,348
83,037
149,149
212,422
265,147
308,487
346,879
379,875
408,789
435,340
458,472
479,560
499,301
517,588
533,430
547,584
560,704
572,913
584,171
594,726
604,449
613,565
622,167
630,356
638,215
645,766
653,024
660,069
666,967
673,744
680,422
%
0
1
3
7
12
18
22
26
29
32
35
37
39
40
41
43
44
44
45
45
46
46
46
47
47
47
47
47
47
47
47
47
47
NOx
Tons
0
0
2,318
17,451
32,587
46,679
56,977
66,353
74,082
80,250
85,550
90,363
94,954
99,293
103,472
107,488
111,322
115,031
118,660
122,198
125,625
128,904
131,559
133,888
136,087
138,182
140,201
142,161
144,061
145,914
147,727
149,501
151,237
%
0
0
1
8
15
21
25
29
32
34
36
37
38
39
40
41
42
43
44
45
46
46
47
47
47
47
48
48
48
48
48
48
48
PM2.5
Tons
0
0
317
634
1,143
1,644
2,113
2,504
2,870
3,213
3,532
3,834
4,089
4,323
4,541
4,743
4,905
5,040
5,164
5,277
5,376
5,465
5,545
5,619
5,688
5,753
5,814
5,872
5,927
5,979
6,029
6,079
6,127
%
0
0
1
2
3
5
6
7
8
9
10
11
11
12
12
13
13
13
13
14
14
14
14
14
14
14
14
14
14
14
14
14
14
PM10
Tons
0
0
344
689
1,243
1,787
2,297
2,722
3,120
3,493
3,839
4,168
4,444
4,698
4,936
5,155
5,331
5,479
5,613
5,736
5,843
5,940
6,027
6,107
6,183
6,253
6,320
6,383
6,442
6,499
6,554
6,607
6,660
%
0
0
1
2
3
5
6
7
8
9
10
11
11
12
12
13
13
13
13
14
14
14
14
14
14
14
14
14
14
14
14
14
14
CO
Tons
0
0
15,540
174,444
389,337
586,836
730,206
846,379
938,623
1,007,991
1,062,836
1,110,258
1,152,325
1,189,759
1,224,712
1,257,637
1,288,972
1,319,604
1,349,882
1,379,596
1,408,463
1,436,465
1,461,222
1,484,233
1,506,542
1,528,288
1,549,616
1,570,618
1,591,273
1,611,647
1,631,803
1,651,746
1,671,479
%
0
0
0
2
3
5
6
7
8
9
9
9
9
10
10
10
10
10
10
10
10
10
10
10
11
11
11
11
11
11
11
11
11
-------
Final Regulatory Impact Analysis
Figure 3.5-1: 50-State Annual THC Exhaust and Evaporative Emissions for Small SI and
Marine SI Engines
2,000,000 - -».
1,500,000
in
O
500,000
2005 2010 2015 2020 2025 2030 2035 2040
Figure 3.5-2: 50-State Annual VOC Exhaust and Evaporative Emissions for
Small SI and Marine SI Engines
2,500,000
2,000,000
ra
£
(A
1,500,000
1,000,000
500,000
2000 2005 2010 2015 2020 2025 2030 2035 2040
Year
3-38
-------
Emission Inventory
Figure 3.5-2: 50-State Annual NOx Exhaust Emissions for Small SI and Marine SI
Engines
300,000
2000 2005 2010 2015 2020 2025 2030 2035 2040
Figure 3.5-3: 50-State Annual PM2.5 Exhaust Emissions for Small SI and
Marine SI Engines
50,000
45,000
5,000
2000
2005
2035
2040
3-39
-------
Final Regulatory Impact Analysis
Figure 3.5-4: 50-State Annual PM10 Emissions for Small SI and Marine SI
Engines
60,000
50,000
40,000
cu
-------
Emission Inventory
3.5.2 Results for Hazardous Air Pollutants
Table 3.5-4 presents the 50-state exhaust and evaporative hazardous air pollutant
emission reductions for small nonroad SI engines that are expected as a result of the final
standards. Table 3.5-5 provides the same information for marine SI engines. Table 3.5-6
summarizes the combined hazardous air pollutant reductions for the final rule. These results are
displayed for 2020 and 2030, when most or all of the engines subject to the standards are
represented in the respective fleets.
3-41
-------
Table 3.5-4: 50-State Air Toxic Emission Reductions for
Small Nonroad Spark-Ignition Engines (short tons)
Year
2020
2030
Benzene
Tons
7,803
9,200
%
34
34
1,3 Butadiene
Tons
965
1,140
%
28
29
Formaldehyde
Tons
1,614
1,907
%
28
29
Acetaldehyde
Tons
777
917
%
28
29
Acrolein
Tons
81
96
%
28
29
Napthalene
Tons
17
21
%
3
29
POM
Tons
4
4
%
3
3
Table 3.5-5: 50-State Air Toxic Emission Reductions for
Marine Snark-Tgnition Engines (short tnns>
Year
2020
2030
Benzene
Tons
3,964
5,893
%
47
70
1,3 Butadiene
Tons
373
535
%
49
71
Formaldehyde
Tons
261
374
%
49
71
Acetaldehyde
Tons
400
577
%
49
71
Acrolein
Tons
13
19
%
49
71
Napthalene
Tons
23
37
%
55
80
POM
Tons
9
13
%
55
80
Table 3.5-6: 50-State Air Toxic Emission Reductions for
Small Nonroad and Marine Snark-Tgnition Engines (short tnns>
Year
2020
2030
Benzene
Tons
11,767
15,093
%
37
43
1,3 Butadiene
Tons
1,338
1,675
%
32
35
Formaldehyde
Tons
1,875
2,281
%
30
32
Acetaldehyde
Tons
1,176
1,494
%
33
37
Acrolein
Tons
94
115
%
30
32
Napthalene
Tons
41
57
%
6
7
POM
Tons
9
57
%
11
10
-------
Emission Inventory
3.6 Emission Inventories Used for Air Quality Modeling
This section briefly summaries the methodology we used for air quality modeling
purposes and to develop the emission inventories for that modeling. It also describes the
changes to our emission inventory modeling inputs and resulting emission inventories that were
made between the preliminary baseline and control scenarios used for the air quality modeling,
and the updated final baseline and control scenarios for the final Phase 3 rule. These differences
often occur because the emission inputs for the air quality modeling are required early in the
analytical process to ensure there is adequate time to complete the analysis and incorporate the
results into the rulemaking. Given that lead time requirement, air quality modeling is often
based on analytical methods and inputs that may be superceded, or on a control scenario that
does not specifically match the final set of emission standards.
3.6.1 General Description of the Air Quality Modeling
Air quality modeling was performed for the 48 contiguous states and the District of
Columbia to estimate the effect of the final rule on future annual fine particulate matter (PM25)
concentrations, 8-hour ozone concentrations, and visibility levels. The analysis was performed
for calendar years 2002, 2020, and 2030 using the Community Multiscale Air Quality (CMAQ)
model. The model simulates the multiple physical and chemical processes involved in the
formation, transport, and destruction of fine parti culate and ozone.
The air quality modeling for the final rule was predominately taken directly from the
work performed for EPA's recent final rulemaking to control air emissions from locomotive
engines and marine compression-ignition engines less than 20 liters per cylinder (the
locomotive/marine rule). This approach was adopted to ensure that the air quality modeling for
this rule included the effects all EPA's most recent air pollution control regulations and to
conserve resources by taking advantage of the existing inventory preparation (i.e., input files),
analytical methods, and results.
More specifically, we used the locomotive/marine rule's "control" scenario as our starting
point. The resulting baseline for the Phase 3 final rule was then modified to include more recent
modeling updates for small nonroad and marine SI engines based on the NONROAD2005d core
model and a set of input files that closely matches those used in this final rule. (The differences
in the input files used in the preliminary inventories and those for the final rule, which also used
NONROAD 2005d core model, are specifically described later in this section.) For air quality
modeling purposes, the nonroad model was executed within the framework of EPA's National
Mobile Inventory Model (NMEVI) that links a county-level database to the model and collates the
output into a single database table. The resulting NMIM inventory estimates for nonroad and
marine SI engines account for local differences in fuel characteristics and temperatures. By
contrast, if NONROAD2005d is run as a stand alone model, results are based on a somewhat less
accurate, but much less resource intensive approach that uses national average daily temperatures
and fuel characteristics.
3-43
-------
Final Regulatory Impact Analysis
The NMIM emissions inventory methodology and results for the highway vehicles and all
nonroad sources (including those for small nonroad and marine SI engines) that were used in the
air quality assessment are in the docket for this final rule.24
3.6.2 Methodology for Comparing the Preliminary Air Quality Emission Inventories and
the Final Inventories
The simplest method for comparing the preliminary emission inventories and the
inventories represented in the final rule is to compare the emission results from using the
NONROAD2005d core model with the national average inputs for temperature and fuel that
reflect the modeling assumptions for the preliminary and final rule inventories. This is possible
because of the similarities in the underlying use of the NONROAD2005d model. More
specifically, even though the two modeling approaches, i.e., NMIM and the stand alone
NONROAD2005 model, use different temperature and fuel characteristics, the computational
method is the same/ This consistency means that the results of the two modeling approaches
will be proportional in nature, i.e., the relative changes in the inventories will be similar. Also,
as is explained in more detail later, the other basic modeling scenario inputs, e.g., emission
factors and deterioration rates, are nearly identical. Taken together, the modeling results from
using the NONROAD2005d model with national average inputs for temperature and fuels will
closely mirror the differences in inventories produced with NMIM. This avoids the more time
and resource intensive approach of rerunning the NMIM model for the final rule scenarios, while
still providing a good comparison of the differences in the absolute inventories, i.e., tons, and
more importantly for air quality considerations, the percent reduction between the baseline and
control cases. Therefore, the comparisons of the preliminary and final emission scenarios that
are presented in the following sections are based on comparing 50-state inventories using the
nonroad model with national average inputs.
3.6.3 Comparison of the Baseline Scenario Emission Inventories
As described in Section 3.2., the final emission inventories for the Phase 3 rule are based
on the use of a special version of the nonroad model, i.e., NONROAD2005d. Similarly, the
preliminary emission inventories for air quality modeling were also constructed using the same
version NONROAD2005d core model. Therefore, the only difference between the preliminary
and final baseline scenarios are the modeling inputs. These differences and a comparison of the
respective inventory results are presented below.
3.6.3.1 Differences Between the Preliminary and Final Baseline Scenarios
The modeling inputs for the final baseline scenario are described in Section 3.2.1. The
only difference in the inputs for small nonroad SI engines between the preliminary and the final
baseline scenarios is that the preliminary results did not include correction of the running loss
f The difference between the preliminary emission inventories using NMIM and final rule emission
inventories using NONROAD2005d is that the NMIM results, which use county-level data for temperatures and fuel
characteristics, are generally 10-15 percent greater depending on the pollutant.
3-44
-------
Emission Inventory
emission factors for Class 1 snowblowers to account for cold weather applications as described
in Section 3.2.1.1.1, number 8. There were no differences in the inputs for marine SI engines.
3.6.3.2 Comparison of Preliminary and Final Baseline Emission Inventories
Table 3.6-1 compares the preliminary and final 50-state baseline scenario inventories for
small nonroad and marine SI engines. As shown, the differences in the baseline scenarios are
insignificant.
3-45
-------
Table 3.6-1: Comparison of 50-State Baseline Scenario Emissions for
Preliminary Air Quality Modeling and Final Rule
Applications
Small Nonroad
SI Subject to
the Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
VOC [short tons]
Final
728,853
842,970
460,481
458,656
1,189,334
1,301,626
Preliminary
729,235
843,410
460,481
458,656
1,189,716
1,302,066
Difference
(382)
(440)
0
0
(382)
(440)
NOX [short tons]
Final
137,002
158,840
111,525
123,335
248,527
282,175
Preliminary
137,002
158,840
111,525
123,335
248,527
282,175
Difference
0
0
0
0
0
0
PM25 [short tons]
Final
30,009
34,535
5,908
5,719
35,917
40,254
Preliminary
30,009
34,535
5,908
5,719
35,917
40,255
Difference
0
0
0
0
0
0
-------
Emission Inventory
Table 3.6-1 (Cont'd)
Comparison of 50-State Baseline Scenario Emissions for
Preliminary Air Quality Modeling and Final Rule
Applications
Small Nonroad
SI Subject to the
Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
PM10 [short tons]
Final
32,618
37,538
6,422
6,217
39,041
43,755
Preliminary
32,618
37,538
6,422
6,217
39,041
43,755
Difference
0
0
0
0
0
0
CO [short tons]
Final
10,645,870
12,310,505
1,638,114
1,671,627
12,283,983
13,982,132
Preliminary
10,645,870
12,310,505
1,638,114
1,671,627
12,283,983
13,982,132
Difference
0
0
0
0
0
0
3.6.4 Comparison of the Control Scenario Emission Inventories
As noted above, the preliminary and final emission inventories for the Phase 3 rule are
based on the same version of the nonroad model, i.e., NONROAD2005d. Therefore, the only
difference between the scenarios are the modeling inputs. These differences and a comparison
of the respective inventory results are presented below.
3.6.4.1 Differences Between the Preliminary and Final Control Scenarios
The modeling inputs for the final control scenario are described in Section 3.4.1. The
only difference in the inputs for small nonroad SI engines between the preliminary and the final
control scenarios is that the preliminary results excluded the following:
1. Update of the Phase 3 Class II zero-hour emission factor for CO from 391.13 to
431.72 g/kW-hr (321.9 g/hp-hr); and
2. Update of the Phase 3 Class II brake specific fuel consumption values from 0.666
to 0.735 Ib/hp-hr to reflect a lower use of electronic fuel injection systems.
For marine SI engines, the only difference in the inputs between the preliminary and the
final control scenarios is that the preliminary results excluded the following:
1. Revised several of the HC and NOx emission factors for outboards, personal
watercraft, and sterndrive/inboards;
2. Revised the Phase 3 standard phase-in dates for high performance
sterndrive/inboards (>600 hp);
3. Delayed by one year the implementation dates for outboards, personal watercraft,
and stern drive/inboards: and
3-47
-------
Final Regulatory Impact Analysis
4. Delayed by one year the implementation of diurnal Phase 3 controls for portable
fuel tanks to 2010 and installed fuel tanks to 2011.
3.6.4.2 Comparison of Preliminary and Final Control Emission Inventories
Table 3.6-2 compares the preliminary and final 50-state control scenario inventories for
small nonroad and marine SI engines. As shown, the difference in the control scenarios are
insignificant.
3-48
-------
Table 3.6-2: Comparison of 50-State Control Scenario Emissions for
Preliminary Air Quality Modeling Scenario and Final Rule (Tons/Year)
Applications
Small Nonroad
SI Subject to
the Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
VOC [short tons]
Final
487,905
558,094
242,957
139,083
730,862
697,177
Preliminary
487,663
557,805
241,216
136,650
728,879
694,455
Difference
242
289
1,741
2,433
1,983
2,722
NOX [short tons]
Final
72,175
81,977
81,398
68,639
153,573
150,616
Preliminary
72,175
81,977
81,162
68,538
153,336
150,515
Difference
0
0
236
101
237
101
PM2 , [short tons]
Final
29,189
33,572
2,640
1,137
31,829
34,709
Preliminary
29,189
33,572
2,640
1,137
31,829
34,709
Difference
0
0
0
0
0
0
-------
Final Regulatory Impact Analysis
Table 3.6-2 (Cont'd)
Comparison of 50-State Control Scenario Emissions for
Preliminary Air Quality Modeling and Final Rule
Applications
Small Nonroad
SI Subject to the
Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
PM10 [short tons]
Final
31,727
36,492
2,869
1,236
34,596
37,728
Preliminary
31,727
36,492
2,869
1,236
34,596
37,728
Difference
0
0
0
0
0
0
CO [short tons]
Final
9,679,462
11,166,921
1,452,196
1,353,989
11,131,658
12,520,910
Preliminary
9,029,001
10,393,508
1,447,553
1,345,079
10,476,554
11,738,587
Difference
650,461
773,413
4,643
8,910
655,104
782,323
3.6.5 Comparison of the Emission Reduction Inventories
Table 3.6-3 compares the emission reductions for preliminary and final 50-state
inventories for small nonroad and marine SI engines. As shown, the differences are
insignificant.
3-50
-------
Table 3.6-3: Comparison of 50-State Emissions Reductions for
Preliminary Air Quality Modeling Scenario and Final Rule (Tons/Year)
Applications
Small Nonroad
SI Subject to
the Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
VOC [short tons]
Final
240,948
284,876
217,524
319,573
458,472
604,449
Preliminary
241,572
285,605
219,265
322,006
460,837
607,611
Difference
(624)
(729)
(1,741)
(2,433)
(2,365)
(3,162)
NOX [short tons]
Final
64,827
76,863
30,128
54,696
94,955
131,559
Preliminary
64,827
76,863
30,364
54,797
95,191
131,660
Difference
0
0
(236)
(101)
(236)
(101)
PM2 5 [short tons]
Final
820
963
1,287
3,269
4,082
5,545
Preliminary
820
963
3,269
4,582
4,089
5,545
Difference
0
0
(1,982)
(1,313)
(7)
0
-------
Final Regulatory Impact Analysis
Table 3.6-3 (Cont'd)
Comparison of 50-State Emission Reductions for
Preliminary Air Quality Modeling and Final Rule
Applications
Small Nonroad
SI Subject to the
Final rule
Marine SI
Total
Year
2020
2030
2020
2030
2020
2030
PM10 [short tons]
Final
892
1,047
3,553
4,981
4,445
6,028
Preliminar
y
892
1,047
3,553
4,981
4,445
6,028
Difference
0
0
0
0
0
0
CO [short tons]
Final
966,407
1,143,584
185,917
317,638
1,152,325
1,461,222
Preliminar
y
1,616,868
1,916,997
190,560
326,548
1,807,428
2,243,545
Difference
(650,461)
(773,413)
(4,643)
(8,910)
(655,103)
(782,323)
3-52
-------
Emission Inventory
Chapter 3 References
1. "NONROAD2005a Emissions Inventory Model and Documentation," Memorandum and attachment from
Richard S. Wilcox to Docket EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, February 5, 2007. Docket Identification EPA-HQ-OAR-
2004-0008-0517.
2. "NONROAD2005d missions Inventory Model," Memorandum and attachment from Richard S. Wilcox to Docket
EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of Transportation and Air Quality, Ann
Arbor, Michigan, February 5, 2007. Docket Identification EPA-HQ-OAR-2004-0008-0719.
3. "Updates to Phase 2 Technology Mix, Emission Factors, and Deterioration Rates for Spark-Ignition Nonroad
Nonhandheld Engines at or below 19 Kilowatts for the NONROAD Emissions Inventory Model," Memorandum
from Phil Carlson to Docket EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, March 6, 2000. Docket Identification EPA-HQ-OAR-2004-
0008-0543.
4. "Phase 3 Technology Mix, Emission Factors, and Deterioration Rates for Spark-Ignition Nonroad Nonhandheld
Engines at or below 19 Kilowatts for the NONROAD Emissions Inventory Model," Memorandum from Phil Carlson
to Docket EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of Transportation and Air
Quality, Ann Arbor, Michigan, May 21, 2008. Docket Identification EPA-HQ-OAR-2004-0008-0842.
5. "Exhaust Emission Factors for Nonroad Engine Modeling; Spark-Ignition, Report No. OlOe," U.S. Environmental
Protection Agency, Office of Transportation and Air Quality, Ann Arbor, Michigan, EPA420-R-05-019, December
2005. Docket Identification EPA-HQ-OAR-2004-0008-0398.
6. "Updates to Technology Mix, Emissions Factors, Deterioration Rates, Power Distribution, and Fuel Consumption
Estimates for SI Marine Engines in the NONROAD Emissions Inventory Model," Memorandum from Michael
Samulski to Docket EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of Transportation
and Air Quality, Ann Arbor, November 30, 2005. Docket Identification EPA-HQ-OAR-2004-0008-0361.
7. "Nonroad Evaporative Emission Rates, Report No. 012c," U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, December 2005. Docket Identification EPA-HQ-OAR-2004-
0008-0362.
8. "Modeling of Ethanol Blend Effects on Nonroad Fuel Hose and Tank Permeation - Updated," Memorandum from
Craig Harvey to Docket EPA-HQ-OAR-2004-0008, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, February 22, 2008. Docket Identification EPA-HQ-OAR-
2004-0008-0698.
9. "Domenici-Barton Energy Policy Act of 2005," 109th Congress, U.S. House of Representatives and U.S. Senate,
July 26, 2005. Docket Identification EPA-HQ-OAR-2004-0008-0268.
10. "Annual Energy Outlook 2007; With Projections to 2030, Table 11: Liquid Fuels Supply and Disposition,"
Energy Information Administration, U.S. Department of Energy, Report # DOE/EIA-0383(2007), 2007.
Http://www.eia.doe.gov/oiaf/archive/aeo07/pdf/aeotab_l 1 .pdf. Docket Identification EPA-HQ-OAR-2004-0008-
0695.
11. "Annual Energy Outlook 2007; With Projections to 2030, Table 17: Renewable Energy, Consumption, by
Sector and Source," Energy Information Administration, U.S. Department of Energy, Report #
DOE/EIA-0383(2007), 2007. Http://www.eia.doe.gov/oiaf/archive/aeo07/pdf/aeotab 17.pdf. Docket Identification
EPA-HQ-OAR-2004-0008-0696.
3-53
-------
Final Regulatory Impact Analysis
12. "Fuel and Oil Hoses," Recommended Practice J30, Society of Automotive Engineers, June 1998. Docket
Identification EPA-HQ-OAR-2004-0008-0176.
13. "Personal Watercraft Fuel Systems," Recommended Practice J2046, Society of Automotive Engineers, January,
19, 2001. Docket Identification EPA-HQ-OAR-2004-0008-0179.
14. "Marine Fuel Hoses," Recommended Practice J1527, Society of Automotive Engineers, Revised February 1993.
Docket Identification EPA-HQ-OAR-2004-0008-0177.
15. "Handheld Fuel Lines," email from William Guerry, representing Outdoor Power Equipment Institute, to Glenn
Passavant and Mike Samulski, U.S. Environmental Protection Agency, Office of Transportation and Air Quality,
Ann Arbor, Michigan, July 6, 2005. Docket Identification EPA-HQ-OAR-2004-0008-0126.
16. Letter from Jim Hardin, Grady-White Boats, to Phil Carlson, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, July 25, 2006. Docket Identification EPA-HQ-OAR-2004-
0008-0437.
17. "Permeation of Gasoline and Gasoline-Alcohol Fuel Blends Through High-Density Polyethylene Fuel Tanks
with Different Barrier Technologies," D. Kathios and R. Ziff, SAE Paper 920164, 1992, Docket Identification
EPA-HQ-OAR-2004-0008-0172.
18. "List of 177 Pollutants in the Assessment," National-Scale Air Toxics Assessment for 1999, U. S.
Environmental Protection Agency, 2006. Http://www.epa.gov/ttn/atw/natal999/177poll.html.
19. "Comparison of Small Nonroad and Marine Spark-Ignition Emissions to Stationary and Other Mobile Source
Emission Inventories," Memorandum and attachment from Richard Wilcox, U.S. Environmental Protection Agency,
Office of Transportation and Air Quality, Ann Arbor, Michigan, March 12, 2007. Docket Identification EPA-HQ-
OAR-2004-0008-0540.
20. "Technical Support Document: Preparation of Emissions Inventories for the 2002-Based Platform, Version 3,
Criteria Pollutants," U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Point, North Carolina, January 2008.
Http://www.epa.gov/scram001/reports/Emissions%20TSD%20Voll_02-28-08.pdf. Docket Identification EPA-HQ-
OAR-2004-0008-0699.
21. "Regulatory Impact Analysis: Control of Emissions or Air Pollution from Locomotives and Marine
Compression-Ignition Engines Less Than 30 Liters per Cylinder," U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Ann Arbor, Michigan, May 2008. EPA Report Number: EPA420-R-08-001a.
22. "Control of Emissions From New Marine Compression-Ignition Engines at or Above 30 Liters per Cylinder,"
Advanced Notice of Proposed Rulemaking, U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, Ann Arbor, Michigan, 72 Federal Register 69521, December 7, 2007.
23. "Control of Hazardous Air Pollutants from Mobile Sources: Regulatory Impact Analysis," U.S. Environmental
Protection Agency, Office of Transportation and Air Quality, Ann Arbor, Michigan, February 2007. EPA Report
Number: EPA420-R-07-002.
24. "Electronic Media Supporting Development of Air Quality Modeling Emissions Inventories for the Small
Nonroad and Marine Spark-Ignition Engine Final Rulemaking," Memorandum and attachments from Harvey
Michaels, U.S. Environmental Protection Agency, Office of Transportation and Air Quality, Ann Arbor, Michigan,
July 22, 2008. Docket Identification EPA-HQ-OAR-2004-0008-0694.
3-54
-------
-------
Feasibility of Exhaust Emission Control
CHAPTER 4: Feasibility of Exhaust Emission Control
Section 213(a)(3) of the Clean Air Act presents statutory criteria that EPA must evaluate
in determining standards for nonroad engines and vehicles including marine vessels. The
standards must "achieve the greatest degree of emission reduction achievable through the
application of technology which the Administrator determines will be available for the engines
or vehicles to which such standards apply, giving appropriate consideration to the cost of
applying such technology within the period of time available to manufacturers and to noise,
energy, and safety factors associated with the application of such technology." This chapter
presents the technical analyses and information that form the basis of EPA's belief that the
exhaust emission standards are technically achievable accounting for all the above factors.
The exhaust emission standards for Small SI engines and Marine SI engines are
summarized in the Executive Summary. This chapter begins with a current state of technology
for spark-ignition (SI) engines and the emission control technologies expected to be available for
manufacturer and continues with a presentation of available emissions data on baseline
emissions and on emission reductions achieved through the application of emission control
technology. In addition, this chapter provides a description of new test procedures including
not-to-exceed requirements.
4.1 General Description of Spark-Ignition Engine Technology
The two most common types of engines are gasoline-fueled engines and diesel-fueled
engines. These engines have very different combustion mechanisms. Gasoline-fueled engines
initiate combustion using spark plugs, while diesel fueled engines initiate combustion by
compressing the fuel and air to high pressures. Thus these two types of engines are often more
generally referred to as "spark-ignition" and "compression-ignition" (or SI and CI) engines, and
include similar engines that use other fuels. SI engines include engines fueled with liquid
petroleum gas (LPG) and compressed natural gas (CNG).
4.1.1 Basics of Engine Cycles
Spark ignition engines may be of two-stroke or four-stroke which refers to the number of
piston strokes per combustion cycle. Handheld Small SI equipment typically use two-stroke
engines while larger non-handheld equipment use four-stroke engines. Outboard and personal
watercraft (OB/PWC) engines, until the advent of recent environmental regulations, were
generally two-stroke engines. They are now a mix of two- and four-stroke engines. Sterndrive
and inboard (SD/I) engines are primarily SI four-stroke engines.
4.1.1.1 Two-Stroke Engines
"Two-stroke" refers to the number of piston strokes per combustion cycle. These two
strokes, compression and expansion, occur in one revolution of the crankshaft. During the
4-1
-------
Final Regulatory Impact Analysis
expansion stroke the piston moves downward. As the piston nears its lowest position, the intake
and exhaust ports are opened. While these ports are open, a fresh charge of fuel and air is
pushed into the cylinder which, in turn, helps force the burned gases from the previous cycle out
of the exhaust port. During the compression stroke, the intake and exhaust ports close and the
fresh charge is compressed. As the piston approaches its highest position, a spark-plug ignites
the fresh charge to generate combustion. The force from the combustion acts on the piston to
move it downward, thereby causing the expansion stroke and generating power.
In traditional two-stroke engine designs, the engines are crankcase-scavenged and
carbureted with intake and exhaust ports on the cylinder walls. The advantage of this engine
design is simplicity (low number of moving parts) and a high power to weight ratio of the
engine. In this design, the carburetor meters fuel into the intake air which is then routed to the
crankcase. The motion of the drive shaft then pressurizes the charge. Oil is typically blended
into the fuel to provide cylinder and reciprocating assembly lubrication. When the piston lowers,
it exposes the intake port on the side of the cylinder wall which allows the pressurized fuel/air
charge to enter the cylinder. At the same time, the exhaust port is exposed allowing burned
gases to escape the cylinder. Because both ports are open at the same time, some of the fresh
charge can exit the exhaust port. These fuel losses are known as "short-circuiting" or
"scavenging" losses and can result in 25 percent or more of the fuel passing through the cylinder
unburned. As the piston moves up, the intake and exhaust ports are covered and combustion is
initiated.
An emerging technology for reducing emissions and scavenging losses from two-stroke
engines is direct-injection. This is used primarily on larger outboard and personal watercraft
engines (37 kW and up) to meet exhaust emission standards. In a direct-injected engine, charge
air is used to scavenge the exhaust gases. Once the exhaust valve closes, fuel is injected into the
charge air and ignited with a spark-plug. Because the exhaust valve is closed during most or all
of the injection event, short-circuiting losses are minimized. Also, because the fuel is not used to
lubricate the crankcase, oil does not need to be blended into the fuel. As a result, much less oil is
used.
4.1.1.2 Four-Stroke Engines
Four-stroke engines are used in many different applications. Virtually all gasoline-
powered highway motorcycles, automobiles, trucks and buses are powered by four-stroke
engines. Four-stroke engines are also common in off-road motorcycles, all-terrain vehicles
(ATVs), boats, airplanes, and numerous nonroad applications such as lawn mowers, lawn and
garden tractors, and generators, pressure washers and water pumps to name just a few.
A "four-stroke" engine gets its name from the fact that the piston makes four passes or
strokes in the cylinder to complete an entire cycle. The strokes are intake, compression,
expansion or power, and exhaust. Two of the strokes are downward (intake & expansion) and
two of the strokes are upward (compression & exhaust). The four strokes are completed in two
revolutions of the crankshaft. Valves in the combustion chamber open and close to route gases
into and out of the combustion chamber or create compression.
4-2
-------
Feasibility of Exhaust Emission Control
Figure 4.1-1: 4-Stroke Cycle
2, Compression 3. Expansion
4, Exhaust
The first step of the cycle is for an intake valve to open during the intake stroke allowing
a mixture of air and fuel to be drawn into the cylinder while an exhaust valve is closed and the
piston moves down the cylinder. The piston moves from top dead center (TDC) or the highest
piston position to bottom dead center (BDC) or lowest piston position. This displacement of the
piston draws air and fuel past the open intake valve into the cylinder.
During the compression stroke, the intake valve closes and the momentum of the
crankshaft moves the piston up the cylinder from BDC to TDC, compressing the air and fuel
mixture. As the piston nears TDC, at the very end of the compression stroke, the air and fuel
mixture is ignited by a spark plug and the air and fuel mixture begins to burn. As the air and fuel
mixture burns, pressures and temperatures increase and the products of combustion expand in the
cylinder, which causes the piston to move back down the cylinder, transmitting power to the
crankshaft during the expansion or power stroke. Near the bottom of the expansion stroke, an
exhaust valve opens and as the piston moves back up the cylinder, exhaust gases are pushed out
through the exhaust valve to the exhaust manifold to complete the exhaust stroke, finishing a
complete four-stroke cycle.
4.1.2 Exhaust Emissions from Nonroad SI Engines
Hydrocarbon (HC) and carbon monoxide (CO) emissions are products of incomplete
combustion. The level of CO exhaust emissions is primarily a function of the air-to-fuel ratio at
which an engine is operated. Hydrocarbon emissions formation mechanisms are somewhat more
complex, and appear to be primarily related to:
1. Quenching of the air/fuel mixture at the walls of the combustion chamber
2. Filling of crevice volumes with the air/fuel mixture that remains unburned due to flame
quenching at the entrance to the crevice
3. Lubricant absorption and desorption of fuel compounds
4. Partial combustion during an operating cycle or even complete misfiring of the air/fuel
mixture during the cycle
5. Entrainment and incomplete combustion of lubricant
4-3
-------
Final Regulatory Impact Analysis
As a result, a number of design and operational variables have an impact on HC
emissions, including air-to-fuel ratio; combustion chamber design and geometry; homogeneity of
the air/fuel charge; intake port geometry and the degree of induced air/fuel charge motion;
ignition energy, dwell, and timing; the effectiveness of the cooling system; and oil consumption.
NOx emissions from SI engines are primarily emissions of nitric oxide (NO). Nitrogen
in the intake air reacts with oxygen at high temperatures primarily via the Zeldovich mechanism
to form NO. Thus variables that impact combustion temperatures can have a significant impact
on NO formation and NOx exhaust emissions. These include air-to-fuel ratio, spark timing and
the quantity of residual exhaust gases carried over between engine firing cycles (either through
external exhaust gas recirculation or inefficient cylinder scavenging).
Particulate matter (PM) emissions from SI engines consists primarily of semi-volatile
organic compounds from the engine lubricant together with elemental-carbon soot formed from
pyrolysis of fuel and lubricant during combustion.
4.1.2.1 Air-to-fuel ratio
The calibration of engine air-to-fuel ratio affects torque and power output, fuel
consumption (often indicated as Brake Specific Fuel Consumption or BSFC), engine
temperatures, and emissions for SI engines. The effects of changing the air-to-fuel ratio on
emissions, fuel consumption and torque (indicated as Brake Mean Effective Pressure or BMEP,
which is torque corrected for engine volumetric displacement) are shown in Figure 3-1.l
In the past, manufacturers have calibrated fuel systems of nonroad SI engines for rich
operation. This was done in part to reduce the risk of lean misfire due to imperfect mixing of the
fuel and air and variations in the air-fuel mixture from cylinder to cylinder. Rich operation at
between approximately 12.5:1 and 13:1 air-to-fuel ratio also generally increased engine torque
output (figure 4.1-1) and prevented lean air-to-fuel ratio excursions during application of
transient loads to the engine. Rich operation also has been used to reduce piston, combustion
chamber, cylinder and exhaust port temperatures, thus reducing the thermal load on the cooling
system, a particularly important issue with air-cooled engines. Operation at air-to-fuel ratios
richer than approximately 13:1 or 13.5:1 can limit the effectiveness of, or pose design challenges
for, post-combustion catalytic exhaust emission controls for HC and CO emissions but work well
for catalytic reduction of NOx. At the same time, because a rich mixture lacks sufficient oxygen
for complete combustion, it results in increased fuel consumption rates and higher HC and CO
emissions.
As can be seen from the figure, the best fuel consumption rates occur when the engine is
running lean of the stoichiometric air-to-fuel ratio (approximately 14.6:1 air-to-fuel ratio for
typical gasolines), but lean operational limits are bounded by the onset of abnormal combustion
(e.g., lean misfire and combustion knock), the ability to pick up load, and exhaust port
temperatures (particularly with air-cooled engines). Many air-cooled engines are limited by
heat-rejection to operation that starts approximately at stoichiometry for light loads, and is rich
4-4
-------
Feasibility of Exhaust Emission Control
of stoichiometry as load is increased.
With the use of more advanced fuel systems, manufacturers would be able to improve
control of the air-fuel mixture in the cylinder. This improved control allows for leaner operation
that is closer to a stoichimetric air-to-fuel ratio without increasing the risk of abnormal
combustion. This can be enhanced through careful selection of intake port geometry and
combustion chamber shape to induce turbulence into the air/fuel cylinder charge. The leaner air-
to-fuel ratios (e.g., operating just rich of stoichiometry) resulting from advanced fuel systems
and intake charge turbulence can significantly reduce HC and CO emissions and fuel
consumption, and can provide more oxygen in the exhaust for improved catalytic control of HC
and CO. Leaner air-to-fuel ratios, however, can increase NOx emissions due to higher
combustion temperatures, particularly for engines that are not equipped with exhaust catalysts.
More advanced fuel systems would allow tailoring of the air to fuel ratio to allow good transient
response and to add enrichment at higher load conditions for engine and catalyst protection and
to reduce engine-out NOx emissions. High-load enrichment is particularly important for air-
cooled engines, since high-load operation at leaner air-to-fuel ratios could also increase
hydrocarbon emissions and PM emissions if the higher cylinder temperatures encountered result
in a significant increase in cylinder-bore distortion and lubricating oil consumption.
Figure 4.1-2: Effects of Air-to-Fuel Ratio on Torque Output, Fuel Consumption and
Emissions for Naturally Aspirated Spark Ignition Engines.
4.1.2.2 Spark-timing
For each engine speed and air-fuel mixture, there is an optimum spark-timing that results
in peak torque ("Maximum Brake Torque" or "MET" timing). If the spark is advanced from
MET, more combustion occurs during the compression stroke. If the spark is retarded from
MET, peak cylinder pressure is decreased because too much combustion occurs later in the
expansion stroke generating less useable torque. Timing retard may be used as a strategy for
reducing NOx emissions, because it suppresses peak cylinder temperatures that lead to high NOx
levels. Timing retard also results in higher exhaust gas temperatures, because less mechanical
work is extracted from the available energy. This may have the benefit of warming catalyst
material to more quickly reach the temperatures needed to operate effectively during light-load
operation.2 Some automotive engine designs rely on timing retard at start-up to reduce cold-start
4-5
-------
Final Regulatory Impact Analysis
emissions.
Advancing the spark-timing at higher speeds gives the fuel more time to burn. Retarding
the spark timing at lower speeds and loads avoids misfire. With a mechanically controlled
engine, a fly-weight or manifold vacuum system adjusts the timing. Mechanical controls,
however, limit the manufacturer to a single timing curve when calibrating the engine. This
means that the timing is not completely optimized for most modes of operation.
4.1.3 Marinization
Gasoline sterndrive and inboard (SD/I) engines are generally derived from land-based
counterparts. Engine marinizers buy automotive engine blocks and modify them for use on
boats. Because of the good power/weight ratio of gasoline engines, most SD/I engines are not
modified to produce more power than the base engines were originally designed to produce. In
some airboat applications, aircraft engines are used.
4.1.3.1 Typical SD/I marinization process
Marine SI engines are typically built from base engines designed for use in cars and
trucks. Currently, the vast majority of base engines are General Motor (GM) engines that range
in size from a 3.0 L in-line four cylinder engine to an 8.1 L V8 engine and range in power from
about 100 to 300 kW. These engines are sold without front accessory drives or intake and
exhaust manifolds. Also, no carbureted versions of these engines are offered; they are either
sold with electronic fuel injection, or no fuel system at all. Relatively small numbers of custom
blocks and Mazda rotary engines are also used.
Marinizers convert the base engines into marine engines in the following ways:
- Choose and optimize the fuel management system.
- Configure a marine cooling system.
- Add intake and exhaust manifolds, and accessory drives and units.
Fuel and air management: Historically, Marine SI engines have been carbureted. Today
this technology seems to be going away but is still offered as cheaper alternative to electronic
fuel injection. Less than half of new engines are sold with carburetors. GM does not offer
carburetors or their associated intake manifolds because they are not used in the higher volume,
automotive applications. Therefore, marinizers who produce carbureted engines must purchase
the fuel systems and intake manifolds elsewhere.
The 3.0 L and 4.3 L base engines are offered with throttle body fuel injection systems as
an option. All of the larger engines are offered with multi-port fuel injection as an option.
Although GM offers a base marine calibration for its electronic control module, it also offers
software allowing marinizers to perform their own engine calibrations. For most engines sold,
the marinizers will alter the calibrations to optimize engine operation. Except for some small
market niches, the marinizers do not calibrate the engines for more power.
4-6
-------
Feasibility of Exhaust Emission Control
Cooling system: Marine SI engines are generally packaged in small compartments
without much air flow for cooling. In addition, Coast Guard safety regulations require that
surface temperatures be kept cool on the engine and exhaust manifold. Typically, marine
exhaust systems are designed with surface temperatures below 93°C (200°F). To do this,
manufacturers use ambient (raw) water to cool the engine and exhaust. Most sterndrive and
inboard engines use raw water to cool the engine. This water is then used, in a water jacket, to
cool the exhaust manifold. Finally, the water is dumped into the exhaust stream.
Most Marine SI engines are cooled with raw water. This means that ambient water is
pumped through the engine, to the exhaust manifold, and mixed with the exhaust. The
exhaust/water mixture is then dumped under water. Mixing the water with exhaust has three
advantages:
- cools the exhaust and protects rubber couplings in sterndrives
- acts as a muffler to reduce noise
- helps tune the exhaust back pressure
An alternative to raw water cooling is fresh water cooling. In a fresh water system, raw
water is used to cool the recirculated engine coolant ("fresh water"). The raw water is generally
still used to cool the exhaust manifold and exits the engine with the exhaust. However, some
systems use the engine coolant to cool the exhaust manifold.
Some gasoline engines, mostly inboards, have fresh water cooling systems which provide
two advantages. 1) Engine corrosion problems are reduced, especially when the boat is used in
saltwater. Fresh water systems keep saltwater, which can be corrosive, out of the engine.
Because salt emulsifies at about 68°C, thermostats in fresh water systems are set around 60-
62°C. 2) Marinizers can achieve much better control of the engine temperature. By reducing
variables in engine operation, combustion can be better optimized.3
There are trade-offs with using a fresh water system. The fresh water system costs more
because of the added pump and heat exchanger. Also, this system is not as efficient for cooling
the engine as pumping raw water directly to the engine
Other additions: As mentioned above, marinizers add intake manifolds to carbureted
engines. As part of the cooling system, marinizers must add water jacketed exhaust manifolds,
pumps, and heat exchangers. SD/I engines may also have larger oil pans to help keep oil
temperatures down. Because of the unique marine engine designs, marinizers also add their own
front accessory drive assembly. Finally, sterndrive engines also must be coupled with the lower
drive unit.
4.1.3.2 High performance SD/I marinization process
There is a niche in the SD/I market where customers are willing to sacrifice engine
durability for a high power to weight ratio. Marinizers who address this niche do so by
increasing the fueling of the engine, optimizing the spark-timing for power, increasing the peak
4-7
-------
Final Regulatory Impact Analysis
engine speed (rpm), and modifying the exhaust manifold for better tuning. In some cases, the
marinizers may actually increase the displacement of the engine by boring out the cylinders.
Other components such as camshafts and pistons may also be modified. Superchargers may also
be added. As an example, GM's largest base engine for this market is rated at 309 kW. One
high performance SD/I engine with a bored cylinder, a high performance fuel injection
calibration, and a supercharger achieves more than 800 kW.
4.1.4 Gaseous Fuels
Engines operating on LPG or natural gas carry compressed fuel that is gaseous at
atmospheric pressure. The technical challenges for gasoline related to an extended time to
vaporize the fuel do not apply to gaseous-fuel engines. Typically, a mixer introduces the fuel
into the intake system. Manufacturers are pursuing new designs to inject the fuel directly into
the intake manifold. This improves control of the air-fuel ratio and the combustion event, similar
to the improvements in gasoline injection technology.
4.2 General Description of Exhaust Emission Control Technologies
HC and CO emissions from spark-ignition engines are primarily the result of poor in-
cylinder combustion. This is intensified in carbureted two-stroke engines with the very high HC
emissions due to short-circuiting losses. Higher levels of NOx emissions are the result of leaner
air-fuel ratios and the resulting higher combustion temperatures. Combustion chamber
modifications can help reduce HC emission levels, while using improved air-fuel ratio and spark
timing calibrations, as discussed in Sections 4.1.2.1 and 4.1.2.2, can further reduce HC emissions
and lower CO emissions. The conversion from carburetor to electronic fuel injection will also
help reduce HC and CO emissions. Exhaust gas recirculation could be used to reduce NOx
emissions. The addition of secondary air into the exhaust can significantly reduce HC and CO
emissions. Finally, the use catalytic converters can further reduce all three emissions.
4.2.1 Combustion chamber design
Unburned fuel can be trapped momentarily in crevice volumes (especially the space
between the piston and cylinder wall) before being released into the exhaust. Reducing crevice
volumes decreases this amount of unburned fuel, which reduces HC emissions. One way to
reduce crevice volumes is to design pistons with piston rings closer to the top of the piston. HC
may be reduced by 3 to 10 percent by reducing crevice volumes, with negligible effects on NOx
emissions.4
HC emissions also come from lubricating oil that leaks into the combustion chamber.
The heavier hydrocarbons in the oil generally do not burn completely. Oil in the combustion
chamber can also trap gaseous HC from the fuel and prevent it from burning. For engines using
catalytic control, some components in lubricating oil can poison the catalyst and reduce its
effectiveness, which would further increase emissions over time. To reduce oil consumption,
manufacturers can tighten tolerances and improve surface finishes for cylinders and pistons,
improve piston ring design and material, and improve exhaust valve stem seals to prevent
4-8
-------
Feasibility of Exhaust Emission Control
excessive leakage of lubricating oil into the combustion chamber.
4.2.2 Fuel injection
Fuel injection has proven to be an effective and durable strategy for controlling emissions
and reducing fuel consumption from highway gasoline engines. Comparable upgrades are also
available for gaseous fuels. This section describes a variety of technologies available to improve
fuel metering.
Throttle-body gasoline injection: A throttle-body system uses the same intake manifold
as a carbureted engine. However, the throttle body replaces the carburetor. By injecting the
fuel into the intake air stream, the fuel is better atomized than if it were drawn through with a
venturi. This results in better mixing and more efficient combustion. In addition, the fuel can be
more precisely metered to achieve benefits for fuel economy, performance, and emission control.
Throttle-body designs have the drawback of potentially large cylinder-to-cylinder
variations with multi-cylinder engines. Like a carburetor, TBI injects the fuel into the intake air
at a single location upstream of all the cylinders. Because the air-fuel mixture travels different
routes to each cylinder, and because the fuel "wets" the intake manifold, the amount of fuel that
reaches each cylinder will vary. Manufacturers account for this variation in their design and may
make compromises such as injecting extra fuel to ensure that the cylinder with the leanest
mixture will not misfire. These compromises affect emissions and fuel consumption.
Port gasoline injection: As the name suggests, port (single cylinder) or multi-port (multi-
cylinder-port) fuel injection means that a fuel injector is placed in close proximity to each of the
intake ports. The intake manifold, if used, flows only air. Sequentially-timed systems inject a
quantity of fuel each time the intake valve opens for each cylinder, but multi-port injection
systems can also be "batch fired" (all injectors pulsed simultaneously on a multicylinder engine)
or continous (e.g., the Bosch CIS automotive systems of the 1970's and 80's). Port injection
allows manufacturers to more precisely control the amount of fuel injected for each combustion
event. This control increases the manufacturer's ability to optimize the air-fuel ratio for
emissions, performance, and fuel consumption. Because of these benefits, multi-port injection is
has been widely used in automotive applications for decades.
Sequential injection has further improved these systems by more carefully timing the
injection event with the intake valve opening. This improves fuel atomization and air-fuel
mixing, which further improves performance and control of emissions.
A newer development to improve injector performance is air-assisted fuel injection. By
injecting high pressure air along with the fuel spray, greater atomization of the fuel droplets can
occur. Air-assisted fuel injection is especially helpful in improving engine performance and
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. On a highway 3.8-liter engine with sequential fuel injection, the air assist was shown
4-9
-------
Final Regulatory Impact Analysis
to reduce HC emissions by 27 percent during cold-start operating conditions. At wide-open-
throttle with an air-fuel ratio of 17, the HC reduction was 43 percent when compared with a
standard injector.5
4.2.3 Exhaust gas recirculation
Exhaust gas recirculation (EGR) has been in use in cars and trucks for many years. The
recirculated gas acts as a diluent in the air-fuel mixture, slowing reaction rates and absorbing
heat to reduce combustion temperatures. These lower temperatures can reduce the engine-out
NOx formation rate by as much as 50 percent.6 HC is increased slightly due to lower
temperatures for HC burn-up during the late expansion and exhaust strokes.
Depending on the burn rate of the engine and the amount of recirculated gases, EGR can
improve fuel consumption. Although EGR slows the burn rate, it can offset this effect with
some benefits for engine efficiency. EGR reduces pumping work of SI engines because the
addition of nonreactive recirculated gases forces larger throttle openings for the same power
output. Because the burned gas temperature is decreased, there is also less heat loss to the
exhaust and cylinder walls. In effect, EGR allows more of the chemical energy in the fuel to be
converted to useable work.7
Electronic EGR control: Many EGR systems in today's automotive applications 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 vehicles in the past, more stringent NOx standards and
emphasis on controlling off-cycle emission levels may require more precise EGR control and
additional EGR during heavy throttle operation to reduce NOx emissions. Automotive
manufacturers now use electronic control of EGR. By using electronic solenoids to directly
open and close the EGR valve or by modulating the vacuum signal to vacuum actuated valves,
the flow of EGR can be precisely controlled.
Stratified EGR. Another method of increasing the engine's tolerance to EGR is to
stratify the reicirculated gases in the cylinder. This stratification allows high amounts of dilution
near the spark plug for NOx reduction while making undiluted air available to the crevices, oil
films, and deposit areas so that HC emissions may be reduced. Stratification may be induced
radially or laterally through control of air and mixture motion determined by the geometry of the
intake ports. Research on a one cylinder engine has shown that stratified EGR will result in
much lower fuel consumption at moderate speed and load (6 percent EGR at 2400 rpm, 2.5 bar
BMEP) while maintaining low HC and NOx emissions when compared to homogeneous EGR.8
For catalyst systems with high conversion efficiencies, the benefit of using EGR becomes
proportionally smaller, although it can offer cost savings by reducing catalyst rhodium loadings.
Including EGR as a design variable for optimizing the engine can add significantly to the
development time needed to fully calibrate the electronic controls of engines or vehicles.
4-10
-------
Feasibility of Exhaust Emission Control
4.2.4 Multiple valves and variable valve timing
Four-stroke engines generally have two valves for each cylinder, one for intake of the air-
fuel mixture and the other for exhaust of the combusted mixture. The duration and lift (distance
the valve head is pushed away from its seat) of valve openings is constant regardless of engine
speed. As engine speed increases, the aerodynamic resistance to pumping air in and out of the
cylinder for intake and exhaust also increases. Automotive engines have started to use two
intake and two exhaust valves to reduce pumping losses and improve their volumetric efficiency
and useful power output.
In addition to gains in volumetric efficiency, four-valve designs allow the spark plug to
be positioned closer to the center of the combustion chamber, which decreases the distance the
flame must travel inside the chamber. This decreases the likelihood of flame-quenching
conditions in the areas of the combustion chamber farthest from the spark plug. In addition, the
two streams of incoming gas can be used to achieve greater mixing of air and fuel, further
increasing combustion efficiency and lowering engine-out emissions.
Control of valve timing and lift take full advantage of the four-valve configuration for
even greater improvement in combustion efficiency. Engines normally use fixed-valve timing
and lift across all engine speeds. If the valve timing is optimized for low-speed torque, it may
offer compromised performance under higher-speed operation. At light engine loads, for
example, it is desirable to close the intake valve early to reduce pumping losses. Variable-valve
timing can enhance both low-speed and high-speed performance with less compromise.
Variable-valve timing can allow for increased swirl and intake charge velocity, especially during
low-load operating conditions where this is most problematic. 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. Automotive engines with valve timing have also replaced external EGR systems with
"internal EGR" accomplished via variable valve overlap, generally with improved EGR rate
control over external systems and improved engine-out NOx emissions.
4.2.5 Secondary air
Secondary injection of air into exhaust ports or pipes 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 or mechanical pump, or by using a passive venturi or check-valve,
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, net oxidizing conditions ahead of the catalyst can bring about an efficient increase in the
exhaust temperature which helps the catalyst to heat up quicker. The exothermic reaction that
occurs is dependent on several parameters (secondary air mass, location of secondary air
injection, engine A/F ratio, engine air mass, ignition timing, manifold and headpipe construction,
etc.), and ensuring reproducibility demands detailed individual application for each vehicle or
engine design.
4-11
-------
Final Regulatory Impact Analysis
Secondary air injection was first used as an emission control technique in itself without a
catalyst, and still is used for this purpose in many highway motorcycles and some off-highway
motorcycles to meet federal and California emission standards. For motorcycles, air is usually
provided or injected by a system of check valves which uses the normal pressure pulsations in
the exhaust manifold to draw in air from outside, rather than by a pump.9
Secondary air injection can also be used in continuous operation with rich-jetted
carbureted engines to a achieve an exhaust chemistry just rich of stoichiometry to improve the
efficiency of 3-way catalysts.10'11
4.2.6 Catalytic Aftertreatment
Over the last several years, there have been tremendous advances in exhaust
aftertreatment systems. Catalyst manufacturers have increased the use of palladium (Pd),
particularly for close-coupled positions in automotive catalyst applications.12 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 advances made in catalyst
technology.13 Current Pd catalysts are capable of withstanding prolonged exposure to
temperatures approaching 1100°C.14 The light-off temperature of these advanced catalysts is in
the range of 250 to 270°C.
There are two types of catalytic converters commonly used: oxidation and three-way.
Oxidation catalysts use platinum and/or palladium to increase the rate of reaction between
oxygen in the exhaust and unburned HC and CO. Ordinarily, this reaction would proceed very
slowly at temperatures typical of engine exhaust. The effectiveness of the catalyst depends on its
temperature, on the air-fuel ratio of the mixture, and on the mix of HC present. Highly reactive
species such as formaldehyde and olefins are oxidized more effectively than less-reactive
species. Short-chain paraffins such as methane, ethane, and propane are among the least reactive
HC species, and are more difficult to oxidize.
Three-way catalysts use a combination of platinum and/or palladium and rhodium. In
addition to promoting oxidation of HC and CO, these metals also promote the reduction of NO to
nitrogen and oxygen. In order for the NO reduction to occur efficiently, an overall rich or
slightly-rich of stoichiometric air-fuel ratio is required. The NOx efficiency drops rapidly as the
ai-fuel ratio becomes leaner than stoichiometric. If the air-fuel ratio can be maintained precisely
at or just rich of stoichiometic, a three-way catalyst can simultaneously oxidize HC and CO and
reduce NOx. The window of air-fuel ratios within which this is possible is very narrow and
there is a trade-off between NOx and HC/CO control even within this window. The window can
be broadened somewhat through the use of oxygen storage components, such as cerium oxide,
within the catalyst washcoating. Cerium oxide also promotes CO and HC removal via steam
reformation with water vapor in the exhaust, and the hydrogen liberated by these reactions
promotes further NOx reduction.
Manufacturers are developing catalysts with substrates that utilize thinner walls in order
to design higher cell density, low thermal mass catalysts for close-coupled applications
4-12
-------
Feasibility of Exhaust Emission Control
(improves mass transfer at high engine loads and increase catalyst surface area). The cells are
coated with washcoat which contain the noble metals which perform the catalysis on the exhaust
pollutants. The greater the number of cells, the more surface area with washcoat that exists,
meaning there is more of the catalyst available to convert emissions (or that the same catalyst
surface area can be put into a smaller volume). Cell densities of 900 cells per square inch (cpsi)
have already been commercialized, and research on 1200 cpsi catalysts has been progressing.
Typical cell densities for conventional automotive catalysts are 400 to 600 cpsi.
There are several issues involved in designing catalytic control systems for the engines
covered by this rule. The primary issues are the cost of the system, packaging constraints, and
the durability of the catalyst. This section addresses these issues.
4.2.6.1 System cost
Sales volumes of recreational vessels are small compared to automotive sales and while
sales of Small SI engines <19kW are similar, the price of equipment is much less than
automotive. Manufacturers therefore have a limited ability to recoup large R&D expenditures
for these applications. For these reasons, we believe it is not appropriate to consider highly
refined catalyst systems that are tailored specifically to nonroad applications. Catalyst
manufacturers have assured us that automotive-type catalysts can easily be built to any size
needed for Small SI and marine applications. We are considering catalyst packaging designs
that do not require the manufactures to incur the costs of reworking the entire exhaust system
and, for Marine SI engines, the lower power unit. The cost of these systems will decrease
substantially when catalysts become commonplace. Chapter 6 describes the estimated costs for
nonroad catalyst systems for Small SI and Marine SI engines.
4.2.6.2 Differences in emission control system application and design by engine
category
One challenge in the use of catalytic control for Small SI and Marine SI engines lies in
acceptable design and packaging of the exhaust catalysts onto a wide variety of different types of
equipment. This section discusses specific issues related to these applications.
4.2.6.2.1 Small SI Class I engines
Class I engines typically are equipped with integral exhaust and fuel systems and are
air-cooled. Significant applications include walk-behind lawn mowers (largest segment),
pressure washers, generator sets and pumps. There are both overhead valve (OHV) and
side-valve (SV) engines used in Class I, but side-valve engines are the predominant type in Class
I, particularly in lawn mower applications. They currently represent about 60 percent of Class I
sales. Exhaust catalyst design for Class I engines must take into account several important
factors that differ from automotive applications:
1. Air-cooled engines run rich of stoichiometry to prevent overheating when under load.
Because of this, CO and HC emissions can be high. Catalyst induced oxidation of a high
4-13
-------
Final Regulatory Impact Analysis
percentage of available reactants in the exhaust in the presence of excess oxygen (i.e.,
lean of stoichiometric conditions) can result in highly exothermic exhaust reactions and
increase heat rejection from the exhaust. For example, approximately 80 to 90 percent of
the energy available from catalyst-promoted exhaust reactions is via oxidation of CO.
2. Air-cooled engines have significant HC and NOx emissions that are typically much
higher on a brake-specific basis than water-cooled automotive engine types. Net heat
available from HC oxidation and NOx reduction at rich of stoichiometric conditions is
considerably less than that of oxidation of CO at near stoichiometric or lean of
stoichiometric conditions due to the much lower concentrations of NO and HC in the
exhaust relative to CO.
3. Most Class I engines do not have 12-volt DC electrical systems to power auxiliaries and
instead are pull start. Electronic controls relying on 12-volt DC power would be difficult
to integrate onto Class I engines without a significant cost increase.
4. Most Class I engines use inexpensive stamped mufflers with internal baffles. Mufflers
are typically integrated onto the engine and may or may not be placed in the path of
cooling air from the cooling fan.
5. The regulatory emission test cycles (A-cycle, B-cycle), manufacturer's durability cycles
and some limited in-use operation data indicate that emissions control should focus
primarily on light and part load operation for the highest volume applications
(lawnmowers).
These factors would lead to exhaust catalyst designs for small engines that should differ
somewhat from those of light duty gasoline vehicle exhaust catalyst designs. Design elements
specific to Class I Phase 3 exhaust catalysts would include:
1. Catalyst substrate volume would be sized relatively small so as to be space-velocity
limited. Catalyst volume for Class I Phase 3 engines would be approximately 18 to 50
percent of the engine cylinder displacement, depending on cell count, engine-out
emission levels, and oil consumption. Catalyst substrate sizes would be compact, with
typical catalyst substrate volumes of approximately 2 to 5 cubic inches. This would
effectively limit mass transport to catalyst sites at moderate-to-high load conditions and
reduce exothermic reactions occurring when exhaust temperature is highest. This is
nearly the opposite of the case of typical automotive catalyst designs. Automotive
catalyst volume is typically 50 to 100 percent of cylinder displacement, with the chief
constraints on catalyst volume being packaging and cold-start light-off performance.
2. Catalyst precious metal loading (Pt-platinum, Pd-palladium, Rh-rhodium) would be kept
relatively low, and formulations would favor NOx and HC selectivity over CO
selectivity. We estimate that typical loading ratios for Phase 3 would be approximately
in the range of 40 to 50 g/ft3 (approximately 50 percent of typical automotive loadings at
light-duty vehicle Tier 2 emission levels) and can be Pt:Rh, Pd:Rh or tri-metallic.
Tri-metallic platinum group metal (PGM) loadings that replace a significant fraction of Pt
with Pd would be less selective for CO oxidation and would also reduce the cost of the
catalyst. Loading ratios would be similar or higher in Rh than what is typically used for
automotive applications (20-25 percent of the total PGM mass in Small SI) to improve
NOx selectivity, improve rich of stoichiometry HC reactions and reduce CO selectivity.
4-14
-------
Feasibility of Exhaust Emission Control
3. Catalysts would be integrated into the muffler design. Incorporating the catalyst into the
muffler would reduce surface temperatures, and would provide more surface area for heat
rejection. This is nearly the opposite of design practice used for automotive systems,
which generally try to limit heat rejection to improve cold-start light-off performance.
The muffler design for Class I Phase 3 engines would have somewhat higher surface area
and somewhat larger volume than many current Class I muffler designs in order to
promote exhaust heat rejection and to package the catalyst, but would be similar to some
higher-end "quiet" Class I muffler designs. Appropriately positioned stamped
heat-shielding and touch guards would be integrated into Class I Phase 3 catalyst-muffler
designs in a manner similar to many Class I Phase 2 mufflers. A degree of heat rejection
would be available via forced convection from the cooling fan, downstream of cooling
for the cylinder and cylinder head. This is the case with many current muffler designs.
Heat rejection to catalyst muffler surfaces to minimize "hot spots" can also be enhanced
internally by turning the flow through multiple chambers and baffles that serve as sound
attenuation within the muffler, similar to the designs used with catalyst-equipped lawn
mowers sold in Sweden and Germany.
4. Many Class I Phase 3 catalysts would include passive secondary air injection to enhance
catalyst efficiency and allow the use of smaller catalyst volumes. Incorporation of
passive secondary air allows halving of catalyst substrate volume for the same catalyst
efficiency over the regulatory cycle. A system for Class I Phase 3 engines would be sized
small enough to provide minimal change in exhaust stoichiometry at high load conditions
so as to limit heat rejection, but would be provide approximately 0.5 to 1.0 points of
air-to-fuel ratio change at conditions of 50 percent of peak torque and below in order to
lower HC emissions effectively in engines operating at air-to-fuel ratios similar to those
of current Class I Phase 2 engines. Passive secondary air systems are preferred.
Mechanical or electrical air pumps are not necessary. Passive systems include stamped
or drawn Venturis or ejectors integrated into the muffler, some of which may incorporate
an air check-valve, depending on the application. Pulse-air injection is also a form of
passive secondary air injection. Pulse air draws air into the exhaust port through a
check-valve immediately following the closure of the exhaust valve. Active secondary
air (air pump) systems were not considered in this analysis since they may be cost
prohibitive for use in Class I applications due to the need for a mechanical accessory
drive or 12-volt DC power.
5. Catalyst durability in side valve engines can be enhanced through two catalyst design
ideas. First, the use of a pipe catalyst upstream of the main catalyst brick can "catch" the
oil in the exhaust thereby limiting the amount seen in the catalyst and thereby catalyst
poisoning. Second, the catalyst brick can be lengthened to allow poisoning to some
degree yet allow for catalyst conversion for the regulatory life of the engine.
6. Class I engines are typically turned off via a simple circuit that grounds the input side of
the ignition coil. Temperature fail-safe capability could, if appropriate, be incorporated
into the engine by installing a bimetal thermal switch in parallel with the ignition
grounding circuit used for turning the engine off. The switch can be of the inexpensive
bimetal disc type in wide-spread use in numerous consumer products (furnaces,
water-heaters, ovens, hair dryers, etc.). To reduce cost, the bimetal switch could be a
non-contact switch mounted to the engine immediately behind the muffler, similar to the
4-15
-------
Final Regulatory Impact Analysis
installation of bimetal sensors currently used to actuate automatic chokes on current
Phase 2 Class I lawn mower engines.
4.2.6.2.2 Small SI Class II engines
Almost all Class II engines are air-cooled. Unlike Class I engines, Class II engines are
not typically equipped with integral exhaust systems and fuel tanks. Significant applications
include lawn tractors (largest segment), commercial turf equipment, generator sets and pumps.
Overhead valve engines have largely replaced side-valve engines in Class II, with the few
remaining side-valve engines certifying to the Phase II standards using emissions credits or
being used in snow thrower type applications where the HC+NOx standards do not apply. Class
II engines are typically built more robustly than Class I engines. They often use cast-iron
cylinder liners, may use either splash lubrication or full-pressure lubrication, employ high
volume cooling fans and in some cases, use significant shrouding to direct cooling air. Exhaust
catalyst design practice for Class II engines will differ depending on the level of emission
control. Class II engine designs are more suitable for higher-efficiency emission control systems
than most Class I engine designs. The design factors are somewhat similar to Class I:
1. Class II engines are mostly air-cooled, and thus must run rich of stoichiometry at high
loads. The ability to operate at air-to-fuel ratios rich of stoichiometry at high load may
be more critical for some Class II engines than for Class I engines due to the longer
useful life requirements in Class II. The larger displacement Class II engines have better
efficiency combustion and some engines incorporate more advanced fuel metering and
spark control than is typical in Class I, in order to meet the more stringent Class II Phase
2 emission standards (12.1 g/kW-hr HC+NOx in Class II versus 16.1 g/kW-hr in Class I).
The heat energy available from CO oxidation is typically somewhat less than the case in
Class I because of slightly lower average emission rates.
2. Class II engines have HC and NOx emissions that are generally in more equal portions,
or have the potential to be, in the total regulated HC+NOx emissions and lower CO
emissions than is the case for Class I engines.
3. Most Class II engines are equipped with 12-volt DC electrical systems for starting.
Electronic controls relying on 12-volt DC power could be integrated into Class II engine
designs. Low-cost electronic engine management systems are extensively used in motor
scooter applications in Europe and Asia. Both Kohler and Honda have introduced Class
II engines in North America that use electronic engine management systems.
4. Class II engines use inexpensive stamped mufflers with internal baffles similar to Class I,
but the mufflers are often not integrated onto the engine design and may be remote
mounted in a manner more typical of automotive mufflers. Class II mufflers are often not
placed in the direct path of cooling air from the cooling fan.
5. As with Class I, the regulatory cycles (A-cycle, B-cycle), manufacturer's durability cycles
and some limited in-use operation data indicate that emissions control should focus
primarily on light and part load operation for the high volume sales of garden tractor
equipment.
Taking these factors into account would point towards exhaust catalyst designs that differ
4-16
-------
Feasibility of Exhaust Emission Control
from those of light duty gasoline exhaust catalysts and differ in some cases from Class I systems.
Elements specific to Class II Phase 3 emission control system design using carburetor fuel
systems would include:
1. Catalyst substrate volume would be sized relatively small so as to be space-velocity
limited. Catalyst volume for Class II Phase 3 engines would be approximately 33-50
percent of the engine cylinder displacement, depending on cell count, engine-out
emission levels, oil consumption and the useful life hours to which the engine's emissions
are certified. Catalyst substrate sizes would be very compact within typical mufflers used
in Class II, with typical catalyst substrate volumes of approximately 8 to 10 cubic inches
(based on sales weighting within useful life categories). This would effectively limit
mass transport to catalyst sites at moderate-to-high load conditions and reduce
exothermic reactions occurring when exhaust temperature is highest.
2. Catalyst precious metal loading would be kept relatively low, and formulations would
favor NOx and HC selectivity over CO selectivity to minimize heat concerns. We
estimate that typical loading ratios for Phase 3 would be approximately in the range of 30
to 50 g/ft3 (approximately 50 percent of typical automotive loadings) and could be Pt:Rh,
Pd:Rh or tri-metallic. Tri-metallic PGM loadings that replace a significant fraction of Pt
with Pd would be less selective for CO oxidation and would also reduce the cost of the
catalyst. Loading ratios would be similar or higher in Rh than what is typically used for
automotive applications (20-25 percent of the total PGM mass in Small SI).
3. Catalysts would be integrated into the muffler design. Incorporating the catalyst into the
muffler would reduce surface temperatures relative to the use of a separate catalyst
component. The catalyst for Class II Phase 3 engines would be integrated into mufflers
that are similar in volume to today's Class II Phase 2 mufflers. Appropriately positioned
stamped heat-shielding and touch guards would be integrated into Class II Phase 3
catalyst-muffler designs in a manner similar to current product. Class II engines typically
have a much higher volume of cooling air available downstream of the cylinder than
Class I engines. Heat rejection from the cylinder and cylinder head increases the
temperature of the cooling air, but it is still sufficiently below the temperature of exhaust
system components to allow its use for forced cooling. Thus a degree of heat rejection
would be available via forced convective cooling of exhaust components via the cooling
fan. However, this would require some additional ducting to supply cooling air to exhaust
system surfaces along with careful layout of engine and exhaust components within the
design of the equipment that it is used to power. Integrated catalyst-mufflers can also use
exhaust energy for ejector cooling (see chapter 6). Heat rejection to catalyst muffler
surfaces to minimize "hot spots" can also be enhanced internally by turning the flow
through multiple chambers and baffles that serve as sound attenuation within the muffler.
4. Some applications may include secondary air injection to enhance catalyst efficiency.
Incorporation of passive secondary air allows halving of catalyst substrate volume for the
same catalyst efficiency over the regulatory cycle. In many cases, this may not be
necessary due to the lower engine-out emissions of Class II engines. In cases where
secondary air is used, it could either be a passive system similar to the previously
described Class I systems, or an active system with an engine driven pump. Pump drive
for active systems could be either 12-volt DC electric or via crankcase pulse, and pump
4-17
-------
Final Regulatory Impact Analysis
actuation could be actively controlled using an electric solenoid or solenoid valve. The
use of active systems is an option but seems unlikely. The most likely control scenario
for Class II would be a combination of engine out emission control, use of a small
catalyst, and no use of secondary air.
Higher catalyst efficiency, considerably lower exhaust emissions levels, and improved fuel
consumption are possible with Class II engines, but heat rejection and safety considerations
might necessitate the use of electronic engine management and open-loop fuel injections
systems. In such a case, the design and integration of the emission control system would more
closely resemble automotive applications with the use of electronic engine management and
larger catalyst volumes with higher precious metal loadings.
4.2.6.2.3Marine SI
Due to the design of marine exhaust systems, fitting a catalyst into the exhaust system
raises unique application issues for many boat/engine designs. Often boat builders will strive to
minimize the space taken up in the boat by the engine compartment. In addition, these exhaust
systems are designed, for safety reasons, to avoid hot surface temperatures. For most Marine SI
engines, the surface temperature is kept low by running raw water through a jacket around the
exhaust system. This raw water is then mixed with the exhaust before being passed out of the
engine. To avoid a major redesign of the exhaust system, the catalyst must be placed upstream
of where the water and exhaust mix. In addition, the catalyst must be insulated and/or water-
jacketed to keep the surface temperatures of the exhaust low.
As discussed later in this chapter,
testing has been performed on prototype
systems where small catalysts have been placed
in the exhaust manifolds of SD/I engines.
Figure 4.2-1 illustrates one installation design.
For outboard engines, this packaging
arrangement would be less straightforward
because of the very short exhaust path between
the cylinder exhaust ports and where the
cooling water and exhaust mix. However, it
may be possible to engineer a packaging
solution for outboards as well similar to that
shown for SD/I in Figure 4.2-1.
Figure 4.2-1: Placement of Marine Catalyst
-------
Feasibility of Exhaust Emission Control
Another issue is maintaining high enough temperatures with a water-jacketed catalyst for
the catalyst to react properly. The light-off temperature of these advanced catalysts is in the
range of 250 to 270°C which was low enough for the catalysts to work effectively in our
laboratory tests. However, it could be necessary for manufacturers to retard the spark timing at
idle and low load for some engines to maintain this minimum temperature in the catalyst.
The matching of the catalyst to the engine may have to be compromised to fit it into the
exhaust manifold. However, significant reductions are still achievable. One study on a 4.3 liter
automotive engine looked at three different Pd-only catalyst displacements. The smallest of
these catalysts had a displacement ratio of 0.12 to 1. The HC+NOx downstream of the catalyst
was measured to be from 1.2 to 2.6 grams per mile, depending on the severity of the catalyst
aging.15 This is equivalent to about 1.5 to 3.2 g/kW-hr based on highway operation.16 This work
suggests that significant reductions are achievable with an "undersized" catalyst. As discussed
later in this chapter, significant reductions in exhaust emissions have been demonstrated for
catalysts packaged in SD/I exhaust systems.
4.2.6.3 Catalyst Durability
Two aspects of marine applications that could affect catalyst durability are thermal load
and vibration. Because the catalyst would be coupled close to the exhaust ports, it would likely
see temperatures as high as 750 to 850°C when the engine is operated at full power. The bed
temperature of the catalyst would be higher due to the reactions in the catalyst. However, even
at full power, the bed temperature of the catalyst most likely would not exceed the exhaust
temperature by more than 50-100°C. In our laboratory testing, we minimized the temperature at
full load by operating the engine with a rich air-fuel mixture. The temperatures seen were well
within the operating range of new Pd-only catalysts which are capable of withstanding prolonged
exposure to temperatures approaching 1100°C.17
In on-highway applications, catalysts are designed to operate in gasoline vehicles for
more than 100,000 miles. This translates to about 4,000-5,000 hours of use on the
engine/catalyst. We estimate that, due to low annual hours of operation, the average useful life
of Small SI and Marine SI engines is only a fraction percent of this value. This suggests that
catalysts designed for automotive use should be durable over the useful life of a Small SI and
Marine SI engines. Use of catalysts in automotive, motorcycle, and hand-held equipment
applications suggests that catalysts can be packaged to withstand the vibration in the exhaust
manifold. As discussed later in this chapter, catalysts have recently been demonstrated, through
in-use testing, to be durable over the useful lives of SD/I marine vessels.
4.2.6.4 Water Reversion
Another aspect of marine applications that could affect catalyst durability is the effect of
water contact with the catalyst. There is concern that, in some designs, water could creep back
up the exhaust passages, due to pressure pulses in the exhaust, and damage the catalyst and
oxygen sensor. This damage could be due to thermal shock from cold water coming into contact
with a hot catalyst or due to salt deposition on the catalyst. One study was performed, using a
4-19
-------
Final Regulatory Impact Analysis
two-stroke outboard equipped with a catalyst, to investigate the effect of water exposure on a
catalyst.18 The results of this study are summarized in Table 4.2-1.
Table 4.2-1: Summary of Marine Catalyst Durability Study
Issue
high catalyst
temperatures
saltwater effects
fresh water effects
thermal shock of hot
catalyst with cold
water
deterioration factor
Investigation
- compared base catalyst to catalysts aged for 10
hrs at 900 and 1050°C
- soaked catalysts in two seawater solutions and
compared to base catalyst
- used intake air with a salt-water mist
- soaked catalyst in fresh water and compared to
base catalyst
- flushed out catalyst with fresh water that was
soaked in saltwater
- as part of the catalyst soaking tests, 900°C
catalysts were soaked in both salt and fresh
water
- operated engine with catalyst for 300 hours of
E4 operation
Result
- little change in conversion efficiency
observed
- large drop in conversion efficiency
observed
- no effect on catalyst
- little change in conversion efficiency
observed
- washing catalyst removes salt and
restores some performance
- no damage to the catalysts was
reported
- 20% loss in conversion efficiency for a
2-stroke engine
The above study on catalysts in marine applications was performed supplemental to an
earlier study.19 The earlier study also showed that immersing the catalysts in saltwater would
hurt the conversion efficiency of the catalyst, but that operating in a marine environment would
not. In addition, this earlier study showed that much of the efficiency loss due to salt on the
catalyst could be reversed by flushing the catalyst with water. This paper also showed that with
the catalyst activated, temperatures at full power were less than at mid power because the space
velocity of the exhaust gases at rated speed was high enough to reduce the conversion efficiency
of the catalyst.
A study of water reversion was performed on a vessel powered by a sterndrive engine.20
However, it was found that the water found in the exhaust system upstream of where the exhaust
and water mix was due to condensation. This condensation was a result of cool surfaces in the
exhaust pipe due to the water-jacketing of the exhaust. This study found that the condensation
could be largely resolved by controlling the exhaust cooling water temperature with a thermostat.
Since that time, data has been collected on a number of catalyst-equipped SD/I vessels operated
either in salt or fresh-water. This data, which showed no significant catalyst deterioration, is
discussed later in this chapter. These engines were designed to prevent water reversion by
placing the catalyst near the engine and away from the water/exhaust mixing point. In addition,
some of the prototype designs used either a water dam or mist barrier to help limit any potential
water reversion.
4-20
-------
Feasibility of Exhaust Emission Control
4.2.7 Advanced Emission Controls
On February 10, 2000, EPA published new "Tier 2" emissions standards for all passenger
vehicles, including sport utility vehicles (SUVs), minivans, vans and pick-up trucks. The new
standards will ensure that exhaust VOC emissions be reduced to less than 0.1 g/mi on average
over the fleet, and that evaporative emissions be reduced by at least 50 percent. Onboard
refueling vapor recovery requirements were also extended to medium-duty passenger vehicles.
By 2020, these standards will reduce VOC emissions from light-duty vehicles by more than 25
percent of the projected baseline inventory. To achieve these reductions, manufacturers will
need to incorporate advanced emission controls, including: larger and improved close-coupled
catalysts, optimized spark timing and fuel control, improved exhaust systems.
To reduce emissions, gasoline-fueled vehicle manufacturers have designed their engines
to achieve virtually complete combustion and have installed catalytic converters in the exhaust
system. In order for these controls to work well for gasoline-fueled vehicles, it is necessary to
maintain the mixture of air and fuel at a nearly stoichiometric ratio (that is, just enough air to
completely burn the fuel). Poor air-fuel mixture can result in significantly higher emissions of
incompletely combusted fuel. Current generation highway vehicles are able to maintain
stoichiometry by using closed-loop electronic feedback control of the fuel systems. As part of
these systems, technologies have been developed to closely meter the amount of fuel entering the
combustion chamber to promote complete combustion. Sequential multi-point fuel injection
delivers a more precise amount of fuel to each cylinder independently and at the appropriate time
increasing engine efficiency and fuel economy. Electronic throttle control offers a faster
response to engine operational changes than mechanical throttle control can achieve, but it is
currently considered expensive and only used on some higher-price vehicles. The greatest gains
in fuel control can be made through engine calibrations - the algorithms contained in the
powertrain control module (PCM) software that control the operation of various engine and
emission control components/systems. As microprocessor speed becomes faster, it is possible to
perform quicker calculations and to increase response times for controlling engine parameters
such as fuel rate and spark timing. Other advances in engine design have also been used to
reduce engine-out emissions, including: the reduction of crevice volumes in the combustion
chamber to prevent trapping of unburned fuel; "fast burn" combustion chamber designs that
promote swirl and flame propagation; and multiple valves with variable-valve timing to reduce
pumping losses and improve efficiency. These technologies are discussed in more detail in the
RIA for the Tier 2 FRM.21
As noted above, manufacturers are also using aftertreatment control devices to control
emissions. New three-way catalysts for highway vehicles are so effective that once a TWC
reaches its operating temperature, emissions are virtually undetectable.22 Manufacturers are now
working to improve the durability of the TWC and to reduce light-off time (that is, the amount of
time necessary after starting the engine before the catalyst reaches its operating temperature and
is effectively controlling VOCs and other pollutants). EPA expects that manufacturers will be
able to design their catalyst systems so that they light off within less than thirty seconds of
engine starting. Other potential exhaust aftertreatment systems that could further reduce cold-
start emissions are thermally insulated catalysts, electrically heated catalysts, and HC adsorbers
4-21
-------
Final Regulatory Impact Analysis
(or traps). Each of these technologies, which are discussed below, offers the potential for VOC
reductions in the future. There are technological, implementation, and cost issues that still need
to be addressed, and at this time, it appears that these technologies would not be a cost-effective
means of reducing nonroad emissions on a nationwide basis.
Thermally insulated catalysts maintain sufficiently high catalyst temperatures by
surrounding the catalyst with an insulating vacuum. Prototypes of this technology have
demonstrated the ability to store heat for more than 12 hours.23 Since ordinary catalysts typically
cool down below their light-off temperature in less than one hour, this technology could reduce
in-use emissions for vehicles that have multiple cold-starts in a single day. However, this
technology would have less impact on emissions from vehicles that have only one or two cold-
starts per day.
Electrically-heated catalysts reduce cold-start emissions by applying an electric current to
the catalyst before the engine is started to get the catalyst up to its operating temperature more
quickly.24 These systems require a modified catalyst, as well as an upgraded battery and
charging system. These can greatly reduce cold-start emissions, but could require the driver to
wait until the catalyst is heated before the engine would start to achieve optimum performance.
Hydrocarbon adsorbers are designed to trap VOCs while the catalyst is cold and unable
to sufficiently convert them. They accomplish this by utilizing an adsorbing material which
holds onto the VOC molecules. Once the catalyst is warmed up, the trapped VOCs are
automatically released from the adsorption material and are converted by the fully functioning
downstream three-way catalyst. There are three principal methods for incorporating an adsorber
into the exhaust system. The first is to coat the adsorber directly on the catalyst substrate. The
advantage is that there are no changes to the exhaust system required, but the desorption process
cannot be easily controlled and usually occurs before the catalyst has reached light-off
temperature. The second method locates the adsorber in another exhaust pipe parallel with the
main exhaust pipe, but in front of the catalyst and includes a series of valves that route the
exhaust through the adsorber in the first few seconds after cold start, switching exhaust flow
through the catalyst thereafter. Under this system, mechanisms to purge the adsorber are also
required. The third method places the trap at the end of the exhaust system, in another exhaust
pipe parallel to the muffler, because of the low thermal tolerance of adsorber material. Again a
purging mechanism is required to purge the adsorbed VOCs back into the catalyst, but adsorber
overheating is avoided. One manufacturer who incorporates a zeolite hydrocarbon adsorber in
its California SULEV vehicle found that an electrically heated catalyst was necessary after the
adsorber because the zeolite acts as a heat sink and nearly negates the cold start advantage of the
adsorber. This approach has been demonstrated to effectively reduce cold start emissions.
4.3 Feasibility of Small SI Engine Standards
We are establishing new, more stringent HC+NOx standards for Small SI engines
(<19kW) used in nonhandheld, terrestrial applications (we are also setting a CO std for Small SI
engines used in marine applications that is discussed in Section 4.4). The standards differ by
engine size. Class I engines have a total cylinder displacement of < 225cc. Class II engines
4-22
-------
Feasibility of Exhaust Emission Control
have a total displacement of >225cc. We are also making changes to the emission certification
protocols for durability testing and test fuel specifications for both classes. The new certification
requirements will improve emissions performance of these engines over their regulatory lifetime
and better align the test fuel with in-use fuel characteristics.
Table 4.3-1 shows the existing Phase 2 exhaust emission standards for Class I and II
small spark ignition engines as well as the new Phase 3 standards. The Phase 3 standards
represent a nominal 35-40 percent reduction from current standards.
Table 4.3-1: Comparison of Phase 2 and Phase 3
Standards for Small Spark-Ignition Engines
Engine Class
Class I (<225 cc)
Class II (>225cc)
Phase 2 Standards
(HC+NOx g/kW-hr)
16.1
12.1
Phase 3 Standards
(HC+NOx g/kW-hr)
10.0
8.0
Percent Reduction
(%)
38
34
The following sections present the technical analyses and information that support our
view that the Phase 3 exhaust emission requirements are technically feasible. We begin with a
review of the current state of compliance with the Phase 2 standards relative to the Phase 3
standards and conclude with a more in depth assessment of the technical feasibility of the
requirements for Class I gasoline-fueled engines, Class II single-cylinder gasoline-fueled
engines, Class II multi-cylinder gasoline-fueled engines, and both classes of gaseous-fueled (e.g.,
liquid propane gas) engines.
4.3.1 Current Technology and 2008 Certification Test Data
In the 2008 model year manufacturers certified engines to the Phase 2 standards using a
variety of engine designs and emission control technology. Table 4.3-2 shows manufacturers'
projected engine sales by technology type. For Class I engines, side-valve designs represent the
majority of sales, although there are also a significant number of overhead-valve sales. An
extremely small number of engines used catalyst-based emission control technology. Class II is
dominated by overhead-valve engine designs. A limited number of these engines used catalyst
technology, electronic fuel injection, or were water cooled.
4-23
-------
Final Regulatory Impact Analysis
Table 4.3-2: 2005 Engine Sales by Technology Market Mix
Engine Technology
Side Valve
Overhead Valve
With Catalyst
With Other (Electronic Fuel
Injection and/or water cooled)
Class I
66%
34%
0.003%
0
Class II
2%
98%
0.4%
1%
Looking at the industry from an engine family rather than a sales perspective, shows that
68 and 212 engine families were emission certified in Class I and II, respectively for 2008. The
range of technology types is shown in Table 4.3-3. The majority of engine families in Class I are
overhead-valve, carbureted engines, with only 14 families using side-valve, carbureted designs
(the side-valve engines still account for the bulk of Class I sales). Five families utilized catalytic
exhaust aftertreatment.
Table 4.3-3: 2005 Small Spark-Ignition
Engine Technology Types and Number of Engine Families
Engine
Class
I
II
Side-Valve
Single-
Cylinder
Carburetor
yes (14)
yes (2)
Single-
Cylinder
Carburetor
w. Catalyst
no
yes (0)
Overhead Valve
Single-
Cylinder
Carburetor
yes (48)
yes (43)
Single-
Cylinder
Carburetor
w. Catalyst
yes (4)
no
Multi-
Cylinder
Carburetor
no
yes (105)
Multi-
Cylinder
Carburetor
w. Catalyst
no
yes (6)
Multi-
Cylinder
Fuel
Injection
no
yes (6)
Multi-
Cylinder
Fuel
Injection
w.
Catalyst
no
yes (8)
In Class II, the majority of the engine families use multi-cylinder (predominately v-twins)
designs incorporating overhead-valve technology. Most of these multi-cylinder families utilized
carburetors, with a few using catalytic exhaust aftertreatment, fuel injection, or electronic engine
controls. There are relatively fewer single-cylinder engine families using the older, less
sophisticated side-valve technology. None of these engines were certified with catalytic
aftertreatment.
Figures 4.3-1 and 4.3-2 present the 2008 certification results at full life for Class I and II
4-24
-------
Feasibility of Exhaust Emission Control
engine families, respectively, by technology type.1 One striking feature of these figures, especially
Figure 4.3-2, is that there are a number of engine families displaying emission levels well above
the existing, i.e., Phase 2, standards. Generally, these families represent somewhat older
technology, low production engines that have been certified using preexisting emission credits.
Under the conditions of the final Phase 3 rules, these engine families will be unable to be certified
using existing credits. As a consequence, we expect these families will be eliminated in favor of
newer designs when the Phase 3 standards become effective.
Looking at the remaining engine families, several families were certified at levels
necessary to comply with the Phase 3 standards. Also, a number of families are very close to the
requisite emission levels. This suggests that, even accounting for the relative increase in
stringency associated with our certification protocols, a number of families will either not need to
do anything or require only modest reductions in their emission performance to meet the new
standards.
1 The data presented in Figure 4.3-1 and Figure 4.3-2 are consistent with the 2008 certification data used for the cost
analysis in Chapter 6. This data does not include certification data from the nearly 90 Chinese manufacturers that have started
certifying nonhandheld engines with EPA in the last few years. (As noted in Chapter 6, EPA has chosen not include data from
Chinese manufacturers because we have no information on actual sales of their engines in the United States. Based on
discussions with nonhandheld engine manufacturers that have been certifying with EPA for over ten years now, it is our
understanding that sales of nonhandheld engines from Chinese manufacturers are relatively small at this time.) The certification
levels of the engines certified by Chinese manufacturers generally fall within the same range as the engines presented in Figure
4.3-1 and Figure 4.3-2.
4-25
-------
Final Regulatory Impact Analysis
Figure 4.3-l:Class I HC+NOx Full Life Certification Results for
2008
100
120
140 160
Displacement (cc)
180
200
220
-
1 10
-- _— o
o
? 6
O
I
9
n
(
Figure 4.3-2: Class II HC+NOx Full Life Certification Results for
2008
X
4 ^ Phase 2
^S^ Standard
ff
A*^AA A A *
AA AA/ * A * A ^ A A Phase 3
A AA AAA|AAA A .X'' Standard
..A A A A ^^ ^ ^ ^f
AA A ^^ A AA A^IA ^ A v- *
AAA AAA
^ A *SVCarb
A A OHV Harh
BOHV Fuel Inj
X *£,
X
D 200 400 600 800 1000 1200 1400
Displacement (cc)
4-26
-------
Feasibility of Exhaust Emission Control
4.3.2 Technology Assessment and Demonstration
As described above, a number of engine families already are certified to emission levels
that likely would comply with the Phase 3 standards. However, many engine families clearly will
have to do more to improve emission performance. Generally, we believe the new requirements
will require many engine manufacturers to adopt exhaust aftertreatment technology using
catalyst-based systems. Other likely changes include improved engine designs and fuel delivery
systems. Finally, adding electronic controls or fuel injection systems may obviate the need for
catalytic aftertreatment for some engine families, with the most likely candidates being
multi-cylinder engine designs.
Many of the technical design considerations for adapting advanced emission controls to
Small SI engines were presented in Section 4.2. These included redirected air from the cooling
fan, redirected exhaust flow through multiple chamber and baffles within the catalyst muffler, or
other design considerations. (These are also the kinds of design elements that engine
manufacturers will need to consider for safe and durable emission control systems.) In the
remainder of this section we describe the specific results of our emission control assessment based
on engine testing of exhaust catalyst systems, as well as a more specific discussion of other
potential emission reduction technology for certain engine types such as electronic engine controls
and fuel injection. The results of our safety assessment are described later in section 4.8 of this
chapter.
4.3.2.1 Overview of Technology Assessment
Our feasibility assessment began by evaluating the emissions performance of current
technology for Small SI engines and equipment. These initial efforts focused on developing a
baseline for emissions and general engine performance so that we could assess the potential for
new emission standards for engines and equipment in this category. This process involved
laboratory and field evaluations of the current engines and equipment. We reviewed engineering
information and data on existing engine designs and their emissions performance. We also
reviewed patents of existing catalyst/muffler designs for Class I engines. We engaged engine
manufacturers and suppliers of emission control-related engine components in discussions
regarding recent and expected advances in emissions performance beyond that required to comply
with the current Phase 2 standards. Finally, we purchased catalyst/muffler units that were already
in mass production by an original equipment manufacturer for use on European walk-behind lawn
mowers and conducted engineering and chemical analysis on the design and materials of those
units.
We used the information and experience gathered in the above effort along with the
previous catalyst design experience of our engineering staff to design and build prototype
catalyst-based emission control systems that were capable of effectively and safely achieving the
Phase 3 requirement based on dynamometer and field testing. We also used the information and
the results of our engine testing to assess the potential need for improvements to engine and fuel
system designs, and the selective use of electronic engine controls and fuel injection on some
engine types. A great deal of this effort was conducted in association with our more exhaustive
4-27
-------
Final Regulatory Impact Analysis
study regarding the efficacy and safety of implementing advanced exhaust emission controls on
Small SI engines, as well as new evaporative requirements for these engines.25 In other testing, we
evaluated advanced emission controls on a multi-cylinder Class II engine with electronic fuel
injection.26
In designing our engine testing program, we selected engines certified to the Phase 2
emission standards that were expected to remain compliant with those standards for the duration of
their useful life based on our low-hour emission testing and the manufacturer's declared
deterioration factor from the certification records for that engine family. We also selected engine
families that represented: 1) a cross section of Class I and Class II side-valve and overhead-valve
technologies; and 2) higher sales volume families. Each engine was maintained based on the
manufacturer's specifications.2 The results of our specific technical feasibility assessment are
presented below.
4.3.2.2 Class I Gasoline-Fueled Engines
We tested six side-valve and six overhead-valve Class I engines that used gasoline fuel
with prototype catalyst/muffler control systems. The primary design target for selecting the
catalyst configuration, e.g., volume, substrate, platinum group metal (PGM), was to achieve
emission levels below the limit of 10 g/kW-hr HC+NOx for this class at 125 hours of engine
operation. That time period represents the useful life requirement for the most common
application in this category, i.e., residential walk-behind lawn mowers. A maximum of about 7
g/kW-hr HC+NOx was set as the low-hour performance target with a catalyst system to allow for
engine and emission control degradation over the engine's useful life. This level assumes a
certification cushion at low hours of 1 g/kW-hr HC+NOx and a multiplicative deterioration factor
of 1.3. Secondary design targets were primarily safety related and included minimizing CO
oxidation at moderate to high load conditions to maintain exhaust system surface temperatures
comparable to those of the original Phase 2 compliant systems. The test engine, size, and salient
catalyst features are shown in Table 4.3-4.
Table 4.3-5 presents the results of our catalyst testing on Class I engines.27'28 Three of the
engines were tested at high hours. The high-hour results for the remaining engines were projected
from their low-hour emission performance. We projected high-hour emission results for these
engines by applying the multiplicative deterioration factor from the manufacturer's Phase 2
certification application to the low-hour emission test results. The certification deterioration
factors ranged from 1.097 to 1.302 g/kW-hr HC+NOx.3 As shown, each of the engines achieved
the requisite emission limit of 10 g/kW-hr HC+NOx at the end of their useful lives.
2 The specific test engines were generally used in residential lawn mower and lawn tractor applications.
These applications were chosen for field testing as part of our safety study because they represented certain
potentially unique and challenging safety concerns connected with operation and storage in environments with
combustible debris.
3 These results were taken from the 2005 certification results.
4-28
-------
Table 4.3-4: Class I Test Engine and Control Technology Description
Engine ID
236
246
248
249
6820
258
241
255
2982
243
244
245
Displace-
ment
(L)
0.20
0.20
0.20
0.20
0.19
0.19
0.19
0.19
0.19
0.16
0.16
0.16
Valve
Train
Side
Side
Side
Side
Side
Side
Overhead
Overhead
Overhead
Overhead
Over-head
Overhead
Fuel
Metering
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Carburetor
Passive
(Venturi)
Secondary
Air?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Catalyst Type
Metal monolith
Metal monolith
Metal monolith
Wire-mesh
Cordierite
Ceramic
Monolith
Cordierite
Ceramic
Monolith
Cordierite
Ceramic
Monolith
Coated tube pre-
catalyst, Metal
monolith main-
body catalyst
Metal monolith
Cordierite
Ceramic
Monolith
Metal monolith
Metal monolith
Catalyst
Volume
44 cc
44 cc
44 cc
60 cc
40 cc
40 cc
40 cc
20 mm dia. X 73
mm long exhaust
tubing, 22 cc
metal monolith
34 cc
30 cc
44 cc
44 cc
Catalyst Cell
Density
200 cpsi
200 cpsi
200 cpsi
N/A
400 cpsi
400 cpsi
400 cpsi
Tube: 2 channels
(annular shape),
Main body: 200
cpsi
100 cpsi
400 cpsi
200 cpsi
200 cpsi
PGM Loading
(mass/catalyst volume,
Pt:Pd:Rh ratio)
30 g/ft3, 4:0:1
30 g/ft3, 4:0:1
30 g/ft3, 0.33:3.66:1
proprietary, 0:0:1
30 g/ft3, 5:0:1
30 g/ft3, 5:0:1
30 g/ft3, 5:0:1
Tube: Proprietary
Main body: 30 g/ft3,
3:1:1
50 g/ft3, 5:0:1
30 g/ft3, 5:0:1
30 g/ft3, 1:3:1
30 g/ft3, 3:1:1
-------
Final Regulatory Impact Analysis
Table 4.3-5: Class I Emission Results with Advanced Catalytic Control Technology
Engine
236
246
248
249
6820
258
241
255
2982
243
244
245
Age
(hours)1
10-20
Projected High
10-20
Projected High
10-20
Projected High
10-20
Projected High
Not Tested
>110
10-20
>110
10-20
>110
10-20
Projected High
10-20
>110
10-20
Projected High
10-20
Projected High
10-20
Projected High
HC+NOx
(g/kW-hr)
4.9±0.62
6.1
5.6
7.0
4.6
5.7
6.3
7.8
na
9.4
6.7
8.2
3. 9 ±0.2
6.6 ± 0.2
5.0
6.5
4.9 ±0.3
7.0 ± 0.4
7± 1
7.7
7.2
7.9
5.6
6.1
1 Projected high hour results estimated by multiplying the low hour test results by the
manufacturer's certification deterioration rate.
2 "±" values represent the 95% confidence intervals of 3 tests using a 2-sided t-test.
4-30
-------
Feasibility of Exhaust Emission Control
The above method for projecting high-hour emission results using a certification
deterioration factor assumes that the catalyst system will control engine-out emissions to the same
extent, i.e., proportional reduction, over the useful life of the engine. For some engines this may
not always be the case depending on oil consumption, air-to-fuel ratio and other factors that may
change the effectiveness of the catalyst over time.4 Our approach also did not explicitly account
for the fact that manufacturers will generally design the engine and catalyst to provide some
certification cushion. It appears that most of the engines in Tables 4.3-5 would accommodate the
above design considerations. However, the projected high-hour results are uncomfortably close to
the 10 g/kW-hr HC+NOx standard for engine number 6820. In these cases, such factors can be
accounted for by the engine manufacturer in the engine family's research and design phase by
either improving the durability of the engine (see the discussion below) or designing the catalyst to
account for degradation in catalyst effectiveness over time, e.g, more precious metal loading,
larger catalyst volume, dividing the catalyst into two separate pieces within the exhaust stream,
etc.
The technical feasibility of the Phase 3 standard for Class I engines is supported by a
number of Small SI engine manufacturers.29'30'31'32 Also, a manufacturer of emission controls
specifically indicated the types of hardware that may be needed to comply with new standards.33
That manufacturer concluded that, depending on the application and engine family, either catalyst
or electronic engine controls should be able to achieve emission standards as low as 9 g/kW-hr
HC+NOx. As demonstrated above, we believe the standard of 10 g/kW-hr HC+NOx can be
achieved using catalysts only. However, based on our engineering judgment, we agree that it may
be possible to achieve the standard with the sole use of electronic engine controls because of the
more precise management of air-fuel mixtures and ignition spark timing offered by that
technology.
We conducted a design and process Failure Mode and Effects Analysis study to assess the
safety of implementing advanced exhaust emission controls on Small SI engines.34 That work,
which was based in part on our engine test program, suggests that manufacturers of Class I may
need to improve the durability of basic engine designs, ignition systems, or fuel metering systems
for some engines in order to comply with the emission regulations at full useful life. Some of
these emission-related improvements may include:
1. Adding a fuel filter or improving the needle and seat design in the carburetor to
minimize fuel metering problems caused by debris from the fuel tank;
2. Improving intake manifold design or materials to reduce air leaks;
3. Upgrading the ignition system design for better ignition spark reliability and durability;
4. Improving design and manufacturing processes for carburetors to reduce the production
variability in air-fuel mixtures; and
4 Catalyst performance degradation can occur from thermal sintering and catalyst poisoning due to oil
consumption. Catalyst performance can also improve as engine air-to-fuel ratio slowly drifts towards stoichiometry
over the useful life of the engine. Air-cooled engines are typically designed with air-to-fuel ratio calibrations that
take into account lean-drift with extended operation, and are designed with a sufficiently rich air-to-fuel ratio to
prevent net-lean operation at high hours that could result in engine damage or deteriorating engine performance.
4-31
-------
Final Regulatory Impact Analysis
5. Enhancing exhaust manifold design for better reliability and durability.
4.3.2.3 Class II Single-Cylinder Gasoline-Fueled Engines
Class II single-cylinder engines that use gasoline fuel are currently certified and sold under
the Phase 2 standard in both side-valve and overhead-valve configurations. In 2008, only 2 out of
107 Class II single-cylinder engine families used side-value designs. Manufacturers certified these
families under the averaging provisions of the applicable regulations with emission credits that
were generated by (low emitting) overhead-valve engines. We believe that the Phase 3 standard
will reduce the number of emission credits available for the certification of side-valve technology.
As a result, we assume that a number of the remaining Class II side-valve engines may be phased
out of applicable manufacturer's product line in the future.
Based on the above, we did not directly assess the technical feasibility of the standard for
side-valve Class II engines in our test program. Instead we assessed only single-cylinder,
overhead-valve Class II engines with prototype catalyst/muffler control systems. The primary
design target for selecting the catalyst configuration for these engines, e.g., volume, substrate,
design and PGM loading, was to achieve emission levels well below the limit of 8 g/kW-hr
HC+NOx for this class to accommodate the longer useful life of many of these engines. The
emission regulations allow useful lives ranging from 250 tolOOO hours. For two of the engines
families, we selected emission control technology with a target of meeting a 3.5 g/kW-hr
HC+NOx. This included the use of electronic engine and fuel controls to improve the
management of air-fuel mixtures and ignition spark timing that allow, among other advantages, the
use of larger catalyst volumes and higher precious metal loading. Secondary design targets were
primarily safety related and included minimizing CO oxidation at moderate to high load conditions
to maintain exhaust system surface temperatures comparable to those of the original Phase 2
compliant systems. The test engines, size, salient catalyst parameters, and use of electronic engine
controls are shown in Table 4.3-6.
4-32
-------
Feasibility of Exhaust Emission Control
Table 4.3-6: Class II Single-Cylinder Test Engine and Control Technology Description
Engine
142
231
251
253
254
232
Displac
ement
(L)
0.40
0.50
0.50
0.50
0.59
0.49
Useful
Life
500 hour
250 hour
250 hour
250 hour
250 hour
1,000 hour
Fuel
Metering
Carburetor
Electronic
Fuel
Injection
Carburetor
Carburetor
Carburetor
Electronic
Fuel
Injection
Catalyst
Type
Cordierite
Ceramic
Monolith
Metal
monolith
Cordierite
Ceramic
Monolith
Cordierite
Ceramic
Monolith
Cordierite
Ceramic
Monolith
Metal
monolith
Catalyst
Volume
250 cc
280 cc
250 cc
250 cc
250 cc
250 cc
Catalyst
Cell
Density
400 cpsi
200 cpsi
400 cpsi
400 cpsi
400 cpsi
200 cpsi
Catalyst Loading
40 g/ft3, 5:0:1'
70 g/ft3, 0:5:1
40 g/ft3, 5:0:1
40 g/ft3, 5:0:1
40 g/ft3, 5:0:1
40 g/ft3, 5:0:1
1 Metal loading expressed as a ratio of platinum:palladium:rhodium.
Table 4.3-7 shows the results of our catalyst testing on single cylinder Class II engines.
Only one of the engines was tested at high hours. As explained above for the Class I engines, the
high-hour results for the remaining engines were projected from their low-hour emission
performance. We projected high-time emission results for these engines by applying the
multiplicative deterioration factor from the manufacturer's Phase 2 certification application to the
low-hour emission test results. The certification deterioration factors ranged from 1.033 to 1.240
g/kW-hr HC+NOx.5 As shown, each of the engines achieved the requisite emission limit of 8
g/kW-hr HC+NOx.
5 These results were taken from the 2005 certification results.
4-33
-------
Final Regulatory Impact Analysis
Table 4.3-7: Class II Single-Cylinder Emission Results
with Advanced Catalytic Control Technology
Engine
231 (w/EFI)
232 (w/ EFI)
251
253
254
142
Age
(hours)1
10-40
Projected High
10-40
Projected High
10-40
Projected High
10-40
Projected High
10-40
Projected High
50
500
HC+NOx
(g/kW-hr)
1.8±0.42
2.2
2.2 ±0.1
2.3
3.1 ±.3
3.8
4. 5 ±0.1
5.6
4.0 ±0.3
4.5
2.5 ±0.6
2.8
1 Projected high-hour results estimated by multiplying the low-hour test results by the manufacturer's 2005
certification deterioration rate.
2 "±" values represent the 95% confidence intervals of 3 tests using a 2-sided t-test.
Again, as with Class I engines, the technical feasibility of the Class II standard was
supported by a number of Small SI engine manufacturers.35363738 Also, a manufacturer of emission
controls specifically indicated the types of hardware that may be needed to comply with new
standards.39 That manufacturer concluded that, depending on application and engine family, a
catalyst and electronic engine controls should be capable of achieving emission standards as low
as 7 g/kW-hr HC±NOx. Also, as described above, that same manufacturer concluded that, again
depending on the application and engine family, either catalyst or electronic engine controls
should be able to achieve emission standards as low as 9 g/kW-hr HC±NOx. Our standard of 8
g/kW-hr HC±NOx is in between these two regions. Therefore, based solely on that manufacturer's
conclusions, complying with the standard may require control technology ranging from either a
catalyst or electronic engine controls, or a combination of both.
Based on the above information, especially our testing as discussed previously, we
conclude that catalysts do not necessarily need to be used in conjunction with electronic engine
controls to achieve our standard of 8 g/kW-hr HC±NOx. Either one of those technologies appear
sufficient. In fact, market forces may cause some manufacturers to shift to electronic controls in
the absence of more stringent emission standards. Nonetheless, we can not discount the possibility
that both technologies may be used by some manufacturers to meet the standard on single-cylinder
Class II engines. (See section 4.2.3.4 for more on electronic engine control and fuel injection.)
The design and process Failure Mode and Effects Analysis study mentioned previously
suggests that manufacturers of Class II may need to improve the durability of basic engine designs,
4-34
-------
Feasibility of Exhaust Emission Control
ignition systems, or fuel metering systems for some engines in order to comply with the emission
regulations at full useful life.40 Some of these emission-related improvements may include:
1. Reducing the variability in air-fuel mixtures with tighter manufacturing tolerances for
fuel metering components; and
2. Improving the ignition system design for better ignition spark reliability and durability.
4.3.2.4 Class II Multi-Cylinder Gasoline-Fueled Engines
Gasoline-fueled Class II multi-cylinder engines are very similar to their single-cylinder
counterparts. Beyond the difference in the number of cylinders, several more Class II multi-
cylinder engine families are currently certified with catalysts and electronic engine control
technology (either with or without a catalyst). Because of the direct similarities and the use of
more sophisticated emission control-related technology on some engine families, we find that our
conclusions regarding the technical feasibility of the 8 g/kW-hr HC+NOx standard for single-
cylinder Class II engines is directly transferable to multi-cylinder Class II engines.
Nonetheless, we also tested two twin-cylinder gasoline-fueled Class II engines from
different engine families by the same manufacturer.41 The engines were basically identical except
for their fuel metering systems, i.e., carbureted or electronic fuel injection. We tested both without
modification and tested the electronic fuel injected engine with a catalyst system that we
developed. All the tests were conducted when the engines had accumulated 10-15 total hours of
operating time.
The results of this testing are shown in Table 4.3-8. As was done for the Class I and II
single-cylinder engines discussed earlier, we projected emission levels at the end of each engine's
useful life using the multiplicative deterioration factors for each engine family as reported in the
manufacturer's Phase 2 certification application. As shown, the carbureted engine is projected to
have end of life emissions of approximately 9.1 g/kW-hr. Based on our experience with single-
cylinder engines, compliance with the new standard may require the use of a catalyst for this
engine family. The unmodified engine with electronic fuel injection is projected to achieve about
7.3 g/kW-hr. This engine is very close to complying with the standard and will most likely require
only additional fuel-air mixture and injection timing calibration changes for compliance.
-------
Final Regulatory Impact Analysis
Table 4.3-8: Class II Multi-Cylinder
Emission Results with Advanced Catalytic Control Technology
(V-Twin, Approximately 0.7 Liter Displacement, 3-Way Catalyst)
Engine
Configu
r-ation
OEM
OEM
OEMw/
catalyst
Fuel
Metering
Carburet
EFI
EFI
Age
(hours)1
10-40
Projected
10-40
Projected
10-40
Projected
HC+NO
X
(g/kW-
hr)
7.2
9.1
5.9
7.3
1.8
2.2
Catalyst
Type
—
—
—
—
Cordi erite
same
Catalyst
Volume
—
—
—
—
700cc
same
Catalyst
Cell
Density
—
—
—
—
400 cpsi
same
Catalyst
Loading
—
—
—
—
60 g/ft3,
same
1 Projected high-hour results estimated by multiplying the low-hour test results by the manufacturer's 2004
certification deterioration rate.
2 Metal loading expressed as a ratio of platinum:palladium:rhodium.
Finally, the combination of electronic fuel injection and catalytic exhaust aftertreatment
clearly has the potential to reduce emissions well below the Phase 3 standard as shown in the table.
We also evaluated emission control technology for twin-cylinder Class II engines, and by
analogy all multi-cylinder engines, as part of our safety study.42 Here again we did not find any
unique challenges in designing catalyst-based control systems for these multi-cylinder engines
relative to the feasibility of complying with the exhaust standards under normal engine operation.
However, we did conclude that these engines may present unique concern with the application of
catalytic control technology under atypical operation conditions. More specifically, the concern
relates to the potential consequences of combustion misfire or a complete lack of combustion in
one of the two or more cylinders when a single catalyst/muffler design is used. (A single muffler
is typically used in Class II applications.) In a single-catalyst system, the unburned fuel and air
mixture from the malfunctioning cylinder would combine with hot exhaust gases from the other,
properly operating cylinder. This condition would create high temperatures within the muffler
system as the unburned fuel and air charge from the misfiring cylinder combusts within the
exhaust system. This could potentially destroy the catalyst.
One solution is simply to have a separate catalyst/muffler for each cylinder. Another
solution is to employ electronic engine controls to monitor ignition and either put the engine into
"limp-mode" or shut the engine down until the condition clears on re-start or until necessary
repairs are made, if appropriate. For engines using carburetors, this would effectively require the
addition of electronic controls. For engines employing electronic fuel injection that may need to
also employ a small catalyst, it would require that the electronic controls incorporate ignition
misfire detection if they do not already utilize the inherent capabilities within the engine
management system.
4-36
-------
Feasibility of Exhaust Emission Control
We expect some engine families will use electronic fuel injection to meet the Phase 3
standard without employing catalytic aftertreatment. As described earlier, engine families that
already use these fuel metering systems and are reasonably close to complying with the
requirement are likely to need only additional calibration changes to the engine management
system for compliance. In addition, we expect that some engine families which currently use
carbureted fuel systems will convert directly to electronic fuel injection. Manufacturers may adopt
this strategy to couple achieving the standard without a catalyst and realizing other advantages of
using fuel injection such as easier starting, more stable and reliable engine operation, and reduced
fuel consumption. A few engine manufacturers have confidentially confirmed their plans to use
electronic fuel injection on some engine families in the future as part of an engine management
strategy in lieu of using catalysts.
Our evaluation of electronic fuel injection systems that could be used to attain the new
standard found that a rather simple, low cost system should be sufficient. We demonstrated this
proof of concept as part of the engine test program we conducted for our safety study. In that
program, we fitted two single-cylinder Class II engines with an electronic control unit and fuel
system components developed for Asian motor-scooters and small-displacement motorcycles.
The sensors for the system were minimized to include a throttle position sensor, air charge
temperature sensor, oil temperature sensor, manifold absolute pressure sensor, and a crankshaft
position sensor. This is in contrast to the original equipment manufacturer (OEM) fuel injection
systems currently used in some two-cylinder Class II engine applications that employ more
sophisticated and expensive automotive-based components.
Regarding the electronic control unit and fuel system components referenced above and in
previous sections, at least two small engine manufacturers have developed simplified, compact,
low-cost electronically controlled fuel injection systems for small motorcycles and scooters.4344
One manufacturer has also developed a general purpose small engine with electronic engine speed
control technology that eliminates the need for a battery.45'46 These manufacturers have generally
reported a number of benefits for these advanced systems, including lower emissions and better
fuel economy.
4.3.2.5 Class II Gaseous-Fueled Engines
Engine manufacturers and equipment manufacturers certify engines to run on liquid
propane gas (LPG) or compressed natural gas (CNG) in a number of applications including indoor
floor buffers which require low CO emissions. The technology to reduce emissions to the Phase 3
levels is catalyst due the fact that most engines run closer to stoichiometry than gasoline engines
and further enleanment to reduce emissions may not be feasible. Due to the high amount of NOx
compared with HC, as seen from engine data in the certification database, the catalysts may need
to be designed to reduce NOx and oxidize a limited amount of CO. The EPA 2008 Certification
Database lists 6 multi-cylinder engine families in the Class II 1000 useful life category as having
catalysts. Due to this fact, it is assumed that gaseous engines do not have the same concerns with
multi-cylinder engines and catalysts as gasoline engines.
4-37
-------
Final Regulatory Impact Analysis
4.4 Feasibility of Outboard/Personal Watercraft Marine Engine Standards
Outboard and personal watercraft (OB/PWC) engines are subject to exhaust emission
standards which require approximately a 75 percent reduction in hydrocarbon emissions compared
to conventional carbureted, crankcase-scavenged two-stroke engines. Because of the emission
credit program included in these requirements, manufacturers are able to sell a mix of old and new
technology engines to meet the standards on average.
We are finalizing new exhaust emission standards for OB/PWC engines based on the
emissions results achievable from the newer technology engines. These technologies have
primarily been two-stroke direct injection and four-stroke engine designs. For a few model years,
one manufacturer certified PWC engines with catalytic aftertreatment. This section presents
emission data for 2004 model year outboard and personal watercraft engines and includes a
description of the various emission control technologies used. In addition, the possibility of using
catalytic aftertreatment on OB/PWC engines is discussed.
4.4.1 OB/PWC Certification Test Data
When engine manufacturers apply for certification to exhaust emission standards, they
submit exhaust emission test data. In the case of the OB/PWC engines, the emission standards are
based on the sum of hydrocarbons and oxides of nitrogen (HC+NOx). Manufacturers submit
emission test data on HC and NOx to demonstrate their emission levels. Although carbon
monoxide (CO) emissions are not currently regulated, manufacturers submit data on CO emissions
as well.
Three primary technologies are used on Marine SI engines: conventional two-stroke
engines, direct injection two-stroke engines, and four-stroke engines. Conventional two-stroke
engines are primarily carbureted, but larger engines may have indirect fuel injection systems as
well (IDI). Four stroke engines come in carbureted, throttle-body fuel injected (TBI), and multi-
port fuel injection (MPI) versions. These technologies are discussed in more detail in
Section 4.4.2.
4.4.1.1 HC+NOx Certification Data
Figure 4.4-1 presents HC+NOx certification levels for 2006 model year outboard engines
and compares this data to the existing and new exhaust emission standards. These certification
levels are based on test data over the ISO E4 duty cycle with an adjustment for emissions
deterioration over the regulatory useful life. The certification data set includes engines well above
and below the emission standard. Manufacturers are able to certify to the standard by meeting it
on average. In other words, clean engines generate emission credits which offset the debits
incurred by the engines emitting above the standard. Figure 4.4-2 presents only the data from
engines that meet the 2006 standard. As shown in these figures, two-stroke direct injection
engines and four-stroke engines easily meet the 2006 standard.
4-38
-------
Feasibility of Exhaust Emission Control
Figure 4.4-1: 2006 MY Outboard HC+NOx Certification Levels
f
.c
i
2
X
O
z
+
O
400 -i
"^n
?nn -
9ciO
900
150
100 -
n -
C
•
#
1
^
% »
^ ^ V
*
h . • ^ ._•_ _• • • •. • -
^•^ • ^™w"™«r^ w^ ~^w~~ p^^^^^
) 50 100 150 2(
Rated Power [kW]
DO
• 2s Carb
• 2s Dl
A 4s Carb
• Ac. PFI
'to cn
~ ™ • BasGlinG
2006
Figure 4.4-2: 2006 MY New Technology Outboard HC+NOx Certification Levels
50 100
Rated Power [kW]
150
200
• 2s Carb
• 2s Dl
A 4s Carb
• 4s EFI
^— proposed
Figures 4.4-3 and 4.4-4 present similar data for personal watercraft engines. These engines
4-39
-------
Final Regulatory Impact Analysis
use similar technology, but the HC+NOx emissions are a little higher on average, presumably due
to higher average power densities for PWC engines. This difference in emissions is reflected in
the new HC+NOx standards.
Figure 4.4-3: 2006 MY Personal Watercraft HC+NOx Certification Levels
400
ocn
^ 250
3)
X
9 i*n
Z 150
Q 100
50 -
n
(
1
1
*
^^
J0v ^^V
D 50 100 150 2(
Rated Power [kW]
DO
• 2s Carb
• 2s Dl
^ 2s Catalyst
• Ac PFI
'ts tri
2006
2010
Figure 4.4-4: 2006 MY New Technology PWC HC+NOx Certification Levels
1
2
x
O
O
I
50 100
Rated Power [kW]
150
200
2s Carb
2s Dl
2s Catalyst
4s EFI
• Baseline
•2006
•2010
4-40
-------
Feasibility of Exhaust Emission Control
4.4.1.2 CO Certification Data
Although no exhaust emission standards for CO are currently in place for Marine SI
engines, the technological advances associated with the HC+NOx standards have resulted in lower
CO emissions for many engines. Figures 4.4-5 and 4.4-6 present reported CO exhaust emission
levels for certified outboard and personal watercraft engines. These engines use similar
technology as outboard engines and show similar emission results.
Figure 4.4-5: Reported CO Emission Levels for 2006 MY Outboard Engines
600
500 -
1=1 400
=£ "300
•2
R 200
100
n
(
V
X
•*
» • .
•
•
%•• •
) 50 100 150 2(
Rated Power [kW]
DO
• 2s Carb
• 2s Dl
ZS L/alaiyST
• 4s EFI
2010
Figure 4.4-6: Reported CO Emission Levels for 2006 MY PWC Engines
50 100 150
Rated Power [kW]
200
4-41
-------
Final Regulatory Impact Analysis
4.4.2 OB/PWC Emission Control Technologies
This section discusses the how general technologies discussed above apply to outboard and
PWC applications and discusses specific OB/PWC technology.
4.4.2.1 Conventional Two-Stroke Engines
As discussed earlier in this chapter, hydrocarbon emissions from two-stroke engines are
primarily the result of short-circuiting losses where unburned fuel passes through the engine and
out the exhaust during cylinder charging. Even with an indirect injection system, the air and fuel
are mixed prior to entering the cylinder. Therefore, even though there is better metering of fuel
and air than with a carbureted engine, short-circuiting losses still occur. Because of the very rich
and cool conditions, little NOx is formed. As shown in Figures 4.4-1 and 4.4-2, HC emissions can
range from 100 to 400 g/kW-hr. CO is formed as a product of incomplete combustion. As a
result, CO emissions range from 200 to 500 g/kW-hr from these engines.
4.4.2.2 Direct Injection Two-Stroke Engines
The primary advantage of direct-injection (DI) for a two-stroke is that the exhaust gases
can be scavenged with fresh air and fuel can be injected into the combustion chamber after the
exhaust port closes. As a result, hydrocarbon emissions, fuel economy, and oil consumption are
greatly improved. Some users prefer direct-injection two-stroke engines over four-stroke engines
due to the higher power to weight ratio. Today, this technology is used on engines with power
ratings ranging from 35 to 220 kW. One manufacturer has recently stated its plans to manufacture
DI two-stroke engines as low as 7.4 kW.
Most of the DI two-stroke engines currently certified to the current OB/PWC emissions
standards have HC+NOx emissions levels somewhat higher than certified four-stroke engines.
These engines also typically have lower CO emissions due to the nature of a heterogeneous
charge. By injecting the fuel directly into a charge of air in the combustion chamber, localized
areas of lean air/fuel mixtures are created where CO is efficiently oxidized. PM emissions may be
higher for DI two-stroke engines than for four-stroke engines because oil is burned in the
combustion chamber and because of localized rich areas in the fuel injection stream.
Recently, one manufacturer has introduced a newer technology DI two-stroke engine that
has comparable HC+NOx emission results as many of the certified four-stroke engines.47 This
engine makes use of a low-pressure fuel injection nozzle that relies on high swirl to produce
uniform fuel flow rates and droplet sizes. Also, significant improvements have been made in oil
consumption. As with the older DI two-stroke designs, CO emissions are much lower than
comparable four-stroke engines. What is unique about this design is that the manufacturer has
reported lower PM emissions than for a comparable four-stroke engine.
4-42
-------
Feasibility of Exhaust Emission Control
4.4.2.3 Four Stroke Engines
Manufacturers currently offer four-stroke Marine SI engines with power ratings ranging
from 1.5 to 224 kW. These engines are available with carburetion, throttle-body fuel injection, or
multi-point fuel injection. Carbureted engines are offered from 1.5 to 60 kW while fuel injected
engines are offered from 22 to 224 kW. One manufacturer has stated that the fuel injection
systems are too expensive to use on the smaller engine sizes. Most of the four-stroke outboard
engines above 19 kW have HC+NOx emissions below 16 g/kW-hr and many have emissions
below 13 g/kW-hr. CO emissions for these engines range from 150 to 250 g/kW-hr. Based on the
certification data, whether the engine is carbureted or fuel injected does not have a significant
effect on combined HC+NOx emissions. For PWC engines, the HC+NOx levels are somewhat
higher. However, many of the four-stroke PWC engines are below 16 g/kW-hr. CO emissions for
these engines are similar as those for four-stroke outboards.
4.4.2.4 Catalysts
One manufacturer has certified two PWC
engine models with oxidation catalysts. One engine
model uses the oxidation catalyst in conjunction with
a carburetor while the other uses throttle-body fuel
injection. The engine with throttle-body fuel
injection has an HC+NOx emission rate of 25 g/kW-
hr which is significantly below the EPA 2006
standard. In this application, the exhaust system is
shaped in such a way to protect the catalyst from
water and is nearly as large as the engine (see Figure
4.4-7). Manufacturers have recently begun efforts to
develop a three-way catalyst system for PWC
engines used in jet boats.
Figure 4.4-7: PWC Engine with Catalyst
^K *v
;
Catalysts have not yet been packaged into the
exhaust system of production outboard marine engines. In current designs, water and exhaust are
mixed in the exhaust system to help cool the exhaust and tune the engine. Water often works its
way up through the exhaust system because the lower end in under water and due to pressure
pulses. As discussed above, salt-water can be detrimental to catalyst performance and durability.
In addition, the lower unit of outboards are designed to be as thin as possible to improve the ability
to turn the engine on the back of the boat and to reduce drag on the lowest part of the unit.
Certainly, the success of packaging catalysts in sterndrive and inboard boats in recent development
efforts (see below) suggests that catalysts may be feasible for outboards. However, this has not yet
been demonstrated and significant development efforts would be necessary.
4.5 Feasibility of Sterndrive/Inboard Marine Engine Standards
We are establishing exhaust emission standards for spark-ignition sterndrive and inboard
(SD/I) engines. These new emission standards are supported by data collected on SD/I engines
4-43
-------
Final Regulatory Impact Analysis
equipped with catalysts. This section presents exhaust emission data from baseline SD/I engines
as well as data from SD/I engines equipped with lean calibrations, exhaust gas recirculation, and
catalytic control.
4.5.1 Baseline SD/I Emissions Data
The vast majority of SD/I engines are four-stroke reciprocating piston engines similar to
those used in automotive applications. The exceptions are small sales of air boats using aircraft
piston-type engines and at least one marinizer that uses rotary engines. More than half of the new
engines sold are equipped with electronic fuel injection while the rest still use carburetors. The
majority of the electronic fuel injection systems are multi-port injection; however, throttle-body
injection is also widely used, especially on smaller engines.
Table 4.5-1 presents baseline emissions for four-stroke SD/I engines built up from
automotive engine blocks.48'49'50'51'52'53'54 All these data were collected during laboratory tests over
the ISO E4 duty cycle. Five of these engines are carbureted, one uses throttle-body fuel injection,
and four use multi-port fuel injection. One of the multi-port fuel injected engines was tested with
three calibrations. Note that without emissions calibrations performed specifically for low
emissions, the HC+NOx emissions are roughly equal for the carbureted and fuel injected engines.
Using the straight average, HC+NOx from the carbureted engines is 15.6 g/kW-hr while it is 16.0
g/kW-hr from the fuel injected engines (15.1 g/kW-hr if the low HC calibration outlier is
excluded).
4-44
-------
Feasibility of Exhaust Emission Control
Table 4.5-1: Baseline SD/I Exhaust Emission Data
Engine
#
1
2
3
4
5
6
7
8
9
9
9
10
11
Power
[kW]
79
91
121
153
158
167
196
159
185
181
191
219
229
Fuel Delivery System
carburetor
carburetor
carburetor
multi-port electronic fuel injection
carburetor
carburetor
carburetor
throttle -body fuel injection
multi-port electronic fuel injection
#9, low CO calibration
#9, low HC calibration
multi-port electronic fuel injection
multi-port electronic fuel injection
HC
[g/kW-hr]
11.2
4.4
8.5
4.9
7.3
8.0
4.4
2.9
5.2
5.8
3.3
4.7
2.7
NOx
[g/kW-hr]
8.0
13.9
6.0
11.7
6.0
5.7
10.3
8.7
9.7
11.7
18.2
9.4
13.1
CO
[g/kW-hr]
281
98
247
111
229
174
101
42
149
48
72
160
44
A distinct class of SD/I engines are the high-performance engines. These engines are
similar to SD/I engines except that they are designed for high power output at the expense of
engine durability. This high power output is typically achieved through higher fuel and air rates,
larger combustion chambers, and through higher peak engine speeds. In most cases, custom
engine blocks are used. Even in the engines that use an automotive block, few stock automotive
engine components are used. Table 4.5-2 presents emission data collected by EPA on five high-
performance engines.55'56'57 This data also includes data submitted by a high performance engine
manufacturer in its public comments on the proposed rule.58
4-45
-------
Final Regulatory Impact Analysis
Table 4.5-2: Baseline High Performance SD/I Exhaust Emission Data [g/kW-hr]
Power
[kW]
391
550
634
778
802
466
410
466
Fuel Delivery System
multi-port electronic fuel injection
carburetor
multi-port electronic fuel injection, supercharger
throttle-body fuel-injection, supercharger,
intercooler
multi-port electronic fuel injection, supercharger
electronic fuel injection
electronic fuel injection
electronic fuel injection, low emission
calibration13
HC
14.7
13.23
16.9
7.6
16.1
15.43
14.83
4.3
NOx
3.8
8.4
9.1
4.9
9.4
3.2
3.9
10.8
CO
243
253
135
349
102
257
325
104
BSFC
354
376
348
448
299
~
~
~
3 HC concentration at idle was out of measurement range
b 15% load factor at idle
4.5.2 Exhaust Gas Recirculation Emission Data
We collected data on three engines over the ISO E4 marine test cycle with and without the
use of exhaust gas recirculation (EGR).59'60'61 The first engine was a 6.8 L Ford heavy-duty
highway engine. Although this was not a marine engine, it uses the same basic technology as SD/I
engines. The second and third engines were the 7.4 L and 4.3 L SD/I engines used in the catalyst
development program described below. These engines are marinized versions of GM heavy-duty
highway engines. The baseline emissions from the 7.4 L engine are a little different than
presented below in the catalyst discussion because engine head was rebuilt prior to the catalyst
development work.
This test data suggests that, through the use of EGR on a SD/I marine engine, a 40-50
percent reduction in NOx (30-40 percent reduction in HC+NOx) can be achieved. EGR was not
applied at peak power in this testing because the throttle is wide open at this point and displacing
fresh air with exhaust gas at this mode of operation would reduce power. We also did not apply
EGR at idle because the idle mode does not contribute significantly to the cycle weighted NOx.
4-46
-------
Feasibility of Exhaust Emission Control
Table 4.5-3: Exhaust Emission Data Using EGR on the E4 Marine Duty Cycle
EGR Scenario
6.8 L Engine: baseline
with EGR
7.4 L Engine: baseline
with EGR
4.3 L Engine: baseline
with EGR
HC
[g/kW-hr]
2.7
2.7
4.5
4.5
4.9
4.2
NOx
[g/kW-hr]
13.4
7.1
8.4
4.8
11.7
5.3
CO
[g/kW-hr]
26.5
24.3
171
184
111
92
Power
[kW]
145
145
209
209
153
148
BSFC
[g/kW-hr]
326
360
349
356
329
350
4.5.3 Catalytic Control Emission Data
4.5.3.1 Engine Testing
In a joint effort with the California Air Resources Board (ARE), we contracted with
Southwest Research Institute to perform catalyst development and emission testing on a SD/I
marine engine.62 This test program was performed on a 7.4 L electronically controlled Mercruiser
engine with multi-port fuel injection. Figure 4.5-1 illustrates the three primary catalyst packaging
configurations used in this test program. The upper right-hand picture shows a catalyst packaged
in a riser extension which would be placed between the lower exhaust manifold and the exhaust
elbow. This riser had the same outer dimensions as the stock riser extension produced by Mercury
Marine. The upper left-hand picture shows a catalyst packaged in the elbow. The lower picture
shows a larger catalyst that was packaged downstream of the exhaust elbow. All of these catalyst
configurations were water jacketed to prevent high surface temperatures.
4-47
-------
Final Regulatory Impact Analysis
Figure 4.5-1: Three Catalyst Configurations Used in SD/I Test Program
Table 4.5-4 presents the exhaust emission results for the baseline test and three catalyst
packaging configurations. In each case a pair of catalysts were used, one for each exhaust
manifold. For the riser catalyst configuration, we tested the engine with two cell densities, 60 and
300 cells per square inch (cpsi), to investigate the effects of back-pressure on power. The catalysts
reduced in HC+NOx in the range of 42 to 77 percent and reduced CO in the range of 46 to 54
percent. There were no significant impacts on power, and fuel consumption actually improved due
to the closed-loop engine calibrations necessary to optimize the catalyst effectiveness. At the full
power mode, we left the engine controls in open-loop and allowed it to operate rich to protect the
catalysts from over-heating.
4-48
-------
Feasibility of Exhaust Emission Control
Table 4.5-4: Exhaust Emission Data on a 7.4 L SD/I Engine with Various Catalysts
Catalyst Scenario*
(cell density, volume, location)
baseline (no catalyst)
60 cpsi, 0.7 L, riser
300 cpsi, 0.7 L, riser
400 cpsi, 1.3 L, elbow
200 cpsi, 1.7 L, downstream
HC
[g/kW-hr]
4.7
2.5
1.7
2.8
2.1
NOx
[g/kW-hr]
9.4
5.7
1.9
1.1
1.2
CO
[g/kW-hr]
160
81
87
81
83
Power
[kW]
219
214
213
217
221
BSFC
[g/kW-hr]
357
345
349
337
341
*Multiply volume by two for total catalyst volume per engine.
Additional reductions in HC+NOx and CO can be achieved by using EGR in addition to a
catalyst. However, the added benefit of EGR is small compared to the emission reductions
achieved by the catalysts. Regardless, the use of EGR could give manufacturers some flexibility
in the design of their catalyst. In the catalyst testing work described above on the 7.4 L SD/I
marine engine, each of the catalyst configurations were tested with and without EGR. Table 4.5-5
presents these test results.
Table 4.5-5: Exhaust Emission Data on a 7.4 L SD/I Engine with Catalysts and EGR
Catalyst Scenario*
(cell density, volume, location)
60 cpsi, 0.7 L, riser
300 cpsi, 0.7 L, riser
400 cpsi, 1.3 L, elbow
200 cpsi, 1.7 L, downstream
HC+NOx [g/kW-hr]
catalyst
8.2
3.6
3.9
o o
J.J
catalyst + EGR
6.8
2.8
3.3
2.5
CO [g/kW-hr]
catalyst
81
87
81
83
catalyst + EGR
74
77
76
73
* Multiply volume by two for total catalyst volume per engine.
4.5.3.2 Freshwater Boat Testing
The catalyst testing described above was a first step in developing and demonstrating
catalysts that can reduce emissions from Marine SI engines. However, this program only looked at
catalysts operating in a laboratory. Additional efforts have been made to address issues with using
catalyst in marine applications by operating engines in boats with catalysts. When the California
Air Resources Board finalized their catalyst-based emission standards for SD/I engines, they
agreed to further assessment of the durability of catalyst used in boats through technology review.
To that end, ARB, industry and the U.S. Coast Guard recently performed a cooperative in-
4-49
-------
Final Regulatory Impact Analysis
boat demonstration program designed to demonstrate the feasibility of using catalysts in SD/I
applications.63'64 This testing included four boats, two engine types, and four catalysts. The
catalysts were packaged in the exhaust emission manifold in such a way that they were water-
jacketed and capable of fitting within the existing boat design. Each of the boats were operated by
the U.S. Coast Guard for 480 hours on a fresh water lake. This service accumulation period,
which was intended to represent the useful life of typical SD/I engines, began in December of
2003 and was completed in September of 2004. Table 4.5-6 presents a description of the boats
that were used in the test program.
Table 4.5-6: Vessel Configurations for Full Useful Life Catalyst Testing
Boat
Inboard Straight-Drive Ski Boat
Inboard V-Drive Runabout
22 ft, Sterndrive Bowrider
19 ft. Sterndrive Runabout
Engine
5.7L, V-8
5.7L, V-8
5. 7 L, V-8
4.3 L, V-6
Catalyst
Type
metallic
ceramic
metallic
ceramic
Catalyst
Volume*
1.4 L
1.7L
1.4 L
0.7 L
Catalyst Cell
Density
300 cpsi
400 cpsi
200 cpsi
400 cpsi
*Multiply volume by two for total catalyst volume per engine.
Exhaust emissions were measured for each catalyst before and after the durability testing.65
No significant deterioration was observed on any of the catalysts. In fact, all of the 5.7 L engines
were below the standard of 5 g/kW-hr HC+NOx even after the durability testing. Although the
zero hour emissions for the 4.3 L engine were less than half of the HC+NOx standard, the final
emissions for the 4.3 L engine were 15 percent above the HC+NOx standard. However, it should
be noted that the 4.3L engine was determined to have excessive fuel delivered to one cylinder bank
and low compression in one of the cylinders. These problems did not appear to be related to the
catalyst installations and would account for the increase in emissions even without catalyst
deterioration. Once the calibration on this engine was corrected, a level of 5 g/kW-hr HC+NOx
was achieved. In addition, no deterioration was observed in the oxygen sensors which were
installed upstream of the catalysts.
Significant carbon monoxide emission reductions were achieved, especially at lower power
modes. At wide-open-throttle, the engines operated in open-loop to prevent the exhaust valves
from overheating. Additional reductions in CO could be achieved through better fuel air ratio
control. For instance, although the engines in this test program were fuel injected, batch injections
were used. In other words, all of the fuel injectors for each bank were firing at the same time
rather than timing the fuel injection with the valve timing for each individual cylinder. Because of
this strategy, the engine would need to be calibrated somewhat rich. The next generation of
electronics for these engines are expected to have more sophisticated control which would allow
for optimized timing for each fuel injector.
4-50
-------
Feasibility of Exhaust Emission Control
Table 4.5-7: Vessel Configurations for Full Useful Life Catalyst Testing
Boat
5.7 L engine
4.3 L engine
Inboard Straight-
Drive Ski Boat
Inboard V-Drive
Runabout
22 ft, Sterndrive
Bowrider
19 ft. Sterndrive
Runabout
Catalyst Aging
baseline (no catalyst)
baseline (no catalyst)
0 hours
480 hours
0 hours
480 hours
0 hours
480 hours
0 hours
480 hours*
HC
[g/kW-hr]
5.4
4.9
1.7
2.1
1.8
1.7
1.8
1.5
1.9
2.9
NOx
[g/kW-hr]
6.7
11.7
1.0
1.7
0.5
1.0
0.5
0.9
0.5
2.1
CO
[g/kW-hr]
193
111
100
117
87
102
74
93
106
116
* after calibration corrected
4.5.3.3 Saltwater Boat Testing
Two test programs were initiated to investigate the feasibility of using catalysts on boats
used in saltwater. In the first program, a small boat with a catalyst was operated over a set of
operation conditions, developed by industry, to represent the worst case conditions for water
reversion. In the second test program, three boats were equipped with catalysts and operated for
an extended period similar to the fresh water testing.
4.5.3.3.1 Safety, Durability, and Performance Testing
We contracted with SwRI to test catalysts on a Sterndrive engine before and after operation
on a boat in saltwater.66 The purpose of the testing was to determine if the catalyst would be
damaged by water reversion in the exhaust manifold. This testing was performed on a 19 foot
runabout with a 4.3 L Sterndrive engine. On previous testing on this boat without a catalyst, SwRI
found that the only water collected in the exhaust manifold was due to condensation. They were
able to prevent this condensation by fitting the water jacket around the exhaust system with a
thermostat to keep the manifold walls from becoming too cool.
The 4.3 L engine was fitted with a pair of riser catalysts similar to the one illustrated in
Figure 4.5-1. These catalysts had a cell density of 300 cpsi and a combined volume of 1.4 L. The
catalysts were water-jacketed to maintain low surface temperatures and, to prevent any possible
water reversion, cones were inserted in the exhaust elbows. These cones were intended to increase
the difficulty for water to creep up the inner walls of the exhaust manifold. The water jacketing
system was fitted with a 82°C thermostat to keep the manifold wall temperatures above the dew
4-51
-------
Final Regulatory Impact Analysis
point of the exhaust gas (~50°C) thereby preventing water condensation in the exhaust manifold.
Prior to testing, the catalysts were aged using a rapid aging cycle designed to represent
50,000 miles of vehicle operation. SwRI estimated that this would likely be more severe than
would be seen over the useful life of an SD/I engine. The engine was then tested for emissions, in
a test cell, with and without the aged catalysts installed in the exhaust manifold risers. In addition
to adding the catalysts, the engine fueling was optimized using closed-loop electronic emission
control.
After the baseline emission tests, the catalysts were installed on a 19 foot runabout
equipped with a similar 4.3 L engine used in the emissions test cell. The boat was operated on
saltwater over a number of safety, durability, and performance tests that were developed by
industry for heat soak, water ingestion, and engine exhaust back-pressure. In addition, SwRI
operated the boat over tests that they designed to represent operation and use that would most
likely induce water reversion. After this boat testing, the catalyst was returned to the laboratory
for a repetition of the baseline emission tests.
Table 4.5-8 presents the baseline, aged catalyst, and post boat operation catalyst emission
test results. No significant deterioration of the catalysts were observed. Prior to boat testing, the
aged catalysts achieved a 75 percent reduction in HC+NOx and a 36 percent reduction in CO.
After the boat operation in saltwater, the catalysts achieved a 73 percent reduction in HC+NOx
and a 34 percent reduction in CO. As described in Chapter 3, if saltwater had reached the catalyst,
there would have been a large reduction in catalyst efficiency. No salt deposits were observed on
the catalysts when they were removed from the boat.
Table 4.5-8: Exhaust Emission Data on a 4.3 L SD/I Engine with Catalysts
Catalyst Scenario
open-loop, no catalyst
closed-loop, no catalyst
aged catalyst pre boat
aged catalyst post boat
HC
[g/kW-hr]
4.9
4.5
2.1
2.2
NOx
[g/kW-hr]
11.7
10.4
2.0
2.3
CO
[g/kW-hr]
111
101
70
73
Power
[kW]
153
153
154
150
BSFC
[g/kW-hr]
329
327
321
327
4.5.3.3.2 Extended Period In-Use Testing
We engaged in a test program with the California Resources Board, United States Coast
Guard, National Marine Manufacturers Association, the Texas Department of Parks and Wildlife,
and Southwest Research Institute to evaluate three additional engines with catalysts in vessels
operating on salt-water. Early in the program, two of the three manifolds experienced corrosion in
the salt-water environment resulting in water leaks and damage to the catalyst. These manifolds
were rebuilt with guidance from experts in the marine industry and additional hours were
4-52
-------
Feasibility of Exhaust Emission Control
accumulated on the boats. Although the accumulated hours are well below the 480 hours
performed on fresh water, the completed operation showed no visible evidence of water reversion
or damage to the catalysts. Table 4.5-9 presents initial exhaust emission results for the three
engines, equipped with catalysts, included in this test program.
Table 4.5-9: Baseline Emission Data for Engines/Catalysts in Saltwater Test Program
Catalyst Scenario
Maxum, 4.3L V6, ceramic catalysts
Sea Ray, 5.7L V8, metal catalysts
Malibu, 5.7L V8, ceramic catalysts
HC
[g/kW-hr]
2.1
1.3
0.5
NOx
[g/kW-hr]
0.7
0.3
0.4
CO
[g/kW-hr]
136
114
107
Power
[kW]
150
191
194
BSFC
[g/kW-hr]
345
351
348
4.5.3.4 Production Engines
At the time of proposal, only one manufacturer was selling inboard Marine SI engines
equipped with catalysts. These engines are being sold nationwide. The engines are based on 5.7L
automotive blocks and use electronically controlled fuel injection, twin catalysts, and onboard
diagnostics. The manufacturer, Indmar, has also performed extended durability testing in a
saltwater environment. Test data from this engine is presented in Table 4.5-10, with and without
an applied deterioration factor.67 One advantage that Indmar has promoted with this engine is very
low CO at part throttle. Part throttle operation is associated with lower boat speeds where the risk
of CO poisoning is highest. The measured CO over the marine duty cycle is primarily due to
emissions at wide open throttle, where the engine goes to open loop rich operation to protect the
exhaust valves from overheating.
Table 4.5-10: Exhaust Emission Data on a 5.7L Production SD/I Engine with Catalysts
measured test results
with deterioration factor applied
HC
[g/kW-hr]
1.8
2.0
NOx
[g/kW-hr]
2.0
2.3
CO
[g/kW-hr]
46.6
51.8
At this time, three manufacturers have certified engines, equipped with catalysts, to the 5
g/kW-hr HC+NOx California standards for SD/I engines. Table 4.5-11 presents the certification
data available from the California Air Resources Board's Off-Road Certification Database.68
4-53
-------
Final Regulatory Impact Analysis
Table 4.5-11: Catalyst-Equipped SD/I Engines Certified in California
Manufacturer
Indmar
Mercury Marine
Volvo Penta
Engine Disp.
[liters]
5.7
5.7
1.6
3.0
5.0
5.7
8.1
8.1
5.0
8.1
Rated Power
[kW]
230
230
75
101
194
246
280
317
239
298
HC+NOx
[g/kW-hr]
3.7
4.6
4.2
2.7
1.8
3.4
2.8
4.6
3.3
2.6
4.5.3.5 CO Emissions Reductions at Low versus High Power
Under stoichiometric or lean conditions, catalysts are effective at oxidizing CO in the
exhaust. However, under very rich conditions, catalysts are not effective for reducing CO
emissions. SD/I engines often run at high power modes for extended periods of time. At these
temperatures, engine marinizers must calibrate the engine to run rich as an engine protection
strategy. If the engine were calibrated for a stoichometric air-fuel ratio at high power, high
temperatures could lead to failures in exhaust valves and cylinder heads.
All of the data presented above on SD/I engines equipped with catalysts were based on
engines that used open-loop engine control at high power. As a result, the catalysts achieved little
reduction in HC and CO at full power (test mode 1). However, NOx reductions were achieved at
mode 1 because NOx is effectively reduced under rich conditions.
The catalysts were effective in reducing CO in modes 2 through 5 of the test procedure. In
these lower power modes, the engines described above saw CO reductions on the order of 80
percent. However, the weighted values over the test cycle only show about a 50 percent reduction
in CO because of the high contribution of mode 1 to the total weighted CO value. Studies have
shown that there is a higher risk of operator exposure to CO at lower boat speeds69 which would
correspond to lower engine power modes. This suggests that CO reductions at lower power modes
may be more beneficial than CO reductions at full power.
To look at the effect of mode 1 on the cycle weighted CO levels, we performed an analysis
in which we recalculated the CO level for ten catalyst-equipped SD/I engines without mode 1. To
determine the weighted value without mode 1, the weighting factor for mode 1 was set to zero
percent and the weighting factors for modes 2 and 3 were each increased so that weighting factors
would sum to 100 percent. Figure 4.5-2 compares the CO emissions with and without including
4-54
-------
Feasibility of Exhaust Emission Control
mode 1 for these engines. Although mode 1 is only weighted as 6 percent of the test cycle, but
makes up the majority of the cycle weighted CO value. Based on this analysis, the weighed CO
level would be 70-90 percent lower if mode 1 were not included in the test procedure.
Figure 4.5-2: CO Emissions for SD/I Engines Equipped with Catalysts
with and without Including Mode 1 in the Weighted Results
140
120
100
DE4 Weighting
• Without Mode 1
10
4.6 Feasibility of Standard for Marine Generator Sets
Currently, SI marine generator sets are regulated as Small SI or Large SI engines,
depending on their size. Most SI marine generators are less than 25 hp and are therefore classified
as Small SI engines. Generator sets in marine applications are unique in that they use liquid-
cooled engines. Liquid cooling allows manufacturers to minimize the temperature of hot surfaces
on marine generators, thereby reducing the risk of fires on a boat. For marine applications, liquid
cooling is practical because of the nearly unlimited source of cooling water around the boat.
Another safety issue that has become apparent in recent years is carbon monoxide
poisoning on boats. Studies have shown that exhaust emissions from engines on boats can lead to
user exposure of high levels of carbon monoxide.70 The marine industry, Coast Guard, American
Boat and Yacht Council, and other stakeholders have been meeting regularly over the past several
years in an attempt to mitigate the risk of CO poisoning in boating.71'72 Mitigation strategies that
have been discussed at these meetings include labeling, education, diverting the exhaust flow with
smoke stacks, CO detectors, low CO emission technologies, and emission standards.
The vast majority of gasoline marine generators are produced by two engine
manufacturers. Recently, these two manufacturers have announced that they are converting their
4-55
-------
Final Regulatory Impact Analysis
marine generator product lines over to low CO engines.73'74 They have stated that this is to reduce
the risk of CO poisoning and that this action is a result of boat builder demand. Both
manufacturers are using a combination of closed-loop electronic fuel injection and catalytic
control. To date, both of these manufacturers have certified some low CO engines and have stated
their intent to convert their full product lines in the near future. These manufacturers also make
use of the electronic controls to monitor catalyst function. Table 4.6-1 presents the 2005 model
year certification levels for these engines.
Table 4.6-1: 2005 MY Certification Levels for Low CO Marine Generator Engines
Engine
Manufacturer
Kohler Power
Systems
Westerbeke
Power
[kW]
10.2
7.5
17.9
Emission Control System
throttle-body injection, O2 sensor, catalyst
throttle-body injection, O2 sensor, catalyst
throttle-body injection, O2 sensor, catalyst
HC+NOx
[g/kW-hr]
7.2
2.0
4.4
CO
[g/kW-hr]
5.2
0.01
0.0
In-use testing has been performed on two marine generator engine equipped with catalysts.
These engines were installed on rental houseboats and operated for a boating season. Testing was
first performed with low hours of operation; 108 hours for the 14 kW engine and 159 hours for the
20 kW.75 The CO performance was reported to be "impressive with exhaust stack CO emissions of
approximately 200 ppm for a fully warmed generator." The emissions measured around the boat
were much lower due to dilution. According to the manufacturer, no significant deterioration has
been found in the emission performance of the catalysts. Note that the manufacturer recommends
changing the catalysts at 2000 hours and inspecting for CO at 1000 hours.
4.7 Test Procedures
We are making several technical amendments to the existing exhaust emission test
procedures for Small SI and OB/PWC engines. These amendments are part of a larger effort to
develop uniform test procedures across all of our programs. We including SD/I engines in these
test procedures. In addition we are establishing not-to-exceed requirements for Marine SI engines.
These new procedures are discussed in this section.
4.7.1 SD/I Certification Test Procedure
We are using the same certification duty cycle and test procedures for all Marine SI
engines, including sterndrives and inboards. Table 4.5-6 presents the certification test duty cycle.
This duty cycle is commonly referred to as the E4 duty cycle and was developed using operational
data on outboard and sterndrive marine gasoline engines.76 In addition, the E4 duty cycle is
recommended by the International Standards Organization for use with all spark-ignition
pleasurecraft less than 24 meters in length.77 Although some Marine SI engines may be used for
commercial activities, these engines would not likely be made or used differently than those used
4-56
-------
Feasibility of Exhaust Emission Control
for pleasure.
Table 4.7-1: SI Marine Certification Steady-State Test Duty Cycle
Mode
1
2
3
4
5
% of Maximum Test
Speed (MES)
100
80
60
40
idle
% of Maximum Torque at
MES
100
71.6
46.5
25.0
Ob
% of Maximum Power2 at
MES
100
57.2
27.9
10.1
Weighting
Factor
0.06
0.14
0.15
0.25
0.40
a o,
% power = (% speed) * (% torque)
15% of maximum torque at MES for high-performance engines
For high-performance engines, the above test procedure is modified slightly. These
engines typically have substantial auxiliary loads and parasitic losses even when the vessel does
not need propulsion power. In addition, these engines are not designed to operate at a low load
idle and survey data suggests that operators do not spend significant time at zero power idle.78 To
account for this, for high-performance engines, we revised the test torque at idle speed to be 15
percent of maximum torque at maximum test speed.
4.7.2 SI Marine Not-To-Exceed Requirements
EPA is concerned that if a marine engine is designed for low emissions on average over a
low number of discrete test points, it may not necessarily operate with low emissions in-use. This
is due to a range of speed and load combinations that can occur on a vessel which do not
necessarily lie on the test duty cycle. For instance, the test modes on the E4 duty cycle lie on an
average propeller curve. However, a propulsion engine may never be fitted with an "average
propeller." In addition, a light planing hull boat may operate at much lower torques than a heavily
loaded boat.
It is our intent that an engine operate with low emissions under all in-use speed and load
combinations that can occur on a boat, rather than just the discrete test modes in the five-mode
duty cycle. To ensure this, we have requirements that extend to typical in-use operation. We are
establishing not-to-exceed (NTE) requirements similar to those established for marine diesel
engines. Under this approach, manufacturers would design their engines to comply with a not-to-
exceed limit, tied to the standard, for HC+NOx and CO, within the NTE zone. In the cases where
the engine is included in averaging, banking, and trading of credits, the NTE limits would be tied
to the family emission limits. We would reserve the right to test an engine in a lab or installed in a
boat to confirm compliance to this requirement.
We believe there are significant advantages to taking this approach. The test procedure is
very flexible so it can represent the majority of in-use engine operation and ambient conditions.
Therefore, the NTE approach takes all of the benefits of a numerical standard and test procedure
and expands it to cover a broad range of conditions. Also, laboratory testing makes it harder to
perform in-use testing because either the engines would have to be removed from the vessel or
4-57
-------
Final Regulatory Impact Analysis
care would have to be taken that laboratory-type conditions can be achieved on the vessel. With
the NTE approach, in-use testing and compliance become much easier because emissions may be
sampled during normal vessel use. Because this approach is objective, it makes enforcement
easier and provides more certainty to the industry of what is expected in use versus over a fixed
laboratory test procedure.
Even with the NTE requirements, we believe it is still important to retain standards based
on the steady-state duty cycle. This is the standard that we expect the certified marine engines to
meet on average in use. The NTE testing is more focused on maximum emissions for segments of
operation and should not require additional technology beyond what is used to meet the new
standards. We believe basing the emission standards on a distinct cycle and using the NTE zone to
ensure in-use control creates a comprehensive program. In addition, the steady-state duty cycles
give a basis for calculating credits for averaging, banking, and trading.
We believe that the same technology that can be used to meet the standards over the five-
mode certification duty cycle can be used to meet the NTE caps in the NTE zone. We therefore do
not expect the NTE standards to cause marinizers to need additional technology. We do not
believe the NTE concept results in a large amount of additional testing, because these engines
should be designed to perform as well in use as they do over the steady-state five-mode
certification test.
4.7.2.1 Shape of the NTE Zone
The NTE zone is intended to capture typical in-use operation for marine vessels. We used
two data sources to define this operation. The first data source was the collection of data on
marine engine operation that was used to develop the ISO E4 steady-state duty cycle.79 Speed and
torque data were collected on 33 outboards and three sterndrives. This data showed that the
marine engines generally operated along a propeller curve with some variation due to differences
in boat design and operation. A propeller curve defines the relationship between engine speed and
torque for a marine engine and is generally presented in terms of torque as a function of engine
speed in RPM raised to an exponent. The paper uses an exponent of 1.5 as a general fit, but states
that the propeller curves for Marine SI applications range from exponents of 1.15 to 2.0.
The second source of data was a study of marine engine operation recently initiated by the
marine industry.80 In this study, sixteen boats were tested in the water at various engine speeds.
These boats included seven sterndrives, three inboards, four outboards, and two personal
watercraft. To identify the full range of loads at each engine speed, boats were operated both fully
loaded and lightly loaded. Boats were operated at steady speeds to identify torque at each speed.
In some cases, the operation was clearly unsafe or atypical. We did not include these operating
points in our analysis. An example of atypical operation would be with a boat so highly loaded
that it was operating in an unstable displacement mode with its bow sticking up into the air.
Figure 4.7-1 presents test data from the two studies as well as the NTE zone for Marine SI
engines. This zone includes operation above and below the theoretical propeller curve used in the
E4 duty cycle. Operation below 25 percent of rated speed is excluded because brake-specific
4-58
-------
Feasibility of Exhaust Emission Control
emissions at low loads becomes very high due to low power in the denominator. This approach is
consistent with the marine diesel NTE zone. The upper and lower borders of the NTE zone are
designed to capture all of the typical operation that was observed in the two studies. The curve
functions for these boarders are presented in Figure 4.7-1.
Figure 4.7-1: NTE Zone and Marine Engine Operation Data
D
k.
O
0)
_N
re
1.5 x Speed-0.16
» Outboard
• PWC
A Inboard
• Sterndrive
x SAE901596
O E4 Setpoints
Prop Curve
Torque Curve
NTE
Zone
Atypical Points
S peed A1.5-0.08
25.3% Torque
20%
40% 60% 80%
Normalized Speed
100%
120%
When testing the engine within the NTE zone, only steady-state operation would be
considered. It is unlikely that transient operation is necessary under the NTE concept to ensure
that emissions reductions are achieved. We designed the NTE zone to contain the operation near
an assumed propeller curve that the steady-state duty cycle represents. We believe that the vast
majority of the operation in the NTE zone would be steady-state. When bringing a boat to plane,
marine engine operation would be transient and would likely be above the NTE zone. However
we do not have enough information to quantify this. Also we do not believe that the NTE zone
should be extended to include areas an engine may see under transient operation, but not under
steady-state operation. For this reason, we do not believe that adding transient operation to the
NTE requirements is necessary at this time. We would revise this opinion in the future if there
were evidence that in-use emissions were increased due to insufficient emission control under
transient operation
4-59
-------
Final Regulatory Impact Analysis
4.7.2.2 Modal Emission Test Data within the NTE Zone
The NTE zone has emissions caps which represent a multiplier times the weighted test
result used for certification. Although ideally the engine should meet the certification level
throughout the NTE zone, we understand that a cap of 1.0 times the standard is not reasonable
because there is inevitably some variation in emissions over the range of engine operation. This is
consistent with the concept of a weighted modal emission test such as the E4 duty cycle.
In developing the emission caps in the NTE zone, we collected modal HC+NOx and CO
emission data on a large number of OB, PWC, and SD/I engines. Because limited modal data is
available in published literature,81'82'83 most of the modal data on outboards and personal watercraft
was provided confidentially by individual manufacturers. Data on SD/I engines with catalysts was
collected as part of the catalyst development efforts discussed earlier in this chapter.84'85'86'87 Our
analysis focuses only on engines using technology that could be used to meet the new standards.
The modal data is presented in Figures 4.7-2 through 4.7-9 in terms of the modal emission rate
divided by the weighted E4 average for that engine. Each color bar represents a different engine.
Because of the large volume of data and differences in engine operation and emissions
performance, data is presented separately for carbureted 4-stroke, fuel-injected 4-stroke, and
direct-injected 2-stroke OB/PWC, and for catalyst-equipped SD/I engines.
Figures 4.7-2 and 4.7-4 present normalized HC+NOx modal data for carbureted and EFI 4-
stroke OB/PWC engines. Note that most of the data points are near or below the E4 weighted
average (represented by bars near or below 1.0). This is largely due to the exclusion of idle
operation from the NTE zone compared to the E4 duty cycle that is 40 percent weighted at idle.
As mentioned above, idle is excluded because brake-specific emissions become very large at low
power due to a low power figure in the denominator (g/kW-hr). Especially for the carbureted
engines, higher normalized HC+NOx emissions are observed at the low power end of the NTE
zone (40 percent speed, 25 percent torque). As shown in Figures 4.7-3 and 4.7-5, a similar trend is
observed with normalized CO emissions from these engines.
4-60
-------
Feasibility of Exhaust Emission Control
Figure 4.7-2: Normalized Modal HC+NOx for Carbureted 4-Stroke OB/PWC
LLJ
"re
•o
o
X
O
O
I
-a
o>
_N
"re
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
Figure 4.7-3: Normalized Modal CO for Carbureted 4-Stroke OB/PWC
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
4-61
-------
Final Regulatory Impact Analysis
Figure 4.7-4: Normalized Modal HC+NOx for EFI 4-Stroke OB/PWC
2n
.u
1 R
UJ 1 R
™ I .O
re
i, n
x, 1 .2
o i f r n n
2 1 n l FT Hi fl
+ t ' i
0)
£ 0.6
§ 0.4
o
2 0.2
On
rTTl nfffl IHn rfl
' -Hjhl ntylgLl
|J| f fj
-pi
I
.U i i
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
Figure 4.7-5: Normalized Modal CO for EFI 4-Stroke OB/PWC
2.5
2.0
i
ra
•o
1.5
O
o
a 1-0
0.5
0.0
f
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
4-62
-------
Feasibility of Exhaust Emission Control
Figures 4.7-6 through 4.7-9 present normalized HC+NOx and CO modal data for direct-
injected 2-stroke OB/PWC engines. Based on the data collected, there appear to be two distinct
types of direct-injection 2-stroke engines. One manufacturer uses a higher pressure fuel system
with a unique combustion chamber design for low emissions. Because the modal variation in
emission results are significantly different for the two engine designs, we designate them headings
of Type 1 and Type 2 engines and look at them separately for the purposes of this analysis. As
shown in Figure 4.7-6 and 4.7-7, Type 1 engines tend to have relatively high HC+NOx at low
power, then fairly low emissions over the rest of the modes. For CO, these engines show much
less variability between modes. For Type 2 engines, HC+NOx is below the E4 average in the mid-
speed range as shown in Figure 4.7-8. However, there is a wide degree of variation in how these
engines behave at low and high speed. Most of these engines seem to have high normalized
HC+NOx emissions either at low or at high speed. Figure 4.7-9 presents CO values for Type 1
engines. These engines tend to have high CO at full power with decreasing CO at lower power
modes.
Figure 4.7-6: Normalized Modal HC+NOx for Type 1 DI 2-Stroke OB/PWC
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
4-63
-------
Final Regulatory Impact Analysis
Figure 4.7-7: Normalized Modal CO for Type 1 DI 2-Stroke OB/PWC
0.0
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
Figure 4.7-8: Normalized Modal HC+NOx for Type 2 DI 2-Stroke OB/PWC
2C
.0
51
LLJ 9 D
™ Z.U
re
•o
o
s
1"~l -I c
X 1 .O
O
•2.
+
O
I 1 n
I .U
•o
0)
_N
"«5
§n ^
u.o
o
•2.
n n
PI
PI
1
r
•
1 n-
-•
Tl
J
ni4l
T
f
• I
U.U I
40%, 25% 60%, 47% 80%, 72%
Normalized Speed, Torque [% of
1
r
III
rn
100%, 100%
rated]
4-64
-------
Feasibility of Exhaust Emission Control
Figure 4.7-9: Normalized Modal CO for Type 2 DI 2-Stroke OB/PWC
0.0
40%, 25% 60%, 47% 80%, 72% 100%, 100%
Normalized Speed, Torque [% of rated]
Figures 4.7-10 and 4.7-11 present normalized HC+NOx and CO modal data for SD/I
engines equipped with catalysts. All of these engines demonstrated HC+NOx emissions below the
E4 average in the mid-speed range. However, some of these engines show somewhat higher
normalized HC+NOx emissions at either the low-power or full power mode. These differences are
likely a function of catalyst design and location as well as air/fuel calibration. At wide open
throttle, all of these engines were calibrated to run rich as an engine protection strategy, so
emission reductions at this mode are due to NOx reductions in the catalyst. Because these engines
are designed to run rich at full power, high CO emissions were observed at this mode. For the rest
of the power range, CO emissions were generally below the E4 average for these engines. As part
of the catalyst development work for SD/I engines, one engine was tested over 26 modes, most of
which are contained in the NTE zone.88 This engine was tested in its baseline configuration (open-
loop fuel injection) as well as with three catalyst configurations. The three catalyst configurations
included one close-coupled to the engine (in the riser), one a little farther downstream (in the
exhaust elbow), and a larger catalyst external to the existing exhaust manifold. This data provided
insight into how exhaust emissions throughout the NTE zone for Marine SI engines compare to the
modal test data on the theoretical propeller curve. This data is presented in Appendix 4A.
4-65
-------
Final Regulatory Impact Analysis
Figure 4.7-10: Normalized
1 R
I .O
1 A
„ 1.4
LLJ
"J5 -| 9
-§ 1-^
O
u 1 n
1 — ' 1 .U
X,
O
O
n R
•o u.o
0)
_N
_ U.4
1 n?
Z. U.Z
n n
Modal HC+NOx for SD/I with Catalysts
i—
- n
_ _
Rw
r _
r -
L
Jl Iff 1
" 41- 1 " fl-
: fl - :
: ..I.:...
i
-
40%, 25% 50%, 35% 60%, 47% 70%, 59% 80%, 72% 90%, 85% 100%, 100%
Normalized Speed, Torque [% of rated]
Figure 4.7-11: Normalized Modal CO for SD/I with Catalysts
•2 C
o n
O.U
HI 9 c
^ z.o
ra
•o
o
s 9 n
!=. Z.U
0
o
__ -I C
~ 1 .0
0)
o
Oc
.0
On
.
I
n n
II
IflaM [
u
n ^rn PI r-L_ F
1" idlL It
-
-
.U i i i
40%, 25% 50%, 35% 60%, 47% 70%, 59% 80%, 72% 90%, 85% 100%, 100%
Normalized Speed, Torque [% of rated]
4-66
-------
Feasibility of Exhaust Emission Control
4.7.2.3 NTE Zones and Standards
As described above, the emissions characteristics of marine engines are largely dependent
on technology type. Four-stroke engines tend to have relatively constant emission levels
throughout the NTE zone. In contrast, two-stroke engine tend to have high variability in
emissions, not only within the NTE zone but between different engine designs as well. Catalyst-
equipped engines tend to have relatively flat emissions profiles, in the NTE zone, when operating
in closed-loop engine control mode. However, at higher engine power, the engines are calibrated
to operate with a rich fuel-air ratio, in open-loop control, to protect the exhaust valves and catalyst
from high exhaust temperatures. This reduces the catalyst efficiency at high power for oxidizing
HC and CO.
Since the NPRM was published, we have worked with engine manufacturers to better
understand the design constraints, operation, and emissions characteristics of marine engines.
Based on this understanding, and the emission data presented above, we decided to develop
different NTE standards for three distinct types of engines; 4-stroke, 2-stroke, and catalyst-
equipped. These standards are discussed below.
We used the modal data presented above and the data on additional operation points
presented in Appendix 4A to develop NTE limits. These limits are presented as a multiplier times
the Family Emission Limit (FEL) developed using the 5-mode test procedure. The limits represent
the levels that can be met by the majority of the marine engines tested. In the case of engines that
have modal emissions that are somewhat higher than the NTE limits, we believe that these engines
can be calibrated to meet these limits. In addition, the limits are based on the FELs chosen by
manufacturers at certification. Therefore, manufacturers would have the option of increasing their
FELs, in some cases, to bring otherwise problem engines into compliance with the NTE limits.
4.7.2.3.1 4-Stroke Engines
The emissions data, presented above, for 4-stroke marine engines suggests that brake-
specific emissions rates are relatively constant throughout the NTE zone. One exception is slightly
higher HC+NOx emissions at low power. To account for this, we are subdividing the NTE zone to
have a low power subzone below 50 percent of maximum test speed. In this low power subzone,
the HC+NOx NTE limit is 1.6, while it is 1.4 for the remainder of the NTE zone. The CO NTE
limit is 1.5 throughout the NTE zone. Figure 4.7-12 presents the NTE zone and subzones. These
limits would apply to all non-catalyzed four-stroke marine engines.
4-67
-------
Final Regulatory Impact Analysis
Figure 4.7-12: 4-Stroke Engine NTE Zone and Subzones
120%
0%
25.3% Torque
SpeedA1.5 (theoretical propeller curve)
20%
40% 60% 80%
Normalized Speed
100%
120%
4.7.2.3.2 2-Stroke Engines
The emissions data presented above, for 2-stroke direct-injection marine engines suggests
that these engines have high variability in emissions, not only within the NTE zone but between
different engine designs as well. Due to this variability, we do not believe that a flat (or stepped)
limit in the NTE zone could be effectively used to establish meaningful standards for these
engines. At the same time, we believe that NTE standards are valuable for facilitating in-use
testing. Therefore, we developed a weighted NTE approach specifically for these engines. In the
long term, we may consider further emission reduction based on catalytic control applied to
OB/PWC engines. In this case, we would revisit the appropriateness of the weighted NTE
approach in the context of those standards.
Under the weighted NTE approach, emissions data is collected at five test points. These
test points are idle, full power, and the speeds specified in modes 2-4 of the 5-mode duty cycle.
Similar to the 5-mode duty cycle, the five test points are weighted to achieve a composite value.
This composite value must be no higher than 1.2 times the FEL for that engine.
The difference in this approach, from the 5-mode duty cycle, is that the test torque is not
specified. During an in-use test, the engine would be set to the target speed and the torque value
would be allowed to float. In addition, at wide open throttle, the engine speed would be based on
4-68
-------
Feasibility of Exhaust Emission Control
actual performance on the boat. Because in-use engines installed in boats do not generally operate
on the theoretical propeller curve used to define the 5-mode duty cycle, this NTE approach helps
facilitate NTE testing.
At each test mode, limits are placed on allowable engine operation. These limits are
generally based on the NTE zone presented above for 4-stroke engines, but there are two
exceptions. First, the lower torque limit at 40 percent speed is lowered slightly to better ensure
that an engine on an in-use boat is capable of operating within the NTE zone. Second, the speed
range is extended at wide-open throttle for the same reason. Figure 4.7-13 presents this approach.
Figure 4.7-13: 2-Stroke Engine Weighted NTE Concept
120%
100%
= 80%
I
re
o
60%
40%
20%
0%
Idle
SpeedA1.5 (theoretical propeller curve)
o%
20%
40% 60% 80%
Normalized Speed
100%
120%
During laboratory testing, any point within each of the four non-idle subzones may be
chosen as test points. These test points do not necessarily need to lie on a propeller curve. It
should be noted that the actual power measured would be used in the calculation of the weighted
brake-specific emissions.
4.7.2.3.3 Catalyst-Equipped Engines
SD/I engines are anticipated to make use of three-way catalysts to meet the new exhaust
emission standards. These catalysts are most effective when the fuel-air ratio in the exhaust is near
stoichiometry. Engine manufacturers use closed-loop control to monitor and maintain the proper
fuel-air ratio in the exhaust for optimum catalyst efficiency. However, at high power, engine
4-69
-------
Final Regulatory Impact Analysis
manufacturers must increase the fueling rate to reduce the temperature of the exhaust. Otherwise,
if the exhaust temperature becomes too high, exhaust valves and catalysts may be damaged.
During rich, open-loop operation at high power, the catalyst is oxygen-limited and less effective at
oxidizing HC and CO. To address the issue of open-loop catalyst efficiency, we created a high
power subzone in the NTE zone for catalyst-equipped engines.
The majority of SD/I engines are based on engine blocks produced by General Motors. To
determine the appropriate threshold for the high power-subzone, we used temperature data
supplied by General Motors.89 This data was consistent with confidential data supplied by engine
marinizers on engine control strategies for catalyst-equipped SD/I engines. Figure 4.7-14 presents
the stoichiometric limits for engine protection, based on the General Motors study, for three
different engine designs.
Figure 4.7-14: GM Study-Stoichiometric Limits for Engine Protection
Engine Torque Curve
100%
0)
3
I- 90%
0)
_N
"re
£ 80%
o
70%
5.7L NA
k 6.0L SC Base
—76.0L SC Hi Pert
\?
70%
80% 90%
Normalized Speed
100%
Figure 4.7-15 presents the NTE zone and high power subzone for catalyst-equipped marine
engines. The shape of the high power subzone is based on the General Motors engine protection
data. The emission limits are based on data discussed above on SD/I engines, with catalysts,
operating in open-loop versus closed-loop engine control.
For catalyst-equipped engines, the largest contribution of emissions over the 5-mode duty
cycle comes from open-loop operation at Mode 1. In addition, the idle point (Mode 5) is weighted
4-70
-------
Feasibility of Exhaust Emission Control
40 percent in the 5-mode duty cycle, but not included in the NTE zone. For this reason,
brake-specific emissions throughout most of the NTE zone are less than the weighted average from
the steady-state testing. For most of the NTE zone, we are therefore requiring a limit equal to the
duty-cycle standard (i.e., NTE multiplier = 1.0).
Emission data on catalyst-equipped engines also show higher emissions near full-power
operation. As discussed above, this is due to the need for richer fuel-air ratios under high-power
operation to protect the engines from overheating. We are therefore establishing higher NTE
limits for subzone 1 based on emissions performance during open-loop operation. Specifically, we
are establishing an HC+NOx limit of 1.5 times the duty-cycle standard. Some HC+NOx control is
expected in subzone 1 because a three-way catalyst will effectively reduce NOx emissions under
rich conditions. However, for subzone 1, we are not setting a CO limit. Under rich conditions, a
three-way catalyst is not at all effective for oxidizing HC or CO emissions. In addition, the cycle
weighted emission level for CO is primarily driven by Mode 1.
Figure 4.7-15: Catalyst-Equipped Engine NTE Zone and Subzones
120%
100%
g. 80%
ft 60%
"re
Z 40%
20%
0%
N 25.3% Torque
SpeedA1.5 (theoretical propeller curve)
o%
20%
40% 60% 80%
Normalized Speed
100%
120%
4.7.2.4 Ambient Conditions
Ambient air conditions, including temperature and humidity, may have a significant effect
on emissions from marine engines in-use. To ensure real world emissions control, the NTE zone
testing should include a wide range of ambient air conditions representative of real world
conditions. Because these engines are used in similar environments as marine diesel engines, we
4-71
-------
Final Regulatory Impact Analysis
are applying the same ambient ranges to the Marine SI NTE requirements as already exist for
marine diesel engine NTE requirements.
We believe that the appropriate ranges should be 13-30°C (55-86°F) for air temperature
and 7.1-10.7 grams water per kilogram dry air (50-75 grains/pound of dry air) for air humidity.
The air temperature ranges are based on temperatures seen during ozone exceedences, except that
the upper end of the temperature range has been adjusted to account for the cooling effect of a
body of water on the air above it.90 We are also aware, however, that marine engines sometimes
draw their intake air from an engine compartment or engine room such that intake air temperatures
are substantially higher than ambient air temperatures. In this case, we would retain 35°C as the
end of the NTE temperature range for engines that do not draw their intake air directly from the
outdoor ambient.
For NTE testing in which the air temperature or humidity is outside the specified range, we
require that the emissions must be corrected back to the specified air temperature or humidity
range. These corrections would be consistent with the equations in 40 CFR Part 91, Subpart E
except that these equations correct to 25°C and 10.7 grams per kilogram of dry air while the NTE
corrections would be to the nearest outside edge of the specified ranges. For instance, if the
outdoor air temperature were higher than 30°C for an engine that drew fresh outdoor air into the
intake, a temperature correction factor could be applied to the emissions results to determine what
emissions would be at 30°C.
Ambient water temperature also may affect emissions due to it's impact on engine cooling.
For this reason, we are requiring that the NTE testing include a range of ambient water
temperatures from 5 to 27°C (41 to 80°F). The water temperature range is based on temperatures
that marine engines experience in the U.S. in-use. At this time, we are not aware of an established
correction for ambient water temperature, therefore the NTE zone testing would have to be withing
the specified ambient water temperature range.
4.8 Impacts on Safety, Noise, and Energy
Section 213 of the Clean Air Act directs us to consider the potential impacts on safety,
noise, and energy when establishing the feasibility of emission standards for nonroad engines.
Furthermore, section 205 of Public Law 109-54 requires us to assess potential safety issues,
including the risk of fire and burn to consumers in use, associated with the emission standards for
nonroad spark-ignition engines under 50 horsepower. As further detailed in the following
sections, we expect that the exhaust emission standards will either have no adverse affect on
safety, noise, and energy or will improve certain aspects of these important characteristics.
4.8.1 Safety
We conducted a comprehensive, multi-year safety study of nonroad SI engines that focused
on the following areas where we are finalizing new exhaust standards.91 These areas are:
New catalyst-based HC+NOx exhaust emission standards for Class I and II
4-72
-------
Feasibility of Exhaust Emission Control
nonhandheld (NHH) engines; and
New HC+NOx exhaust emission standards for outboard and personal watercraft
(OB/PWC) engines and vessels, and a new CO exhaust emission standard for NHH
engines used in marine auxiliary applications.
Each of these four areas is discussed in greater detail in the next sections.
4.8.1.1 Exhaust Emission Standards for Small Spark-Ignition Engines
The technology approaches that we assessed for achieving the Small SI engine standards
included exhaust catalyst aftertreatment and improvements to engine and fuel system designs. In
addition to our own testing and development effort, we also met with engine and equipment
manufacturers to better understand their designs and technology and to determine the state of
technological progress beyond EPA's Phase 2 standards.
The scope of our safety study included Class I and Class II engine systems that are used in
residential walk-behind and ride-on lawn mower applications, respectively. Residential lawn
mower equipment was chosen for the following reasons.
Lawn mowers and the closely-related category of lawn tractors overwhelmingly
represent the largest categories of equipment using Class I and Class II engines. We
estimate that over 47 million walk-behind mowers and ride-on lawn and turf equipment
are in-use in the US today.
These equipment types represent the majority of sales for Small SI engines.
- Consumer Product Safety Commission (CPSC) data indicates that more thermal burn
injuries associated with lawn mowers occur than with other NHH equipment; lawn
mowers therefore represent the largest thermal burn risk for these classes of engines.
- General findings regarding advanced emission control technologies for residential lawn
and garden equipment carry over to commercial lawn and turf care equipment as well
as to other NHH equipment using Class I and Class II engines. Lawn mower design
and use characteristics pose unique safety implications not encountered by other NHH
equipment using these engines (i.e. a mower deck collects debris during operation
whereas a pressure washer collects no debris). Thus, other NHH equipment may
employ similar advanced emission control technologies for meeting the standards
without a corresponding concern regarding the safety issues analyzed in this study.
We conducted the technical study of the incremental risk on several fronts. First, working
with the CPSC, we evaluated their reports and databases and other outside sources to identify
those in-use situations which create fire and burn risk for consumers. The outside sources
included meetings, workshops, and discussions with engine and equipment manufacturers. The
following scenarios were identified for evaluation:
- Thermal burns due to inadvertent contact with hot surface on engine or equipment;
- Fires from grass and leaf debris on the engine or equipment;
4-73
-------
Final Regulatory Impact Analysis
Fires due to fuel leaks on hot surfaces;
Fires related to spilled fuel or refueling vapor;
- Equipment or structure fire when equipment is left unattended after being used;
Engine malfunction resulting in an ignitable mixture of unburned fuel and air in the
muffler (engine misfire); and
- Fire due to operation with richer than designed air-fuel ratio in the engine or catalyst.
These scenarios cover a comprehensive variety of in-use conditions or circumstances
which potentially could lead to an increase in burns or fires. They may occur presently or not at
all, but were included in our study because of the potential impact on safety if they were to occur.
The focus of the analysis was, therefore, on the incremental impact on the likelihood and severity
of the adverse condition in addition to the potential causes as it related to the use of more advanced
emissions control technology.
Second, we conducted extensive laboratory and field testing of both current technology
(Phase 2) and prototype catalyst-equipped advanced-technology engines and equipment (Phase 3)
to assess the emission control performance and thermal characteristics of the engines and
equipment. This testing included a comparison of exhaust system, engine, and equipment surface
temperatures using thermal imaging equipment.
Third, we contracted with Southwest Research Institute (SwRI) to conduct design and
process Failure Mode and Effects Analyses (FMEA).92 The SwRI FMEA focused on comparing
current Phase 2 and Phase 3 compliant engines and equipment to evaluate incremental changes in
risk probability as a way of evaluating the incremental risk of upgrading Phase 2 engines to meet
Phase 3 emission standards. This is an engineering analysis tool to help engineers and other
professional staff on the FMEA team to identify and manage risk. In a FMEA, potential failure
modes, causes of failure, and failure effects are identified and a resulting risk probability is
calculated from these results. This risk probability is used by the FMEA team to rank problems
for potential action to reduce or eliminate the causal factors. Identifying these causal factors is
important because they are the elements that a manufacturer can consider reducing the adverse
effects that might result from a particular failure mode.
Our technical work and subsequent analysis of all of the data and information strongly
indicate that effective catalyst-based standards can be implemented without an incremental
increase in the risk of fire or burn to the consumer either during or after using the equipment.
Similarly, we did not find any increase in the risk of fire during storage near typical combustible
materials. In many cases, the designs used for catalyst-based technology can lead to an
incremental decrease in such risk.
More specifically, our work included taking temperature measurements and infrared
thermal images of both OEM mufflers and prototype catalyst/mufflers on six Class 1 engines and
three Class 2 engines as part of the safety study. We integrated the emission reduction catalyst
into the muffler. In doing so, we generally designed heat management features into the
catalyst/muffler and cooling system. These heat management design elements, all of which were
not used on every prototype, included: 1) positioning the catalyst within the cooling air flow of the
4-74
-------
Feasibility of Exhaust Emission Control
engine fan or redirecting some cooling air over the catalyst area with a steel shroud; 2) redirecting
exhaust flow through multiple chambers or baffles within the catalyst/muffler; 3) larger
catalyst/muffler volumes than the original equipment muffler; and 4) minimizing CO oxidation at
moderate to high load conditions to maintain exhaust system surface temperatures comparable to
those of the OEM systems. The measurements and images were taken during various engine
operating conditions and as the engines cooled down after being shut off. This latter event,
termed "hot soak," is an important consideration since it is often when the operator is in close
proximity to the engine either performing maintenance or refueling the equipment.
Figures 4.8-1 and 4.8-2 are an example of the measurements and images taken to compare
Class 1 engine original equipment (OEM) mufflers to the same engines equipped with prototype
catalyst/mufflers. The first figure depicts surface temperatures from engine number 244 while
operated on a laboratory dynamometer over three modes of EPA's A-cycle steady-state test cycle.
The second figure shows surface temperatures for the same engine at different times during hot
soak. The prototype catalyst/muffler system shown in these figures uses one of the most effective
heat management designs in the safety study. As shown, the catalyst system in this example has
much lower surface temperatures during both engine operation and hot soak.
Similar information was collected in the laboratory for Class 2 engines used in lawn
tractors. However, those tests were conducted on the "raw" engines without the chassis, which is
an integral part of the overall engine cooling system for most residential Class 2 applications.
Because of this, we believe it is more appropriate to compare the thermal measurements from field
testing of the integrated unit.
The test results for engine 251 are fairly typical of the Class 2 lawn tractor test results.
During engine operation, the OEM muffler configuration had exposed surface temperatures of
approximately 200 °C as viewed from both sides of the tractor when cutting moderate to heavy
grass and peak temperatures as high as 300 to 365 °C. The lawn tractor equipped with engine 253,
which is from the same engine family as number 251, was fitted with a prototype catalyst/muffler
exhibited exposed surface temperatures of approximately 115 to 130 °C and peaks of 160 to 190
°C. The lower temperatures for the prototype catalyst system is in part due to the more effective
cooling of the catalyst/mufflers due to the re-routing of cooling air through the chassis and other
heat management design elements.
The hot soak results for the above engines and two other related Class 2 lawn mowers are
shown in Figure 4.8-3. The two-minute nominal refueling point after engine shut-down following
30 minutes of grass-cutting operation is shown for reference. In these tests, both of the engines
with prototype catalyst/mufflers had lower peak surface temperatures than the OEM muffler
configurations.
4-75
-------
Final Regulatory Impact Analysis
Figure 4.8-1: Surface Temperature Infrared Thermal Images of Exhaust System
Components for Class 1 Engine 244 with a Catalyst/Muffler (left) and an OEM Muffler
(right) at Various Operating Modes.
ygtfun f HCinduy *ir
>oi ThncCle
f3;e temperature: 230 °C
50% Load-Mode 3
Maximum surface temperature: 102
10% Load-Mode 5
OEMttufflar
aximum surface temperature: 551 °C
50% Load-Mode 3
Maximum surface temperature: 421 °C
10% Load-Mode 5
ace temperature: 363 °C
4-76
-------
Feasibility of Exhaust Emission Control
Figure 4.8-2: Hot Soak Surface Temperature Infrared Thermal Images of Exhaust System
Components for Class 1 Engine 244 After Sustained Wide Open Throttle and 100 Percent
Load.
Modified
Maximum, surfeice t£mf lenrjjre
Ms-dmum surfeiie temf icntijre
M^odmium. £UEfi.ce
OEMIiluffler
Mmnoim. suEfece taifiOTimK: 4-83 * C
mm
Maximumeurfiice temperature: 4-18 *
4-77
-------
Final Regulatory Impact Analysis
Figure 4.8-3: Hot Soak Peak Surface Temperatures Infrared Thermal Images for Class 2
Lawn Tractors Following After Approximately 30-Minutes of Grass Cutting.
o
£
-------
Feasibility of Exhaust Emission Control
use. Based on the applicability of Coast Guard and ABYC safety standards and the good in-use
experience with advanced-technology engines in the current vessel fleet, we believe new emission
standards would not create an incremental increase in the risk of fire or burn to the consumer.
4.8.2 Noise
As automotive technology demonstrates, achieving low emissions from spark-ignition
engines can correspond with greatly reduced noise levels. Direct-injection two-stroke and four-
stroke OB/PWC have been reported to be much quieter than traditional carbureted two-stroke
engines. Catalysts in the exhaust act as mufflers which can reduce noise. Additionally, adding a
properly designed catalyst to the existing muffler found on all Small SI engines can offer the
opportunity to incrementally reduce noise.
4.8.3 Energy
Adopting new technologies for controlling fuel metering and air-fuel mixing, particularly
the conversion of some carbureted engines to advanced fuel injection technologies, will lead to
improvements in fuel consumption. This is especially true for OB/PWC engines where we expect
the standards to result in the replacement of old-technology two stroke engines with more fuel
efficient technologies such as two-stroke direct injection or four-stroke engines. Carbureted
crankcase-scavenged two-stroke engines are inefficient in that 25 percent or more of the fuel
entering the engine may leave the engine unburned. We estimate a fuel savings of about 61
million gallons of gasoline from marine engines in 2030, when most boats would be using engines
complying with the standard.
The conversion of some carbureted Small SI engines to fuel injection technologies is also
expected to improve fuel economy. We estimate approximately 7 percent of the Class II engines
will be converted to fuel injection and that this will result in a fuel savings of about 10 percent for
each converted engine. This translates to a fuel savings of about 22 million gallons of gasoline in
2030 when all of the Class II engines used in the U.S. will comply with the Phase 3 standards. By
contrast, the use of catalyst-based control systems on Small SI engines is not expected to change
their fuel consumption characteristics. These estimates are discussed in more detail in Chapter 6.
4-79
-------
Final Regulatory Impact Analysis
APPENDIX 4A: Normalized Modal Emissions for a 7.4 L MPI SD/I
Figure 4A-1: HC+NOx Ratios for 7.4L MPI Engine, Baseline
140%
120%
°«6
0.6 a »
°«' 0.9 1.1
05 o«
a 0.6
0.8
o
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
Figure 4A-2: CO Ratios for 7.4L MPI Engine, Baseline
140%
120%
-------
Feasibility of Exhaust Emission Control
Figure 4A-3: HC+NOx Ratios for 7.4L MPI Engine, Riser Catalysts
140%
120%
CD 100%
3
E
I— 80%
^
N
ID 60%
o
~Z. 40%
20%
0%
0
07
0.7
O
0.6
0.8
n- °-9 ^ no
<* oa7 a9 0^9
°d5 07 07
0.9 ... ° °66 1.0
0 0.4 1 0 0
1.6 a V
f> Q 0
<> 0.4 05
V 0.3 ° °
1.3
o
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
Figure 4A-4: CO Ratios for 7.4L MPI Engine, Riser Catalysts
140%
120%
0 100%
3
E
I— 80%
^
N
ID »*
i_
o
Z 40%
20%
0%
0
1 8
18
1.9
0.7
° 0.9 V
°o° 0.2 04 '-4
a '
°o4 n^ 02
°o3 n^ ° °^1 168
« 0.3 n ^ °
0.3 a uod
Q 0
<> 0.4 04
^5 o.o * °
°c-6
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
4-81
-------
Final Regulatory Impact Analysis
Figure 4A-5: HC+NOx Ratios for 7.4L MPI Engine, Elbow Catalysts
140%
120%
CD 100%
E
I— 80%
N
ID 60%
o
~Z. 40%
20%
0%
0
Dfi
° 0.5
O
0.6
03 °"6 °'5 05
0 °a3 064 6
0X6 0.5 °63
°<55 05 0.2 05
H 0.5 o <5
17 05 °
15 °o4 o
b 0,3
167
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
Figure 4A-6: CO Ratios for 7.4L MPI Engine, Elbow Catalysts
140%
120%
0 100%
3
E
I— 80%
^
N
ro s""/0
i_
o
Z 40%
20%
0%
0
1 R
1.7
O
1.8
0.5
1.2
Oo 0,7 c>
6 rf9 0.6
° ^0^4°
0,4 0.3 0.1
^ V 04 ai
0,4 0 o
0 04 0.5
04 <> o
^ OJ
°68
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
4-82
-------
Feasibility of Exhaust Emission Control
Figure 4A-7: HC+NOx Ratios for 7.4L MPI Engine, External Catalysts
CD
E
£
N
O
140%
120%
100%
80%
60%
40%
20%
0%
0
0.4
0.3
0.3
°o3 °
0.3
0.3 o
n ^ o n °
0^3 0^3 o
° 0.2 04 o c
14 H
i..t n ^
0<56 °<>6
A A
a 0.0
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
Figure 4A-8: CO Ratios for 7.4L MPI Engine, External Catalysts
140%
120%
0 100%
3
E
I— 80%
^
N
ID »*
i_
o
Z 40%
20%
0%
0
0.3
0.3
o
0.5
0,4 "
0.4
0.2 *
n ? o n ^
V 0,1 04 V
°c8 0.3 CL1
0.4 " 0.2 06
« 0 4 0.5 o
0,9 Q, d °
" ^ 0,7 0,7
a5 °6°
°,1
NA
% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110%
Normalized Speed
4-83
-------
Final Regulatory Impact Analysis
Chapter 4 References
1. Internal Combustion Engine Fundamentals, Heywood, J., McGraw-Hill, Inc., New York, 1988, pp.829-836,
Docket Identification EPA-HQ-OAR-2004-0008-0203.
2. Internal Combustion Engine Fundamentals, Heywood, J., McGraw-Hill, Inc., New York, 1988, pp.827-829,
Docket Identification EPA-HQ-OAR-2004-0008-0203.
3. Conversation between Tim Fileman at Flagship Marine and Mike Samulski at U.S. EPA, February 10, 1999.
4. "Benefits and Cost of Potential Tier 2 Emission Reduction Technologies," Energy and Environmental Analysis,
Final Report, November 1997, Docket Identification EPA-HQ-OAR-2004-0008-0198.
5. "Study on Air Assist Fuel Injector Atomization and Effects on Exhaust Emission Reduction," Saikalis, G., Byers,
R., Nogi., T.,SAE Paper 930323, 1993, Docket Identification EPA-HQ-OAR-2004-0008-0217.
6. "Three-Way Catalyst Technology for Off-Road Equipment Powered by Gasoline and LPG Engines," Southwest
Research Institute, prepared for CARB, CEP A, and SCAQMD, (SwRI 8778), April 1999, Docket Identification
EPA-HQ-OAR-2004-0008-0224.
7. Internal Combustion Engine Fundamentals, Heywood, J., McGraw-Hill, Inc., New York, 1988, pp. 836-839,
Docket Identification EPA-HQ-OAR-2004-0008-0203.
8. "Improving the NOx/Fuel Economy Trade-Off for Gasoline Engines with the CCVS Combustion System,"
Stokes, J., Lake, T., Christie, M., Denbratt, I., SAE Paper 940482, 1994, Docket Identification EPA-HQ-OAR-2004-
0008-0498.
9. Wu, H., Yang, S., Wang, A., Kao, H., "Emission Control Technologies for 50 and 125 cc Motorcycles in
Taiwan," SAE Paper 980938, 1998, Docket Identification EPA-HQ-OAR-2004-0008-0502.
10. Borland, M., Zhao, F., "Application of Secondary Air Injection for Simultaneously Reducing Converter-In
Emissions and Improving Catalyst Light-Off Performance," SAE Paper 2002-01-2803, 2001, Docket Identification
EPA-HQ-OAR-2004-0008-0497.
11. Koehlen, C., Holder, E., Guido, V., "Investigation of Post Oxidation and Its Dependency on Engine Combustion
and Exhaust Manifold Design," SAE Paper 2002-01-0744, 2002.
12. "Benefits and Cost of Potential Tier 2 Emission Reduction Technologies," Energy and Environmental Analysis,
Final Report, November 1997, Docket Identification EPA-HQ-OAR-2004-0008-0198.
13. "Reduction of Exhaust Gas Emissions by Using Pd-based Three-way Catalysts," Lindner, D., Lox, E.S.,
Kreuzer, T., Ostgathe, K., van Yperen, R., SAE Paper 960802, 1996, Docket Identification EPA-HQ-OAR-2004-
0008-0500.
14. "Overview of Recent Emission Control Technology Developments," Manufacturers of Emission Controls
Association, November 18, 1997, Docket Identification EPA-HQ-OAR-2004-0008-0206.
15. "Application of Accelerated Rapid Aging Test (RAT) Schedules with Poisons: The Effects of Oil Derived
Poisons, Thermal Degradation and Catalyst Volume on FTP Emissions," Ball, D., Mohammed, A., Schmidt, W.,
SAE Paper 972846, 1997, Docket Identification EPA-HQ-OAR-2004-0008-0501.
4-84
-------
Feasibility of Exhaust Emission Control
16. "Update Heavy-Duty Engine Emission Conversion Factors for MOBILE6; Analysis of Fuel Economy, Non-
Engine Fuel Economy Improvements, and Fuel Densities," prepared by Arcadis for U.S. EPA, EPA420-R-006,
January 2002, Docket Identification EPA-HQ-OAR-2004-0008-0251.
17. "Overview of Recent Emission Control Technology Developments," Manufacturers of Emission Controls
Association, November 18, 1997, Docket Identification EPA-HQ-OAR-2004-0008-0206.
18. "Catalytic Converter Applications for Two Stroke, Spark-Ignited Marine Engines," Fujimoto, H., Isogawa, A.,
Matsumoto, N., SAE Paper 951814, 1995, Docket Identification EPA-HQ-OAR-2004-0008-0220.
19. "Catalytic Converter Applications for Two Stroke, Spark-Ignited Marine Engines," Fujimoto, H., Isogawa, A.,
Matsumoto, N., SAE Paper 941786, 1994, Docket Identification EPA-HQ-OAR-2004-0008-0218.
20. "Public Meeting to Consider a Status Report on Catalyst Testing of Spark-Ignition Inboard/Sterndrive
Pleasurecraft: Staff Report," State of California; Air Resources Board, October 28, 2004, Docket Identification EPA-
HQ-OAR-2004-0008-0236.
21. "Regulatory Impact Analysis—Control of Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements," U.S. EPA, December 1999, EPA420-R-99-023,
http://www.epa.gov/otaq/regs/ld-hwy/tier-2/frm/ria/r99023.pdf
22. "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty Trucks," J. McDonald and L. Jones, SAE
2000-01-1957, Docket Identification EPA-HQ-OAR-2004-0008-0496.
23. "SULEV and 'Off-Cycle' Emissions Benefits of a Vacuum-Insulated Catalytic Converter," Burch, S.D., and J.P.
Biel, SAE 1999-01-0461, Docket Identification EPA-HQ-OAR-2004-0008-0495.
24. "Development of an Alternator-Powered Electrically-Heated Catalyst System," Laing, P.M., SAE 941042,
Docket Identification EPA-HQ-OAR-2004-0008-0494.
25. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower", U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
26. "Nonroad SI Manufacturer Discussion," Slide 26-700cc Class II/Phase 2 Horizontal V-Twin. U.S.
Environmental Protection Agency, November 22, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0087.
27. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower", U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
28. "Test Results for Small Spark-Ignition Engine Number 258," Memorandum and Excel spreadsheet from Richard
Wilcox, Office of Transportation and Air Quality, U.S. Environmental Protection Agency, January 23, 2007, Docket
IdentificationEPA-HQ-OAR-2004-0008-0530.
29. Letter from David Raney, American Honda Motor Co., Inc., to Margo T. Oge, U. S. Environmental Protection
Agency, September 1, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0304.
30. Letter from Kent B. Herrick, Tecumseh Products Company, to Margo T. Oge, U. S. Environmental Protection
Agency, September 7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0306.
31. Letter from Jeffrey Cl Shetler, Kawasaki Motors Corp, U.S.A., to Margo T. Oge, U. S. Environmental
Protection Agency, September 7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0305.
4-85
-------
Final Regulatory Impact Analysis
32. Letter from James P. Doyle, Kohler Co., to Margo T. Oge, U. S. Environmental Protection Agency, September
7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0303.
33. "Reducing Emissions From Gasoline Platforms at Less Then 20 kW," Slide 9, B. Walker and A. Gopalkrishna,
Delphi Automotive, Asian Vehicle Emission Control Conference, Jaipur, Japan, September 22, 2006. EPA-HQ-
OAR-2004-0008-0535.
34. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower", Appendix C - FMEA of Small SI Equipment and Engines, U.S. Environmental Protection Agency,
EPA420-R-06-006, March 2006, Docket Identification EPA-HQ-OAR-2004-0008-0329.
35. Letter from David Raney, American Honda Motor Co., Inc., to Margo T. Oge, U. S. Environmental Protection
Agency, September 1, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0304.
36. Letter from Kent B. Herrick, Tecumseh Products Company, to Margo T. Oge, U. S. Environmental Protection
Agency, September 7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0306.
37. Letter from Jeffrey Cl Shetler, Kawasaki Motors Corp, U.S.A., to Margo T. Oge, U. S. Environmental
Protection Agency, September 7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0305.
38. Letter from James P. Doyle, Kohler Co., to Margo T. Oge, U. S. Environmental Protection Agency, September
7, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0303.
39. "Reducing Emissions From Gasoline Platforms at Less Then 20 kW," Slide 9, B. Walker and A. Gopalkrishna,
Delphi Automotive, Asian Vehicle Emission Control Conference, Jaipur, Japan, September 22, 2006. EPA-HQ-
OAR-2004-0008-0535.
40. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower", Appendix C - FMEA of Small SI Equipment and Engines, U.S. Environmental Protection Agency,
EPA420-R-06-006, March 2006, Docket Identification EPA-HQ-OAR-2004-0008-0329.
41. "Nonroad SI Manufacturer Discussion," Slide 26-700cc Class II/Phase 2 Horizontal V-Twin. U.S.
Environmental Protection Agency, November 22, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0087.
42. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
43. "Research on the New Control Method Using Crankshaft Rotational Changes for Electronically Controlled FI
System of Small Motorcycle Single Cylinder Engine," Toichiro Hikichi, Tetsuya Kaneko, and Toshimitsu Nakajima,
Honda R&D Co., Ldt, SAE Paper 2006-32-0108, Small Engine Technology Conference and Exhibition, San
Antonio, Texas, November 13-16, 2006, Docket Identification EPA-HQ-OAR-2004-0008.
44. "Fuel Injection System for Small Motorcycles," Michihisa Nakamure, Yuuichirou Sawada, and Shigeki
Hashimoto, Yamaha Motor Co., Ltd., SAE Paper 2003-32-0084, September 2003,Docket Identification EPA-HQ-
OAR-2004-0008.
45. "Honda Releases the iGX440 Next-Generation General Purpose Engine with Electronic Control Technology-a
World's First," Honda Motor Co., Ldt., Press Release, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-
0008.
46. "Honda's iGX Intelligent Engine," Honda Motor Co., Ldt., Product Information, April 2007, Docket
IdentificationEPA-HQ-OAR-2004-0008.
4-86
-------
Feasibility of Exhaust Emission Control
47. "Evinrude E-TEC Presentation; Ann Arbor Michigan," presented to U.S. EPA by Fernando Garcia of BRP on
September 30, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0246.
48. "Exhaust Emissions from Marine Engines Using Alternate Fuels," Michigan Automotive Research Corporation,
Prepared for NMMA, October 8, 1992, Docket Identification EPA-HQ-OAR-2004-0008-0205.
49. "Effects of Transience on Emissions from Inboard Marine Engines," EPA memo by Mike Samulski, May 30,
1996, Docket Identification EPA-HQ-OAR-2004-0008-0209.
50. Letter from Jeff Carmody, Santa Barbara Air Quality management District, to Mike Samulski, U.S. EPA, July
21, 1997, Docket Identification EPA-HQ-OAR-2004-0008-0193.
51. "National Marine Manufacturers Association's Small Business Boat Builder and Engine Manufacturers
Comments in Response to EPA's Initial Regulatory Flexibility Analysis Regarding EPA's Plans to Propose
Emission Regulations for Recreational Marine Gas and Diesel Powered Sterndrive/Inboard Engines," July 12, 1999,
Docket Identification EPA-HQ-OAR-2004-0008-0213.
52. "Emissions from Marine Engines with Water Contact in the Exhaust System," Mace, et. al., SAE Paper 980681,
1998, Docket Identification EPA-HQ-OAR-2004-0008-0219.
53. "Marine Gasoline Engine Testing," Carroll, J., White, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2001, Docket A-2000-01, Docket Identification EPA-HQ-OAR-2004-0008-0253.
54. "Marine Gasoline Engine and Boat Testing," Carroll, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0254.
55. "Exhaust Emission Testing of Two High-Performance SD/I Marine Engines," EPA Memo from Mike Samulski
and Matt Spears, June 23, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0349.
56. Email from Mark Reichers, Mercury Marine, to Mike Samulski, US EPA, "1075 / 850 Test Data," June 9, 2005,
Docket Identification EPA-HQ-OAR-2004-0008-0295.
57. Email from Mark Reichers, Mercury Marine, to Mike Samulski, US EPA, "525 EFI Baseline Emissions Data,"
June 10, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0296.
58. Ray, P., Brown, R., " Ilmor Comments on the Notice of Proposed Rulemaking for Control of Emissions from
Nonroad Spark-Ignited Engines and Equipment 72 Fed. Reg. 28,908 (May 18,2007)," August 3, 2007.
59. "EGR Test Data from a Heavy-Duty Gasoline Engine on the E4 Duty Cycle," EPA Memo from Joe McDonald
and Mike Samulski, July 12, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0199.
60. "Marine Gasoline Engine Testing," Carroll, J., White, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0253.
61. "Marine Gasoline Engine and Boat Testing," Carroll, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0254.
62. "Marine Gasoline Engine Testing," Carroll, J., White, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0253.
63. "Summary of SwRI Project for CARD entitled, 'Development of Low Emissions SD/I Boats," White, J.,
Southwest Research Institute, August 14, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0263.
4-87
-------
Final Regulatory Impact Analysis
64. "Development of Low Emissions; SD/I Boats; Project Update," Carrol, I, Southwest Research Institute,
September 11, 2003, Docket Identification EPA-HQ-OAR-2004-0008-0264.
65. "Public Meeting to Consider a Status Report on Catalyst Testing of Spark-Ignition Inboard/Sterndrive
Pleasurecraft: Staff Report," State of California; Air Resources Board, October 28, 2004, Docket Identification EPA-
HQ-OAR-2004-0008-0236.
66. "Marine Gasoline Engine and Boat Testing," Carroll, I, Southwest Research Institute, Prepared for U.S. EPA,
September, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0254.
67. E-mail from Rich Waggoner, Indmar, to Mike Samulski, U.S. EPA, "SBREFA notes from Indmar," July 24,
2006, Docket Identification EPA-HQ-OAR-2004-0008-0438.
68. http://www.arb.ca.gov/msprog/offroad/cert/cert_results.php?order=0, February 14, 2008.
69. Garcia, A., Beamer, B., Earnest, G., "In Depth Survey Report of Carbon Monoxide Emissions and Exposures on
Express Cruisers Under Various Operating Conditions," National Institute for Occupational Safety and Health,
Report No. EPHB 289-1 la, January 2006, Docket Identification EPA-HQ-OAR-2004-0008-0396.
70. Garcia, A., Beamer, B., Earnest, G., "In Depth Survey Report of Carbon Monoxide Emissions and Exposures on
Express Cruisers Under Various Operating Conditions," National Institute for Occupational Safety and Health,
Report No. EPHB 289-1 la, January 2006, Docket Identification EPA-HQ-OAR-2004-0008-0396.
71. "Summary of Government/Industry Carbon Monoxide Workshop and Two Follow-Up Meetings," Memo from
Mike Samulski, U.S. EPA to Docket OAR-2004-0008, March 31, 2004, Docket Identification EPA-HQ-OAR-2004-
0008-0008.
72. "Carbon Monoxide Follow-up Meeting; Miami FL - IBEX Show; Minutes," American Boat and Yacht Council,
October 19, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0286.
73. "News Release: Floating Environmental Laboratory Demonstrates Westerbeke's Safe-CO™ Ultra-Low CO
Emissions," Westerbeke Engines & Generators, September 26, 2005, Docket Identification EPA-HQ-OAR-2004-
0008-0298.
74. "Create the Right Atmosphere; Kohler Marine Generators Helping Make Boating Safer," Presented by Kohler
Power Systems at the 2005 International Boatbuilders' Exhibition and Conference, October 20, 2005, Docket
IdentificationEPA-HQ-OAR-2004-0008-0292.
75. Zimmer, A., Earnest, G., Kurimo, R., "An Evaluation of Catalytic Emission Controls and Vertical Exhaust
Stacks to Prevent Carbon Monoxide Poisonings from Houseboat Generator Exhaust," National Institute for
Occupational Safety and Health, EPHB 171-36a, September 2005, Docket Identification EPA-HQ-OAR-2004-0008-
0397.
76. "Duty Cycle for Recreational Marine Engines," Morgan, E., Lincoln, R., SAE Paper 901596, 1990, Docket
IdentificationEPA-HQ-OAR-2004-0008-0216.
77. "Reciprocating Internal Combustion Engines - Exhaust Emission Measurement - Part 4: Test Cycles for
Different Engine Applications," ISO/DIS 8178-4, International Standards Organization, Docket Identification EPA-
HQ-OAR-2004-0008-0503.
78. Ray, P., Brown, R., " Ilmor Comments on the Notice of Proposed Rulemaking for Control of Emissions from
Nonroad Spark-Ignited Engines and Equipment 72 Fed. Reg. 28,908 (May 18,2007)," August 3, 2007.
4-8
-------
Feasibility of Exhaust Emission Control
79. Morgan, E., Lincoln, R., "Duty Cycle for Recreational Marine Engines," SAE Paper 901596, 1990, Docket
IdentificationEPA-HQ-OAR-2004-0008-0216.
80. Carroll, I, "Determination of Operating Ranges of Marine Engines," prepared by Southwest Research Institute
for the National Marine Manufacturers Association, August 2004, Docket Identification EPA-HQ-OAR-2004-0008-
0255.
81. Samulski, M., "Exhaust Emission Testing of a Two-Stroke and a Four-Stroke Marine Engine; Results and
Procedures," U.S. EPA, Memo to Docket #A-92-98, Docket Identification EPA-HQ-OAR-2004-0008-0097.
82. Wasil, I, Montgomery, D., Strauss, S., Bagley, S., "Life Assessment of PM, Gaseous Emissions, and Oil Usage
in Modern Marine Outboard Engines," SAE-GRAZ 43, Society of Automotive Engineers International, 2004,
Docket Identification EPA-HQ-OAR-2004-0008-0262.
83. Klak, Joe, "Marine NTE Zones," Bombardier Recreational Products, October 26, 2006, Docket Identification
EPA-HQ-OAR-2004-0008-0508.
84. "Marine Gasoline Engine Testing," Carroll, I, White, I, Southwest Research Institute, Prepared for U.S. EPA,
September, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0253.
85. "Marine Gasoline Engine and Boat Testing," Carroll, I, Southwest Research Institute, Prepared for U.S. EPA,
September, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0254.
86. "Summary of SwRI Project for CARB entitled, 'Development of Low Emissions SD/I Boats," White, J.,
Southwest Research Institute, August 14, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0263.
87. "Development of Low Emissions; SD/I Boats; Project Update," Carrol, J., Southwest Research Institute,
September 11, 2003, Docket Identification EPA-HQ-OAR-2004-0008-0264.
88. "Marine Gasoline Engine Testing," Carroll, J., White, J., Southwest Research Institute, Prepared for U.S. EPA,
September, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0253.
89. Mantey, C., "GMPT Study - Stoich Limits for Engine Protection," Presented by General Motors to EPA on
November 19, 2007.
90. Memorandum from Mark Wolcott to Charles Gray, "Ambient Temperatures Associated with High Ozone
Concentrations," U.S. EPA, September 6, 1984, Docket Identification EPA-HQ-OAR-2004-0008-0235.
91. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
92. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
4-89
-------
Feasibility of Evaporative Emission Control
CHAPTER 5: Feasibility of Evaporative Emission Control
Section 213(a)(3) of the Clean Air Act presents statutory criteria that EPA must evaluate
in determining standards for nonroad engines and vehicles including marine vessels. The
standards must "achieve the greatest degree of emission reduction achievable through the
application of technology which the Administrator determines will be available for the engines
or vehicles to which such standards apply, giving appropriate consideration to the cost of
applying such technology within the period of time available to manufacturers and to noise,
energy, and safety factors associated with the application of such technology." This chapter
presents the technical analyses and information that form the basis of EPA's belief that the new
evaporative emission standards are technically achievable accounting for all the above factors.
The evaporative emission standards for Small SI equipment and Marine SI vessels are
summarized in the Executive Summary. This chapter presents available emissions data on
baseline emissions and on emission reductions achieved through the application of emission
control technology. In addition, this chapter provides a description of the test procedures for
evaporative emission determination.
Evaporative emissions from equipment and vessels using spark-ignition (SI) engines can
be very high. This is largely because Small SI and Marine SI applications generally have fuel
tanks that are vented to the atmosphere and because materials used in the construction of the
plastic fuel tanks and hoses generally have high permeation rates. Evaporative emissions can be
grouped into five categories:
DIURNAL: Gasoline evaporation increases as the temperature rises during the day,
heating the fuel tank and venting gasoline vapors. We also include, under this heading, diffusion
losses which are vapors that will escape from an open vent even without a change in
temperature.
PERMEATION: Gasoline molecules can saturate plastic fuel tanks and rubber hoses,
resulting in a relatively constant rate of emissions as the fuel continues to permeate through these
components.
RUNNING LOSSES: The hot engine and exhaust system can vaporize gasoline when the
engine is running.
HOT SOAK: The engine remains hot for a period of time after the engine is turned off
and gasoline evaporation continues.
REFUELING: Gasoline vapors are always present in typical fuel tanks. These vapors are
forced out when the tank is filled with liquid fuel.
5-1
-------
Final Regulatory Impact Analysis
5.1 Diurnal Emissions
In an open fuel tank, the vapor space is at atmospheric pressure (typically about 14.7 psi),
and contains a mixture of fuel vapor and air. At all temperatures below the fuel's boiling point,
the vapor pressure of the fuel is less than atmospheric pressure. This is also called the partial
pressure of the fuel vapor. The partial pressure of the air is equal to the difference between
atmospheric pressure and the fuel vapor pressure. For example, in an open-vented fuel tank at
60°F, the vapor pressure of typical gasoline is about 4.5 psi. In this example, the partial pressure
of the air is about 10.2 psi. Assuming that the vapor mixture behaves as an ideal gas, then the
mole fractions (or volumetric fractions) of fuel vapor and air is equal to their respective partial
pressures divided by the total pressure; thus, the fuel would be 31 percent of the mixture
(4.5/14.7) and the air would be 69 percent of the mixture (10.2/14.7).
Diurnal emissions occur when the fuel temperature increases, which increases the
equilibrium vapor pressure of the fuel. For example, assume that the fuel in the previous
example was heated to 90°F, where the vapor pressure that same typical fuel is about 8.0 psi. To
maintain the vapor space at atmospheric pressure, the partial pressure of the air would need to
decrease to 6.7 psi, which means that the vapor mixture must expand in volume. This forces
some of the fuel-air mixture to be vented out of the tank. When the fuel later cools, the vapor
pressure of the fuel decreases, contracting the mixture, and drawing fresh air in through the vent.
When the fuel is heated again, another cycle of diurnal emissions occurs. It is important to note
that this is generally not a rate-limited process. Although the evaporation of the fuel can be
slow, it is generally fast enough to maintain the fuel tank in an essentially equilibrium state.
As fuel is used by the engine, and the liquid fuel volume decreases, air is drawn into the
tank to replace the volume of the fuel. (Note: the decrease in liquid fuel could be offset to some
degree by increasing fuel vapor pressure caused by increasing fuel temperature.) This would
continue while the engine was running. If the engine was shut off and the tank was left
overnight, the vapor pressure of the fuel would drop as the temperature of the fuel dropped. This
would cause a small negative pressure within the tank that would cause it to fill with more air
until the pressure equilibrated. The next day, the vapor pressure of the fuel would increase as the
temperature of the fuel increased. This would cause a small positive pressure within the tank
that would force a mixture of fuel vapor and air out. In poorly designed gasoline systems, where
the engine or exhaust is very close to the fuel tank the engine/exhaust heating may cause large
amounts of gasoline vapor to be vented directly to the atmosphere.
Several emission-control technologies can be used to reduce diurnal evaporative
emissions. Many of these technologies also control running loss and hot soak emissions and
some could be used to control refueling emissions. We believe manufacturers will have the
opportunity use a wide variety of technology approaches to meet the evaporative emission
standards. The advantages and disadvantages of the various possible emission-control strategies
are discussed below. This section summarizes the data and rationale supporting the diurnal
emission standard for Marine SI vessels and Small SI equipment presented in the Executive
Summary.
5-2
-------
Feasibility of Evaporative Emission Control
5.1.1 Baseline Emissions
5.1.1.1 Marine Vessels
We tested two aluminum marine fuel tanks in their baseline configurations for diurnal
emissions. Aluminum fuel tanks were used so that permeation emissions would not occur during
the testing. The 17 gallon aluminum tank was constructed for this testing, but is representative
of a typical marine fuel tank; the 30 gallon aluminum tank was removed from an 18 foot
runabout. The fuel tanks were tested with the venting through a length of 5/8 inch hose to ensure
that the emissions measured were a direct result of the fuel temperature heating and not diffusion
through the vent (see Section 5.1.3). The advantage of using the aluminum fuel tanks for this
testing was to exclude permeation emissions from the measured results. All of the testing was
performed with fuel tanks filled to 40 percent of capacity with 9RVP1 test fuel.
The diurnal test results are presented in units of grams per gallon capacity of the fuel tank
per day. These units are used because gallons capacity is a defining characteristic of the fuel
tank. Diurnal vapor formation itself is actually a function of the vapor space above the fuel in
the fuel tank rather than the total capacity.
Table 5.1-1 presents the test results compared to anticipated results. The anticipated
results are based on the Wade model which is a set of theoretical calculations for determining
diurnal emissions based of fill level, fuel RVP, and temperature profile. These calculations are
presented in Chapter 3. Although the Wade model over-predicts the vapor generation, it does
show a similar trend with respect to temperature. To account for this over prediction, we use a
correction factor of 0.78. This correction factor is based on empirical data1, has historically been
used in our automotive emission models, and appears to be consistent with the data presented
below.
1 Reid Vapor Pressure (psi). This is a measure of the volatility of the fuel. 9 RVP represents a typical
summertime fuel in northern states.
5-3
-------
Final Regulatory Impact Analysis
Table 5.1-1; Baseline Diurnal Evaporative Emission Results (varied temperature)
Temperatures
22-36°C (72-96°F)
22-36°C (72-96°F)
24-33°C (74-91°F)
22 - 30°C (71 - 86°F)
25-3TC (77-88°F)
26 - 32°C (78 - 90°F)
28-31°C(82-87°F)
Capacity
[gallons]
17
30
30
30
30
30
30
Measured
[g/gallon/day]
1.40
1.50
1.13
0.88
0.66
0.85
0.47
Wade Model
[g/gallon/day]
2.30
2.30
1.33
1.02
0.88
1.04
0.43
Corrected Wade
[g/gallon/day]
1.79
1.79
1.04
0.80
0.69
0.81
0.34
5.1.1.2 Small SI Equipment
We contracted with an outside lab for the testing of thirteen Small SI fuel tanks over
various test temperature profiles.2'3 This testing was performed with the tanks filled to 50
percent capacity with certification gasoline and is discussed in more detail below in the
Section 5.2.1. This data is presented in Table 5.1-2. In addition, in cases where the fuel
temperature profiles were within the input range of the Wade model for diurnal emissions,
theoretical emissions were also calculated using the same correction factor discussed above for
marine fuel tanks. As shown below, the measured values are fairly consistent with the
theoretical values.
5-4
-------
Feasibility of Evaporative Emission Control
Table 5.1-2: Fuel Temperature Measurements During Operation of Small SI Equipment
Equipment Type
Riding mower
Walk-behind
mower
Generator set
Pressure washer
Fuel Capacity
[gallons]
1.1
1.4x2
1.7
2.5
3.0
6.5
6.5x2
0.34
0.25
0.22
8.5
7.0
1.8
Temperature
Profile °C
15.7-28.4
21.9-29.7
19.5-30.3
27.0-35.0
26.6-28.4
24.3 -33.2
20.5-23.9
23.3 -33.0
28.7-46.7
28.7-59.7
20.6-25.8
25.8-50.0
19.0-50.6
Measured HC
grams/gallon
0.92
0.88
0.82
1.29
0.25
1.20
0.26
0.76
4.92
36.9
0.45
9.90
11.6
Theoretical HC
grams/gallon
0.91
0.71
0.94
1.16
0.17
1.08
0.23
1.18
NA*
NA*
0.38
NA*
NA*
* outside the temperature range of the model
The California Air Resources Board performed diurnal testing on seven pieces of
handheld equipment and 20 pieces of non-handheld equipment by placing the whole equipment
in a SHED.4 They filled the fuel tanks to 50 percent with 7 RVP fuel and tested over their 65-
105° F summer day test cycle. Because the entire piece of equipment was included in these
tests, not only were diurnal venting emissions measured, but tank and hose permeation as well
(plus any potential leaks). Average test results by equipment type are presented in Table 5.1-3.
Table 5.1-3: ARB Measurement of Evaporative Emissions from Small SI Equipment
(7 RVP California Certification Fuel, 50% Fill, 65-105°F)
Equipment Type
Handheld equipment
Walk-behind lawnmowers
Generators
Riding Mowers
Edgers
Tiller
Number of Data Points
7
12
2
O
2
1
Average Measured HC [grams/day]
1.04
3.51
11.2
8.70
1.53
4.12
5-5
-------
Final Regulatory Impact Analysis
ARB also performed tests on a subset of the equipment using fuel containing MTBE and
fuel containing ethanol to investigate fuel effects. They observed nearly a 50 percent increase in
emissions when an ethanol blend was used compared to an MTBE blend. The reason for this
increase was not discussed, but may have due to increases in permeation caused by the ethanol
or due to differences in fuel volatility. On five pieces of equipment, a California wintertime
cycle (51.6-69.5° F) was used as well. As would be expected, the emissions were reduced
significantly. The theoretical models predict about an 85 percent reduction in diurnal venting
emissions and about a 60 percent reduction in permeation. The observed results were about a 70
percent reduction which is in this range.
5.1.2 Insulation of the Fuel Tank
The diurnal vapor generated in a fuel tank is directly related to the diurnal temperature
profile of the fuel in the tank. A reduction in temperature variation causes less vapor to be
formed. To investigate this effect, we used insulation around the fuel tank to reduce the effect of
the ambient air temperature variation on the fuel temperature variation. In our preliminary
testing, we insulated a 23 gallon rotationally molded marine fuel tank using 3 inch thick
construction foam with an R-value of 15 as defined by 16 CFR 460.5. This testing was
performed with the fuel tank vent open to atmosphere. Table 5.1-4 presents the fuel
temperatures and evaporative emissions over the three day test.
We tested this fuel tank over a three day diurnal test with an ambient temperature of 72-
96°F. This experiment resulted in a 50 percent reduction in emissions from baseline on the
highest of these three test days. The baseline emissions were measured to be 2.5 g/gallon/day;
however, it should be noted that for both the baseline test and the insulated tank tests we did not
control for permeation or diffusion. Over this test, the emissions decreased for subsequent days.
We believe this was due to the fuel temperature cycle stabilizing. Although we did not control
for permeation or diffusion, the results from this preliminary experiment directionally show the
effect of insulation on diurnal emissions.
Table 5.1-4: Evaporative Emission Results for Insulated Flat, Plastic Tank
Test Day
Day#l
Day #2
Day #3
SHED Temperature
22-36°C (72-96°F)
22-36°C (72-96°F)
22-36°C (72-96°F)
Fuel Temperature
22-28°C (72-82°F)
26-30°C (78-86°F)
26-30°C (80-86°F)
Evaporative HC
1.2g/gal/day
l.Og/gal/day
0.8 g/gal/day
In boats with installed fuel tanks, the fuel tank is generally hidden beneath the deck. As a
result, there is a certain amount of "inherent" insulation caused by the boat itself. This effect is
increased for a boat that is stored in the water. The water acts as a cooling medium for the fuel
tank, especially if it is installed in the bottom of the boat. In addition, the thermal inertia of the
fuel in the tank can act to dampen temperature variation imposed from the diurnal heating of the
5-6
-------
Feasibility of Evaporative Emission Control
ambient air. To investigate this effect, we tested several boats by recording the ambient air
temperature and fuel temperatures over a series of days. Two boats were tested on trailers
outside in the summer, two boats were tested on trailers in a SHED, and two boats were tested in
the water on summer days. Table 5.1-5 presents the average results of this testing. The
temperature traces are presented in Appendix 5 A.
Table 5.1-5: Ratio of Fuel to Ambient Temperature Swing for Boats
Boat Type
9 ft. personal watercraft
16 ft. jet boat
18 ft. runabout
16 ft. jet boat
18 ft. runabout
21 ft. deck boat
Test Conditions
outside, on trailer
outside, on trailer
in SHED, on trailer
in SHED, on trailer
outside, in water
outside, in water
Capacity
[gallons]
13
40
30
40
30
20
Fuel Tank
Fill Level
50%
50%
40%
90%
100%
90%
Temperature
Ratio*
66%
52%
68%
33%
19%
27%
* Average ratio of change in fuel temperature to change in ambient air temperature over test days.
In their comments on the 2002 proposed rule, the National Marine Manufacturers
Association presented temperature data on 18 foot runabout, with a 32 gallon tank, tested in a
SHED with an ambient temperature of 72-96°F.5 The average fuel to ambient temperature ratio
was 54 percent for this testing. This ratio is in the range of EPA test results for boats tested on a
trailer. Brunswick also included temperature data in their comments.6 The average days test on
a boat on the water was 19 percent, which is consistent with our water tests. Brunswick's
average for boats tested while stored out of the water was 27 percent which is considerably lower
than the EPA and NMMA testing. Combining all of the EPA and industry data, the average fuel
to ambient temperature ratio (based on test days) is about 20 percent for boats in the water and
50 percent for boats stored out of the water.
During diurnal testing of lawnmowers, ARB found that the fuel and tank skin
temperature follow the ambient temperature closely.7 This same phenomenon would be expected
for other Small SI equipment as well (and portable fuel tanks) because of the small fuel volumes
and because these tanks are generally exposed to ambient air. One issue that we considered was
that Small SI equipment is often stored in garages or sheds. In that case, we were interested in if
the garage or shed acts to insulate the fuel tanks from ambient temperature swings. ARB
collected data on four garages and one shed. This data included summer and winter California
temperature measurements. For each test, the inside and outside temperature were measured for
five days. This data is presented in Table 5.1-7. For the garages, the inside temperature was
generally warmer than outside, but the variable temperature swings were smaller. For the shed,
the inside temperature was warmer and showed higher heat builds than the outside temperature.
5-7
-------
Final Regulatory Impact Analysis
Table 5.1-6 also presents an estimate of the effect on diurnal emissions using the theoretical
equations presented in Chapter 3. No conclusive evidence of was observed to suggest that these
fuel tanks are generally subject to inherent insulation.
Table 5.1-6: Comparison of Ambient to Inside Diurnal Temperature Swings
Season
Winter
Summer
Enclosure
garage D
garage G
garage J
garage A
garage D
garage G
garage J
shed
Inside Temperature °C
Avg T Avg Delta T
13.8
12.1
13.5
27.4
35.9
27.4
27.6
27.1
6.4
9.2
2.4
3.6
11.7
15.7
8.9
20.1
Outside Temperature °C
Avg T Avg Delta T
10.1
5.8
8.0
22.4
30.3
21.3
23.7
23.6
9.3
14.3
7.3
12.2
15.6
19.5
20.3
14.1
Emission
Effect
-8%
-9%
-55%
-63%
20%
23%
-61%
119%
Some of the variance between the fuel temperature and ambient temperature, especially
for larger fuel tanks, is likely due to the thermal inertia of the fuel in the tank. The fuel has mass
and therefore takes time to heat up. ARB performed a study in which the fuel temperature and
ambient temperature were recorded for aboveground storage fuel tanks.8'9 Three fuel tanks sizes
were included in the study: 350, 550, and 1000 gallons. Because of the large size of these tanks,
the thermal inertia effects would be expected to be larger than for typical fuel tanks used in
Marine SI and Small SI applications. For the 350 gallon fuel tank, ARB also measured the effect
of insulating the fuel tank on temperature. Table 5.1-7 presents the results of this testing. Note
that the test results are the average of five days. Ambient temperature on these test days
typically had a minimum in the 60-70°F range and a maximum temperature in the 95-105T
range.
EPA performed testing on 17 gallon marine fuel tank in a SUED over a single 72-96T
diurnal test and measured both ambient and fuel temperature.10 This data is also included in
Table 5.1-7. Note that for the smaller tank, there is little difference between the ambient and fuel
temperature profiles. However, for larger tanks, the fuel temperature has about a 25-30 percent
smaller temperature swing than the ambient temperature. Note that the insulated fuel tank had a
temperature ratio similar to the fuel tank stored in a boat in the water.
5-8
-------
Feasibility of Evaporative Emission Control
Table 5.1.7: Ratio of Fuel to Ambient Temperature for Uninsulated Fuel Tanks
Fuel Tank Type
marine fuel tank
aboveground storage tank
(with insulation)
aboveground storage tank
aboveground storage tank
Tank Capacity [gallons]
17
350
550
1000
Temperature Ratio*
95%
75%
(18%)
70%
76%
* Average ratio of change in fuel temperature to change in ambient air temperature over test days.
5.1.3 Diffusion Effect
For the purposes of this discussion, diffusion refers to the process in which gasoline
vapor penetrates air in an attempt to equalize the concentration throughout the gas mixture. This
transport phenomenon is driven by the concentration gradient and by effective area. In the case
of a mobile source fuel system that has a vent to atmosphere, the fuel vapor concentration is near
saturation in the fuel tank and near zero outside of the fuel system. Therefore, the diffusion rate
is primarily a function of the path between the fuel tank and atmosphere. The following equation
describes the relationship between the flux of gasoline vapor out of the tank, the concentration
gradient, and the vent path:
Flux =
mass
area x time
= Dx
AC
A*
where: D = diffusion coefficient (constant)
AC = concentration gradient
Ax = path length
area = cross sectional area of vent
Based on the above equation, diffusion from a tank through a vent hose would be a
function of the cross-sectional area divided by the length of the hose. Therefore a longer hose
would theoretically limit fuel vapor venting due to diffusion. Whenever a hydrocarbon (HC)
molecule escapes from the fuel tank, a new molecule of air enters the fuel tank to replace the
escaped HC. This brings the concentration of HC vapor in the fuel tank out of equilibrium. To
balance the partial pressures in the fuel tank, more HC must evaporate as HC in the vapor space
is depleted. In this way, the vapor concentration in the fuel tank remains saturated.
5.1.3.1 Marine Fuel Tank Data
In testing diurnal emissions from fuel tanks with open vents, the configuration of the vent
can have a significant effect on the measured emissions due to the diffusion of vapor out of any
5-9
-------
Final Regulatory Impact Analysis
opening in the fuel tank. Depending on the size and configuration of the vent, diffusion can
actually occur when the fuel temperature is cooling. Most marine vessels with an installed fuel
tank vent through a hose. As shown below this configuration can minimize diffusion.
To quantify the diffusion component for a typical fuel tank, we ran four 72-96°F diurnal
tests on a 17 gallon aluminum marine fuel tank using various configurations for venting The
first configuration was with the fuel cap cracked open and the vent sealed, the second
configuration was with a 68 cm length of vent hose, and the third configuration was with a 1000
micron (1 mm) limiting flow orifice in the vent opening. This 1000 micron orifice was large
enough to allow venting without any measurable pressure increase in the fuel tank during the
diurnal test. The fourth configuration was a combination of the limited flow orifice and the vent
hose. Table 5.1-8 presents the results of this testing.
Table 5.1-8: Diurnal Test Results with Varied Venting Configurations
Vent Configuration
cracked fuel cap
68 cm of 5/8" fuel hose
1000 micron orifice
1000
micron orifice + 68 cm of 5/8" fuel hose
Evaporative HC [g/gallon/day]
2.05
1.40
1.47
1.34
The above testing showed a 50 percent higher emission rate for the tank vented through a
cracked fuel cap compared to one vented through a hose. In the test with the cracked fuel cap,
an increase in HC concentration in the SHED was observed throughout the test, even when the
fuel temperature was cooling. For the other three tests, the HC concentration leveled off when
the temperature began to cool. This suggests that the difference in measured emissions of 0.6 -
0.7 g/gal/day was due to diffusion losses.
To further investigate this diffusion effect, we tested the 17 gallon aluminum tank with
several venting configuration, at two constant temperature settings. Under these conditions, all
of the measured evaporative emissions would be expected to be due to diffusion. As seen in
Table 5.1-9, diffusion can be very high with too large of a vent opening unless a vent hose is
used. The two lengths of vent hose tested did not show a significant difference in diffusion
emissions. We believe that the vent hose limits diffusion by creating a gradual gradient in fuel
vapor concentration.
5-10
-------
Feasibility of Evaporative Emission Control
Table 5.1-9: Constant Temperature Test Results with Varied Venting Configurations
Vent Configuration
1/2" ID. fitting
68 cm of 5/8" fuel hose
137 cm of 5/8" fuel hose
1000 micron orifice
22 °C (72°F)
Evaporative HC [g/gal/day]
5.65
0.11
0.07
0.28
36°C (96°F)
Evaporative HC [g/gal/day]
10.0
0.18
0.24
0.41
The above data suggest that, at least for open vent fuel systems, the size and
configuration of the venting system can have a significant effect on evaporative emissions. In
marine applications, there is typically a vent hose attached to the fuel tank. Diffusion emissions
appear to be minimal if the fuel tank is vented through a length of hose. This is probably
because the long residence times in the hose cause more opportunities for molecular collisions
which direct the HC molecules back towards the fuel tank.
One study looked at the evaporation of liquids from a tube filled to various fill heights.11
As the fill height decreased (effectively increasing the length of the tube above the liquid
surface) the evaporation quickly decreased. These results are consistent with the observed
effects of venting through a hose in our testing. Installed marine fuel tanks typically vent
through a hose to the outside of the boat; therefore, diffusion losses are likely relatively small for
these applications. Another study was performed on automotive fuel caps which suggests that a
crack in the gasket on the fuel cap of 1 percent of the gasket area can result in more than 2 grams
of HC emissions per day.12
5.1.3.2 Small SI Fuel Tank Data
For Small SI applications (and portable marine fuel tanks), the tanks are typically vented
through an opening in the fuel cap. Therefore, unless the cap is sealed, we would expect
diffusion emissions to occur. The above data suggest that diffusion can account for a significant
portion of the evaporative HC emissions measured from a metal tank with a small vent in the cap
over a 72-96°F diurnal test. Because diffusion would still occur at constant temperature, the
contribution of diffusion to measured diurnal emissions would increase, on a percentage basis, as
the diurnal temperature swing approached zero.
To investigate the effect of fuel cap design on diffusion for Small SI applications, we
implemented a test program which included four fuel tank configurations (one metal and three
plastic) and the corresponding fuel caps. These four fuel tanks were taken from lawnmowers
using engines from the three lawnmower engine manufacturers with the highest U.S. sales and
represent the majority of lawnmower fuel tanks on the market. Table 5.1-10 presents a
description of these fuel tanks.
5-11
-------
Final Regulatory Impact Analysis
Table 5.1-10: Lawnmower Fuel Tanks Used in Diurnal/Diffusion Testing
Tank
BM
BP
HP
TP
Tank description
metal, 800 ml
plastic, 1175 ml
plastic, 950 ml
plastic, 920 ml
Fuel Cap Vent Description
Three 1/16" dia. holes drilled in top of cap. Four similar holes
drilled in fibrous gasket
Three torturous pathways through plastic gasket, with venting
between tank/cap threads. (Also performed test using a
modified cap similar to the cap used on the metal tank.)
Pinhole in gasket center leading to two indentations in rubber
gasket at mating surface, with venting between tank/cap threads
Four indentations in rubber gasket at mating surface, with
venting between tank/cap threads
We contracted with two outside laboratories to perform the diurnal/diffusion tests for the
Small SI equipment fuel tanks shown above.13'1445'16 In this effort, the fuel tanks were sealed,
except for the vents in the fuel cap, and filled to 40 percent of capacity with 9 RVP fuel. These
tanks were then tested in a mini-SHED over the EPA 72-96°F 24-hour diurnal test procedure.
To minimize the effect of permeation on the test results, new fuel caps and plastic fuel tanks
were used for each test that had not been exposed to fuel or fuel vapor prior to the test.
Under this testing, emissions continued to climb even when temperature was cooling
back from 96°F to 72°F. These emissions were clearly not driven by temperature, so they were
determined to represent diffusion emissions. Total diffusion for the test was determined by
recording the HC emissions that occurred during the last 12 hours of the test (during the cooling
event) and then multiplying these emissions by two to represent 24 hours. Although the peak
temperature occurs after nine hours, only the last 12 hours were used to ensure that the fuel in
the tank was not still heating due to a thermal time lag. Diffusion was then subtracted off the
total HC measurement to determine non-diffusion diurnal emissions. For the fuel cap with the
three holes drilled straight through it, the emissions were so high that it went out of measurement
range near the end of the tests performed by one of the contractors. However, all of the observed
diffusion rates were linear, making it simple to extrapolate the data where necessary. Table 5.1-
11 presents the diurnal and diffusion data from these tests and compares it to the theoretical
diurnal emissions using the corrected Wade equations discussed above. Charts in Appendix 5B
present the time series of the measured HC compared to the mini-SHED temperature.
5-12
-------
Feasibility of Evaporative Emission Control
Table 5.1-11: Diurnal and Diffusion Emissions from Lawnmower Fuel Tanks (g/gal/day)
over a 72-96-72 °F (22.2-35.6-22.2 °C) Temperature Profile
Tank
BM
BP
BP cap 2*
HP
TP
Total HC
47.8
2.1
24.1
1.6
2.1
Diffusion
43.6
0.1
19.3
0.1
0.2
Diurnal
4.2
2.0
4.8
1.5
2.0
Wade Diurnal
1.8
1.8
1.8
1.8
1.8
modified to be similar to cap on metal tank (BM)
The fuel caps in the above table for the lawnmower tanks labeled as BM and BP cap 2
resulted in very high diffusion emissions. Although this fuel cap type is a common design used
in Small SI applications, it may represent one of the worst case configurations for diffusion.
There are three small holes in the cap itself, and four small holes in the fibrous material
imbedded in the inside of the cap. Presumably, this design was intended to minimize fuel from
splashing out of the tank while still allowing the tank to breathe to prevent pressure or vacuum
from occurring in the tank. Because the carburetor on this lawnmower is gravity fed, too much
vacuum in the fuel tank could cause the engine to stall from lack of enough fuel. The reason that
this may be a worst case configuration is that there is a direct (and relatively large) path for fuel
vapor to escape from the fuel tank.
The other three fuel cap designs were also from stock lawnmower fuel systems. In all
three of these designs, the venting occurred through small grooves in the gasket that seals the
mating between the fuel cap and the fuel tank. The venting then occurs through the thread paths
between the cap and tank. As a result, vapor and air must pass through a tortuous pathway to
enter or leave the tank. This tortuous pathway appears to limit diffusion in much the same way
as venting through a long hose does.
The above emission testing was repeated except that the vents in the fuel cap were sealed
and the tank was vented through a 8 inch length of 1/4" ID. hose. A lawnmower air intake filter
was attached to the end of this hose in order to simulate the venting configuration on a
lawnmower with running loss control. To minimize the effect of permeation, a low permeation
barrier hose was used that had never before been exposed to fuel or fuel vapor. The test results
in which the tanks were vented through hoses are presented in Table 5.1-12.
5-13
-------
Final Regulatory Impact Analysis
Table 5.1-12: Diurnal and Diffusion Emissions from Lawnmower Fuel Tanks (g/gal/day)
with Modified Venting Through Hose/Air Filter to Simulate Running Loss Control
over a 72-96-72 °F (22.2-35.6-22.2 °C) Temperature Profile
Tank
BM
BP
BP cap 2*
HP
TP
Total HC
vent through stock cap
47.8
2.1
24.1
1.6
2.1
Total HC
vent through hose/filter
12.9
1.9
1.9
2.0
2.9
Reduction in
Total HC
34.8
0.2
22.2
(0.4)
(0.7)
* modified to be similar to cap on metal tank (BM)
As shown in the table above, venting through the hose greatly reduced the measured
emissions compared to the BM cap vent. When vented through the hose configuration, diffusion
emissions were on roughly the same order as when the tortuous cap vents were used. This is
consistent with the data presented earlier on marine fuel tanks vented through a hose. In an in-
use running loss system, a valve or limited flow orifice would likely also be in the vent line.
These components would likely further reduce, or even eliminate, diffusion emissions.
There was some concern that diffusion may have been underestimated in the above tests
because air flowing back into the fuel tank during the cooling period may have limited diffusion
by pulling HC molecules back into the fuel tank. In addition, we believed that testing at constant
temperature would allow us to more directly measure diffusion. Therefore, the above testing
was repeated at a constant temperature of 29°C.17'18'19 However, it should be noted that this
testing may have overestimated diffusion somewhat because of small temperature fluctuations
(less than 0.5 °C) around the average during the test. Therefore, any HC measurements from the
"constant" temperature testing may have overstated diffusion due to vapor generated by the
repeated mini-diurnal cycles during in the test. These test results are presented in Table 5.1-13.
5-14
-------
Feasibility of Evaporative Emission Control
Table 5.1-13: Isothermal [29 °C] Diurnal and Diffusion Emissions from
Lawnmower Fuel Tanks (g/gal/day) with Modified Venting
Through Hose/Air Filter to Simulate Running Loss Control
Tank
BM
BP
BP cap 2*
HP
TP
Total HC
vent through stock cap
43.2
1.3
29.3
1.0
0.9
Total HC
vent through hose/filter
8.9
1.0
1.0
0.8
0.9
Reduction in
Total HC
34.3
0.3
28.3
0.2
0.0
* modified to be similar to cap on metal tank (BM)
At constant temperature, the relationship between measured diffusion emissions between
the venting configurations was consistent with the variable temperature testing. However, the
indicated diffusion results were somewhat higher. These higher results were influenced by two
effects. In the variable temperature testing, the diffusion was measured during the cooling
period when air was being drawn into the fuel tank. This would reduce diffusion into the SHED
because escaping HC molecules would need to overcome the air flow into the tank. At the same
time, the constant temperature test may have overstated diffusion due to the measured small
fluctuations in temperature that may have caused mini-diurnal cycles. Likely, the actual
diffusion rates are somewhere in-between the results presented in Tables 5.1-11 and 5.1-12.
Appendix 5B contains data charts that present the results of the Small SI diffusion testing in
more detail.
Although the results are presented above on a gram per gallon basis for comparison with
diurnal emissions, diffusion appears to be more a function of orifice size that fuel tank size.
Presumably, the diffusion rate on a grams per day basis would be the same through a given
orifice regardless of size of the vapor space. This is reflected in the data above in that the
permeation rates on a gram per gallon basis from the lawnmower fuel tanks with holes in the fuel
cap were much larger than for the marine fuel tank in the testing discussed earlier. At the same
time, larger fuel tanks may be designed with larger orifice sizes to account for higher amounts of
vapor expansion in the tank.
5.1.4 Carbon Canister
The primary diurnal evaporative emission control device used in automotive applications
is a carbon canister. With this technology, vapor generated in the tank is vented through a
canister containing activated carbon (similar to charcoal). The fuel tank must be otherwise
sealed; however, this only results in a minimal amount of pressure in the tank. The activated
carbon collects and stores the hydrocarbons. Once the engine is running, purge air is drawn
through the canister and the hydrocarbons are burned in the engine. These carbon canisters
5-15
-------
Final Regulatory Impact Analysis
generally are about a liter in size for an automotive tank and have the capacity to store three days
of vapor over the test procedure conditions. For automotive applications, this technology
reduces diurnal emissions by more than 95 percent.
In a marine application, the vessel may sit for weeks without an engine purge; therefore,
canisters were not originally considered to be a practical technology for controlling diurnal vapor
from boats. Since that time, however, we have collected information showing that, during
cooling periods, the canister is purged sufficiently enough so that it can be used effectively to
reduce diurnal emissions. When the fuel in the tank cools, fresh air is drawn back through the
canister into the fuel tank. This fresh air will partially purge the canister and return
hydrocarbons back to the fuel tank.20'21 Therefore, the canister will have open sites available to
collect vapor during the next heating event. Test data presented below show that a canister that
starts empty is more than 90 percent effective at capturing hydrocarbons until it reaches
saturation. Once the canister reaches saturation, it is still capable of achieving more than a 60
percent reduction in diurnal emissions due to passive purging. Passive purging occurs as a result
of fresh air that is pulled through the canister during fuel tank cooling periods. With the addition
of an engine (active) purge, greater reductions would be expected.
We tested a 30 gallon aluminum fuel tank over three, multiple-day diurnal cycles with
and without a charcoal canister. The carbon canister was 2.1 liters in size with a butane working
capacity (BWC) of 11 g/dL (based on EPA test) and was aged using multiple 24 hour diurnal
cycles prior to testing. In our first test, the fuel temperature was cycled from 72-96°F using a
heating blanket in a SUED for at total of 28 days. Because we were not able to test over
weekends, we brought the fuel temperature down to 72°F and held it to prevent the generation or
purging of vapors. On Mondays, we saw higher vapor rates than the rest of the week which was
likely due to the vapor redistributing itself equally through the canister over the weekend when
the temperature was held constant. Under normal conditions, the continued diurnal cycles would
maintain a gradient through the canister and this effect would not occur. Appendix 5C contains
graph showing the results of the 28 day test. This test is interesting because we began with a
purged canister and were able to observe the loading of the canister over the first few days. It
took about five test days to achieve canister breakthrough and another ten test days before the
canister loading/purging cycle stabilized.
Once the canister was saturated, the emissions results stabilized. Therefore, for the
subsequent canister tests, we began with a loaded canister and tested for four days. The results
were collected beginning after the first night so that the canister would have a cooling cycle for
back-purge. Table 5.1-14 presents our test results for the baseline and stabilized with canister
diurnal emission rates.
5-16
-------
Feasibility of Evaporative Emission Control
Table 5.1-14: EPA Diurnal Emission Test Results With and Without a Canister
on a 30 Gallon Aluminum Marine Fuel Tank [g/gal/day]
Temperature Range
22.2-35.6°C (72-96°F)
25.6-32.2°C (78-90°F)
27.8-30.6°C (82-87°F)
Baseline
1.50
0.85
0.47
With a Canister
0.52
0.28
0.14
Reduction
65%
67%
71%
Marine manufacturers raised the concern that the high humidity in the areas where boats
are used would be detrimental to this technology. They stated that the carbon could become
saturated with water vapor, thereby reducing the available sites for hydrocarbon capture. These
manufacturers also commented that carbon canisters may not be able to survive shocks and
vibration that would be seen on a boat. Carbon canisters have been used in automotive
applications for decades, which are subject to high humidity (rainy days) and shocks and
vibration. In addition, one manufacturer, who is a primary supplier to the automotive industry,
has developed a new grade of carbon that has low moisture adsorption characteristics and about
40 percent harder than typical automotive carbon.22'23 This carbon has been designed specifically
for marine applications. Based on this manufacturer's testing, more than a 60 percent reduction
in diurnal vapor emissions can be achieved with a passive purge system. This reduction is based
on a canister capacity of 0.03 to 0.04 liters of carbon per gallon of fuel tank capacity.
The National Marine Manufacturers Association performed a test program has to
demonstrate the durability of carbon canisters in marine applications. This test program included
installing carbon canisters on a total of fourteen boats made by four boat builders.24 These boat
types included cruisers, runabouts, pontoon boats, and fishing boats. The carbon canister design
used for these boats is a simple cylinder that can be cut to length with end caps and mounting
brackets. The canisters were installed in the vent lines and a valve was added to prevent liquid
fuel from reaching the canister during refueling. These canisters use marine grade carbon. At
the end of this test program, each of the canisters were tested for working capacity and each
canister showed proper performance.25
Another issue that has been raised has been the ability of carbon canisters to pass the
Coast Guard flame test. The carbon canisters could be made out of a variety of materials,
including metal. Even a thin-walled nylon fuel tank could be manufactured to pass the flame test
if a flame-resistant coating or cover were used. One study attempted to ignite a carbon canister
that was loaded with fuel vapor.26 When an ignition source was applied to the canister vent, the
gases exiting the canister were ignited and burned as a small, steady flame until the canister tube
opening began to melt. No explosion occurred. In any case, as with the vent line, if the carbon
canister is self-draining, then the canister would not likely hold the five ounces of fuel, specified
in 33 CFR 183.558, to trigger the need for flame protection.
Manufacturers have raised the concern that it is common for liquid fuel to pass out the
5-17
-------
Final Regulatory Impact Analysis
Fill neck
Canister
vent line during refueling. If there were a canister in the vent line it would become saturated
with fuel. While this would not likely cause permanent damage to the canister, we believe
marine fuel systems should prevent liquid fuel from exiting the vent line for both environmental
and safety reasons. A float valve or small orifice in the entrance to the vent line from the fuel
tank would prevent liquid fuel from reaching the canister or escaping from the tank. Any
pressure build-up from such a valve would cause fuel to back up the fill neck and shut off the
fuel dispensing nozzle. In addition,
a vapor space should be included to Fi§ure 5A~1: Canister Installation Example
account for fuel expansion. These Vent line
fuel system design considerations are
straight-forward, long used in other
applications,27 are applicable to
boats,28 and would have the added
benefit of minimizing fuel spillage,
from boats, into the water. One
possible design for preventing fuel
from reaching the canister, due to
refueling, sloshing, or expansion, is
shown in Figure 5.1-1.
Recently, the California Air Resources Board (ARE) performed diurnal emission testing
on a commercial mower and a generator with 6 gallon fuel tanks and 0.65 liter canisters.29 Their
testing showed better than 50 percent reductions, on average, in diurnal emissions through the
use of canisters without an engine purge. The testing was performed over two diurnal
temperature ranges, 53-71°F and 65-105°F which are intended to represent an average day and a
high temperature episode.
Over a decade ago, testing performed on a car showed similar results.30 A 1988 Regency
98 with an 18 gallon fuel tank was subjected to an 8 day diurnal without driving. This diurnal
was performed using a 72-96°F temperature profile, a tank filled to 40 percent with 9RVP
gasoline, and a purged canister at the beginning of testing. The test results showed, that the
canister loading/purging cycle began to stabilize after 6 days. Due to the canister back-purge,
the stabilized diurnal emission rate about 11.5 grams per day which was more than a 50 percent
reduction compared to baseline.
A manufacturer of activated carbon performed studies of ethanol fuel blend and carbon
bed temperature on carbon efficiency.31 Testing was performed with carbon canisters using
gasoline, E10, and E85 fuel for onboard vapor refueling emissions efficiency. The emissions
control was similar for each of the test fuels. Testing was also performed to measure gasoline
working capacity for carbon soaked at temperatures ranging from 25 to 80°C. Over this range
only a 10 percent decrease in working capacity was observed with increasing temperature. Over
the 25-40°C range, which is more representative of boat or Small SI equipment use, the effect
was only 1-2 percent. Based on the results from these studies, carbon canister efficiency would
be expected to be effective at reducing diurnal emissions over the range of fuels and
temperatures that may be seen in use.
5-18
-------
Feasibility of Evaporative Emission Control
5.1.5 Sealed System with Pressure Relief
Evaporative emissions are formed when the fuel heats up, evaporates, and passes through
a vent into the atmosphere. By closing that vent, evaporative emissions are prevented from
escaping. However, as vapor is generated, pressure builds up in fuel tank. Once the fuel cools
back down, the pressure subsides. One way to control these emissions is to seal the fuel system.
However, depending on the fuel tank design, a pressure relief valve may be necessary which
would limit the control.
5.1.5.1 Pressure Relief Valve
For most marine applications, U.S. Coast Guard safety regulations require that fuel tanks
be able to withstand at least 3 psi and must be able to pass a pressure impulse test which cycles
the tank from 0 to 3 psi 25,000 times (33 CFR part 183).2 The Coast Guard also requires that
these fuel tanks must be vented such that the pressure in the tank in-use never exceeds 80 percent
of the pressure that the tank is designed to withstand without leaking. The American Boat and
Yacht Council makes the additional recommendation that the vent line should have a minimum
inner diameter of 7/16 inch.32 However, these recommended practices also note that "there may
be EPA or state regulations that limit the discharge of hydrocarbon emissions into the
atmosphere from gasoline fuel systems. The latest version of these regulations should be
consulted."
To prevent pressure from building too high in marine tanks, we first considered a 2 psi
pressure relief valve. This is a typical automotive rating and is below the Coast Guard
requirements. With this valve, vapors would be retained in the tank until 2 psi of pressure is
built up in the tank due to heating of the fuel. Once the tank pressure reached 2 psi, just enough
of the vapor would be vented to the atmosphere to maintain 2 psi of pressure. As the fuel cooled,
the pressure would decrease. In our August 14, 2002 proposal (67 FR 53050) we considered
standards based on a 1 psi valve which would only achieve a modest reduction over the diurnal
test procedure. However this reduction would be significantly greater in use because the test
procedure is designed to represent a hotter than average day. On a more mild day, there would
be less pressure buildup in the tank and the valve may not even need to open. With the use of a
sealed system, a low pressure vacuum relief valve would also be necessary so that air could be
drawn into the tank to replace fuel drawn from the tank when the engine is running.
Manufacturers of larger plastic fuel tanks have expressed concern that their tanks are not
designed to operate under pressure. For instance, although they will not leak at 3 psi,
rotationally molded fuel tanks with large flat surfaces could begin deforming at pressures as low
as 0.5 psi. At 2.0 psi, the deformation would be greater. This deformation would affect how the
tank is mounted in the boat. Also, fuel tank manufacturers commented that some of the fittings
2 These regulations only apply to boats with installed fuel tanks and exclude outboard boats.
However, ABYC recommended practice effectively extends many of these requirements to outboard
boats as well.
5-19
-------
Final Regulatory Impact Analysis
or valves used today may not work properly under 2 psi of pressure. Finally, they commented
that backup pressure-relief valves would be necessary for safety. For smaller fuel tanks, such as
used in personal watercraft, portable fuel tanks, and Small SI equipment, pressure is less of an
issue because of the smaller internal surface area of these fuel tanks. In addition, the
construction of these fuel systems are generally vertically integrated which allows for more
precise control of design parameters. For instance, personal watercraft manufacturers are
already sealing their fuel systems to prevent fuel from spilling into the water. These systems
generally have pressure relief valves ranging from 0.5 to 4.0 psi. In addition, portable fuel tanks
are designed to be sealed without any pressure relief.
We looked at two types of pressure relief strategies: pressure relief valves and limited
flow orifices. Because the Coast Guard requires that fuel systems not exceed 80 percent of their
design capacity of 3 psi, we only looked at pressure relief strategies that would keep the pressure
below 2.4 psi under worst case conditions.
For the pressure relief valve testing, we looked at several pressures ranging from 0.5 to
2.25 psi. The 2.25 psi valve was an off-the-shelf automotive fuel cap with a nominal 2 psi
pressure relief valve and 0.5 psi vacuum relief valve. For the other pressure settings, we used
another automotive cap modified to allow adjustments to the spring tension in the pressure relief
valve. We performed these tests on the 17 gallon aluminum fuel tank to remove the variable of
permeation. Emissions were vented through a hose to prevent diffusion losses from affecting the
measurements. We operated over two temperature profiles. The first set of tests were performed
in a variable temperature SUED with a 72-96°F air temperature profile. This temperature profile
was based on the existing automotive cycle which is intended to represent a typical summer day
on which a high ozone event may occur. The second set of tests were performed using a heating
blanket to create a 78-90°F fuel temperature profile. This testing was intended to represent a
fuel tank in a boat, where the tank may be inherently insulated, during the same ambient
temperature profile. This inherent insulation creates a time lag on the heating and cooling of the
fuel and reduces the amplitude of the temperature profile by half.
As shown in Figure 5.1-2, there was a fairly linear relationship between the pressure
setting of the valve and the emissions measured over the variable-temperature test procedure. In
addition, the slopes of the lines are similar for both test temperature scenarios. This suggests that
over a smaller temperature profile, a greater percent reduction in HC can be achieved at a given
pressure setting. This is reasonable because, in each case, a constant amount of vapor is
captured. In other words, regardless of the temperature profile, the same amount of vapor must
be generated to create a given pressure. For instance, with a 1 psi valve, about 0.4 grams/gallon
of HC are captured over each temperature profile. However, this represents a 50 percent
reduction over a 78-90°F temperature profile while only about a 25 percent reduction over the
72-96°F temperature profile.
5-20
-------
Feasibility of Evaporative Emission Control
Figure 5.1-2: Effect of Pressure Cap on Diurnal Emissions
0.0
0.5 1.0 1.5
pressure relief setting [psi]
2.0
2.5
Portable marine fuel tanks are generally equipped with a valve that allows the tank to be
sealed when not in use. The valve must be opened during engine operation or the draw of fuel
from the fuel tank can create a vacuum in the fuel tank. If the vacuum becomes too great, the
engine will not be able to draw sufficient fuel from the tank and therefore stall. During storage,
the valve may be closed to prevent fuel vapors from escaping. These fuel tanks are designed to
withstand pressurization much greater than would be experienced when a fuel tank is heated by
ambient temperature changes. However, this vapor control strategy relies heavily on user
behavior. According to one survey, the majority of boat owners store their portable marine fuel
tanks with fuel in them, and many do not close the vent during storage.33 Therefore, significant
emission reductions may be achieved with a sealed fuel tank with an automatic vacuum relief
valve. This is discussed further in section 5.1.5.3 below.
The California Air Resources Board tested a lawnmower in a SHED for diurnal
emissions in a baseline configuration, a sealed system, and with various pressure relief settings.34
Because the whole lawnmower was tested, permeation (and potentially leakages) were measured
as well as diurnal venting emissions. The testing was performed over a 65-105°F temperature
cycle with the fuel tank filled to 50 percent with 7 RVP fuel. For the system as a whole, they
measured a 76 percent reduction in emissions when the tank was fully sealed compared to the
open vent configuration. This suggests that diurnal venting made up about 76 percent of the
evaporative emissions measured. Testing using 2, 3, and 4 psi pressure relief valves showed
reductions of 43 percent, 43 percent, and 63 percent respectively. They also collected pressure
data over various diurnal temperature cycles on a lawnmower fuel tank. Over the 65-105°F
cycle, the measured a pressure increase of about 2.5 psi. Even under an extreme cycle of 68-
121°F, the measured increase in tank pressure was about 3.6 psi.
5-21
-------
Final Regulatory Impact Analysis
5.1.5.2 Limited Flow Orifice
Another strategy for maintaining a design pressure is to use a limited flow orifice on the
vent. In our testing, we are looked at three orifice sizes: 25, 75, and 1,000 microns in diameter.
Again, we performed tests over a 72-96°F diurnal using a 17 gallon aluminum tank. To get these
exact orifice sizes, we ordered from a company that specializes in boring holes with a laser
device. These orifices were relatively inexpensive. It should be noted that a smaller tank would
need a smaller orifice and a larger tank could use a larger orifice to build up the same pressure in
the tank. The test results are presented in Table 5.1-15. For all of the tests with the limited flow
orifices, no vent hose was attached.
Table 5.1-15: Diurnal Evaporative Emissions with Limited Flow Orifices
Orifice Diameter (microns)
baseline (open vent with hose)
1000
75
25
Peak Pressure [psi]
0.0
0.0
1.6
3.1
Evaporative HC [g/gallon/day]
1.40
1.47
1.16
0.24
By limiting the flow of the vapor from the tank, emissions were reduced with some
pressure build up in the tank. However, because the vapor is flowing from the tank even at low
pressure, this strategy is less effective for reducing diurnal emissions than a pressure relief valve.
Generally, a higher peak pressure is necessary with the LFO for a given emission reduction. In
addition, the limited flow orifice would have to be sized for worst case conditions to prevent the
tank from reaching too high of a pressure. A LFO sized for worst case conditions would be less
effective under typical conditions because the vapor flow out of the tank could be too low for the
LFO to create a restriction. In comparison, a pressure relief valve would achieve higher percent
reductions under typical conditions than for worst case conditions because the valve would open
less often.
5.1.5.3 Vacuum Relief Valve
For some fuel tanks, pressure relief is not necessary. An example of this is portable
marine fuel tanks which are currently equipped with a manual sealing valve. This valve can be
sealed by the operator during storage to prevent vapor from escaping. Although pressure will
build up during diurnal heating, the fuel tanks are designed to withstand this pressure. However,
the valve must be opened by the operator during engine operation so that a vacuum does not
form in the fuel tank as fuel is drawn to the engine. If this vacuum were to become too high, it
could cause the engine to stall by restricting fuel to the engine.
The existing design requires that the operator close the valve whenever the engine is not
running for diurnal emissions to be controlled. If an automatic vacuum relief valve were used,
5-22
-------
Feasibility of Evaporative Emission Control
then the operator would not need to operate the sealing mechanism. It would always control
diurnal emissions. At the same time, the vacuum relief valve would allow air to be drawn into
the fuel tank when the engine is operating to prevent a significant vacuum from being formed.
One manufacturer's approach to this automatic valve design is to use a diaphragm valve
such as those used in automotive fuel systems.35 This inexpensive design would be able to seal
the tank under pressure, yet open at very low vacuums. This design (or other vacuum relief
valve designs) could be used in any nonroad application where the fuel system is able to
withstand pressure.
5.1.6 Selective Permeability Membrane
Another approach we investigated was fitting a molecular membrane in the vent line.
The theory was that the membrane would allow oxygen and nitrogen to pass through, but block
most longer-chain hydrocarbon molecules. We used a membrane fabricated using Teflon AF®
which is an amorphous fluoropolymer. Because oxygen and nitrogen (and some smaller
hydrocarbons) can pass through the membrane, hydrocarbons can be trapped in the fuel tank.
However, the process for molecules passing through the membrane is slow, so it is important to
size the membrane properly to prevent pressure build-up. This membrane could be placed in the
vent line or directly in an opening in the top of the fuel tank.
Similar membranes are already used for several applications. One manufacturer provides
membranes for a variety of uses such as oxygen or nitrogen enrichment of air or for separation of
hydrocarbons from air.36 One of these uses is to act as a vapor processor to prevent hydrocarbon
vapor from escaping from retail gasoline stations in California.37 Another membrane used for
similar applications allows hydrocarbons to permeate but blocks smaller gases. This membrane
is used in hydrocarbon recovery applications.38 In the above noted applications, the membranes
are typically used with a pump to provide a pressure drop across the membrane which causes
permeation through the membrane. Typically, adequate mixing is needed to maintain an
efficient diffusion rate.
We tested an amorphous fluoropolymer membrane with a surface area of about 40 cm2 in
the vent line of both a 30 and a 17 gallon aluminum fuel tank over three temperature cycles. The
membrane was applied to a wire mesh in a cylindrical shape with the an outside diameter not
much larger than the vent hose. Hydrocarbon emissions and fuel tank pressure were measured.
Over these tests we consistently saw a pressure build up, even over a 24 hour test. To investigate
the impacts of surface area, we increased the surface area by using 3 filters in parallel (single
vent line to assembly). Our test results suggest that the pressures associated with this technology
are comparable with the pressure relief valves needed to achieve the same reductions. However,
this technology may have the potential for meeting our standards if used in conjunction with a
pump to provide a pressure differential across the filter without allowing pressure (and mixing)
to build up in the fuel tank. Our test results are presented in Table 5.1-16.
5-23
-------
Final Regulatory Impact Analysis
Table 5.1-16: Diurnal Venting Emissions with Selective Permeable Membranes
Tank Size
[gallons]
30
17
Venting
open
1 filter
3 filters
open
3 filters
72-96°F
g/gal/day psi
1.50
0.24
0.39
1.40
0.45
0
2.9
2.2
0
2.1
78-90°F
g/gal/day psi
0.85
0.14
-
—
0.30
0
1.5
-
—
1.2
81.6-86.4°F
g/gal/day psi
0.47
0.19
-
—
-
0
0.6
-
—
-
5.1.7 Volume Compensating Air Bag
Another concept for minimizing pressure in a sealed fuel tank is through the use of a
volume compensating air bag.39 The purpose of the bag is to fill up the vapor space in the fuel
tank above the fuel itself. By minimizing the vapor space, less air is available to mix with the
heated fuel and less fuel evaporates. As vapor is generated in the small vapor space, air is forced
out of the air bag, which is vented to atmosphere. Because the bag collapses as vapor is
generated, the volume of the vapor space grows and no pressure is generated.3 Once the fuel
tank cools as ambient temperature goes down, the resulting vacuum in the fuel tank will open the
bag back up.
We tested a 6 gallon portable plastic fuel tank with a 1.5 gallon volume compensating
bag made out of Tedlar. Tedlar is a light, flexible, clear plastic which we use in our labs for
collecting exhaust emissions samples. In our testing, the pressure relief valve never opened
because the volume compensating bag was able to hold the vapor pressure below 0.8 psi for each
of the three days. This testing supports the theory that a volume compensating bag can be used
to minimize pressure in a fuel tank, which in turn, reduces emissions when used in conjunction
with a pressure relief valve.
We did see an emission rate of about 0.4 g/gal/day over the 3 day test. The emission rate
was fairly constant, even when the ambient temperature was cooling during the test. This
suggests that the emissions measured were likely permeation through the tank. Other materials
may be more appropriate than Tedlar for the construction of these bags. The bags would have to
hold up in a fuel tank for years and resist permeation while at the same time be light and flexible.
One such material that may be appropriate would be a fluorosilicon fiber.
3 The Ideal Gas Law states that pressure and volume are inversely related. By increasing the volume of the
vapor space, the pressure can be held constant.
5-24
-------
Feasibility of Evaporative Emission Control
5.1.8 Bladder Fuel Tank
Probably the most effective technology for reducing evaporative emissions from fuel
tanks is through the use of a collapsible fuel bladder. In this concept, a non-permeable bladder
would be installed in the fuel tank to hold the fuel. As fuel is drawn from the bladder, the
vacuum created collapses the bladder. Therefore, there is no vapor space and no pressure build
up. Because the bladder would be sealed, there would be no vapors vented to the atmosphere.
In addition, because there is no vapor space, vapor is not displaced during refueling events. We
have received comments that bladder tanks would be cost prohibitive because its use would
increase tank costs by 30 to 100 percent depending on tank size. However, bladder fuel tanks
have positive safety implications as well and are already sold by at least one manufacturer to
meet market demand in niche applications. Information on this system is available in the
docket.40
We tested a marine bladder fuel tank in our lab for both diurnal and permeation
emissions. Over the diurnal test procedure we saw an emission rate of 0.2 g/gal/day. Because
the system was sealed, this measured emission rate was likely due to permeation through the
bladder and not due to diurnal losses. We later tested the bladder fuel tank for permeation
emissions at 29°C and measured a permeation rate of 0.46 g/gal/day. The bladder used in our
testing was constructed out of polyurethane. The manufacturer of this bladder tank is now
working with a lower permeability material known as THV. THV is a fluoropolymer that can be
used to achieve more than a 95 percent reduction in permeation from current bladder fuel tanks
made out of polyurethane.41 In addition, THV is resistant to ethanol. Permeation rates for these
materials are presented in Appendix 5D.
5.1.9 Floating Fuel and Vapor Separator
Another concept used in some stationary engine applications is a floating fuel and vapor
separator. Generally small, impermeable plastic balls are floated in the fuel tank. The purpose
of these balls is to provide a barrier between the surface of the fuel and the vapor space.
However, this strategy does not appear to be viable for fuel tanks used in mobile sources.
Because of the motion of Small SI equipment and Marine SI vessels, the fuel sloshes and the
barrier would be continuously broken. Even small movements in the fuel could cause the balls
to rotate and transfer fuel to the vapor space. In addition, the unique geometry of many fuel
tanks could case the balls to collect in one area of the tank. However, we do not preclude the
possibility that some form of this approach could be made to work effectively in some mobile
source applications.
5.1.10 Liquid Vapor Trap
One company has developed a
liquid vapor trap that it refers to as a
fuel vapor containment system (VCS).42
The VCS behaves similar to a liquid
trap used in sink drains in that trapped
Figure 5.1-3: Liquid Vapor Trap
II—
fuel tank
L— -^^
VCS
~il Ir
A
B
vent
5-25
-------
Final Regulatory Impact Analysis
liquid creates a barrier to gases. This trap would be placed in the vent line to limit fuel vapor
emitted from the fuel tank. Figure 5.1-3 presents an illustration of the basic concept.
When the temperature in the fuel tank increases, the vapor would expand in the fuel tank.
The fuel vapor would enter chamber A and force more of the liquid into chamber B. This would
provide room for the vapor to expand without allowing vapor to escape through the vent. As the
fuel tank cools, the vapor would condense. This would cause the level of the liquid in chamber A
to rise while the level of the liquid in chamber B would drop. Some pressurization may occur in
the fuel tank with this system, but it would be much less than for a sealed fuel tank due to the
expansion chamber. Any pressure or vacuum in the fuel system would be a function of the VCS
design and would be expected to be less than 0.5 psi. In addition, a pressure relief valve could
be added to the system to protect against any high pressure excursions.
In the initial testing of the VCS, the manufacturer has used water as the liquid barrier.
However, they stated that ethylene-glycol or even oil could be used which would be more stable
liquids and would resist freezing. Diurnal testing was performed on a 25 gallon fuel tank
equipped with a roughly 3 gallon VCS unit.43 Testing was performed in a mini-SHED over the
EPA 72-96°C diurnal cycle for two days. The tank was filled to 50 percent capacity with 9 RVP
certification gasoline. The total weight loss was 1.1 grams on the first day and 2.6 grams on the
second day. Using the higher of the two days, we get a diurnal emission rate of about 0.1
g/gal/day. The peak pressure during this testing was approximately 0.5 psi.
5.2 Running Loss Emissions
Running loss emissions are similar to diurnal emissions except that the fuel temperature
rise is due to heat from the engine or other heat producing components, such as hydraulic
systems, when the engine is running. This section summarizes the data and rationale supporting
the running loss emission standard for Small SI equipment presented in the Executive Summary.
5.2.1 Baseline Emissions
To investigate running loss emissions, we instrumented seven riding lawnmowers, three
walk-behind lawnmowers, two generators, and one pressure washer to measure the fuel
temperature during typical operation. Many of the temperature measurements were made by a
contractor.44 Of the riding mowers, two had fuel tanks in front near the engine, three had fuel
tanks in rear away from engine (but near the hydraulic system), and two were "zero-turn"
mowers that had pairs of side saddle tanks that were relatively close to the rear mounted engine.
All of the riding mowers had plastic fuel tanks. One of the walk-behind mowers had a metal
tank directly mounted to the block while the others had plastic tanks near the top/side of the
engine. Both generators had plastic tanks mounted above the engine while the pressure washer
had a metal tank mounted above the engine. All of the equipment vented through the fuel caps.
The pressure washer had a metal fuel tank mounted above the engine. The equipment was
operated in the field until the fuel temperature stabilized. For lawnmowers, the fuel temperature
stabilized within 20 to 30 minutes while the larger equipment took up to an hour.
5-26
-------
Feasibility of Evaporative Emission Control
By measuring the increase in fuel temperature during operation, we were able to make a
simple determination of the running loss emissions vented from the fuel tank. Other potential
running loss emissions would be from the carburetor, due to permeation increases due to heating
the fuel, or vibration-induced leaks in the fuel system. However, we believed that the majority
of the running loss emissions would be due to breathing losses associated with heating the fuel.
Table 5.2-1 presents the results of the temperature testing.
We contracted with an independent testing laboratory to test fuel tanks from most of the
above pieces of equipment over the measured fuel temperature profiles.45 For three of the tests
on larger fuel tanks, we found that the measured emissions were inconsistent with theoretical
predictions. An investigation of the test data suggested that the test had been ended too soon to
see the full effect of the heat build. Repeat tests were performed with a longer sample time.46
From this data we get the running loss emissions due to the breathing losses associated with the
heating of the fuel tank. New tanks were purchased for this testing that had not been previously
exposed to fuel so permeation emissions would not be included in the emission measurements.
Table 5.2-1 also presents the test results for the above equipment.
Table 5.2-1: Fuel Temperature Measurements During Operation of
Small SI Equipment and Hydrocarbons Measured Over This Temperature Profile
Equipment Type
Riding mower
front tank near engine
Riding mower
rear tank away from engine
Zero-turn riding mower
2 saddle tanks near engine
Walk-behind mower (plastic)
Walk-behind mower (metal)
Generator set
Pressure washer
Fuel Capacity
[gallons]
1.7
1.1
6.5
3.0
2.5
6.5x2
1.4x2
0.34
0.25
0.22
8.5
7.0
1.8
Min. Temp
°C
19.5
15.7
24.3
26.6
27.0
20.5
21.9
23.3
28.7
28.7
20.6
25.8
19.0
Max. Temp
°C
30.3
28.4
33.2
28.4
35.0
23.9
29.7
33.0
46.7
59.7
25.8
50.0
50.6
HC [g/hr]
1.4
1.0
7.8
0.7
3.2
3.4
2.5
0.3
1.2
8.1
1.8
69.3
20.3
The California Air Resources Board performed running loss tests on several pieces of
Small SI equipment.47 This equipment included four lawnmowers (2 new and 2 old), one string
trimmer, two generators, two ATVs, and two forklifts. To measure running loss emissions, the
equipment were operated on California certification fuel in a SFtED and the exhaust was routed
5-27
-------
Final Regulatory Impact Analysis
outside the SHED. Running loss emissions were determined by measuring the HC concentration
in the SHED. Therefore the measurements included all evaporative emissions during operation
including those from fuel heating, permeation, carburetor losses, and, for the two older
lawnmowers, liquid fuel leaks. Although the ATVs and forklifts are not considered to be small
offroad engines, these data can be used as surrogates for equipment that were not tested. Table
5.2-2 presents this data.
Table 5.2-2: Results from ARE Running Loss Tests
Equipment Type
lawnmower
string trimmer
generator
ATV
forklift
Model Year*
2000
2001
1994
1989
1999
1995
2001
2001
2001
1995
1987
Running Loss [g/hr]
0.8
2.6
27.0
12.1
0.6
19.5
1.8
21.4
1.3
1.8
7.4
the 2000 and 2001 equipment were new at the time of testing
5.2.2 Control Technology
Running loss emissions can be controlled by sealing the fuel cap and routing vapors from
the fuel tank to the engine intake. In doing so, vapor generated heat from the engine will be
burned by the engine. It may be necessary to use a valve or limited flow orifice in the purge line
to prevent fuel from entering the line in the case of the equipment turning over and to limit the
vapor to the engine during operation. Depending on the configuration of the fuel system and
purge line, a one way valve in the fuel cap may be desired to prevent a vacuum in the fuel tank
during engine operation. We anticipate that a system like this would eliminate running loss
venting emissions. However, higher temperatures during operation would increase permeation
somewhat. In addition, the additional length of vapor line would increase permeation.
Considering these effects, we still believe that the system described here would result in more
than a 90 percent reduction in running loss emissions from Small SI equipment.
A secondary benefit of running loss control for Small SI equipment has to do with
diffusion emissions. As discussed above, venting a fuel tank through a hose (rather than through
an open orifice) greatly reduces diffusion. In the system discussed above, all venting losses
would occur through the vapor hose to the engine intake rather than through open vents in the
5-28
-------
Feasibility of Evaporative Emission Control
fuel cap. Therefore, the diffusion effect should be largely eliminated.
Another approach to reducing running loss emissions would be to insulate the fuel tank
or move it further from heat sources such as the engine or hydraulic system. With this approach,
the fuel cap vent would likely still be used, but diffusion could be controlled using a tortuous
vent path in the cap as described above.
For marine fuel tanks we are not considering running loss emissions. For portable fuel
tanks and installed fuel tanks on larger vessels, we would not expect there to be significant
heating of the fuel tanks during engine operation due to the distance from the engine and the
cooling effect of operating the vessel in water. For personal watercraft, the fuel tanks have a
sealed system with pressure relief that should help contain running loss emissions. For other
installed fuel tanks, we would expect the diurnal emission control system to capture about half of
any running losses as well.
5.3 Fuel Tank Permeation
The polymeric material (plastic) of which many gasoline fuel tanks manufactured
generally has a chemical composition much like that of gasoline. As a result, constant exposure
of gasoline to these surfaces allows the material to continually absorb fuel. Permeation is driven
by the difference in the chemical potentials of gasoline or gasoline vapor on either side of the
material. The outer surfaces of these materials are exposed to ambient air, so the gasoline
molecules permeate through these fuel-system components and are emitted directly into the air.
Permeation emissions continue at a nearly constant rate, regardless of how much the vehicle or
equipment is used. Because of these effects, permeation-related emissions can therefore add up
to a large fraction of the total emissions from nonroad equipment.
This section summarizes the data and rationale supporting the permeation emission
standard for Small SI and Marine SI fuel tanks presented in the Executive Summary.
5.3.1 Baseline Fuel Tank Technology and Emissions
Fuel tanks may be constructed in several ways. Portable marine fuel tanks and some
small, higher production-volume, installed marine fuel tanks are generally blow-molded using
high-density polyethylene (HDPE). Larger, installed marine fuel tanks are generally either
rotationally-molded using cross-link polyethylene (XLPE) or are constructed out of welded
aluminum. Some boat builders even construct the fuel tanks out of fiberglass as part of the
vessel construction. Fuel tanks on Small SI equipment may be injection molded, blow molded or
rotationally molded. Blow-molded and injection-molded tanks are primarily made of HDPE, but
nylon is used as well in some applications. Rotationally molded fuel tanks are generally made
out of XLPE.
Blow molding is widely used for the manufacture of Small SI, portable marine, and PWC
fuel tanks. Typically, blow molding is performed by creating a hollow tube, known as a parison,
by pushing high-density polyethylene (HDPE) through an extruder with a screw. The parison is
5-29
-------
Final Regulatory Impact Analysis
then pinched in a mold and inflated with an inert gas. In automotive applications, non-
permeable plastic fuel tanks are produced by blow molding a layer of ethylene vinyl alcohol
(EVOH) or nylon between two layers of polyethylene. This process is called coextrusion and
requires at least five layers: the barrier layer, adhesive layers on either side of the barrier layer,
and HDPE as the outside layers which make up most of the thickness of the fuel tank walls.
However, multi-layer construction requires additional extruder screws which significantly
increases the cost of the blow molding machine.
Injection molding can be used with lower production volumes than blow molding due to
lower tooling costs. In this method, a low viscosity polymer is forced into a thin mold to create
each side of the fuel tank. The two sides are then welded together. In typical fuel tank
construction, the sides are welded together by using a hot plate for localized melting and then
pressing the sides together. The sides may also be connected using vibration or sonic welding.
Rotational molding has two advantages over blow molding, which is widely used for
forming automotive parts. First, the tooling cost is an order of magnitude lower than for blow-
molding. Therefore, for small production volumes such as seen for marine applications,
rotational molding is more cost-effective. Manufacturers of rotationally molded plastic fuel
tanks have commented that they could not produce their tanks with competitive pricing in any
other way. The second advantage of rotational molding is that larger parts can generally be
molded on rotational molding machines than on blow-molding machines. Plastic marine fuel
tanks can exceed 120 gallons.
Installed plastic marine fuel tanks are often produced in many shapes and sizes to fit the
needs of specific boat designs. These fuel tanks tanks are generally rotationally-molded out of
cross-link polyethylene. Cross-link polyethylene, which has a permeation rate comparable to
HDPE, is used in larger marine applications because of its ability to pass the U.S. Coast Guard
flame resistance requirements (33 CFR 183.590). Rotational-molding is also used in some Small
SI applications where there are low production volumes of unique fuel tanks. XLPE is used in
these fuel tanks as well because the fuel tank is often exposed and must be able to withstand
impacts such as flying debris.
5.3.1.1 Baseline permeation test data
5.3.1.1.1 Marine fuel tanks
To determine the baseline permeation emissions from marine fuel tanks, we have
collected permeation data on several plastic fuel tanks. Because gasoline does not permeate
through aluminum, we did not perform permeation testing on aluminum fuel tanks.
We tested ten plastic fuel tanks that were either intended for marine use or are of similar
construction. This permeation testing was performed at 29°C with gasoline. Prior to testing, the
fuel tanks were stored with gasoline in them for about 20 weeks to ensure stable permeation
rates. Table 5.3-1 presents the measured permeation rates for these fuel tanks in grams per
gallon of fuel tank capacity. Where the internal surface area was either easily determined or
5-30
-------
Feasibility of Evaporative Emission Control
supplied by the manufacturer, we also calculated the permeation rate in terms of grams per
square meter of inside surface area. The 31 gallon tank showed much lower permeation than the
other fuel tanks. This was likely due to the thickness of the walls in this tank. Even after
stabilization, permeation is a function of material thickness. According to Pick's Law, if the
wall thickness of a fuel tank were double, the permeation rate would be halved.48
Table 5.3-1: Permeation Rates for Plastic Marine Fuel Tanks Tested by EPA at 29°C
Tank Capacity
[gallons]
3.3
6.0
6.0
6.0
6.6
6.6
6.0
23
31
Permeation
[g/gal/day] [g/m2/day]
0.96
0.61
1.18
0.75
0.83
0.77
0.60
0.64
0.44
12.7
6.8
13.1
8.4
9.1
8.4
8.3
8.1
5.5
Construction
HOPE
HOPE
HOPE
HOPE
HOPE
HOPE
cross-link
cross-link
cross-link
Application
portable marine
portable marine
portable marine
portable marine
portable marine
portable marine
marine test tank
installed marine
installed marine
The Coast Guard tested three rotationally-molded, cross-link polyethylene marine fuel
tanks at 40°C (104°F) for 30 days.49 The results are presented in Table 5.3-2. Because
permeation emissions are a function of surface area and wall thickness, there was some variation
in the permeation rates from the three tanks on a g/gal/day basis. These results are not directly
comparable to the EPA testing because of the difference in test temperature. However, we can
adjust the permeation rates for temperature using Arrhenius' relationship50 combined with
empirical data collected on permeation rates for materials used in fuel tank constructions
(described below). These adjusted permeation rates are shown in Table 5.3-2 and are consistent
with the EPA test data.
Table 5.3-2: Permeation Rates for Cross-Link Marine Fuel Tanks at 40°C
Tank Capacity
[gallons]
12
18
18
Measured Permeation
Loss [g/gal/day]
1.48
1.39
1.12
Average Wall
Thickness [mm]
5.3
5.6
6.9
Adjusted to 29°C
[g/gal/day]
0.71
0.67
0.54
5.3.1.1.2 Small SI equipment fuel tanks
The California Air Resources Board (ARE) investigated permeation rates lawn & garden
equipment fuel tanks. The ARE data is compiled in several data reports on their web site and are
5-31
-------
Final Regulatory Impact Analysis
included in our docket.51'52'53'54'55 Table 5.3-3 presents a summary of this data which was
collected using the ARB Test Method 513.56 Where multiple tests were run on a given tank or
tank type, the average results are presented. Although the temperature in the ARB testing is
cycled from 18 - 41°C rather than held at a constant temperature, the average temperature is
29°C which is similar to the EPA testing. Therefore, the permeation results would likely be
similar if the data were collected at the average temperature of 29°C used in the EPA testing.
Variation in permeation rates on a gram per square meter basis is likely due to differences in the
wall thicknesses. Note that surface area measurements were not available for all of the fuel
tanks. Smaller fuel tanks would be expected to have higher emissions on a gram per gallon basis
due to the increased surface area to volume ratio. However, lower permeation rates were
observed for the fuel tanks less than 1 quart, potentially due to relatively thicker walls or due to a
difference in material used for these applications.
5-32
-------
Feasibility of Evaporative Emission Control
Table 5.3-3: Permeation Rates for Plastic Lawn and
Garden Fuel Tanks Tested by ARE Over a 18-41°C Diurnal
Tank Capacity
[gallons]
0.06
0.08
0.09
0.09
0.10
0.12
0.15
0.16
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
0.29
0.38
0.38
0.38
0.38
0.50
0.50
0.55
0.74
1.4
1.7
1.8
3.9
5.0
5.0
5.0
7.5
Permeation Loss
[g/gal/day]
0.20
0.26
0.12
0.19
0.28
0.53
0.42
0.29
1.32
0.73
0.67
0.74
0.86
0.68
1.06
1.24
0.99
0.67
0.66
0.62
1.39
1.26
1.27
0.27
1.30
0.92
0.08
1.39
1.04
1.24
1.82
1.72
1.14
1.47
3.28
3.20
2.75
3.82
2.07
Permeation Loss
[g/m2/day]
5.39
6.67
—
5.88
—
9.01
7.32
4.79
11.56
10.65
9.75
10.75
12.54
9.91
9.24
10.84
8.68
9.80
9.65
9.07
12.17
11.03
15.00
—
10.66
9.18
—
12.69
8.53
—
—
7.81
—
6.19
4.84
—
—
8.80
2.86
5-3
-------
Final Regulatory Impact Analysis
Some handheld equipment, primarily chainsaws, use structurally-integrated fuel tanks
where the tank is molded as part of the body of the equipment. In these applications the frames
(and tanks) are typically molded out of nylon for strength. We tested structurally-integrated fuel
tanks from four handheld equipment manufacturers at 29°C on both gasoline and s 10 percent
ethanol blend. The test results suggest that these fuel tanks are capable of meeting the standards
using their current materials. In the cases where the permeation rates were higher than the
standards, it was observed that the fuel cap seals had large exposed surface areas on the O-rings,
which were not made of low permeation materials. Emissions could likely be reduced
significantly from these tanks with improved seal designs. Table 5.3-4a presents the results of
this testing. Note that permeation emissions are 20 to 70 percent higher on E10 than on gasoline
for these fuel tanks.
Table 5.3-4a: Permeation Rates for Handheld Fuel Tanks Tested by EPA at 29°C
Tank ID
Rl
R2
R3
Bl
B2
B3
B4
Wl
W2
W3
Gl
G2
G3
Application
clearing saw
(0.24 gallons)
hedge clipper
(0.05 gallons)
chainsaw
(0.06 gallons)
chainsaw
(0.06 gallons)
Material
nylon 6
nylon 6, 33% glass
nylon 6, 30% glass
nylon 6, 30% glass
Test Fuel
gasoline
E10
E10
gasoline
E10
E10
E10
gasoline
E10
E10
gasoline
E10
E10
Permeation Loss
[g/m2/day]
0.34
0.42
0.48
0.62
1.01
1.12
0.93
1.45
2.18
2.46
1.30
1.41
2.14
The handheld industry also tested a number of fuel tanks for their products.57 In this
testing, they investigated the effect of fuel type and gasket material on the permeation results.
These test results suggested that permeation can be reduced significantly by using a low
permeation material, such as FKM, for the seal on the fuel cap. In addition, data on aged tanks
suggested that NBR o-rings may deteriorate in-use such that the permeation rate (or vapor leak
rate) through the seal increases greatly. This test data is presented in Table 5.3-4b.
-------
Feasibility of Evaporative Emission Control
Table 5.3-4b: Permeation Rates for Handheld Fuel Tanks Tested by Industry
Tank
Capacity
350cc
260cc
773cc
400cc
1538cc
1538cc
530cc
400cc
Tank Material
nylon 6, 30% glass
nylon 6, 30% glass
nylon 6, 30% glass
nylon 6, 30% glass
nylon 6, 33% glass
nylon 6, 30% glass
nylon 6, 30% glass
HOPE
O-ring
Material
NBR
NBR
Aged NBR*
New NBR*
NBR
NBR
FKM
NBR
NBR
FKM
NBR
NBR (gasket)
Test
Temp.
28°C
28°C
40°C
40°C
40°C
40°C
40°C
40°C
Test Fuel
CE10
CE10
CE10
CE10
gasoline
CE10
CE10
CE10
gasoline
gasoline
Permeation Loss
[g/m2/day]
1.00
0.64
45.2
0.92
1.31
0.64
3.58
2.60
1.10
0.40
1.11
1.42
0.37
1.55
17.5
* these units were used in the field prior to testing
5.3.1.1.3 Portable fuel tanks
The California Air Resources Board (ARE) investigated permeation rates from portable
fuel containers. Although this testing was not on Small SI or marine fuel tanks, the fuel tanks
tested are of similar construction.58'59 The ARB data is compiled in several data reports on their
web site and is included in our docket. Table 5.3-5 presents a summary of this data which was
collected using the ARB Test Method 513.60 Due to the increasing surface to volume ratio with
decreasing fuel tank sizes, data presented in terms of grams per gallon for smaller tanks would be
expected to be higher for the same grams per surface area permeation rate. Although the
temperature in the ARB testing is cycled from 18 - 41°C rather than held at a constant
temperature, the results would likely be similar if the data were collected at the average
temperature of 29°C which is used in the EPA testing.
5-35
-------
Final Regulatory Impact Analysis
Table 5.3-5: Permeation Rates for HDPE Portable
Fuel Containers Tested by ARE Over a 18-41°C Diurnal
Tank Capacity
[gallons]
1.0
1.0
1.0
1.0
1.0
1.0
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
2.1
2.1
2.1
2.1
2.5
2.5
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.6
Permeation Loss
[g/gal/day]
1.63
1.63
1.51
0.80
0.75
0.75
0.50
0.49
0.51
0.52
0.51
0.51
1.51
1.52
1.88
1.95
1.91
1.78
1.46
1.09
0.89
0.62
0.99
1.39
1.46
1.41
1.47
1.09
5.3.1.2 Effect of temperature on permeation rate
It is well known that the rate of permeation is a function of temperature. For most
materials, permeability increases by about a factor of 2 for every 10°C increase in temperature.61
To determine this relationship for nonroad fuel tanks, we performed permeation testing on nine
HDPE Small SI fuel tanks at both 29°C and 36°C (85°F and 96°F). This sample set included
both baseline and surface treated fuel tanks. On average (excluding the outlier), the temperature
effect was equivalent to nearly a factor of 2 increase in permeation per 10°C increase in
temperature. The one outlier likely resulted from measurement error due to the very low
permeation levels (0.5 grams lost over 2 weeks). Table 5.3-6 presents the test results.
5-36
-------
Feasibility of Evaporative Emission Control
Table 5.3-6: Effect of Temperature on Permeation from HOPE Small SI Fuel Tanks
Tank
A
B
C
D
E
F
H
I
J
Treatment
untreated
sulfonated
fluorinated
29°C [g/m2/day]
11.5
11.4
11.2
2.48
2.73
2.24
0.56
0.62
0.22
36°C [g/m2/day]
17.1
16.6
17.0
4.10
3.98
3.42
0.75
0.68
0.31
Increase per 10°C
92%
86%
97%
127%
85%
100%
60%
17%
80%
Published data collected on HDPE samples at four temperatures62'63 suggest that the
permeation of gasoline through HDPE increases by about 80 percent for every 10°C increase in
temperature. This relationship is presented in Figure 5.3-1, and the numeric data can be found in
Appendix 5D.
Figure 5.3-1: Effect of Temperature on HDPE Permeation
450
0
0
10 20 30 40
degrees Celsius
50
60
70
Another study was performed on the permeation from complete automotive fuel
systems.64 These fuel systems, which included fuel tanks, hoses, and other components, were
tested at both 29°C and 40°C on three fuel types (gasoline, ethanol blend, and MTBE blend).
5-37
-------
Final Regulatory Impact Analysis
The effect of temperature on permeation did not appear to be significantly affected by fuel type.
Table 5.3-7 presents this data for ten automotive fuel systems tested on gasoline. This data
showed more than a factor of 2 increase in permeation per 10°C increase in temperature.
Table 5.3-7: Effect of Temperature on Permeation from Automotive Fuel Systems
Fuel System
2001 Toyota Tacoma
2000 Honda Odyssey
1999 Toy ota Corolla
1997 Chrysler Town & Country
1995 Ford Ranger
1993 Chevrolet Caprice Classic
1991 Honda Accord LX
1989 Ford Taurus GL
1985 Nissan Sentra
1978 Olds Cutlass Supreme
Fuel Tank
Metal
Plastic (enhanced
evap)
Metal
Plastic (enhanced
evap)
HOPE
Fluorinated HOPE
Metal
Metal
Metal
Metal
29°C
[mg/hr]
9
21
10
23
309
95
40
24
53
57
40°C
[mg/hr]
20
55
24
52
677
255
110
52
148
122
Increase
per 10°C
101%
136%
124%
110%
102%
143%
148%
100%
152%
99%
5.3.1.3 Units for reporting the permeation rate (g/gal/day vs. g/m2/day)
Much of the permeation data presented in this chapter is in units of grams of
hydrocarbons lost in a day divided by the capacity of the fuel tank (g/gal/day). For diurnal
emissions, these units are used because the vapor generation is a function of fuel tank volume.
For permeation emissions, we considered using these units because the capacity of the fuel tank
is generally readily available; either identified on the fuel tank or readily measured. However,
although volume is generally used to characterize fuel tank emission rates, permeation is actually
a function of surface area. Because the surface to volume ratio of a fuel tank changes with
capacity and geometry of the tank, two similar shaped tanks of different volumes or two different
shaped tanks of the same volume could have different g/gal/day permeation rates even if they
were made of the same material and used the same emission control technology. For this reason,
the final standards are based on units of grams per square meter of inside surface area
(g/m2/day).
This chapter presents permeation data for a large number of Small SI, marine, and other
fuel tanks. For many of these fuel tanks, we had information on both the volume and inside
surface area. Figure 5.3-2 presents the relationship between fuel tank volume in gallons and
inside surface area in square meters. As a fuel tank becomes smaller, its surface to volume ratio
increases. This relationship can be seen better in the chart to the right which presents only data
for fuel tanks less than 1 gallon. A hyperbolic curve is fit through the data in Figure 5.3-2 to
represent this relationship. This is seen better in the right-side chart which presents only smaller
tank sizes. In addition to fuel tank volume, the surface to volume ratio is affected by geometry
5-38
-------
Feasibility of Evaporative Emission Control
of the fuel tank. A long flat-fuel tank would have a higher surface to volume than a cube or
spherical design. Larger plastic fuel tanks, used primarily in marine vessels, tend to have
somewhat high surface to volume ratios for this reason.
Figure 5.3-2: Relationship Between Tank Volume and Inside Surface Area
0.0
0 10 20 30
volume [gallons]
40
0.20
0.00
0.0 0.5 1.0
volume [gallons]
5.3.1.4 Effect of fuel tank fill level on permeation
Permeation is driven by the chemical potential of the fuel or vapor in contact with the
plastic. In a fuel tank, the vapor is essentially at equilibrium with the fuel in a fuel tank.
Therefore, the permeation rate is the same through the surfaces in contact with saturated vapor as
it is through the surfaces in contact with the liquid fuel. Because the permeation rate of saturated
vapor and liquid fuel are the same, the fill level of the fuel tank during a permeation test does not
affect the measured results.
The fact that liquid fuel and saturated fuel vapor result
in the same permeation rates is supported by published
literature.65'66'67'68 In two of these studies, permeation was
measured for material samples using the cup method
illustrated in Figure 5.3-3. In these tests, no significant
difference was seen between the permeation rates for material
samples exposed to liquid fuel or to fuel vapor. To test for
permeation with fuel vapor, the cup was inverted so that the
fuel was on the bottom and the sample was taken off the top.
Table 5.3-8 presents the data from these two reports. In both
cases, the material being tested was a fluoroelastomer.
Figure 5.3-3: Cup Method
5-39
-------
Final Regulatory Impact Analysis
Table 5.3-8: Permeation Measured in Cup Method with Fuel Versus Vapor Fuel Exposure
Paper
SAE 200 1-0 1-1999
SAE 2000-0 1-1096
Fuel
CE10
CE10
CM15
Temperature
40°C
23°C
40°C
23 °C
40°C
Liquid Fuel Exposure
30.5 g/m2/day
0.3 g/test
2.6 g/test
3.1 g/test
9.5 g/test
Fuel Vapor Exposure
29.5 g/m2/day
0.3 g/test
2.5 g/test
2.9 g/test
8.5 g/test
One commenter presented test data suggesting that fill level may affect permeation
emissions.69 They tested four HDPE jugs, two filled to 40 percent and two filled to 100 percent
with gasoline and saw a 15 percent difference in the average permeation results for the two fill
levels (1.3 g/gal/day for 40 percent fill and 1.5 g/gal/day for 100 percent fill). Although this
small measured difference was likely due to test variability, we performed our own testing to
study the effect of fill level. For this testing, we used two 6-gallon FfDPE portable marine fuel
tanks. The fuel tanks were soaked with gasoline for 12 weeks to ensure a stabilized permeation
rate. Each tank was tested at both 50 percent and 90 percent fill. No significant difference in
permeation rate was observed for either tank. Table 5.3-9 presents the results in terms of
g/gal/day at 29°C.
Table 5.3-9: Effect of Fuel Tank Fill Level on Permeation
for Two Portable Marine Fuel Tanks [g/gal/day]
Tank 1
Tank 2
50% fill
1.16
0.77
90% fill
1.21
0.78
Another study showed mixed results. Four automotive fuel systems (including fuel tank,
hose, and other components) were tested for permeation with the fuel tanks filled with Fuel C to
both 20 percent and 100 percent of capacity.70 Prior to the testing, the fuel tanks were soaked
with fuel at the specified fill levels until a stable permeation rate was achieved. It was not clear
what fraction of the permeation came from the fuel tanks compared to other fuel system
components or how the fuel level affected the exposure of the other components. In this study,
two of the fuel systems saw no significant change in permeation as a result of a change in fill
level. These two fuel system were on older vehicles, one with an untreated and one with a
fluorinated HDPE fuel tank. Two other fuel systems, using fuel tanks that meet automotive
enhanced evaporative emission requirements, showed significant reductions in fuel system
permeation (32 percent and 49 percent) when tested with the fuel tank filled to only 20 percent
capacity. The study presented no rationale for this effect; however, it should be noted that these
were very low permeation systems and measurement error would presumably be larger. These
data are presented in Table 5.3-10. In addition, it is possible that the change in fill level affected
whether or not there was fuel in the hoses. As discussed later in this chapter, the vapor
5-40
-------
Feasibility of Evaporative Emission Control
concentration in fuel hoses may be significantly lower than saturated when exposed only to
vapor due to diffusion constraints.
Table 5.3-10: Effect of Fuel Tank Fill Level on Permeation
for Four Automotive Fuel Systems at 29°C [g/hour]
Rig 2
Rig 4
Rig 5
Rig 6
Description of Fuel Tank
enhanced evap system
enhanced evap system
HDPE fuel tank
fluorinated HDPE fuel tank
20% fill
0.013
0.021
0.350
0.095
80% fill
0.019
0.041
0.349
0.094
The California Air Resources Board also performed testing on three pairs of portable fuel
tanks.71 All of the fuel tanks were identical 1 gallon tanks made out of HDPE. Each pair was
filled to a different level with California certification fuel (30 percent, 50 percent and 70 percent
fill). The fuel tanks were then sealed and subjected to five days of the California diurnal test
(65-105T) and weight loss was measured daily. Over the five days of testing, the tanks with
lower fill levels actually saw significantly higher permeation than the other tanks. Looking at
the last day of testing, which represents some conditioning of the fuel tanks by the fuel resulting
in more stabilized permeation rates, the permeation rates are similar regardless of the fill level.
This data, which is presented in Table 5.3-11, suggests that the fuel vapor in the tanks permeated
at the same rate as (or higher than) the liquid fuel.
Table 5.3-11: Effect of Fuel Tank Fill Level on Permeation
for Three Pairs of Portable Fuel Tanks [g/day]
Tank
30a
3 Ob
50a
5 Ob
70a
70b
Fill Level
30%
50%
70%
5 -Day Permeation
1.79
1.57
1.53
1.03
1.26
1.08
Last Day Permeation
1.87
1.91
1.91
1.43
1.85
1.43
5.3.1.5 Effect of background concentration on permeation
As discussed above, permeation is driven by the difference in chemical potential between
the inside and outside of the tank. If the concentration of vapor outside the fuel tank were large
enough, it could reduce the permeation rate of fuel through the tank. One commenter presented
test data suggesting that, at very low concentrations of vapor in the boat around the fuel tank,
that the permeation rate would be significantly reduced.72 This test data was based on two three
hour tests on 5 gallon HDPE bottles at 35°C. They measured 0.57 g/hr with a background
5-41
-------
Final Regulatory Impact Analysis
concentration of 26 ppm and 0.36 g/hr with a background of 212 ppm. No repeat tests were run.
It is not clear why the above results were measured. Compared to the concentration of the fuel
vapor in the tank, this difference between 212 and 26 ppm is minuscule (about three orders of
magnitude difference from saturated vapor). It is more likely that this effect was due to test
variation.
To investigate this potential effect on permeation emissions further, we performed our
own testing. First, we measured the concentration of fuel vapor around the fuel tank on a
summer day in a runabout with the tank installed in the hull. This concentration was 1400 ppm.
We then tested two different fuel tanks for permeation with different background concentrations.
The background concentration was maintained by controlling the bleed of fresh air through the
test container or SHED. Each test ran for about two weeks and the permeation rates were
determined using the weight loss method. Prior to the testing, the tanks were soaked until a
stable permeation rate was achieved, then new fuel was added to the tank just prior to beginning
the test. The fuel tank was soaked until the fuel temperature stabilized at 29°C before the
beginning weight was measured. The results, which are presented in Table 5.3-12, showed no
significant difference in permeation as a function of background concentrations of hydrocarbon
vapor.
Table 5.3-12: Effect of Background Concentration on Permeation
Fuel Tank
6 gallon HOPE
23 gallon cross-link PE
Background [ppmC]
30
1500
30
150
1350
Permeation [g/gal/day]
0.77
0.78
0.64
0.67
0.66
5.3.2 Fuel Tank Permeation Reduction Technologies
There are several strategies that can be used to reduce permeation from plastic fuel tanks.
This section presents data collected on five permeation control strategies: sulfonation,
fluorination, non-continuous barrier platelets, coextruded continuous barrier, and alternative
materials.
5.3.2.1 Sulfonation
Sulfonation is a process where the surface of the fuel tank is treated to minimize
permeation. The sulfonation process uses sulfur trioxide is used to create the barrier by reacting
with the exposed polyethylene to form sulfonic acid groups on the surface. Current practices for
sulfonation are to place fuel tanks on a small assembly line and expose the inner surfaces to
sulfur trioxide, then rinse with a neutralizing agent. However, sulfonation can also be performed
off-line. Either of these processes can be used to reduce gasoline permeation by more than 90
5-42
-------
Feasibility of Evaporative Emission Control
percent from new tanks.73
We tested several sulfonated marine fuel tanks at 29°C for permeation. This testing
included both HDPE blow-molded fuel tanks and cross-link polyethylene rotationally-molded
tanks. Both gasoline and alcohol fuel blends were investigated. In some cases, the fuel tanks
were exposed to durability testing as described in Section 5.6.2. The fuel tanks were stored with
fuel in them (soaked) for preconditioning, then they were drained and then filled with fresh fuel
prior to each permeation test. The purpose of the soak periods was to ensure that the fuel
permeation rate had stabilized and the purpose of the pressure cycles and slosh testing was to
evaluate the durability of the barrier treatment.
We also collected data from ARB and other sources on the effectiveness of sulfonation
for reducing permeation emissions from plastic fuel tanks. Most of this research has been
performed on blow-molded HDPE fuel tanks. As shown in these data, it is important that the
resin formulation be matched to the sulfonation process. The following discussions look at
sufonation results on HDPE and on cross-link polyethylene separately.
HDPE fuel tanks
We tested several HDPE fuel tanks that were sulfonated on the internal surfaces. These
included three 6-gallon and one 3.3 gallon portable marine fuel tanks and three all-terrain vehicle
(ATV) fuel tanks. These fuel tanks were sent to a sulfonater for barrier treatment. Multiple fuel
tanks were used so that they could be tested on certification gasoline , E10 (10 percent ethanol),
and M15 (15 percent methanol). The test results, presented in Table 5.3-13, showed more than a
90 percent reduction in permeation emissions from baseline. However, the two fuel tanks that
were subjected to slosh testing saw emission levels above the standard. This may have been a
material compatibility issue as discussed below. The test results are consistent with similar data
collected by the California Air Resources Board.
5-43
-------
Final Regulatory Impact Analysis
Table 5.3-13: EPA Permeation Data on Sulfonated HDPE Fuel Tanks at 29°C
Treatment
Fuel
Soak Period
g/gal/day
g/m2/day
6 gallon portable marine fuel tanks
baseline
sulfonated
sulfonated
sulfonated
sulfonated
gasoline
gasoline
gasoline, sloshed
E10
M15
15 weeks
16 weeks
12 weeks
24 weeks
24 weeks
0.77
0.04
0.39
0.14
0.08
8.53
0.45
4.30
1.58
0.84
4 gallon ATV fuel tanks
sulfonated
sulfonated
sulfonated
gasoline
E10
M15
20 weeks
24 weeks
24 weeks
0.13
0.06
0.08
1.05
0.45
0.64
3.3 gallon portable fuel tank
baseline
sulfonated
sulfonated
E10
E10
E10, sloshed
14 weeks
14 weeks
38 weeks
0.96
0.06
0.16
12.7
0.83
2.09
We performed slosh testing on the 6 and 3.3 gallon portable marine fuel tanks with E10
fuel. This slosh testing included 1 million cycles consistent with the durability test procedure.
After the slosh testing, the permeation rates were measured to be 2.0 and 4.3 g/m2/day for the 3.3
and 6 gallon fuel tanks, respectively. As discussed below, we believe that the impact of the
durability testing on the effectiveness of sulfonation can be minimized if the sulfonation process
and material properties are matched properly. However, this data supports the need for the
durability testing requirements.
The California Air Resources Board (ARE) collected test data on permeation rates from
sulfonated portable fuel containers using California certification fuel.74 The results show that
sulfonation can be used to achieve significant reductions in permeation from plastic fuel
containers. This data was collected using a diurnal cycle from 18-41°C which is roughly
equivalent to steady-state permeation testing at 29°C. The average emission rate for the 32
sulfonated fuel tanks is 0.35 g/gal/day; however, there was a wide range in variation in the
effectiveness of the sulfonation process for these fuel tanks. Some of the data outliers were
actually higher than baseline emissions. This was likely due to leaks in the fuel tank which
would result in large emission increases due to pressure built up with temperature variation over
the diurnal cycle. Removing these five outliers, the average permeation rate is 0.17 g/gal/day
with a minimum of 0.01 g/gal/day and a maximum of 0.64 g/gal/day. This data suggests that
more than a 90 percent reduction in permeation from HDPE fuel tanks is possible through
sulfonation. This data is presented in Table 5.3-14.
5-44
-------
Feasibility of Evaporative Emission Control
Plastic
Table 5.3-14: Permeation Rates for Sulfonated
Fuel Containers Tested by ARE Over a 18-41°C Diurnal
Tank Capacity
[gallons]
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5
5
5
5
5
5
5
Permeation Loss
[g/gal/day]
0.05
0.05
0.05
0.06
0.06
0.06
0.08
0.12
0.14
1.23
1.47
1.87
0.02
0.02
0.48
0.54
1.21
0.03
0.08
0.32
0.38
0.42
0.52
0.64
0.80
0.01
0.04
0.05
0.06
0.11
0.13
0.15
Variation can occur in the effectiveness of this surface treatment if the sulfonation
process is not properly matched to the plastic and additives used in the fuel tank material. For
instance, if the sulfonater does not know what UV inhibitors or plasticizers are used, they cannot
maximize the effectiveness of their process. Earlier data collected by ARB showed consistently
high emissions from sulfonated fuel tanks; however, ARB and the treatment manufacturers agree
that this was due to inexperience with treating fuel tanks and that these issues have since been
5-45
-------
Final Regulatory Impact Analysis
largely resolved.75
ARB also investigated the effect of fuel slosh on the durability of sulfonated surfaces.
Three half-gallon fuel tanks used on Small SI equipment were sulfonated and tested for
permeation before and after being sloshed with fuel in them 1.2 million times.76'77 These fuel
tanks were blow-molded HDPE tanks used in a number of Small SI applications including
pressure washers, generators, snowblowers, and tillers. The results of this testing show that an
85 percent reduction in permeation was achieved on average even after the slosh testing was
performed. Table 5.3-15 presents these results which were recorded in units of g/m2/day. The
baseline level for Set #1 is an approximation based on testing of similar fuel tanks, while the
baseline level for Set #2 is based on testing of those tanks.
The sulfonater was not aware of the materials used in the fuel tanks sulfonated for the
slosh testing. After the tests were performed, the sulfonater was able to get some information on
the chemical make up of the fuel tanks and how it might affect the sulfonation process. For
example, the UV inhibitor used in some of the fuel tanks is known as HALS. HALS also has the
effect of reducing the effectiveness of the sulfonation process. Two other UV inhibitors, known
as carbon black and adsorber UV, are also used in similar fuel tank applications. These UV
inhibitors cost about the same as HALS, but have the benefit of not interfering with the
sulfonation process. The sulfonater claimed that if HALS were not used in the fuel tanks, a 97
percent reduction in permeation would have been seen.78 To confirm this, one manufacturer
tested a sulfonated tank similar to those in Set #2 except that carbon black, rather than HALS,
was used as the UV inhibitor. This fuel tank showed a permeation rate of 0.88 g/m2/day at
40°C79 which was less than half of what the CARB testing showed on their constant temperature
test at 40°C.80 A list of resins and additives that are compatible with the sulfonation process is
included in the docket.81'82
5-46
-------
Feasibility of Evaporative Emission Control
Table 5.3-15: Permeation Rates for Sulfonated Fuel Tanks
with Slosh Testing by ARE Over a 18-41°C Diurnal
Technology
Configuration
Set #1 Approximate
Baseline
Set #1 Sulfonated
Set#l Sulfonated &
Sloshed
Set #2 Average Baseline
Set #2 Sulfonated
Set #2 Sulfonated &
Sloshed
Units
g/m2/day
g/m2/day
%
reduction
g/m2/day
%
reduction
g/m2/day
g/m2/day
%
reduction
g/m2/day
%
reduction
Tankl
10.4
0.73
93%
1.04
90%
12.1
1.57
87%
2.09
83%
Tank 2
10.4
0.82
92%
1.17
89%
12.1
1.67
86%
2.16
82%
Tank3
10.4
1.78
83%
2.49
76%
12.1
1.29
89%
1.70
86%
Average
10.4
1.11
89%
1.57
85%
12.1
1.51
88%
1.98
84%
About a year and a half after the California ARE tests on the Set #2 fuel tanks, we
performed permeation tests on these fuel tanks. During the intervening period, the fuel tanks
remained sealed with California certification fuel in them. We drained the fuel tanks and filled
them with fresh California certification fuel. We then measured the permeation rate at 29°C.
Because this is roughly the average temperature of the California variable temperature test,
similar permeation rates would be expected. The untreated fuel tanks showed slightly lower
permeation over the constant temperature test. This difference was likely due to the difference in
the temperature used for the testing. However, the Sulfonated fuel tanks showed an increase in
permeation. This increase in permeation appears to be the result of the 1.5 year additional fuel
soak. After this long soak, the average permeation reduction changed from 84 to 78 percent.
Table 5.3-13 presents this comparison.
5-47
-------
Final Regulatory Impact Analysis
Table 5.3-16: Permeation Rates [g/m2/day] for Sulfonated Fuel Tanks Tested by
ARE and EPA on CA Certification Gasoline with a 11A Year Fuel Soak Differential
Technology
Configuration
Baseline, CARB testing
Baseline, EPA testing
after 1.5 year additional
fuel soak
Sulfonated, CARB
testing
Sulfonated, EPA testing
after 1.5 year additional
fuel soak
Temperatu
re
18-4FC
29°C
% change
18-4FC
29°C
%
reduction
Tankl
12.1
11.5
-5%
2.09
2.48
78%
Tank 2
12.1
11.4
-6%
2.16
2.73
76%
Tank3
12.1
11.2
-7%
1.70
2.24
80%
Average
12.1
11.4
-6%
1.98
2.5
78%
After the above testing, we drained the fuel tanks and filled them with certification
gasoline splash-blended with 10 percent ethanol (E10). We then soaked the fuel tanks for 20
weeks to precondition them on this fuel. Following the preconditioning, we tested these fuel
tanks for permeation at 29°C (85°F). Table 5.3-17 presents these emission results compared to
the emission results for three baseline tanks (untreated) that were subject to the same
preconditioning. Percent reductions are presented based on the difference between the
sulfonated fuel tanks and the average results of the three untreated fuel tanks.
Table 5.3-17: Permeation Rates for Sulfonated Fuel Tanks on E10 Fuel at 29°C
Technology
Configuration
Baseline (untreated)
Sulfonated
Units
g/m2/day
g/m2/day
%
reduction
Tankl
13.9
3.91
72%
Tank 2
13.7
4.22
70%
Tank3
14.4
2.92
79%
Average
14.0
3.69
74%
An in-use durability testing program was also completed for sulfonated HDPE fuel tanks
and bottles.83 The fuel tank had a 25 gallon capacity and was removed from a station wagon that
had been in use in southern California for five years (35,000 miles). The fuel tank was made of
FIDPE with carbon black used as an additive. After five years, the sulfonation level measured on
the surface of the plastic fuel tank did not change. Tests before and after the aging both showed
a 92 percent reduction in gasoline permeation due to the sulfonation barrier compared to the
permeation rate of a new untreated tank. Testing was also done on 1 gallon bottles made of
FIDPE with 3 percent carbon black. These bottles were shown to retain over a 99 percent barrier
5-48
-------
Feasibility of Evaporative Emission Control
after five years. This study also looked at other properties such as yield strength and mechanical
fatigue and saw no significant deterioration.
One study looked at the effect of alcohol in the fuel on permeation rates from sulfonated
fuel tanks.84 In this study, the fuel tanks were tested with both gasoline and various methanol
blends. No significant increase in permeation due to methanol in the fuel was observed.
XLPE fuel tanks
We tested eight sulfonated cross-link polyethylene (XLPE) fuel tanks for permeation
emissions. These tanks were produced by marine fuel tank manufacturers specifically for this
testing. The fuel tanks were then treated by a sulfonater. For the first four tanks tested, the fuel
tanks were molded using the resin formulation and processes currently used by the fuel tank
manufacturers. When the sulfonation was applied, we observed that the barrier was soft and
could be scraped off easily. When tested, the barrier on these fuel tanks was not as effective as
had been seen on HDPE fuel tanks.
Because the barrier could be scratched off, the sulfonater ascertained that the sulfonation
had poor surface penetration and the darkness of the barrier suggested heavy oxidation. For the
next batch of four test tanks, the sulfonater worked with the material supplier and roto-molder
and attempted to develop a formulation that may be more compatible with sulfonation. They
decided to use the same material, but bake it in the oven longer to remove more oxygen from the
surface of the fuel tank. Four bake times were used to produce the four 6-gallon test tanks: 11,
12, 14, and 16 minutes. It was observed that the sulfonation barrier could not easily be scratched
off these fuel tanks. We tested the four sulfonated on E10 (10 percent ethanol) using the same
procedures as for the HDPE tanks discussed above. The test results did not show a significant
improvement.
Another approach may be to mold an inner liner of HDPE inside a XLPE shell. These
materials readily bond with each other and sulfonation has been demonstrated for HDPE. This
construction, which is currently used in chemical storage applications, is performed in the oven
through the use of a "drop box" in the mold containing the HDPE. This drop-box is opened part
way through the oven cycle allowing for a HDPE layer to be molded on the inside of the fuel
tank.
5.3.2.2 Fluorination
Another barrier treatment process is known as fluorination. The fluorination process
causes a chemical reaction where exposed hydrogen atoms are replaced by larger fluorine atoms
which form a barrier on the surface of the fuel tank. In this process, fuel tanks are generally
processed post production by stacking them in a steel container. The container is then voided of
air and flooded with fluorine gas. By pulling a vacuum in the container, the fluorine gas is
forced into every crevice in the fuel tanks. As a result of this process, both the inside and outside
surfaces of the fuel tank would be treated. As an alternative, fuel tanks can be fluorinated on-
line by exposing the inside surface of the fuel tank to fluorine during the blow molding process.
5-49
-------
Final Regulatory Impact Analysis
However, this method may not prove as effective as off-line fluorination which treats the inside
and outside surfaces.
We tested several fluorinated marine fuel tanks at 29°C for permeation. This testing
included both HDPE blow-molded fuel tanks and cross-link polyethylene rotationally-molded
tanks. Both gasoline and alcohol fuel blends were investigated. In some cases, the fuel tanks
were exposed to durability testing as described in Section 5.6.2. The fuel tanks were stored with
fuel in them (soaked) for preconditioning, then they were drained and then filled with fresh fuel
prior to each permeation test. The purpose of the soak periods was to ensure that the fuel
permeation rate had stabilized and the purpose of the pressure cycles and slosh testing was to
evaluate the durability of the barrier treatment.
We also collected data from ARB and other sources on the effectiveness of fluorination
for reducing permeation emissions from plastic fuel tanks. Most of this research has been
performed on blow-molded HDPE fuel tanks. However, we believe that fluorination can also be
applied effectively for injection-molded HDPE tanks as well. The following discussion looks at
each material separately as well as rotationally-molded cross-link polyethylene.
Blow-molded HDPE fuel tanks
We tested one fluorinated HDPE fuel tank which we bought off the shelf and sent to a
fluorinater for barrier treatment. The fuel tank type used was a 6-gallon portable marine fuel
tank. The fuel tank was soaked for 20 weeks with certification gasoline prior to testing. We
measured a permeation rate of 0.05 g/gal/day (0.56 g/m2/day) which represents more than a 95
percent reduction from baseline. We then began soaking this fuel tank on E10, subjected it to
the pressure and slosh testing, and retested the fuel tank. The post durability testing result
showed a permeation rate of 0.6 g/gal/day (6.8 g/m2/day). As discussed below, we believe that
the impact of the durability testing on the effectiveness of fluorination on can be minimized if
the fluorination process and material properties are matched properly. In addition, this fuel tank
was treated to a significantly lower level of fluorination than is now available. However, this
data supports the need for the durability testing requirements.
The California Air Resources Board (ARB) collected test data on permeation rates from
fluorinated fuel containers using California certification fuel.85'86 The results show that
fluorination can be used to achieve significant reductions in permeation from plastic fuel
containers. This data was collected using a diurnal cycle from 18-41°C which is roughly
equivalent to steady-state permeation testing at 30°C. For the highest level of fluorination, the
average permeation rate was 0.04 g/gal/day which represents a 95 percent reduction from
baseline. Earlier data collected by ARB showed consistently high emissions from fluorinated
fuel tanks; however, ARB and the treatment manufacturers agree that this was due to
inexperience with treating fuel tanks and that these issues have since been largely resolved.87
The ARB data is presented in Table 5.3-18.
5-50
-------
Feasibility of Evaporative Emission Control
Table 5.3-18: Permeation Rates for Fluorinated
Plastic Fuel Containers Tested by ARE Over a 18-41°C Diurnal
Barrier Treatment*
Level 4
(average =0.09 g/gal/day)
Level 5
(average =0.07 g/gal/day)
SPAL
(average =0.04 g/gal/day)
Tank Capacity
[gallons]
1
1
1
5
5
5
1
1
1
1
1
1
1
1
1
2.5
2.5
2.5
2.5
2.5
5
5
5
5
5
5
Permeation Loss
[g/gal/day]
0.05
0.05
0.06
0.11
0.11
0.15
0.03
0.04
0.05
0.05
0.07
0.08
0.11
0.11
0.12
0.04
0.04
0.05
0.07
0.07
0.05
0.10
0.11
0.04
0.04
0.04
* designations used in ARB report; shown in order of increasing treatment
All of the data on fluorinated fuel tanks presented above were based on fuel tanks
fluorinated by the same company. Available data from another company that fluorinates fuel
tanks shows a 98 percent reduction in gasoline permeation through a HDPE fuel tank due to
fluorination.88
ARB investigated the effect of fuel slosh on the durability of fluorinated surfaces. Two
sets of three fluorinated fuel tanks were tested for permeation before and after being sloshed with
fuel in them 1.2 million times.89'90 These fuel tanks were 0.5 gallon, blow-molded HDPE tanks
used in a number of Small SI applications including pressure washers, generators, snowblowers,
and tillers. The results of this testing show that an 80 percent reduction in permeation was
achieved on average even after the slosh testing was performed for Set #1. However, this data
5-51
-------
Final Regulatory Impact Analysis
also showed a 99 percent reduction for Set #2. This shows the value of matching the barrier
treatment process to the fuel tank material. Table 5.3-19a presents these results which were
recorded in units of g/m2/day. The baseline level for Set #1 is an approximation based on testing
of similar fuel tanks, while the baseline for Set #2 is based on testing of those tanks.
Table 5.3-19a: Permeation Rates for Fluorinated Fuel Tanks
with Slosh Testing by ARE Over a 18-41°C Diurnal
Technology
Configuration
Set #1 Approximate
Baseline
Set #1 Fluorinated
Set #1 Fluorinated &
Sloshed
Set #2 Approximate
Baseline
Set #2 Fluorinated
Set #2 Fluorinated &
Sloshed
Units
g/m2/day
g/m2/day
%
reduction
g/m2/day
%
reduction
g/m2/day
g/m2/day
%
reduction
g/m2/day
%
reduction
Tankl
10.4
1.17
89%
2.38
77%
12.1
0.03
>99%
0.07
99%
Tank 2
10.4
1.58
85%
2.86
73%
12.1
0.00
>99%
0.11
99%
Tank3
10.4
0.47
96%
1.13
89%
12.1
0.00
>99%
0.05
>99%
Average
10.4
1.07
90%
2.12
80%
12.1
0.01
>99%
0.08
99%
About a year and a half after the California ARB tests on the Set #2 fuel tanks, we
performed permeation tests on these fuel tanks. During the intervening period, the fuel tanks
remained sealed with California certification fuel in them. We drained the fuel tanks and filled
them with fresh California certification fuel. We then measured the permeation rate at 29°C.
Because this is roughly the average temperature of the California variable temperature test,
similar permeation rates would be expected. The untreated fuel tanks showed slightly lower
permeation over the constant temperature test. This difference was likely due to the difference in
the temperature used for the testing. However, the fluorinated fuel tanks showed an increase in
permeation. This increase in permeation appears to be the result of the 1.5 year additional fuel
soak. Even after this long fuel soak, the fluorination achieves more than a 95 percent reduction
in permeation. Table 5.3-19b presents this comparison.
5-52
-------
Feasibility of Evaporative Emission Control
Table 5.3-19b: Permeation Rates [g/m2/day] for Fluorinated Fuel Tanks Tested by
ARE and EPA on CA Certification Gasoline with a 11A Year Fuel Soak Differential
Technology
Configuration
Baseline, CARB testing
Baseline, EPA testing
after 1.5 year additional
fuel soak
Fluorinated, CARB
testing
Fluorinated, EPA testing
after 1.5 year additional
fuel soak
Temperature
18-4FC
29°C
% change
18-4FC
29°C
% reduction
Tankl
12.1
11.5
-5%
0.07
0.56
95%
Tank 2
12.1
11.4
-6%
0.11
0.62
95%
Tank3
12.1
11.2
-7%
0.05
0.22
98%
Average
12.1
11.4
-6%
0.08
0.47
96%
After the above testing, we drained the fuel tanks and filled them with certification
gasoline splash-blended with 10 percent ethanol (E10). We then soaked the fuel tanks for 20
weeks to precondition them on this fuel. Following the preconditioning, we tested these fuel
tanks for permeation at 29°C (85°F). Table 5.3-21 presents these emission results compared to
the emission results for three baseline tanks (untreated) that were subject to the same
preconditioning. Percent reductions are presented based on the difference between the
fluorinated fuel tanks and the average results of the three untreated fuel tanks. The slight
increase in permeation on the E10 fuel was similar for the baseline and fluorinated fuel tanks and
still resulted in permeation rates well below the standard.
Table 5.3-20: Permeation Rates for Fluorinated Fuel Tanks on E10 Fuel at 29°C
Technology
Configuration
Baseline (untreated)
Fluorinated
Units
g/m2/day
g/m2/day
% reduction
Tankl
13.9
0.43
97%
Tank 2
13.7
0.62
96%
Tank3
14.4
0.62
96%
Average
14.0
0.56
96%
The handheld industry also tested a number of fluorinated fuel tanks for their products.91
This testing included three fuel types and two test temperatures. These test results suggest that
fluorination may be used to significantly reduce permeation from handheld fuel tanks. Higher
permeation rates were observed in CE10 than gasoline; however, it is not clear whether these
impacts were due to increased permeation through the gaskets or through the tank. As shown
earlier in Table 5.3-4b, the use of low permeation gasket materials can significantly improve
permeation rates, especially on fuel CE10. This test data is presented in Table 5.3-21.
5-53
-------
Final Regulatory Impact Analysis
Table 5.3-21: Permeation Rates for Handheld Fuel Tanks Tested by Industry
Tank Capacity
695
420
400*
252
Gasket Material
NBR
NBR
NBR
FKM
NBR
Test Temp.
28°C
28°C
40°C
40°C
Test Fuel
CE10
CE10
gasoline
CE10
gasoline
CE10
E10
Permeation Loss
[g/m2/day]
2.01
2.88
1.46
1.32
0.56
0.64
2.94
0.84
1.95
1.40
* A similar, untreated, tank was measured to have a permeation rate of 17.5 g/m2/day.
Another study also looked at the effect of alcohol in the fuel on permeation rates from
fluorinated fuel tanks.92 In this study, the fuel tanks were tested with both gasoline and various
methanol blends. No significant increase in permeation due to methanol in the fuel was
observed.
Under their rule for small offroad equipment, California may issue executive orders to
manufacturers with low emission products. As of August, 2006, ARB has issued 5 executive
orders for low permeation fuel tanks.93 Under these executive orders, three fluorination
approaches have been approved. The California fuel tank permeation standard is 1.5 g/m2/day
tested at 40°C on California certification fuel. Table 5.3-22 presents the test results for the fuel
tanks with ARB executive orders. Note that the reported emissions are the average of five test
samples.
Table 5.3-22: ARB Fuel Tank Executive Orders for Small Offroad Equipment
EO#
C-U-05-015
C-U-06-019
C-U-06-006
Test Fuel
Phase II
Phase II
Phase II
g/m2/day
1.10
0.30
0.38
One automobile manufacturer used fluorination to reduce permeation on HDPE fuel
tanks to meet the LEV I vehicle standards. This manufacturer used similar or more stringent
requirements for fuel soak, durability, and testing than finalized today. At 40°C, this
5-54
-------
Feasibility of Evaporative Emission Control
manufacturer stated that they measured 0.15-0.2 g/day for fluorinated tanks compared to over 10
g/day for untreated HDPE fuel tanks.94
Injection-molded HOPE fuel tanks
The issue has been raised by manufacturers that HDPE intended for injection-molding
has a somewhat different composition than HDPE used for blow-molding. To address this
concern, testing has been performed on fluorinated, injection-molded fuel tanks as well.95 These
fuel tanks were tested using California's TP-901 test procedures which preconditioning steps
including fuel soak, slosh testing, and pressure-vacuum cycling. California Phase II gasoline
was used for this testing.
Three similar fuel tanks were tested also over the Federal test procedure.96 Under this
testing, E10 fuel was used. Weight loss tests were performed before and after the durability tests
in 40 CFR 1501.515.97 These durability tests included slosh testing, pressure vacuum cycling,
and UV exposure. Results from this testing are presented in Table 5.3-23. The permeation was
significantly higher when tested on E10 fuel, especially when accounting for differences in test
temperature. In addition, permeation increased somewhat after the durability testing. However,
the measured permeation rates were well below the fuel tank permeation standard on E10 after
the durability testing.
Table 5.3-23: Permeation Rates for Fluorinated, Injection-Molded Fuel Tanks [g/mVday]
Test Procedure
California TP-901
Federal Baseline
After Durability Testing
Test
Temperature
40°C
28°C
28°C
Tankl
0.28
0.32
0.30
Tank 2
0.26
0.47
0.92
Tank 3
0.27
0.42
0.57
Average
0.27
0.41
0.60
XLPE fuel tanks
We tested several fluorinated cross-link polyethylene (XLPE) fuel tanks for permeation
emissions. The first tank was a 6 gallon test tank produced by a marine fuel tank manufacturer
specifically for this testing. The remaining fuel tanks were purchased on the open market. The
fuel tanks were then treated by a fluorinater. We tested the first tank on certification gasoline.
After a 20 week soak, we observed a permeation rate of 0.11 g/gal/day (1.52 g/m2/day), which
represented more than an 80 percent reduction in permeation.
The remainder of the fluorinated tanks were tested on E10 (10 percent ethanol) using the
same procedures as for the HDPE tanks discussed above. These fuel tanks were treated at a level
equivalent to what the fluorinater uses for automotive applications. All of the fuel tanks were
treated both on the inside and outside. The test results, presented in Table 5.3-24, showed
emission reductions of about 40 percent on average. Emission results from the sloshed fuel
5-55
-------
Final Regulatory Impact Analysis
tanks were not significantly different than from the tanks that were not sloshed.
Table 5.3-24: EPA Permeation Data on Fluorinated Cross-Link Fuel Tanks at 29°C on
E10
Fuel Tank
1
2
3
Capacity
12 gallons
12 gallons
12 gallons
Soak Period
29 weeks
29 weeks
29 weeks
g/gal/day
0.27
0.39
0.32
0.36
0.38
0.39
0.34
0.42
0.32
0.28
0.22
0.22
g/m2/day
4.1
5.9
4.9
5.4
5.8
5.7
5.0
6.2
4.6
3.4
2.6
2.8
slosh test?
no
no
no
no
no
yes
no
no
no
yes
no
no
5.3.2.3 Barrier Platelets
Another approach to creating a permeation barrier in a fuel tank is to blend a low
permeable resin in with the HDPE and extrude it with a single screw. The trade name typically
used for this permeation control strategy is Selar®. The low permeability resin, typically nylon
or EVOH, creates non-continuous platelets in the HDPE fuel tank which reduce permeation by
creating long, tortuous pathways that the hydrocarbon molecules must navigate to pass through
the fuel tank walls. Although the barrier is not continuous, this strategy can still achieve greater
than a 90 percent reduction in permeation of gasoline. EVOH has much higher permeation
resistance to alcohol than nylon; therefore, it would be the preferred material to use for meeting
our standard which is based on testing with a 10 percent ethanol fuel.
We tested several portable gas cans and marine tanks molded with low permeation non-
continuous barrier platelets 29°C. Six of fuel tanks tested were constructed using nylon as the
barrier material. The remainder of the fuel tanks were constructed using ethylene vinyl alcohol
(EVOH) as the barrier material. The advantage of EVOH is that it has much better resistance to
alcohol than nylon. Five of the nylon based fuel tanks were tested on certification gasoline. The
sixth tank was tested on E10 (10 percent ethanol) to evaluate the effectiveness of this material
with alcohol blended fuel. The fuel tanks with the EVOH barrier were all tested on E10.
Testing was performed after the fuel tanks had been filled with fuel and stored at room
temperature. The purpose of the soak period was to ensure that the fuel permeation rate had
stabilized. Although 20 weeks was generally accepted as an acceptable period, we soaked the
tanks with gasoline for 22 weeks and the tanks with E10 for 37 weeks. The fuel tanks were
5-56
-------
Feasibility of Evaporative Emission Control
drained and then filled with fresh fuel prior to the permeation tests. Because the barrier platelets
are integrated in the tank wall material, it did not seem likely that pressure or slosh testing would
significantly affect the performance of this technology.
Table 5.3-25 presents the results of the permeation testing on the fuel tanks with barrier
platelets. These test results show more than an 80 percent reduction for the nylon barrier tested
on gasoline. However, the nylon barrier does not perform as well when a fuel with a 10 percent
ethanol blend is used. Testing on a pair of 2 gallon tanks with nylon barrier showed 80 percent
percent higher emissions when tested on E10 than on gasoline. We also tested fuel tanks that
used EVOH barrier platelets. EVOH has significantly better resistance to permeation on E10
fuel than nylon (see Appendix 5D for material properties). For the fuel tanks blended with 6
percent EVOH, we observed an average permeation rate of about 1.4 g/m2/day on E10 fuel
which meets our permeation standard.
Table 5.3-25: Permeation Rates for Plastic Fuel Containers
with Barrier Platelets Tested by EPA at 29°C
Percent
Selar®*
Tank Capacity
[gallons]
Test Fuel
Nylon barrier platelets
unknown* *
unknown* *
4%
4%
4%
4%
2
2
5
5.3
6.6
6.6
gasoline
E10
gasoline
gasoline
gasoline
gasoline
EVOH barrier platelets
2%
4%
4%
6%
6%
6.6
6.6
6.6
6.6
6.6
E10
E10
E10
E10
E10
Fuel Soak
[weeks]
40
40
22
22
22
22
37
37
37
37
37
g/gal/day
g/m2/day
0.54
0.99
0.35
0.11
0.15
0.14
3.7
6.8
4.1
1.2
1.6
1.5
0.23
0.14
0.15
0.08
0.09
3.0
1.9
2.0
1.4
1.4
*trade name for barrier platelet technology used in test program
** designed to meet California permeation requirement
Manufacturers raised the concern about whether or not a tank using barrier platelets
would have a stabilized permeation rate after 20 weeks. In other words, manufacturers were
concerned that this technology may pass the test, but have a much higher permeation rate in-use.
We tested one of the 4 percent and 6 percent EVOH tanks on E10 again after soaking for a total
of 104 weeks (2 years). The measured permeation rates were 2.0 and 1.4 g/m2/day for the 4
percent and 6 percent EVOH tanks, respectively, which represents no significant changes in
permeation from the 37 week tests. In contrast we measured the 4 percent nylon tanks again
after 61 weeks and measured a permeation rates of 2.8 and 2.7 g/m2/day which represented about
5-57
-------
Final Regulatory Impact Analysis
an 80-90 percent increase in permeation compared to the 22 week tests.
The California Air Resources Board (ARE) collected test data on permeation rates from
portable fuel containers molded with low permeation non-continuous barrier platelets using
California certification fuel. These fuel tanks all used nylon as the barrier resin. The results
show that this technology can be used to achieve significant reductions in permeation from
plastic fuel containers. This data was collected using a diurnal cycle from 18-41°C which is
roughly equivalent to steady-state permeation testing at 30°C. Because the data is reported in
g/gal/day, we only include the data on fuel tanks here that are compatible in size with marine
fuel tanks. This test data showed that more than a 90 percent reduction in permeation is
achievable through the use of nylon barrier platelets. However, all of this testing was performed
on California certification fuel which does not include ethanol.
Table 5.3-26: Permeation Rates for Plastic Fuel Containers
with Barrier Platelets Tested by ARB Over a 18-41°C Diurnal
Percent Selar®*
4%
(average =0.12 g/gal/day)
6%
(average =0.09 g/gal/day)
807
/O
(average =0.07 g/gal/day)
Tank Capacity
[gallons]
5
5
5
5
5
6
6
5
5
5
5
5
5
6
6
5
5
6
6
Permeation Loss
[g/gal/day]
0.08
0.09
0.13
0.16
0.17
0.08
0.10
0.07
0.07
0.07
0.08
0.12
0.17
0.06
0.07
0.08
0.10
0.05
0.06
*trade name for barrier platelet technology used in test program
Dupont, who manufacturers Selar®, has performed testing on HDPE with higher
blends of EVOH (known as Selar RB®). Table 5.3-27 presents permeation rates for HDPE and
three Selar RB® blends when tested at 60°C on xylene.98 Xylene is a component of gasoline and
gives a rough indication of the permeation rates on gasoline. This report also shows a reduction
of 99 percent on naptha and 98 percent on toluene for 8 percent Selar RB®.
5-58
-------
Feasibility of Evaporative Emission Control
Table 5.3-27: Xylene Permeation Results for Selar RB® at 60°C
Composition
100% HOPE
10%RB215/HDPE
10%RB300/HDPE
15%RB421/HDPE
Permeation, g mm/m2/day
285
0.4
3.5
0.8
% Reduction
_
99.9%
98.8%
99.7%
5.3.2.4 Alternative Materials
Permeation can also be reduced from fuel tanks by constructing them out of a lower
permeation material than HDPE. Examples of alternative materials are metal, various grades of
plastic, and new fiberglass construction.
5.3.2.4.1 Metal
Gasoline does not permeation through metal. Therefore, the only permeation from a
metal fuel tank would be through rubber gaskets or O-rings that may be used to seal connections
on the fuel tank. Examples would be the gasket or O-ring in a fuel cap or a bolted-on component
such as a sender unit for a marine tank. Presumably, the exposed surface area of the gaskets
would be small enough that a metal fuel tank would be well below our permeation standard.
One issue with metal fuel tanks, however, is fuel leakage due to corrosion. A study sponsored
by the Coast Guard in 1994 showed that aluminum (and even stainless steel) fuel tanks are prone
to failure, both in salt water and fresh water applications., due to corrosion." Fuel leakages
would not only be an environmental issue, but could be a safety issue as well. Aluminum fuel
tank manufacturers have stated that corroding fuel tanks are typically due to improper
installation.
5.3.2.4.2 A Iternative Plastics
There are grades of plastics other than HDPE that could be molded into fuel tanks. One
material that has been considered by manufacturers is nylon; however, although nylon has
excellent permeation resistance on gasoline, it has poor chemical resistance to alcohol-blended
fuels. As shown in Appendix 5D, nylon could be used to achieve more than a 95 percent percent
reduction in permeation compared to HDPE for gasoline. However, for a 10 percent ethanol
blend, this reduction would significantly less depending on the grade of nylon. For a 15 percent
methanol blend, the permeation would actually be several times higher through nylon than
HDPE.
Some handheld equipment, primarily chainsaws, use structurally-integrated fuel tanks
where the tank is molded as part of the body of the equipment. In these applications, the frames
(and tanks) are typically molded out of nylon for strength. We tested structurally-integrated fuel
tanks from four handheld equipment manufacturers at 29°C on both gasoline and a 10 percent
ethanol blend. The test results suggest that permeation emissions are 20 to 70 percent higher on
5-59
-------
Final Regulatory Impact Analysis
E10 than on gasoline for these fuel tanks. Note these fuel tanks are capable of meeting the
standards using their current materials. In the cases where the permeation rates were higher than
these standards, it was observed that the fuel cap seals had large exposed surface areas on the O-
rings, which were not made of low permeation materials. Emissions could likely be reduced
significantly from these tanks with improved seal designs. Table 5.3-28 presents the results of
this testing.
Table 5.3-28: Permeation Rates for Nylon Handheld Fuel Tanks Tested by EPA at 29°C
Tank ID
Rl
R2
R3
Bl
B2
B3
B4
Wl
W2
W3
Gl
G2
G3
Application
clearing saw
hedge clipper
chainsaw
chainsaw
Material
nylon 6
nylon 6, 33% glass
nylon 6, 30% glass
nylon 6, 30% glass
Test Fuel
gasoline
E10
E10
gasoline
E10
E10
E10
gasoline
E10
E10
gasoline
E10
E10
Permeation Loss
[g/m2/day]
0.34
0.42
0.48
0.62
1.01
1.12
0.93
1.45
2.18
2.46
1.30
1.41
2.14
Other materials which have excellent permeation resistance even with alcohol-blended
fuels are acetal copolymers and thermoplastic polyesters. These polymers can be used to form
fuel tanks in the blow-molding, rotational-molding, and injection-molding processes. An
example of an acetal copolymer is known as Celcon® which has excellent chemical resistance to
fuel and has been shown to be durable based on exposure to automotive fuels for 5000 hours at
high temperatures.100 As shown in Appendix 5D, Celcon would result in more than a 99 percent
reduction in permeation compared to HDPE for gasoline. On a 10 percent ethanol blend, the use
of Celcon would result in more than a 95 percent reduction in permeation. Two thermoplastic
polyesters, known as Celanex® and Vandar®, are also being considered for fuel tank construction
and are being evaluated for permeation resistance by the manufacturer. Celcon has a more
crystalline structure than Vandar resulting in lower permeation but less impact resistance.
We tested a 1-liter blow-molded Vandar fuel tank and three rotationally-molded 3-liter
fuel tanks made of impact toughened Celcon for permeation at 29°C on E10 fuel. Prior to the
permeation testing, the fuel containers were soaked in E10 for more than 20 weeks. These test
results are included in Table 5.3-29 below. For the Celcon tank tests, higher emissions were
observed in the second week than the first week. This behavior was seen in repeat tests and was
5-60
-------
Feasibility of Evaporative Emission Control
likely due to deterioration of the epoxy seal used in this testing. Therefore, the actual emission
rates of the material are likely lower than presented below. More detailed data on this testing is
available in the docket.101
Table 5.3-29: Permeation Results Acetal Copolymer Fuel Tanks at 29°C on E10
Material Name
Vandar VI
Impact CIO
Resistant
Cll
Celcon C13
Material Type
thermoplastic polyester
modified
acetal copolymer
g/gal/day
1.7
0.13
0.09
0.10
g/m2/day
5.6
0.75
0.53
0.59
Fuel tank manufacturers have expressed some concern that the acetal copolymer is not as
tough as cross-link polyethylene. Thermoplastic polyesters have better impact resistance, but
higher permeation. The impact toughened fuel tanks mentioned above were in response to these
concerns. Also, the rotational molding process must be better controlled to use these materials in
comparison to XLPE. The temperature profile must be tightly controlled to uses Celcon, or
formaldehyde gases may form. The moisture level of Vandar must be kept low prior to molding.
Acetal copolymers are also used today to produce many fuel resistant automotive
components such as low permeation fuel caps. This construction has been used for many years
in automotive applications and now acetal copolymers are being used to manufacture low
permeation fuel caps for nonroad equipment as well.
Another low permeation thermoplastic that can be used in the manufacture of fuel tanks
is a polyester/polycarbonate alloy. One example is marketed under the trade name of Xenoy
6620. This engineered plastic is impact modified and is intended for the injection molding
process. The polyester provides good chemical resistance and the polycarbonate provides the
impact resistance. Permeation testing was performed on a fuel tank made of Xenoy 6620
following the California test procedures. At 40°C on California Phase II CERT fuel, the
measured permeation rate was 0.26 g/m2/day.102 The manufacturer of this material also has a
version that is modified slightly so that it can be used in the blow-molding process.
5.3.2.4.3 Low Permeation Fiberglass
One manufacturer has developed a low permeation fiberglass fuel tank construction.103
The composite tanks are fabricated using a glass fiber reinforced closed cell urethane composite
sheet as substrate and assembled with structural urethane adhesive as a fastening medium. These
fuel tanks may be hand constructed, or for larger volume production, they may be molded at
lower cost. Once fully assembled with necessary fuel fittings the tank is coated with fiberglass
reinforced resin, sufficient for H-24 ABYC (American Boat and Yacht Council) and 33 CFR
183.510 standards for fuel systems mechanical strength requirements. A final gel coat finish may
5-61
-------
Final Regulatory Impact Analysis
was applied for aesthetics.
Permeation control is achieved by incorporating fillers into a resin system and coating the
assembled tank interior and exterior. This filler is made up of nanocomposites (very small
particles of treated volcanic ash)4 which are dispersed into a carrier matrix. This construction
creates a tortuous pathway for hydrocarbon migration through the walls of the fuel tank. We
tested a 14 gallon fuel tank provided by this manufacturer and measured a permeation rate of
0.97 g/m2/day on E10 fuel at 29°C. Other advantages of this technology are improved strength
and flame resistance compared to plastic fuel tanks.
5.3.2.5 Multi-Layer Construction
Fuel tanks may also be constructed out of multiple layers of materials. In this way the
low cost and structural advantages of traditional materials can be utilized in conjunction with
higher grade materials which can provide effective permeation resistance. Today, fuel tanks are
made in many ways including higher volume blow-molding, lower volume injection molding,
and very low volume rotational-molding. The discussion below presents data on several multi-
layer fuel constructions.
5.3.2.5.1 Blow-Molded Coextruded Barrier
Coextruded barrier technology has been long established for blow-molded automotive
fuel tanks. Data from one automobile manufacturer showed permeation rates of 0.01-0.03 g/day
for coextruded fuel tanks at 40°C on EPA certification fuel. They are using this technology to
meet LEV II vehicle standards. For comparison, this manufacturer reported permeation rates of
more than 10 g/day for standard HDPE fuel tanks.104
Another study looks at the permeation rates, using ARB test procedures, through multi-
layer fuel tanks.105 The fuel tanks in this study were 6 layer coextruded plastic tanks with EVOH
as the barrier layer (3 percent of wall thickness). The outer layers were HDPE and two adhesive
layers were needed to bond the EVOH to the polyethylene. The sixth layer was made of
recycled polyethylene. The two test fuels were a 10 percent ethanol blend (CE10) and a 15
percent methanol blend (CM15). See Table 5.3-30.
Table 5.3-30: Permeation Results for a Coextruded Fuel Tank Over a 18-41°C Diurnal
Composition
100% HDPE (approximate)
3% EVOH, 10% ethanol (CE10)
3% EVOH, 15% methanol (CM15)
Permeation, g/day
6-8
0.2
0.3
% Reduction
97%
96%
4 Chemically modified montmorillonite for nanocomposite formulation
5-62
-------
Feasibility of Evaporative Emission Control
The California Air Resources Board tested two sets of three 5-gallon portable fuel
containers.106 Each set was manufactured by a different company, but all of the fuel tanks were
blow-molded with a coextruded barrier layer. Testing was performed over the California 18-
41 °C temperature cycle with California Phase II gasoline. Testing was performed with and
without the spouts removed. The test data presented in Table 5.3-31 was after 174 days of fuel
soak with the spouts removed and the openings welded shut. California reported the test results
in grams per gallon. Table 5.3-31 also presents approximate g/m2/day values based on the
relationship between tank capacity and inside surface area used in the NONROAD2005
emissions model.
Table 5.3-31: ARB Permeation Results for a Coextruded Portable Fuel Tanks
Fuel Tank
Bl
B2
B3
Average
Ml
M2
M3
Average
Permeation, g/gal/day
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
Approximate Rate
0.09
0.11
0.11
0.10
0.14
0.21
0.18
0.17
in g/m2/day
The handheld industry also tested a number of fuel tanks for their products that had been
coextruded with an EVOH barier.107 Even with NBR gaskets on the fuel caps, these tanks had
permeation rates well below the new standards. This test data is presented in Table 5.3-32.
Table 5.3-32: Permeation Rates for Handheld Fuel Tanks Tested by Industry
Tank Capacity
1840
470
Gasket Material
NBR
NBR
Test Temp.
28°C
28°C
Test Fuel
CE10
CE10
Permeation Loss
[g/m2/day]
0.23
0.26
0.75
0.54
Another approach has recently been developed in which a multi-layer fuel tank can be
blow-molded with only two layers.108 In this construction, a barrier layer of a polyarylamide
known as Ixef MXD6 is used on the inside of a HDPE fuel tank. Ixef has permeation properties
similar to EVOH. Test results showed a permeation rate of 0.8 g-mm/m2/day at 60°C on CE10
for a test film of Ixef. Unlike EVOH, Ixef can be exposed directly to the fuel which removes the
need for an inner layer of HDPE. In addition, a tie material can be blended into the HDPE which
will allow the polyarylamide to bond directly to the HDPE rather than using an adhesive layer.
5-63
-------
Final Regulatory Impact Analysis
Permeation emissions were measured on five 2.5 liter fuel tanks at 28°C using fuel
CE10.109 During the preconditioning of these fuel tanks, the pressure-vacuum, slosh, and UV-
exposure durability tests were performed. All of the testing was performed using stock fuel caps
and gaskets. One of the fuel tanks showed considerably higher emissions than the others. This
higher emitting tank was found to have an issue with the interface between the tank and the fuel
cap. In general, the permeation rates were well below the tank permeation standard. These test
results are presented in Table 5.3-33.
We also tested three test bottles and three fuel tanks made using this construction for
permeation emissions. The test bottles were about 1.3 liters in volume and used a fluoropolymer
gasket under the caps. The fuel tanks were similar to those described above. Each of the test
bottles and tanks was filled with E10 and soaked for more than 20 weeks. Prior to the two week
weight loss test, fresh fuel was added to each bottle/tank. As shown in Table 5.3-33, the
measured permeation results were well below the new tank permeation standard.
Table 5.3-33: Permeation Results Ixef Barrier Test Bottles at 29°C on E10
Tank
tank 1
tank 2
tankS
tank 4
tank5a
tank 6
tank 7
tankS
bottle lb
bottle 2b
bottle 3b
Conditions
28 °C,
Fuel CE10
29°C,
Fuel E10
29°C,
Fuel E10
Preconditioning
fuel soak, slosh,
pre ssure -vacuum,
and UV exposure
fuel soak
fuel soak
g/gal/day
0.11
0.14
0.19
0.13
0.34
0.04
0.19
0.04
0.05/0.02
0.14/0.02
0.07/0.02
g/m2/day
0.67
0.87
1.17
0.80
2.07
0.36
1.80
0.36
0.26/0.12
0.72/0.12
0.39/0.12
3 interface issue reported with cap
b repeated test with epoxy seal on fuel cap
Tanks 6-8 and bottles 1-3 were tested by EPA using screw on fuel caps with gaskets. For
the test bottles, gaskets were cut at EPA from FKM rubber. It was thought that the permeation
results were affected by the seal at the fuel cap. Therefore, the tests were rerun 6 months later
using epoxy as a secondary seal at the fuel cap. As a result, much lower permeation rates were
measured for all three test bottles. For the fuel tanks, the stock fuel caps and gaskets were used.
One of the tanks had significantly higher permeation emissions than the other two tanks. This
difference may have been a seal issue at the fuel cap.
5.3.2.5.2 Rotational Molded Dual-layer Construction
As discussed above, an inner layer can be molded into the inside of a rotationally molded
5-64
-------
Feasibility of Evaporative Emission Control
fuel tank through the use of a drop-box that opens after the XLPE tank begins to form. Through
this method, a XLPE fuel tank could be molded with a low permeation inner barrier. With this
construction, it may be possible to reduce the amount of XLPE used depending on the structural
characteristics of the inner liner material. For instance, acetal copolymer can be rotationally
molded and could be used as the inner liner. This way, the permeation characteristics of an
acetal copolymer could be achieved through an inner liner while still retaining the toughness of
XLPE . One issue would be that acetal copolymers do not readily adhere to XLPE. Therefore
fitting designs would need to account for this.
Another material that could be used in a multi-layer approach is nylon which comes in
many grades. Typical nylon grades used in Small SI fuel tank constructions may not perform
well in marine applications because of the hygroscopic nature of these nylons. In other words,
typical nylon adsorbs water which can make it brittle. In addition, E10 fuel permeates through
nylon much more readily than gasoline.
One manufacturer is working with a nylon known as Rilsan® polyamide 11 (PA 11) in
constructing low permeation multi-layer rotational-molded fuel tanks.110 Rilsan® polyamide 11
has two advantages to traditional nylons in that it is not hygroscopic and it is more resistive to
alcohol fuels. One manufacturer has manufactured fuel tanks using the PA11 as an inner liner in
a polyethylene shell. The manufacturer using this approach reports a permeation rate of about 3
g-mm/m2/day on fuel CE10 at 28°C compared to about 30 g-mm/m2/day for XLPE. In addition,
the nylon used in multi-layer constructions is formulated with a polyethylene graft that causes it
to adhere well to XLPE. This prevents the layers from separating in use.
We tested two 10 gallon multi-layer rotational molded fuel tanks at 29°C with E10 fuel
after a 35 week soak with two fuel changes during that period.111 One of the tanks was molded
with an outer shell of medium-density polyethylene while the other was molded with an outer
shell of cross-link polyethylene. The long soak period was due to test equipment problems and
the fuel was changed with each test attempt. However, it presents valuable data on the longer
term effectiveness of this technology. This test data is presented in Table 5.3-34. The
manufacturer reported that this tank design passed testing on the Coast Guard burn, pressure,
shock, and impulse test requirements.112'113'114'115 In addition, a tank of this construction was
tested and passed the tank durability tests for snowmobiles specified in SAE J288.116 These tests
include cold (-40°C) and hot temperature (60°C) immersion and drop tests.
Typically, multi-layer rotational-molded fuel tanks are constructed with the use of a drop
box which adds the inner-layer material into the mold after the first material sets. Other
approaches are to use a meltable bag containing the inner-layer material or even to pull the mold
from the oven to add the inner-layer material. However, one manufacturer, that participated in
the SBREFA process, has stated that they have developed a method to mold the inner liner
without the use of a drop box or other approach that lengthens molding cycle time. This fuel
tank manufacturer is selling fuel tanks using this construction for use in Small SI equipment and
is selling mono-layer XLPE rotational-molded tanks for use in boats.
5-65
-------
Final Regulatory Impact Analysis
Table 5.3-34: Permeation Results PA 11/PE Fuel Tanks at 29°C on E10
Tank
1
2
Outer Shell
MDPE
XLPE
g/gal/day
0.05
0.06
g/m2/day
0.71
0.79
Under their rule for small offroad equipment, California may issue executive orders to
manufacturers with low emission products. As of August, 2006, ARB has issued 5 executive
orders for low permeation fuel tanks.117 Under these executive orders, two basic multi-layer
rotomolded (XLPE and nylon) approaches have been approved. The California fuel tank
permeation standard is 1.5 g/m2/day tested at 40°C on California certification fuel. However,
most of the testing was performed on fuel CE10 which is a significantly more aggressive fuel for
permeation. Table 5.3-35 presents the test results for rotational-molded fuel tanks with ARB
executive orders. Note that the reported emissions are the average of 3-5 test samples.
Table 5.3-35: ARB Fuel Tank Executive Orders for Small Offroad Equipment
EO#
C-U-05-005
C-U-06-014
Test Fuel
CE10
Phase II
CE10
CE10
CE10
g/m2/day
0.81
0.18
0.10
0.00
0.09
There is another approach to dual-layer rotomolded fuel tanks under development that
uses a "single shot" approach to molding.118 In this method a material known as polybutylene
terephthalate cyclic oligimor (CBT) is combined with the XLPE in the mold. Because of the
different melt rates and viscosities of the two materials, during the mold process, the CBT®
polymerizes into a thermoplastic known as polybutylene terephthalate (PBT) to form a barrier
layer on the inside of the fuel tank. Adhesion between the PBT and XLPE comes from
mechanical bonding between the two layers. This material can be used without lengthening the
cycle time for rotational molding, and it does not require forced cooling.119 Initial testing shows
a permeation rate of <1 g/m2/day when tested with fuel CE10 at 40°C for a sample with a 3.9 mm
total wall thickness.120 This wall thickness for this testing was composed of 0.9mm CBT and
3.0mm XLPE. PBT itself has a permeation rate on CE10 at 40°C of less than 0.05 g-mm/m2/day.
5.3.2.5.3 Injection-Molded Dual-Layer Construction
To add a barrier layer in the injection molding process, a thin sheet of the barrier material
may be placed inside the mold prior to injection of the poleythylene. The polyethylene, which
generally has a much lower melting point than the barrier material, bonds with the barrier
material to create a shell with an inner liner.
5-66
-------
Feasibility of Evaporative Emission Control
5.3.2.5.4 ThermoformedMulti-Layer Construction
As an alternative, multiple layers can be created through thermoforming.121 In this
process, sheet material is heated then drawn into two vacuum dies. The two halves are then
fused while the plastic is still molten to form the fuel tank. Before the halves are fused together,
it is possible to add components inside of the fuel tank. Low permeation fuel tanks can be
constructed using this process by using multi-layer sheet material. This multi-layer sheet can be
extruded using similar materials to multi-layer blow-molded fuel tank designs. A typical barrier
construction would include a thin EVOH barrier, adhesion layers on both sides, a layer of HDPE
regrind, and HDPE layers on the outside surfaces.
This process has low capital costs compared to blow-molding and should be cost
competitive with injection molding and rotational-molding. Manufacturers have indicated that
this construction could be coated with an intumescent material which would help it pass the
Coast Guard fire test. This coating could be applied directly to the multi-layer plastic sheets
while they are still hot after extrusion. Once the plastic cools, it could be applied using flame
ionization or electric arcing to increase the surface are of the plastic for adhesion.
EPA tested two, 5.6 gallon, thermoformed fuel tanks for permeation. These fuel tanks
were constructed as described above with a thin EVOH barrier and were soaked with E10 for 27
weeks prior to testing. Due to test variability, testing was repeated at 35 and 44 weeks (fresh
fuel was added prior to each weight loss test). From day to day, a constant weight loss was not
always observed, and weight gains were occasionally seen. This variability in measured weight
loss was likely due to the very low permeation rates combined with the effect of atmospheric
conditions on measured weight. The highest variations in weight loss were observed when
storms passed through suggesting that the changes in barometric pressure and relative humidity
were affecting the buoyancy of the fuel tanks (discussed in more detail in Section 5.6.2.3). In
the third round of testing (after 44 weeks), barometric pressure and humidity were measured and
deemed to be relatively stable. In addition, a smaller tank with sand in it (rather than fuel) was
measured simultaneously as a control to give some indication of the buoyancy effect. A small
weight loss was measured for the control tank, suggesting that the measured test results may
slightly overstate the permeation for the thermoformed fuel tanks. Table 5.3-36 presents the test
results for each of the three tests.
5-67
-------
Final Regulatory Impact Analysis
Table 5.3-36: Permeation Results Multilayer Thermoformed Fuel Tanks at 29°C on E10
Soak (weeks)
27
35
44
Average
Tank
#16
#21
#16
#21
#16
#21
#16
#21
g/gal/day
0.01
0.01
0.01
0.01
0.01
0.00
0.01
0.01
g/m2/day
0.15
0.05
0.07
0.09
0.11
0.04
0.11
0.06
5.3.2.5.5 Epoxy Barrier Coating
Another approach that has shown promising results is to coat a plastic fuel tank with a
low permeation epoxy barrier coating. Early attempts at coating a plastic fuel tank resulted in
coatings that eventually wear off due to the difficulty of bonding some materials to HDPE and
XLPE. However, because fluorination increases the surface energy of the plastic, a low level of
fluorination can be used to make it possible to apply an epoxy coating, even to XLPE. Because
this approach is applied to the fuel tank post-molding, it can be used for any plastic fuel tank,
regardless of the production molding method.
We performed permeation testing on six 12 gallon rotationally-molded XLPE fuel tanks
with a thin, low-permeation epoxy coating. This coating was a two-part epoxy that was sprayed
onto the tank and thermally cured in 45 minutes. Prior to the permeation measurements, the fuel
tanks were soaked with E10 fuel at about 25°C for 15 weeks. The tanks were then drained and
fresh E10 was added prior to the 29°C constant temperature permeation test. Inspection of the
externally coated fuel tanks showed that the epoxy was unevenly applied and that some bare
spots existed. This was reflected in the unsatisfactory permeation results. A more careful
coating would be expected to result in similar results as the internal coatings. One of the
externally coated fuel tanks was over-coated with a 1-part epoxy that was cured with a 45 second
UV exposure. This tank was soaked for an additional 6 weeks prior to retesting. These test
results, which are presented in Table 5.3-37, show that this technology can be used to reduce
permeation emissions by more than 90 percent.
5-68
-------
Feasibility of Evaporative Emission Control
Table 5.3-37: EPA Permeation Data on Epoxy Coated XLPE Fuel Tanks at 29°C on E10
Fuel Tank Set
1
2
3
Coating
Inside
Thermocured
Outside
Thermocured*
Outside
UV cured
Soak Period
15 weeks
15 weeks
additional
6 weeks
g/gal/day
0.04
0.001
0.07
0.13
0.23
0.23
0.03
g/m2/day
0.6
0.02
1.0
1.9
3.3
3.3
0.4
slosh test?
no
no
yes
no
no
yes
no
* inspection showed uneven application of the coating which affected permeation results
Since the above testing was performed, the fluorinater and the epoxy manufacturer who
developed this approach have performed more testing on their UV cured, 1-part epoxy. The
testing was performed on epoxy coated HDPE bottles and 2 gallon fuel tanks using the
California ARB test procedure of 40°C with California certification fuel.122 At 29°C, we would
expect the permeation rate to be about half of these levels due to the relationship between
permeation and temperature discussed above in Section 5.3.1.2. The results for this testing were
reported to be 0.3 g/m2/day on average for both the bottles and tanks on gasoline. The bottles
had a permeation rate of 0.5 g/m2/day on gasohol (ethanol blend). This technology resulted in
better than 95 percent reductions in permeation. Table 5.3-38 presents the test results after a 9
week fuel soak at 40°C.
Table 5.3-38: Permeation Data: Epoxy Coated HDPE Fuel Tanks at 40°C on CA Cert Fuel
Fuel Tank
1
2
3
4
g/gal/day
0.04
0.02
0.02
0.08
g/m2/day
0.25
0.09
0.11
0.49
Roto-molders of marine fuel tanks generally use cross-link polyethylene. The advantage
of XLPE is that its cross-link structure causes it to behave like thermoset which helps the fuel
tanks pass the Coast Guard fire test (33 CFR 183.590) by holding their shape longer under
exposure to fire. If a flame retardant were included in the epoxy coating, a less expensive
material, such as FIDPE could be used to make fuel tanks that are subject to the flame test
requirement. The manufacturers who have developed the above approach for permeation have
developed an additive that provides an intumescent coating to allow the fuel tanks to be
produced at a lower cost. Testing on the Coast Guard burn test showed that an FIDPE fuel tank
would fail around after being exposed to a flame for about 1.5 minutes (the standard is 2.5
5-69
-------
Final Regulatory Impact Analysis
minutes). With the intumescent coating, the fuel tank passed the flame test and survived more
than 5 minutes.123
5.4 Fuel/Vapor Hose Permeation
The polymeric materials (plastic or rubber) used in the construction of gasoline fuel and
vapor hoses generally have chemical compositions much like that of gasoline. As a result,
constant exposure of gasoline to these surfaces allows the material to continually absorb fuel.
Permeation is driven by the difference in the chemical potentials of gasoline or gasoline vapor on
either side of the material. The outer surfaces of these materials are exposed to ambient air, so
the gasoline molecules permeate through these fuel-system components and are emitted directly
into the air. Permeation emissions continue at a nearly constant rate, regardless of how much the
vehicle or equipment is used. Because of these effects, permeation-related emissions can
therefore add up to a large fraction of the total evaporative emissions.
This section summarizes the data and rationale supporting the permeation emission
standard for fuel lines presented in the Executive Summary.
5.4.1 Baseline Hose Technology and Emissions
5.4.1.1 Marine Fuel Hose Subject to 33 CFR part 183
The majority of marine fuel hoses are constructed primarily of nitrile rubber with a
chloroprene cover for abrasion and flame resistance. Hoses are designed to meet the Coast
Guard requirements in 33 CFR part 183 which reference SAE J1527.124 Fuel hose for boats with
gasoline engines (excluding outboards) must meet the Class 1, Type A requirements which
specify a maximum permeation rate of 100 g/m2/day at 23°C on ASTM Reference Fuel C125 (50
percent toluene, 50 percent iso-octane). Class 1 refers to hose that is used where liquid fuel is
normally continuously in the hose. Type A refers to hose that will pass a 2V2 minute flame
resistance test.
On a fuel containing an alcohol blend, permeation would likely be higher from these fuel
hoses. In fact, the SAE J1527 standard also requires Class 1 hose to meet a permeation rate of
300 g/m2/day on fuel CM15 (15 percent methanol). Although ethanol is generally less
aggressive than methanol, ethanol in the fuel would still be expected to increase the permeation
rate significantly through most fuel hoses. Based on the data presented in Appendix 5D,
permeation through nitrile rubber is about 50 percent higher when tested on Fuel CE10 (10
percent ethanol) compared to testing on Fuel C.
Fuel fill neck hoses are subject to a less stringent permeation standard under the Coast
Guard specifications because they are not normally continuously in contact with fuel (Class 2).
This relaxed standard is 300 g/m2/day on Fuel C and 600 g/m2/day on Fuel CM15 at 23°C.
Where marine fuel hose is typically extruded, fill neck hose is generally constructed by wrapped
layers on a mandrill. Fill neck hose is constructed with a larger inner diameter (1.5-2") to
accommodate higher fuel rates and with thicker, more heavily reinforced walls, to prevent
5-70
-------
Feasibility of Evaporative Emission Control
buckling and pinching.
Marine fuel hose is typically designed to be somewhat lower than the SAE J1527
requirements. Confidential data by one manufacturer supplying baseline marine fuel hose
suggested that their fuel feed hose is about 25 percent lower than the Class 1, Type A
requirement on Fuel C and about 35 percent lower on Fuel CM15. In their comments on the
2002 proposal for marine evaporative emission control, Lawrence industries stated that the
majority of their fill neck hose permeates in the range of 150 to 180 g/m2/day which is about half
of the 300 g/m2/day requirement required by the Coast Guard.126
We collected test data on marine hose permeation through contracts with outside
laboratories.127'128'129'130'131 Data was also available on a fuel feed hose testing funded by the
marine industry.132 All of the hose were prepared by soaking with liquid fuel for long enough
periods to stabilize the permeation rate. This data is presented in Table 5.4-1. Note that this data
shows somewhat lower permeation than was reported by manufacturers based on their own
testing. Especially in the case of the fuel feed hose, this may be a function of the hose
construction. This hose was purchased by the contractor without any knowledge of the hose
construction. Therefore, it is not known if this is a representative sample of a baseline hose
construction or if it contains some sort of barrier material.
Table 5.4-1: Permeation Rates for Baseline SAE J1527 Marine Fuel Hose
Hose Type
fuel feed hose
vent hose
fill neck hose
fill neck hose
fill neck hose
I.D.
3/8"
5/8"
1.5"
1.5"
1.5"
Fuel Type*
E10
Fuel CE10
E10
FuelC
FuelC
Fuel CE10
FuelC
E10
Fuel CE10
g/m2/day
43
88
37
95
98
109
87
164
123
123
274
Test Temperature
23 °C
28 °C
22-36 °C
temperature cycle
23 °C
23 °C
* E10 refers to gasoline with 10 percent ethanol
Although fuel hose used in personal watercraft is subject to 33 CFR part 183, personal
watercraft manufacturers do not use hose specified in SAE J1527. Fuel hose specifications are
contained in a separate recommended practice under SAE J2046.133 Under this practice, the
permeation requirement is 300 g/m2/day with testing performed in accordance with SAE J1527.
5-71
-------
Final Regulatory Impact Analysis
5.4.1.2 Other Marine Fuel Hose
Fuel hose used with outboard engines is not subject to 33 CFR part 183. This hose
includes the fuel line from the portable fuel tank to the engine and fuel hose on the engine itself
and is generally either constructed out of nitrile rubber with an abrasion resistant cover similar to
hose used in recreational vehicle applications or is constructed out of polyvinyl chloride (PVC).
One manufacturer of marine hose for use in outboard marine engines supplied permeation data
on five hose constructions tested at 23°C.134 This data is presented in Table 5.4-2 for Fuel C,
Fuel CE10, and Fuel CM15 (15 percent methanol). As shown by this data, hose permeation rates
can increase dramatically when tested on fuel blended with alcohol. Fuel lines connected to a
portable fuel tank are also generally fitted with a primer bulb which is also typically constructed
from nitrile rubber.
Table 5.4-2: Permeation Rates for Baseline Fuel Hose [g/m2/day at 23°C1
Fuel Hose
C-464-D11
C-530-D2-CE
ECO/CPE
J30R7
CMC ESI 763
FuelC
195
5
228
426
141
Fuel CE10
420
183
402
279
290
Fuel CM 15
590
546
565
433
314
gasoline*
66
4
53
27
43
E10
192
74
131
126
103
* cited as Marathon 92
5.4.1.3 Small SI Equipment Hose
Fuel hoses produced for use in Small SI equipment are generally extruded nitrile rubber
with a cover for abrasion resistance. This hose is often equivalent to SAE J30 R7 hose which as
a permeation requirement of 550 g/m2/day at 23°C135 on ASTM Fuel C (50 percent toluene, 50
percent iso-octane). On a fuel containing an alcohol blend, permeation would likely be much
higher for these fuel hoses. R7 hose is made primarily of nitrile rubber (NBR). Based on the
data presented in Appendix 5D, permeation through NBR is 50 percent higher when tested on
Fuel CE10 (10 percent ethanol) compared to testing on Fuel C.
One manufacturer performed a study of several hose samples and various fuel types.136
Permeation testing was performed using the methodology in SAE J30. These hose samples
included SAE J30 R7, R8, and R9 hose. The R7 hose samples were constructed with an
acrylonitrile inner tube with a chlorosulfonated polyethylene cover layer. The R8 hose samples
were constructed using a epichlorohydrin ethyleneoxide copolymer. The R9 hose used a
fluoroelastomer barrier for the inner tube with an outer tube made of chlorosulfonated
polyethylene compound reinforced with a polyester braid. Over the two week tests, the study
showed a peak permeation rate after 4-6 days for R7 and R8 hose and a peak permeation rate
after 10-12 days for the lower permeating R9 hose. Table 5.4-3 below presents the two week
averages for each of the hose samples and test fuels. In this study, the hose manufacturers were
not identified, but the hose samples were each given a letter designation.
5-72
-------
Feasibility of Evaporative Emission Control
Table 5.4-3: Permeation Rates for SAE J 30 Fuel Hose [g/m2/day at 23°C1
Fuel Hose
SAE J30 R7 "mfr. D"
SAE J30 R7 "mfr. E"
SAE J30 R8 "mfr. B"
SAE J30 R8 "mfr. F"
SAE J30 R9 "mfr. A"
SAE J30 R9 "mfr. C"
FuelC
450
330
152
130
2
2
Fuel CE10
508
501
385
355
11
6
Fuel CE15
541
433
337
308
10
4
Fuel CM 15
587
707
620
545
73
55
Handheld equipment typically use smaller diameter hose made of a single material with
no cover. This fuel hose may either be extruded straight run hose or may be more complex
injection-molded designs. To determine baseline permeation emission rates from hose on
handheld equipment, testing was performed by industry using a modified SAE J30 weight loss
procedure.137 In this modified procedure, E10 fuel was used and the testing followed a 30 day
fuel soak intended to stabilize the permeation rate. Further testing was later performed by
industry on similar fuel lines exposed to E10 and to CE10.138 This testing showed much higher
permeation on CE10 than on E10 for these hose samples. Table 5.4-4 presents the test results.
Table 5.4-4: Handheld Product Fuel Line Permeation Test Data [23"C1
Hose ID
90014
90015
90016
S3
S4
HI
H2
Bl*
B2*
SI
S2
C*
Bl*
B2*
E*
p*
A*
Construction
extruded
molded
extruded
molded
Test Fuel
E10
E10
CE10
CE10
Material
NBR
NBR
NBR/PVC
NBR/PVC
NBR
NBR
NBR/PVC
NBR/PVC
g/m2/day
198
192
168
165
171
360
455
205
224
198
386
270
742
662
1148
736
975
average of 5 measurements
5-73
-------
Final Regulatory Impact Analysis
5.4.1.4 Fuel Effects on Hose Permeation
As shown in the data above, adding ethanol or methanol to the test fuel significantly
affects the permeation rate through fuel hoses. Because the SAE guidelines typically specify
Fuel C for testing, most of the hose data available in the literature is on Fuel C or some blend of
Fuel C and ethanol or methanol.
One study looked at the effect of fuel composition on the permeation of several materials
used in baseline hose constructions.139 This data suggests that Fuel C is a more aggressive fuel
with respect to permeation than gasoline. In addition, this data shows that permeation for these
materials is very low with diesel fuel. Table 5.4-5 presents the data from this study. Appendix
5D includes a table spelling out the acronyms for the hose materials in this table.
Table 5.4-5: Permeation Rates by Fuel and Fuel and Hose Material [g/m2/day at 21°C]
Material
CFM
CO
ECO
ETER
39%ACNNBR
CSM
CR
FuelC
nil
150
190
230
300
490
640
CE10
35
270
390
400
420
575
690
CM10
nil
255
310
360
360
665
740
Indolene*
0.1
10
55
65
110
210
320
IE10
20
80
180
205
200
240
340
IM10
nil
125
150
165
200
300
385
Diesel
3
2
5
10
15
nil
10
"Indolene" refers to a fuel meeting the EPA specifications for certification gasoline
This difference in permeation between Fuel C and gasoline is likely due to the higher
aromatic content of Fuel C than of certification gasoline. A second study compared three
common fuel system materials on Fuel C and certification gasoline.140 Fuel C is made up of 50
percent toluene and 50 percent isooctane. As a result, it is half aromatics and half aliphatics. In
this study, the certification gasoline was observed to be 29 percent aromatics, 67 percent
aliphatics, and 4 percent olefms. The test results were indicative of the effect of aromatics on
permeation. Table 5.4-6 presents the permeation rate reported in g-mm/m2/day for three sample
materials: a low permeation fluoroelastomer (FKM), two medium permeation epichlorohydrins
(ECO) and two high permeation nitrile rubbers (NBR). This testing, which was performed at
24°C, gives a good comparison of the effect of gasoline versus Fuel C on permeation.
5-74
-------
Feasibility of Evaporative Emission Control
Table 5.4-6: Fuel C Versus Gasoline Permeation by Hose Material [g-mm/mVday]
Material
FKM-1
ECO-1
ECO-3
NBR-1
NBR-2
FuelC
3.3
180
282
570
705
Indolene*
1.2
33
45
255
510
% difference
-64%
-82%
-84%
-55%
-28%
"Indolene" refers to a fuel meeting the EPA specifications for certification gasoline
5.4.1.5 Vent Hose Permeation
Permeation occurs not only through hose walls that are in contact with liquid gasoline,
but also through surfaces exposed to fuel vapor. In the event that the fuel vapor represents a
saturated mix of air and fuel, we would expect permeation to be the same as that for exposure to
liquid fuel. In a fuel tank, the walls of the tank are readily exposed to saturated vapor as
discussed earlier in Section 5.3.1.4. In a fuel system hose not continuously exposed to liquid
fuel, the vapor concentration may be significantly lower than saturation for several reasons.
Clearly, if a hose is open to atmosphere, such as vent hose, there would be a gradient through the
hose ranging from saturated vapor in the fuel tank to fresh air outside of the fuel system. In
addition, if the tank is venting and drawing in air due to diurnal (or other) temperature changes,
then the fuel hose will regularly be exposed to varying vapor concentrations.
To investigate permeation rates for vent hose exposed to gasoline vapor, we contracted
with an outside laboratory to measure the permeation of fuel through marine hoses under various
venting configurations.141'142 The marine hose used in this testing met the USCG requirements
for SD/I vessels in specified in 33 CFR part 183 and SAE Recommended Practice J1527. Each
section of hose was connected to a metal fuel reservoir and exposed to liquid fuel for 8 weeks at
40°C to stabilize the permeation rate. The test fuel was EPA certification gasoline blended with
10 percent ethanol (E10) Each section of hose was then soaked for an additional 2 weeks at
40°C in the planned test configuration. After the soak, fresh fuel was added to the reservoirs and
permeation was measured in a mini-SFIED. Hose sections were tested at constant temperature in
three configurations.
One section of hose was tested exposed to liquid fuel. Two sections of hose (1.5 and 5/8"
ID.) were tested with one end connected to the fuel reservoir and the other opened to
atmosphere through a fitting in the SFIED. This configuration was intended to simulate vent
hose at constant temperature. A third configuration was also tested where three sections of hose
were configured as vent hose and tested over a 22.2-35.6°C one day diurnal sequence. This test
was intended to simulate vent hose in a fuel system exposed to fuel tank breathing caused by
temperature variation. The data in this testing, shown in Table 5.4-7, suggest that permeation
rates for vent lines are much lower than for hose that is regularly exposed to liquid fuel. This
result is likely due to a fuel concentration gradient in the hose which is largely due to one end
5-75
-------
Final Regulatory Impact Analysis
being exposed to fresh air.
Table 5.4-7: Effect of Venting on Hose Permeation with E10 [g/m2/dayl
I.D.
inches
1.5
0.625
0.625
0.625
0.625
Length
feet
1
3
3
3
3
Temperature
28°C (84°F) constant
22-3 6°C (72-96°F) diurnal
Liquid Exposure
123*
37
-
Vented to Atmosphere
3.3
5.8
4.3
4.5
4.9
* taken from Table 5.4-1 on a similar hose for comparison
The marine industry also funded permeation testing on vent hose exposed only to fuel
vapor and air.143 The vent line hose was preconditioned by attaching the hose to a 55 gallon steel
drum containing commercial gasoline containing 10 percent ethanol and setting the drum outside
during the summer. A carbon canister was attached to the end of the hose to simulate a vent line
with diurnal emission control. Permeation was measured after 90, 120, 150, and 180 days of
preconditioning. Because of the large size of the test rig, weight loss testing could not be
performed. Instead, a sleeve was fitted over the hose and nitrogen was flowed through the sleeve
to a carbon trap. The change in the weight of the carbon trap was then measured to determine
the permeation rate. As with the fill neck testing, the hose was configured to run vertically from
the top of the fuel reservoir (55 gallon drum). Repeat testing was performed on this hose and
both values for each hose are presented in Table 5.4-8. The permeation rates for this testing
were lower than for similar hose exposed to liquid fuel. Fuel vapor stratification may have been
caused by a number of factors including breathing of fresh air into the tank during ambient
cooling periods, gravity, and a limiting diffusion rate.
Table 5.4-8: Industry Test Data on Marine Vent Hose Exposed to Fuel Vapor
Hose manufacturer
#1
#2
Permeation [g/m2/day]
2.7,2.2
2.7,2.8
8.9,8.5
5.7,6.6
2.2,2.0
2.5,2.2
2.5,2.6
5-76
-------
Feasibility of Evaporative Emission Control
5.4.1.6 Vapor Hose Permeation
Even in a vapor hose that is sealed at one end, stratification may occur for a fuel system
due to gravity. An example of vapor hose would be fuel fill neck hose with a sealed cap.
Because fuel vapor is heavier than air, even a large diameter hose may see stratification of fuel
vapor concentration if it reaches high enough above the surface of the liquid fuel. The
stratification of vapor molecules happens slowly but would likely be observed under static
conditions. Another cause of low vapor concentration in fuel system hose may occur due to the
properties of diffusion discussed above in Section 5.1.3. If the hose diameter is small compared
to its length, diffusion of vapor into the hose may be the rate limiting step rather than the
permeation rate through the hose. In other words, the fuel vapor may enter the hose much slower
than rate at which it could permeate through the hose. This effect could be combined with the
other effects discussed above to cause lower permeation for fuel hose exposed to vapor rather
than liquid fuel.
The marine industry funded permeation testing on fill neck exposed only to fuel vapor.144
For the fill neck hose, a three foot section of hose was attached to the top of a five gallon metal
fuel reservoir and configured vertically. The fuel reservoir was filled half-way with gasoline
containing 10 percent ethanol. Approximately every 30 days, this hose/reservoir assembly was
weighed for five days in a row. After the fifth day, the fuel in the reservoir was replaced with
fresh fuel. Testing was performed at 23°C. The only liquid fuel exposure was a weekly
inversion of the assembly for about 1 minute. No attempt was made to simulate fuel slosh that
would be likely be seen in a boat in the water. Also the hose was configured straight up and
down rather than in a more representative configuration as seen on a boat that would include
more horizontal orientation for most of the length of the hose. Repeat testing was performed on
the hose.145 During this repeat testing, permeation was also measured for the same fill neck hose
exposed to liquid fuel.
Four of the fill neck hose constructions were specified as meeting the A2 designation in
SAE J1527. The other two fill neck hose samples were not identified except that they are made
by a hose manufacturer that is known to offer fill neck hose with and without a fluoroelastomer
barrier. Table 5.4-9 presents the test results which show much lower permeation rates for fill
neck hose exposed vapor rather than liquid fuel. Because the end of the hose was not exposed to
atmosphere, and because the hose was situated well above the surface of the liquid fuel in a
vertical fashion, stratification may have occurred in the hose largely due to gravity. This
stratification would be expected to lower the vapor concentration in the hose and therefore lower
permeation.
5-77
-------
Final Regulatory Impact Analysis
Table 5.4-9: Industry Permeation Data on Marine Fill Neck Hose [g/mVday]
Hose manufacturer
#1
#2
Vapor Exposure
4.8,4.8
4.5,4.4
4.7,4.8
4.7,4.7
1.3, 1.1
0.6,6.9
Liquid Exposure
129
114
113
121
5.6
8.5
The marine industry testing was all performed on static test rigs with vertically oriented
hose. No consideration was given to how sloshing the test configuration, as would be seen in a
boat in the water, would have affected the results. For in-use equipment, especially boats in the
water, the fuel is sloshed regularly due to operation or waves. This sloshing may mix up the
vapor in the tank and hose. The industry test program also did not consider how a different hose
configuration (i.e. more horizontally oriented) would have affected the results. Fill neck hose in
boats often runs nearly horizontal from the tank to the edge of the boat, then runs more vertically
near the fill port.
Figure 5.4-1: Hose Test Configurations
Liquid
Horizontal
Vertical
We contracted with an outside test lab to
investigate the effects of fuel slosh and hose
configuration on permeation through marine fill
neck hose.146 All of the testing was performed
on 3 foot sections of 1.5" ID. marine fill neck
hose. Testing was performed in each of the
three configurations shown in Figure 5.4-1. For
each fuel vapor exposure test, the hose was first
preconditioned by subjecting it to liquid fuel for
5 weeks followed by fuel vapor for an
additional 5 weeks. For the liquid fuel exposure
tests, the hose was soaked with liquid fuel for
10 weeks. Fuel soaking was performed at 40°C.
A total of eleven tests were run. For each configuration, testing was performed on three
fuels: Fuel C, CE10, and E10. The liquid fuel exposure tests were performed in the static
position, while the fuel vapor exposure tests were performed with the fuel tanks on a slosh table.
Sloshing was performed at 15 cycles per minute with a deviation of+7° to -7° from level to
simulate movement that might be seen on a boat. An additional two tests were performed to
measure permeation through vapor hose in the vertical and horizontal positions without sloshing.
Permeation was measured similar to the industry testing using weight loss measurements of the
entire test rigs at 23 °C.
The test results from this testing are presented in Table 5.4-10. It was observed that
permeation was much lower for vapor fuel exposure than for liquid fuel exposure. Fuel
5-78
-------
Feasibility of Evaporative Emission Control
permeation was significantly higher for the horizontal hose configuration than for the vertical
hose configuration. This suggests that a large amount vapor stratification was occurring for the
vertical hose, while some fuel vapor was collecting in the horizontal hose. The fuel sloshing
applied in this testing doubled the permeation through the horizontal hose. Regardless of fuel
slosh, no measurable permeation was observed through the vertically oriented hose. Permeation
emissions were observed to be about twice as high on fuel CE10 than on Fuel C or E10.
Table 5.4-10: Effect of Hose Configuration, Vapor Exposure,
and Test Fuel on Marine Fill Neck Hose Permeation at 23°C
Hose Configuration
horizontal
vertical
Vapor Exposure
stationary
sloshed
sloshed
sloshed
stationary
sloshed
sloshed
sloshed
liquid soak
Test Fuel
CE10
CE10
E10
FuelC
CE10
CE10
E10
FuelC
CE10
E10
FuelC
Permeation |g/m2/day]
4.6
9.1
4.6
9.1
0.0
0.0
0.0
0.0
273.7
123.2
123.2
In another study, the effects of liquid fuel versus vapor were studied in which the vapor
hose was not open to atmosphere.147 The fuel hose used for this testing was purchased over the
counter and was labeled as SAE J30 R7. Further investigation of the hose revealed that this
particular grade is made of lower permeation materials than typical Small SI hose constructions.
It was constructed of NBR with a relatively high ACN blend (39 percent) and an ECO cover was
used. This construction was originally intended to allow the hose to be painted with a lacquer-
based paint, then dried in an oven. Although this is a somewhat atypical hose construction, the
test results should still reflect the effects of liquid versus vapor on permeation.
In this testing, all of the fuel hose was preconditioned by soaking in liquid fuel for 5
weeks at about 40°C. This soak was then repeated, except that half of the hose sections were
then exposed only to fuel vapor resulting from attaching the hose to the top of a metal fuel
reservoir. Three fuels were used; California certification gasoline (CARB II), EPA certification
gasoline (gasoline), and EPA gasoline blended with 10 percent ethanol (E10). After the soak
period, the fuel was refreshed and weight loss testing was performed at 23°C. Table 5.4-11
presents the test results. Note that each data point in this table is the average of three hose
samples. In this testing, the end of the hose was plugged and the hose was configured
horizontally. The lower permeation rates for vapor exposure were likely the result a low vapor
concentration in the hose. This low vapor concentration may have been caused because the
diffusion into the long narrow hose may have been the rate limiting effect rather than the
5-79
-------
Final Regulatory Impact Analysis
permeation rate through the hose.
Table 5.4-11: Fuel Hose Permeation with Vapor vs. Liquid Exposure [g/m2/day]
Test Fuel
CARBII
Gasoline
E10
Liquid Exposure
35.8
44.5
80.3
Vapor Exposure
0.3
0.1
0.7
5.4.2 Hose Permeation Reduction Technologies
Materials used in current automotive fuel lines are two to three orders of magnitude less
permeable than nitrile hoses.148 In automotive applications, multilayer plastic tubing, made of
fluoropolymers is generally used. An added benefit of these low permeability lines is that some
fluoropolymers can be made to conduct electricity and therefore can prevent the buildup of static
charges.149 Although this technology can achieve more than an order of magnitude lower
permeation than barrier hoses, it is relatively inflexible and may need to be molded in specific
shapes for each Small SI application. For marine applications, this tubing would not likely meet
the Coast Guard or AB YC durability specifications for fuel and vent hose.
Thermoplastic fuel lines for automotive applications are generally built to SAE J2260
specifications.150 Category 1 fuel lines under this specification have permeation rates of less than
25 g/m2/day at 60°C on CM15 fuel (15 percent methanol). One thermoplastic used in
automotive fuel line construction is polyvinylidene fluoride (PVDF). Based on the data
presented in Appendix 5D, a PDVF fuel line with a typical wall thickness (1 mm) would have a
permeation rate of 0.2 g/m2/day at 23°C on CM15 fuel. However, manufacturers involved in the
boat building industry have commented that this fuel line would not be flexible enough to use in
their applications because they require flexible rubber hose to fit tight radii and to resist
vibration. They also commented that the hose they use must pass the Coast Guard flame
resistance requirements.151452
Recreational vehicle manufacturers are required to use hose that meets a permeation
standard of 15 g/m2/day at 23°C on gasoline blended with 10 percent ethanol (E10). Low
permeation hose constructions that have been identified for these applications could also be used
in Small SI equipment. We believe that the same barrier materials that will be used for
recreational vehicle hose can also be used for marine hose constructions. Marine hose
constructions generally meet the Coast Guard flame resistance requirements either through the
use of a flame-resistant cover, or by increasing the wall thickness. Therefore, the addition of an
inner permeation barrier would not be expected to affect the flame resistance of the hose.
Several low permeation hose constructions are discussed below. Even though most of this data
is on hoses not designed for marine applications, the barrier technology can be used in marine
hose.
5-80
-------
Feasibility of Evaporative Emission Control
We are requiring that fuel and vapor hose meet our standards on E10 fuel for two
reasons. First, ethanol is commonly a component of in-use fuels. Second, for many materials
used in hose constructions, permeation would likely be much higher for fuel containing ethanol.
For instance, a typical barrier material used in barrier hose constructions is FKM. Based on the
data presented in Appendix 5D for FKM, the permeation rate is 3-5 times higher on Fuel CE10
than Fuel C. Therefore, a hose meeting 15 g/m2/day at 23°C on Fuel C may actually permeate at
a level of 40-50 g/m2/day on fuel with a 10 percent ethanol blend.
There are lower permeation fuel hoses available today that are manufactured for
automotive applications. These hoses are generally used either as vapor hoses or as short
sections of fuel line to provide flexibility and absorb vibration. One example of such a hose153 is
labeled by General Motors as "construction 6" which is a multilayer hose with an inner layer of a
fluoroplastic known as THV sandwiched in inner and outer layers of a rubber known as ECO.5
A hose of this construction would have less than 8 g/m2/day at 40°C when tested on CE10.
Permeation data on several low permeation hose designs were provided to EPA by an
automotive fuel hose manufacturer.154 This hose, which is as flexible as non-barrier hose, was
designed for automotive applications and is available today. Table 5.4-12 presents permeation
data on three hose designs that use THV 800 as the barrier layer. The difference in the three
designs is the material used on the inner layer of the hose. This material does not significantly
affect permeation emissions through the hose but can affect leakage at the plug during testing (or
connector in use) and fuel that passes out of the end of the hose which is known as wicking. The
permeation testing was performed using the ARB 18-41°C diurnal cycle using a fuel with a 10
percent ethanol blend (E10).
Table 5.4-12: Hose Permeation Rates with THV 800 Barrier over ARB Cycle (g/m2/day)
Hose Name
CADBAR9610
CADBAR9710
CADBAR9510
Inner Layer
THV
NBR
FKM
Permeation
0.16
0.17
0.16
Wicking
0.00
0.29
0.01
Leaking
0.02
0.01
0.00
Total
0.18
0.47
0.18
The data presented above shows that there is hose available that can easily meet the hose
permeation standard on CE10 fuel. Although hose using THV 800 is available, it is produced for
automobiles that must meet the tighter evaporative emission requirements in the Tier 2
standards. Hose produced in mass quantities today uses THV 500. This hose is less expensive
and could be used to meet the hose permeation requirements. Table 5.4-13 presents information
comparing hose using THV 500 with the hose described above using THV 800 as a barrier
layer.155 In addition, this data shows that permeation rates more than double when tested on
CE10 versus Fuel C.
5 THV = tetrafluoroethylene hexafluoropropylene, ECO = epichlorohydrin/ethylene oxide
5-81
-------
Final Regulatory Impact Analysis
Table 5.4-13: Comparison of Hose Permeation Rates with THV 500 and 800 (g/m2/day)*
Hose Inner
Diameter, mm
6
8
10
THV 500
FuelC
0.5
0.5
0.5
Fuel CE10
1.4
1.4
1.5
THV 800
FuelC
0.2
0.3
0.2
Fuel CE10
0.5
0.5
0.5
Calculated using data from Thwing Albert materials testing (may overstate permeation)
We contracted with an independent testing laboratory to test several samples of SAE J30
R9 hose and a sample each of automotive vent line and fill neck hose for
permeation.156'157'158'159'160'161 The fuel and vapor hoses had a six mm inner diameter. The test lab
used the SAE J30 test procedures for R9 hose with both Fuel C and Fuel CE10. Most of the R9
fuel hose was supplied by recreational vehicle manufacturers who also supplied information on
the materials used in the construction of the hose as well. We purchased one sample of the R9
hose (which was labeled as such) from a local auto parts store without knowing its construction.
Two additional R9 hoses were tested by a fuel hose manufacturer on fuel CE10 after a four week
soak.162 The SAE permeation specification for R9 hose is 15 g/m2/day at 23 °C on Fuel C. The
R9 hose tested all met this limit, even on ethanol blend fuels which typically result in higher
permeation. The automotive vent line showed similar results, but the automotive fill neck
showed much lower permeation. Table 5.4-14 presents the test data on the above hose samples.
5-82
-------
Feasibility of Evaporative Emission Control
Table 5.4-14: Test Results on Commercially Available Hose Samples (g/m2/day)
Hose Sample
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
SAE J30 R9
Automotive vent line
Automotive fill neck
Construction
FKM/ECO
FKM/ECO
FKM/NBR/CM
FKM/ECO
FKM/ECO
PVC/EEC
FKM barrier
fluorine/hydrin
unknown
FKM barrier
FKM barrier
unknown
unknown
FuelC
-
-
-
-
-
-
-
-
10.1
-
-
10.9
0.33
Fuel CE10
7.6
2.1
4.2
10.9
5.2
11.6
6.6
9.0
12.1
4.2
6.7
9.0
0.49
Another hose construction that can be used to meet the marine hose permeation standards
is known as F200 which uses Teflon® as a barrier layer. Teflon® has a permeation rate of 0.03-
0.05 g-mm/m2/day on 15 percent methanol fuel. F200 hose is used today to meet SAE J30 Rl 1
and R12 requirements for automotive applications. Table 5.4-15 presents data on permeation
rates for several F200 constructions.163
Table 5.4-15: F200 Typical Fuel Permeation
Film Thickness [mils]
2
2
2
2
2
1
1
Hose Diameter [in.]
0.375
0.275
0.275
0.470
0.625
0.625
1.5
Fuel
TF-2
TF-2
M25
CE10
CE10
CE10
CE10
g/m2/day @23°C
~
~
0.5
~
~
~
1.5
g/m2/day @40°C
0.7
1
4
3
3
4
~
Low permeability hoses produced today are generally constructed with a barrier material
5-83
-------
Final Regulatory Impact Analysis
layer. There are hoses used in some marine applications with a thermoplastic layer (either nylon
or EVOH) between two rubber layers to control permeation. Because the thermoplastic layer is
very thin, on the order of 0.1 to 0.2 mm, the rubber hose retains its flexibility. Through contract
with two independent labs, we tested three samples of marine barrier hose that were available
prior to our 2002 proposal for marine permeation emissions. This hose included two 3/8"
samples and one 5/8" sample which all used nylon as the permeation barrier. These hose
constructions are used in some sterndrive and inboard applications. Table 5.4-16 presents the
permeation test results at 23°C.164'165'166'167'168'169
Table 5.4-16: Test Results on Available Barrier Marine Hose Samples (g/m2/day)
Hose Description
3/8" marine barrier fuel hose
5/8" marine barrier fuel hose
Lab 1
FuelC
0.80
~
Fuel CE10
5.2
11.6
3.4
Lab 2*
FuelC
0.36
0.76
* average of three tests
Similar testing was performed by the marine industry on commercially available low
permeation marine hose.170 In this testing, the 3/8" ID. fuel hose samples were connected to
metal fuel reservoirs and soaked with gasoline containing 10 percent ethanol at 23 °C for 180
days. The weight of the container/hose assembly was measured for five days in a row
approximately every 30 days. The fuel was replaced with fresh fuel after each series of weight
measurements. The test report did not specify details on the hose constructions. However, based
on the manufacturer part numbers, several of the hoses in this test program were determined to
use a nylon barrier layer. One of the hoses included was a baseline rubber construction meeting
Coast Guard requirements for SD/I fuel hose. Repeat testing was performed on the hose.171
During this repeat testing, permeation was also measured for the same hose exposed to fuel
CE10. Although the permeation rate was generally higher on fuel CE10, the barrier hose
permeation rates were still well below the standard. Table 5.4-17 presents the results of this
testing.
Table 5.4-17: Permeation Results for Commercially Available Marine Barrier Hose
Tested at 23°C with Gasoline Containing 10% Ethanol (g/mVday)
Hose Construction
SAE J1527 Al constructions with nylon barrier
not reported
Gasoline with 10% Ethanol
6.2,5.2
5.6,5.1
4.4,3.8
4.4,3.2
0.4,0.1
CE10
6.1
6.7
10.0
12.1
0.0
5-84
-------
Feasibility of Evaporative Emission Control
After the 2002 proposal for marine permeation emissions, two marine hose manufacturers
developed hose samples using the F200 hose construction. In addition, other hose manufacturers
supplied samples of barrier hose using the F200 hose construction and using THV800 as a
barrier layer. These manufacturers stated that they could make marine hose using the same
barrier construction. We contracted to have these hose samples permeation tested on fuel CE10
at 23°C following a four week soak.172 These test results are presented in Table 5.4-18.
Table 5.4-18: Permeation Test Results on New Marine Barrier Hose Constructions
Application
marine fill neck
marine fuel hose
fuel hose
fuel hose
Barrier Material
Teflon (F200)
Teflon (F200)
Teflon (F200)
THV800
I.D. [inches]
P/2
3/8
1/4
1/4
g/m2/day
0.2
5.0
3.8
5.1
Currently, the Coast Guard requires that fuel pumps on engines be located on or near the
engine to minimize the length of high pressure fuel lines on the vessel. However, at least one
manufacturer sells boats with the high pressure fuel pump in the fuel tank. They received a
waiver from the Coast Guard by using fuel lines that use either a glass fiber or stainless steel
braid cover and quick connect end fittings that are designed to withstand very high pressures
(much higher than would be seen on a boat).173 This particular fuel line construction also uses
Teflon® as a barrier layer. Table 5.4-19 presents permeation test data on this hose.174
Table 5.4-19: Permeation Test Data on Reinforced Fuel Hose
Application
Marine
Outdoor Power
Equipment
I.D. [inches]
0.31
0.25
0.19
0.31
0.25
0.19
Temperature
23°C
60°C
Fuel
CE10
CM15
g/m2/day
0.05
0.08
0.05
0.52
0.93
1.08
Primer bulbs are typically injection-molded out of nitrile rubber. Fuel lines for some
handheld equipment are manufactured in a similar manner. Low permeation primer bulbs and
fuel lines could be manufactured using a similar process by molding them from a
fluoroelastomer such as FKM. Fluoroelastomers, such as FKM, have similar physical properties
as nitrile rubber but are much more fuel-resistant. If the primer bulb or fuel line were molded
out of a FKM with a sufficient flurorine concentration, the permeation rate would be less than
fuel line permeation standard. Alternatively, primer bulbs could be manufactured to meet the
standards by molding a fluoroelastomer inner liner with a nitrile shell to reduce costs. Other
5-85
-------
Final Regulatory Impact Analysis
materials may be applicable as well (see tables of material properties in Appendix 5D).
One manufacturer has developed a low-permeation primer bulb replacement that meets
the hose permeation standard.175476 This design uses a rigid, low permeation housing in the
shape of a traditional primer bulb. However, it uses a simple piston displacement pump inside to
take the place of soft elastomer squeeze type bulb. It is designed to have similar dimensions to
existing primer bulbs and to be easily retrofitted into current applications.
Under their rule for small offroad equipment, California may issue executive orders to
manufacturers with low emission products. As of August, 2006, ARB has issued 24 executive
orders for low permeation fuel lines.177 The California fuel line permeation standard is 15
g/m2/day tested at 40°C on California certification fuel. However, many of the manufacturers
tested their products on CE10 fuel which results in significantly higher permeation rates. Some
manufacturers even tested at 60°C. In all cases, the test results were below the 15 g/m2/day
standard, even under the more challenging test conditions. Table 5.4-20 presents the test results
for the fuel lines with ARB executive orders. Note that the reported emissions are the average of
5-6 test samples.
5-86
-------
Feasibility of Evaporative Emission Control
Table 5.4-20: ARB Fuel Hose Executive Orders for Small Offroad Equipment
EO#
C-U-06-016
C-U-06-001
G-05-016
G-05-017
G-05-019
C-U-05-004
C-U-05-010
G-05-019*
G-05-015a
C-U-05-001
C-U-06-001*
C-U-06-001*
C-U-06-020
C-U-05-014
C-U-06-021
C-U-06-002
C-U-06-011
C-U-05-011
C-U-06-017
C-U-05-013
C-U-05-006
C-U-05-012
C-U-05-003
G-05-018
C-U-05-009
C-U-06-010
C-U-05-002
ID. [mm]
4.8
6.0
6.4
6.4
6.4
6.4
6.4
6.4
7.9
8.0
6.0
6.0
4.5
6.4
6.4
6.4
6.4
2.0
3.5
4.0
4.0
4.0
4.5
4.8
4.8
4.8
6.4
Test Fuel
CE10
CE10
CE10
CE10
CE10
CE10
CE10
CE10
CE10
CE10
CM15
FuelC
Indolene
Indolene
Indolene
Indolene
Indolene
Phase II
Phase II
Phase II
Phase II
Phase II
Phase II
Phase II
Phase II
Phase II
Phase II
Temperature
40
40
40
40
40
40
40
60
60
60
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
g/m2/day
3.75
1.42
4.62
5.97
0.02
12.3
10.6
0.26
11.1
8.22
3.77
0.78
3.20
8.20
7.40
5.00
12.7
4.63
10.8
1.22
10.3
7.33
12.3
0.87
3.94
4.69
3.76
* fuel tube
5.4.3 Low Temperature Hose Materials
In some applications, molded fuel hoses are used rather than simple extruded fuel hose.
These fuel hoses are typically molded out of nitrile rubber (NBR) or a fluoroelastomer such as
FKM. FKM is essentially rubber impregnated with fluorine which results in good fuel
permeation resistance. Manufacturers of handheld equipment that may be used in very cold
weather have stated that they must use nitrile rubber because the FKM material may become
brittle at very low temperatures.178 Examples of such equipment are ice augers and chainsaws.
5-87
-------
Final Regulatory Impact Analysis
Industry has not raised an issue with the capability of using extruded multi-layer hose in
cold temperature applications. This type of hose construction has been demonstrated for low
temperature use in automobiles and snowmobiles. Extruded fuel hose meeting SAE and ASTM
standards is available today which meets a widespread set of safety and durability requirements.
Industry has stated that for some applications, such as chainsaws, that extruded fuel hose will not
work. In these applications, injection molding is used to manufacture complex fuel hose
geometries designed to account for high vibration of the equipment. This vibration generally
results in different motion patterns for the carburetor and fuel tank resulting in variable distances
between the two.
Industry presented information on FKM fuel lines that became brittle and cracked at very
low temperatures.179'180 However, this information was was based on an FKM compound without
a low temperature additive package. There are a wide range of FKM products available on the
market. Many of these fluoroelastomers are designed for use at low temperatures.181'182 For
instance, low temperature o-rings are common in automotive applications.183'184'185 Low
temperature grade FKM products are available with a glass transition temperature as low as -
40°C and a brittleness point as low as -60°C.186 However, low temperature grade FKM products
typically cost several times as much as FKM products intended for less severe temperatures. In
addition, these materials have not been demonstrated for use in molded fuel lines for handheld
applications.
A lower cost option may be to blend a standard fluorosilicone such as FVMQ with a
standard grade FKM. The fluorosilicone brings very low temperature characteristics to the
blend. However, the permeation resistance is not nearly as good as for FKM products. The
blended product would be intended to create a balance between cost, permeation, and low
temperature properties.187 This product is currently used in automotive o-rings. However, it is
not clear if this material could be molded into fuel lines that would meet the appropriate design
criteria for handheld applications.
A new material, called F-TPV, has been developed that is a dynamically vulcanized
combination of fluorothermoplastic resin and fluoroelastomer compound.188 The mix of the two
materials can be varied to trade-off permeation resistance with material hardness. This material
has been shown to have a permeation rate ranging from 3 to 30 g-mm/m2/day on fuel CE10 at
60°C. Rubber hose molded out of even the softest version of this material would be expected to
be capable of achieving a permeation rate well below the standard. In addition, the impact
brittleness temperature is below -50°C for the full range of material blends discussed above.
Finally, the cost of this material is much lower than for low-temperature FKM products. Further
development efforts would be necessary to determine the suitability of this material for fuel lines
on handheld equipment.
Table 5.4-4, above, presents permeation data on several samples of NBR fuel lines used
on handheld equipment today. The permeation rates from these fuel lines range from 165 to 455
g/m2/day with E10 fuel at 23 °C. Later discussions with industry revealed that the NBR hose
with the lower permeation rates had higher acrylonitrile (ACN) contents. Although high ACN
rubber cannot achieve the same low permeation rates as FKM or F-TPV, some permeation
5-88
-------
Feasibility of Evaporative Emission Control
reductions could still be achieved with this material.
5.5 Other Evaporative Emissions
5.5.1 Other Venting Losses
Hot soak emissions occur after the engine is turned off, especially during the resulting
temperature rise. The primary source of hot soak emissions is the evaporation of the fuel left in
the carburetor bowl. Other sources can include increased permeation and evaporation of fuel
from plastic or rubber fuel lines in the engine compartment.
Refueling emissions occur when the fuel vapors are forced out when the tank is filled
with liquid fuel. At a given temperature, refueling emissions are proportional to the volume of
the fuel dispensed into the tank. Every gallon of fuel put into the tank forces out one-gallon of
the mixture of air and fuel vapors. Thus, refueling emissions are highest when the tank is near
empty. Refueling emissions are also affected by the temperature of the fuel vapors and
dispensed fuel. At low dispensed fuel temperatures, the fuel vapor content of the vapor space
that is replaced is lower than it is at higher temperatures because of the cooling effect on the
vapor in the fuel tank.
In automotive applications, the carbon canister is sized not only to capture diurnal
emissions, but refueling, hot soak, and running loss emissions as well. With an engine purge, the
canister would effectively capture running loss emissions and hot soak emissions because the
canister would presumably be nearly empty after a short period of operation. For the canister to
be effective at collecting refueling emissions, it would need to be purged before the refueling
event. However, even without a purged canister, refueling emissions could be minimized by
matching the geometry of the fuel fill opening to the fuel pump nozzle. By minimizing the open
space in the fuel fill opening around the nozzle, less air will be entrained which will minimize
vapor generation during the refueling event. This will not help control the expulsion of vapor
that is displaced by liquid fuel.
5.5.2 Refueling Spitback/Spillage
Installed fuel systems on boats are typically open vented. The exception to this is PWC
which have sealed fuel tanks with pressure relief valves, largely to prevent spillage of fuel during
operation. For larger boats, fuel spillage during operation is less of an issue; however, it is
common for fuel to be lost to the environment during refueling or shortly thereafter.189'190 There
are several mechanisms that lead to fuel loss due to a refueling event. These mechanisms
include restrictions in the fill neck, fuel flowing out the vent line, and expansion of fuel in the
tank.
The American Boat and Yacht Council (AB YC) has a voluntary refueling standard
designed to help prevent fuel from backing up the fill neck during a refueling event.191 This test
requires that no fuel back up the fill neck when a fuel tank in a boat is filled from 25 to 75
percent full at a fill rate of 9 gallons per minute. This test is apparently designed to make sure
5-89
-------
Final Regulatory Impact Analysis
that the fill neck does not have a restriction that may cause fuel to back up the fill neck during
refueling. To prevent fill from backing up the fill neck, fill necks are typically made of large
diameter hose which is reinforced to prevent kinking. In addition, the fuel fill opening is
typically positioned higher than the vent line. This test does not consider fuel overflow that may
occur from filling a marine tank to 100 percent full. In addition, the full rate may be too low to
require a design that would work in typical in-use situations. One survey on 19 marinas saw a
range of 8 to 25 gallons per minute for gasoline fill nozzles with an average of 14 gallons per
minute.192
The most common refueling spillage today is overflow out the vent line. Typically the
vent line is the path of least resistance for fuel overflow. Boats typically do not have a
mechanism that prevents fuel tanks from filling all the way to the top. In fact, the fill and vent
hose are attached to the top of the fuel tank and are often filled with fuel in addition to the tank.
Because the vent hose exits the boat lower than the fill neck opening, the tank can be filled until
fuel begins to exit through the vent hose. In addition, fuel may expand in the fuel tank when
cool fuel is pumped into the fuel tank on a warm day. This expansion can cause additional fuel
overflow out the vent line.
A number of devices have been produced to help control fuel spillage during refueling.
These devices include liquid/vapor separators, combination deck fills and vents, and fuel flow
monitoring systems. A study was performed by Boat US Foundation to evaluate the
effectiveness of several of these systems which are currently available on the market.193 The
results of this study are discussed below.
Liquid/vapor separators are valves that are installed in the fuel line. The typical design is
for the valve to contain a ball that rises when liquid fuel reaches it which closes the vent to liquid
fuel. As the tank fills, fuel backs up the fill neck, allowing the automatic shut-off on the nozzle
to stop the fuel flow. The study found that these systems typically worked best at lower fuel fill
rates and that the larger units were more effective. The effectiveness of the larger units was
probably because they essentially included a reservoir, allowing extra room for fuel expansion.
For the smaller units, the testing consistently showed fuel backing up the fill neck too quickly for
the automatic shut-off valve to engage and fuel spit back out of the deck fill.
In a vented deck fill design, the vent line is routed back to the top of the fill neck. The
intent is that the fuel surging out of the vent line would return to the fill neck and back to the
tank. The study found that the combination vented deck fills significantly reduced
spitback/spillage, but still needed to be used with some caution. One issue was that even when
the fuel came back up and shut off the nozzle, pressure in the fuel tank would cause fuel to
continue to rise in the line and spill onto the deck. Another manufacturer has a similar device
except that a clear section of tubing that redirects the fuel overflow from the vent line to the fill
neck. The operator only attaches this tubing during refueling. Because the tubing is clear, the
operator can see when the fuel is coming out of the vent and can manually slow down or stop the
fuel flow.
Fuel flow monitoring systems are designed to keep track of fuel usage by measuring fuel
5-90
-------
Feasibility of Evaporative Emission Control
flowing to the engine. The study did not present definitive results for the use of flow meters to
accurately refuel the tank without overfill.
Where a carbon canister is used in the vent line for diurnal vapor control, it is important
to include a device to prevent liquid fuel from entering the canister. This device could take the
form of a floating ball valve, limited flow orifice, or other liquid/vapor separation mechanism.
In addition, this device could be positioned in such a way as to prevent the tank from filling all
the way to the top. For instance, the vent fitting could reach down into the fuel tank. Leaving a
vapor space in the fuel tank gives room for fuel in the tank to expand.
In automotive applications, carbon canisters have been used for many years in vehicles
that also meet fuel spit-back standards set by EPA. In typical automotive fuel systems, the fuel
shut-off on the nozzle is tripped before the fuel comes back out the fill neck. It is common to
have a narrow tube parallel to the fill neck reach into the fuel tank at the desired peak fill level of
the tank. The narrow tube connects to the fill neck near the top where the small hole on the
nozzle would be. When fuel splashes on this small hole, the vacuum draw is broken and the
shut-off device is triggered. Fuel travels up the narrow tube more quickly than up the fill neck
and triggers the nozzle shut-off well before fuel spit-back can occur.
At least one company is developing a similar design for use in boats. Testing has been
performed on one system by an independent laboratory that also performs ABYC and Coast
Guard tests for the marine industry. During the testing, a fuel tank was filled 30,000 times, using
this fuel system configuration, without any spillage.194 Also, this fuel system configuration
creates a vapor space in the top of the tank which allows fuel to expand during heating, thereby
preventing fuel spillage due to expansion of the fuel in the tank.195 This system has since been
modified to be adaptable to any fuel tank with a fuel sending unit based on the standard SAE 5-
hole pattern. The updated system was tested using a similar methodology as in the Boat US
study discussed above and underwent 25,000 refueling events at 15 gallons per minute without
experiencing any spills.196 Pictures and video of this system are included in the docket.
5.6 Evaporative Emission Test Procedures
This section discusses test procedures for measuring fuel line permeation, fuel tank
permeation, and diurnal emissions.
5.6.1 Fuel line Permeation Testing
Fuel line permeation must be measured at a temperature of 23 ± 2°C using the weight
loss method specified in SAE J30197 and SAE J1527.198 In this method, one end of a specified
length of hose is connected to a metal reservoir while the other end is plugged. Test fuel is then
added to the reservoir at a volume high enough to ensure that the hose is filled with fuel. Once
care has been taken to ensure that no air bubbles are trapped in the fuel line, the reservoir is
sealed and the entire system is weighed. Permeation is determined by weighing the system every
24 hours and noting the weight loss. After each weighing, the fuel is mixed by inverting the
assembly, then returning it to its original position.
5-91
-------
Final Regulatory Impact Analysis
We are including two modifications to SAE J30 that are consistent with our current
requirements for recreational vehicles and highway motorcycles. First, the test fuel must be
ASTM Fuel C199 (50 percent toluene, 50 percent iso-octane) blended with 10 percent ethanol.6
This fuel is known as CE10 and is commonly used in industry standards and test procedures such
as in SAE recommended practices (including SAE J1527). Section 5.4, and Appendix 5D
presents permeation data for several hose constructions and materials used in hose constructions
on fuels with and without ethanol. As shown in this data, adding ethanol to the test fuel
significantly increases permeation. Standard recommended practice for hose testing uses Fuel C,
or some blend of Fuel C and either ethanol or methanol. This test fuel is generally more
aggressive than standard gasoline. Although hoses are not generally exposed to Fuel C in use,
the level of the standard was based on testing using Fuel C and Fuel C blends. In addition, most
of the test data on low permeation hose presented in this Chapter is based on fuel CE10. For
these reasons, we believe that it is appropriate to allow Fuel CE10 for hose testing.
The second modification is that the hose must be preconditioned by filling the hose with
fuel and soaking long enough to ensure that the permeation rate has stabilized. We are using a
soak period of 8 weeks at 23 ± 5°C. If a longer time period is necessary to achieve a stabilized
permeation rate for a given hose design, we expect the manufacturer to use a longer soak period
(and/or higher temperature) consistent with good engineering judgement. For instance, thick-
walled marine fuel hose may take longer to reach a stable permeation rate than thinner-walled
hose used in Small SI applications. In addition, we are clarifying that the weight loss
measurement period should be two weeks.
Alternatively, for purposes of submission of data at certification, permeation could be
measured using alternative equipment and procedures that provide equivalent results. One
alternative approaches that we anticipate manufacturers may use are the recirculation technique
described in SAE J1737.200 To use other alternative methods, such as enclosure-type testing such
as in 40 CFR part 86, manufacturers have to apply to us and demonstrate equivalence. In
enclosure testing, manufacturers would need to show how they would account for the ethanol
fraction of the permeate.
Recommended practice for automotive fuel tubing is defined in SAE J2260.201 The
permeation requirements in this standard are one to two orders of magnitude lower than those
defined for marine hoses. These permeation requirements are based on the same fuels as the
revised SAE J 1527, but at a much higher temperature (60°C). At 60°C, permeation rates for a
given material may be 16 times as high or higher than at 23 °C based on the rule of thumb that
permeation doubles for every 10°C increase in temperature. SAE J2260 refers to the permeation
test procedures in SAE J1737.202
The procedures in SAE J1737 were designed to measure the low permeation rates needed
in automotive applications to meet EPA evaporative emission requirements. There was concern
that the weight loss measurement, such as used in SAE J1527, was not sensitive enough to
6 An exception to this is that fuel IE10 may be used for cold-weather fuel lines.
5-92
-------
Feasibility of Evaporative Emission Control
measure these low permeation rates. In addition, this procedure requires exposing the material to
be tested for hundreds of hours, depending on the material and fuel, to reach a steady-state
permeation rate. In this procedure, fuel is heated to 60°C and circulated through a tube running
through a glass test cell. Nitrogen around the tube in this test cell is used to carry the permeate
to activated charcoal canisters. The canisters are weighed to determine their capture. Because
the canister is much lighter than the reservoir/hose in the SAE J1527 configuration, a much more
accurate measurement of the permeation loss can be made.
Some manufacturers of low permeability product are finding that as their emission rates
decrease, they need more refined test procedures to accurately measure permeation. These
manufacturers are finding that the weight of the charcoal canisters are much higher than the
permeate being measured. As an alternative to the gravimetric approach used in the above two
procedures, even very low permeation emissions can be measured by a flame ionization detector
and a SHED. As discussed earlier, SHED testing is generally used to measure evaporative
emissions from whole automobile systems as well.
5.6.2 Fuel Tank Permeation Testing
We are applying similar fuel tank permeation test procedures to Small SI equipment and
Marine SI vessels as we currently use for recreational vehicles. This testing includes
preconditioning, durability testing, and permeation measurement. The differences in the test
procedure compared to recreational vehicles are minor and are intended to simplify the testing.
For instance, the durability testing is performed during the preconditioning soak period prior to
the weight loss testing rather than testing the tank twice; once before durability testing, and once
after. Figure 5.6-2 provides flow charts for this testing (2) compared to the recreational vehicle
test (1) which includes the calculation of a deterioration factor.
5-93
-------
Final Regulatory Impact Analysis
Figure 5.6-2: Flow Chart of Fuel Tank Permeation Test with and without a
Deterioration Factor (DF) Determination
1: Full Test Procedure
with DF* Determination
2: Short Test without
DF Determination
baseline
permeation test run
ElOfuel
28 + 2 C
Durability Testing
• Pressure Cvclina
! 10,000 x-0.5 to 2.0 psi
UV Exposure
24W/m2
Slosh Testina
1 million cycles
ElOfuel
Durability Testing
Pressure Cvclina
1 0,000 x -0.5 to 2.0
•••
UV Exposure
24W/m
Slosh Testina
1 million cycles
ElOfuel
psi •
final
permeation test run
ElOfuel
28±2C
use final permeation
test result for
certification
use final permeation
test result for
certification
* The deterioration factor (DF) is the difference
between the baseline and final permeation test
runs in the full test procedure. In future tests, the
first 3 steps would be performed, then a DF could
be applied to determine the final test result.
** The length of "soak" during durability testing
may be included in the fuel soak period provided
that fuel remains in the tank. Soak periods can be
shortened to 10 weeks if performed at 43 ±5 C
5-94
-------
Feasibility of Evaporative Emission Control
For the purpose of this testing, "fuel tank" includes the fuel cap and other components
directly mounted to the tank that become part of the barrier for the fuel and vapor. During
testing, fittings and openings in the fuel tank intended for hose connections (or petcock) is sealed
with an impermeable plug. An opening containing a fuel petcock could also be plugged with an
impermeable fitting because this is an opening to the fuel hose which will be required to meet
permeation standards. In many installed marine fuel tanks, the fuel cap is not directly mounted
on the fuel tank. Instead, the fuel cap is usually linked to the fuel tank by a fill neck hose. In
this case, the fill neck opening in the fuel tank may be sealed with an impermeable plug during
permeation testing.
5.6.2.1 Durability Testing
Prior to the weight loss test, the fuel tank must be preconditioned to ensure that the
hydrocarbon permeation rate has stabilized. Under this step, the fuel tank must be filled with a
10 percent ethanol blend (E10), sealed, and soaked for 20 weeks at a temperature of 28 °C ± 5
°C. Once the permeation rate has stabilized, the fuel tank is drained and refilled with E10,
sealed, and tested for a baseline permeation rate. The permeation rate from the fuel tank is
determined by measuring the weight difference the fuel tank before and after soaking at a
temperature of 28 °C ± 2 °C over a period of at least 20 weeks. The soak periods could be
shortened to 10 weeks if performed at 43 °C ± 5 °C. The durability testing described below may
be performed during the soak period. During the slosh testing, a lower tank fill level, consistent
with the slosh test, is acceptable.
To determine a permeation emission deterioration factor, we established three durability
tests: slosh testing, pressure-vacuum cycling, and ultra-violet (UV) light exposure. The purpose
of these deterioration tests is to help ensure that the technology is durable and the measured
emissions are representative of in-use permeation rates. For slosh testing, the fuel tank is filled
to 40 percent capacity with E10 fuel and rocked for 1 million cycles. The pressure-vacuum
testing contains 10,000 cycles from -0.5 to 2.0 psi. The slosh testing is designed to assess
treatment durability as discussed above. These tests are designed to assess surface
microcracking concerns. These two durability tests are based on a draft recommended SAE
practice.203 The third durability test is intended to assess potential impacts of UV sunlight (0.2
|im - 0.4 |im) on the durability of the surface treatment. In this test, the tank must be exposed to
a UV light of at least 0.40 W-hr/m2 /min on the tank surface for 15 hours per day for 30 days.
Alternatively, it can be exposed to direct natural sunlight for an equivalent period of time in
exposure hours.
The order of the durability tests is optional. However, we require that the fuel tank be
soaked to ensure that the permeation rate is stabilized just prior to the weight loss test. If the
slosh test is run last, the length of the slosh test may be considered as part of this soak period.
Where possible, the deterioration tests may be run concurrently. For example, the fuel tank
could be exposed to UV light during the slosh test. In addition, if a durability test can clearly be
shown to not be appropriate for a given product, manufacturers may petition to have this test
waived. Only fuel tanks using surface barrier treatments or barriers consisting of post-
processing coatings are subject to the slosh and pressure-vacuum tests. In addition, a fuel tank
5-95
-------
Final Regulatory Impact Analysis
that is only used in vehicles where an outer shell prevents the tank from being exposed to
sunlight may not be subject UV testing.
After the durability testing, once the permeation rate has stabilized, the fuel tank is
drained and refilled with fresh fuel, sealed, and tested for a final permeation rate. The final
permeation rate from the fuel tank is determined using the same measurement method as for the
baseline permeation rate. The final permeation rate is used for the emission rate from this fuel
tank. The difference between the baseline and final permeation rates could be used to determine
a deterioration factor for use on subsequent testing of similar fuel tanks.
5.6.2.2 Test Fuel
As discussed in Chapter 3, about much of the fuel sold in the U.S. contains ethanol and
this percentage is expected to increase dramatically in 2012 and later. The fuel tank permeation
test fuel is E10, which is a blend of 90 percent certification gasoline (as specified in 40 CFR
1065.210) blended with 10 percent ethanol for permeation testing of fuel tanks. As an
alternative, we would allow testing on ASTM Fuel C blended with 10 percent ethanol (Fuel
CE10). Fuel CE10 is commonly used in industry standards and test procedures such as in SAE
recommended practices.
5.6.2.2.1 Effect of ethanol on fuel tank permeation
Most plastic nonroad fuel tanks today are made out of high-density polyethylene (HDPE)
or cross-link polyethylene (XLPE). For Small SI and Marine SI markets, plastic is much more
widely used than metal for fuel tank constructions. For HDPE, E10 fuel has little effect on
permeation emissions and may even result in slightly lower emissions according to one study.204
We tested three 0.5 gallon Small SI fuel tanks for permeation using both certification gasoline
and E10 and found a slight increase in permeation due to ethanol. ARB also tested several Small
SI fuel tanks on both gasoline and ethanol blends 205'206>207>208 ancj saw a small increase in
permeation. Permeation data was collected on two XLPE marine fuel tanks on E10. The
measured permeation rates were within the range of data from other XLPE marine fuel tanks
tested on gasoline presented earlier in Table 5.3-1. This data is presented in Table 5.6-1.
5-96
-------
Feasibility of Evaporative Emission Control
Table 5.6-1: Effect of Ethanol on Permeation for HDPE Fuel Tanks
Material
HDPE
HDPE
HDPE
XLPE
Test Equipment
material sample
Small SI fuel
tanks
(EPA Testing)
Small SI fuel
tanks
(ARE Testing)
marine tanks
(EPA testing)
Tank
gallons
NA
0.5
0.5
0.5
0.25
0.25
0.25
0.25
0.25
0.5
3.9
12
12
Test
Temp(s)
40°C
29°C
18-41T
29°C
gasoline
[g/m2/day]
90a
11.5
11.4
11.2
11.6
10.7
12.5
9.9
9.2
12.7
4.8
__b
E10
[g/m2/day]
69a
13.9
13.7
14.4
13.6
11.6
11.4
10.3
10.3
14.8
5.0
7.5
8.5
Increase in
Permeation
-23%
21%
21%
28%
18%
7%
-9%
4%
12%
17%
4%
minimal
3 ASTM Fuel C was used as gasoline (50% toluene, 50% isooctane). Units are per mm of thickness
b See Table 5.3-1 for data on similar tanks tested on gasoline.
Although E10 does not have a large effect on permeation through polyethylene, it does
have a large effect on most other materials used in fuel systems, especially those designed for
low permeation. This is supported by the data presented in Appendix 5D of permeation rates for
several fuel system materials on fuel C, CE10, and CIS. In addition, ethanol is commonly
blended into fuels in-use and alcohol fuels may be used more in the future in an effort to use
alternative energy sources. Therefore, we are requiring El0 as a test fuel to ensure that the
permeation standard will be met on in-use fuels.
One study found that permeation from automotive fuel systems increased significantly
when gasoline containing ethanol was used compared to gasoline without ethanol.209 In this
case the ethanol fuel was specifically blended to achieve two weight percent oxygen. This test
fuel represents California reformulated fuel and contains 5.7 percent by volume ethanol. Table
5.6-2 presents the test results at 29°C. The average increase in permeation due to using E5.7 was
60 percent. Presumably, this effect would have been higher on E10. Because most of the fuel
tanks are metal, the effect is largely due to fuel hose/tubing permeation.
5-97
-------
Final Regulatory Impact Analysis
Table 5.6-2: Effect of Ethanol on Permeation from Automotive Fuel Systems
Fuel System
2001 Toyota Tacoma
2000 Honda Odyssey
1999 Toyota Corolla
1997 Chrysler Town & Country
1995 Ford Ranger
1993 Chevrolet Caprice Classic
1991 Honda Accord LX
1989 Ford Taurus GL
1985 Nissan Sentra
1978 Olds Cutlass Supreme
Fuel Tank
Metal
Plastic (enhanced evap)
Metal
Plastic (enhanced evap)
HOPE
Fluorinated HOPE
Metal
Metal
Metal
Metal
Gasoline
10
19
11
40
348
94
39
28
73
73
E5.7
32
53
57
66
342
137
100
73
177
139
Increase
220%
179%
418%
65%
-2%
46%
156%
161%
142%
90%
One significant finding with the above study was that switching from one fuel to another
affects the permeation rate within a few weeks. Although operating on gasoline with ethanol
changes the fuel tank material in such a way that permeation increases, this effect is reversible
when gasoline is used in the fuel tank for a long enough period of time. This study found that
the permeation rate at 40°C typically approached a stabilized level within 1 to 2 weeks of
switching from one fuel to another.
To investigate the potential effects of fuel switching, we tested two pairs of 6.6 gallon
portable marine fuel tanks. These fuel tanks used the barrier platelet technology discussed
above. The first pair used nylon as a barrier material which is highly sensitive to ethanol while
the second pair used EVOH which is much less sensitive to ethanol. All four tanks were soaked
on E10 fuel, then the fuel was drained and replaced for testing. For each pair, one tank was
tested on EPA certification gasoline and the other was tested on E10 fuel (10 percent ethanol, 90
percent gasoline). We continued the test for more than six weeks to observe the effects of fuel
switching on the permeation rates. The results suggest that switching to gasoline significantly
reduces the permeation rate for the nylon barrier tanks, but has no significant effect on the fuel
tanks using EVOH as a barrier. Note that the nylon tanks had permeation rates near the
standards when soaked and tested on gasoline, but have much higher permeation rates when
tested on E10. This data is presented in Figure 5.6-1. The R-squared values for linear fits to the
data are also presented. The fuel tank with a nylon barrier that experienced fuel switching had a
lower R-squared value than the other fuel tanks.
5-98
-------
Feasibility of Evaporative Emission Control
Figure 5.6-1: Effect of Fuel Switching on Permeation from Barrier Platelet Fuel Tanks
Nylon, Gasoline
Nylon, E10
EVOH, Gasoline
EVOH, E10
Fuel tank permeation data on both gasoline and E10 fuel are presented earlier in this
chapter for nylon handheld tanks, fluorinated and sulfonated Small SI tanks, portable tanks with
non-continuous nylon barrier platelets, and rotationally molded tanks with a nylon inner barrier.
This data is repeated here in Table 5.6-3 to better focus on the effect of ethanol on fuel tank
permeation. As shown by this data and the previous discussion, ethanol in the test fuel tends to
increase permeation. However, the effect of ethanol on permeation appears to be highly variable
depending on the materials or surface treatments used in constructing the fuel tank.
5-99
-------
Final Regulatory Impact Analysis
Table 5.6-3: Permeation Rates on Gasoline and E10 for Barrier Fuel Tanks
Permeation Control
nylon 6
nylon 6, 33% glass
nylon 6, 30% glass
nylon 6, 30% glass
fluorination
sulfonation
non-continuous nylon platelets
Rotomolded with PA1 1 liner*
Capacity
[gallons]
0.24
0.05
0.06
0.06
0.5
0.5
2.0
1mm barrier
thickness
Gasoline
[g/m2/day]
0.34
0.62
1.45
1.30
0.56
0.62
0.22
2.5
2.7
2.2
3.7
0.17
0.24
0.12
E10
[g/m2/day]
0.42
0.48
1.01
1.12
0.93
2.2
2.5
1.4
2.1
0.43
0.62
0.62
3.9
4.2
2.9
6.8
0.91
0.72
0.78
0.81
% Increase
32%
65%
60%
37%
19%
49%
84%
350%
based on testing for California (California Phase II gasoline and fuel CE10)
5.6.2.2.2 Effect ofCElO versus E10 on fuel tank permeation
As discussed above, we will allow the use of fuel CE10 as an alternative to E10 for fuel
tank permeation testing. The primary fuel, E10 is representative of in-use fuel and is consistent
with the certification fuel used for recreational vehicles. However, fuel CE10 is widely used by
industry for materials testing. Data presented earlier in this chapter suggests that permeation is
generally significantly higher on fuel CE10 for fuel hoses. We were therefore interested in the
effect of fuel CE10 versus E10 on fuel tank permeation. We tested several fuel tanks and found
that permeation was only slightly higher on CE10 than E10 for most of the fuel tanks tested.
To study the effects of CE10 versus E10 on permeation, we used fuel tanks that had been
previously tested on fuel E10. All of these tanks were drained and refueled with fresh test fuel.
Most of the tanks were filled with fuel CE10; however, with some exceptions, one of each tank
type was filled with fresh E10 for comparison. These fuel tanks were then preconditioned by
soaking them for 12 weeks with the new test fuel. Note that all of the test tanks had been
5-100
-------
Feasibility of Evaporative Emission Control
soaking with E10 fuel for more than a year (and in some cases multiple years) prior to beginning
this preconditioning soak. Following the soak period, each tank was drained, refilled with fresh
fuel, and sealed. Permeation was measured over two weeks at 29°C. The fuel tanks were
weighed on each weekday during this period.
Table 5.6-4 presents the results of this testing. In most cases, emissions were only
slightly higher on CE10 than E10. The exceptions were the nylon 6 and the acetal copolymer
fuel tanks which showed much higher permeation on CE10. However, the permeation rates for
these fuel tanks were still below the standard when tested on fuel CE10. The fuel tank with a
continuous EVOH barrier was well below the standard on fuel CE10. No comparison was made
to E10 results for this technology.
Table 5.6-4: Permeation Rates on Gasoline and E10 for Barrier Fuel Tanks
Permeation Control
nylon 6
HOPE
fluorination
sulfonation
non-continuous platelets (4% nylon)
non-continuous platelets (2% EVOH)
non-continuous platelets (4% EVOH)
non-continuous platelets (6% EVOH)
continuous EVOH barrier
acetal copolymer
Capacity
[gallons]
0.24
0.5
0.5
0.5
6.6
6.6
6.6
6.6
5.6
0.8
E10
[g/m2/day]
0.69
12.5
0.41
3.1
4.5
3.0*
2.2
1.3
~
0.25
CE10
[g/m2/day]
1.4
1.2
13.3
13.5
0.49
0.52
4.2
2.9
5.3
3.3
2.3
1.4
0.05
0.01
0.55
0.65
% Increase
90%
7%
21%
16%
16%
10%
6%
6%
NA
140%
based on previous testing (presented earlier in this chapter)
5.6.2.3 Reference Tank
In cases where the permeation of a fuel tank is low, and the sample tank is properly
sealed, the effect of air buoyancy can have a significant effect the measured weight loss. Air
buoyancy refers to the effect on air density on the perceived weight of an object. As air density
5-101
-------
Final Regulatory Impact Analysis
increases, it will provide an upward thrust on the fuel tank and create the appearance of a lighter
tank. Air density can be determined by measuring relative humidity, air temperature, and air
pressure.210
One testing laboratory presented data to EPA on their experience with variability in
weight loss measurements when performing permeation testing on portable fuel tanks.211 They
found that the variation was due to air buoyancy effects. By applying correction factors for air
buoyancy, they were able to greatly remove the variation in the test data. A technical brief on
the calculations they used is available in the docket.212
A more direct approach to accounting for the effects of air buoyancy is to use a reference
fuel tank. In this approach, an identical fuel tank to that being tested for permeation is tested
without fuel in it and used as a reference fuel tank. Dry sand can be added to this tank to make
up the difference in mass associated with the test tank being full of fuel. The reference tank is
then sealed so that the buoyancy effect on the reference tank is the same as the test tank. The
measured weight loss of the test tank can then be corrected by any measured changes in weight
in the reference tank. The California Air Resources Board uses this approach for measuring
portable fuel tank emissions, and they refer to the reference tank as a "trip blank."213
5.6.2.4 Engineering Design-Based Certification
A metal fuel tank automatically meets the design criteria for a design-based certification
as a low-permeation fuel tank, subject to the restrictions on fuel caps and seals described below.
There is also a body of existing test data showing that co-extruded fuel tanks from automotive
applications have permeation rates that are well below the new standard. We are allowing
design-based certification for co-extruded high-density polyethylene fuel tanks with a continuous
ethylene vinyl alcohol (EVOH) barrier layer. The EVOH barrier layer is required to be at least 2
percent of the wall thickness of the fuel tank. In addition, the ethylene content of the EVOH can
be no higher than 40 mole percent.
To address the permeability of the gaskets, and seals used on metal and co-extruded
tanks, the design criteria include a specification that seals (e.g. gaskets and o-rings) not made of
low-permeation materials must have a total exposed surface area less than 0.25 percent of the
total inside surface area of the fuel tank. For example, consider a four gallon fuel tank with an
inside surface area of 0.40 square meters. The total exposed surface area of these seals on the
fuel tank, in this example, must be smaller than 1000 mm2 (= 0.25%/100 x 0.40m2 x 1,000,000
mm2/m2). This is consistent with the proposed rule and the current requirements for recreational
vehicles, but allows for larger seals for larger tanks. In addition, if a non-metal fuel cap is
directly mounted to the fuel tank, the surface area of the fuel cap (determined by the
cross-sectional area of the fill opening) may not exceed 3.0 percent of the total inside surface
area of the fuel tank.
A metal or co-extruded fuel tank with a fuel cap and seals that meet these design criteria
would be expected to reliably pass the standard. However, we believe it is not appropriate to
assign an emission level to fuel tanks using a design-based certification option that will allow
5-102
-------
Feasibility of Evaporative Emission Control
them to generate emission credits. Given the uncertainty of emission rates from the seals and
gaskets, we will not consider these tanks to be any more effective than other fuel tanks meeting
emission standards.
In the case where the fuel cap is directly mounted on the fuel tank, we consider the cap
and associated seals to be part of the fuel tank. As discussed above, we allow the fuel cap to be
tested either mounted on the fuel tank, or individually. As an alternative to testing the fuel cap,
the manufacturer may opt to use a default permeation rate of 30 g/m2/day. To be eligible for this
default rate, the seal on the fuel cap must be made of a low-permeation material, such as a
fluoroelastomer. The surface area associated with this default value is the cross sectional area of
the opening that is sealed by the fuel cap. If this default value were used, the fuel fill would be
sealed with a non-permeable plug during the tank permeation test, and the default permeation
rate would be factored into the final result.
For the purposes of this provision, we consider low-permeation materials to be those that
haves a permeation rate not more than 10 g-mm/m2/day at 23 °C on CE10 fuel as tested under the
procedures specified in SAE J2659.214
5.6.3 Diurnal Emission Testing
The test procedure for diurnal emissions is to place the fuel tank in a SHED7, vary the
temperature over a prescribed profile, and measure the hydrocarbons escaping from the fuel tank.
The final result is reported in grams per gallon where the grams are the mass of hydrocarbons
escaping from the fuel tank over 24 hours and the gallons are the nominal fuel tank capacity.
The test procedure is based on the automotive evaporative emission test described in 40 CFR
part 86, subpart B, with modifications specific to marine applications.
5.6.3.1 Temperature Profile
For installed marine fuel tanks, we believe that the fuel temperature profile observed in
the tank has a lower variation in temperature due to the inherent insulation provided by the boat
hull. Data discussed earlier in this chapter, and presented in Appendix 5A, suggest that the fuel
temperature in an installed marine tank sees a change in temperature less than that of ambient
air. Based on this data, the fuel temperature change in boats stored on trailers would be expected
to be about half of ambient. For boats stored in the water, the fuel temperature change would be
expected to be about 20 percent of ambient. Based on discussions with industry, we use a boat
length as a surrogate for determining if a boat is a trailer boat. We consider a boat that is at or
below 8.5 feet in width and below 26 feet in length as a trailer boat and larger boats as being
primarily stored in the water.
To account for the differences between ambient and fuel temperature, we established a
7 Sealed Housing for Emission Determination
5-103
-------
Final Regulatory Impact Analysis
test temperature profile of 78-90°F (25.6-32.2°C) for marine fuel tanks installed in boats less
than 26 feet in length and at or below 8.5 feet in width. For larger boats, we established a test
temperature profile of 81.6-86.4°F (27.6-30.2°C). These test temperature profiles are based on
fuel rather than ambient temperature. Figure 5.6-3 presents the three temperature profiles over
24 hours. Numerical values are presented in Appendix 5E.
Figure 5.6-3: Diurnal Temperature Profiles
• portable (air temp)
installed <26ft (fuel temp)
installed >26ft (fuel temp)
8 12 16
Elapsed Time [hours]
20
24
The automotive diurnal test procedure includes a three day temperature cycle. The
purpose of this test length is to ensure that the carbon canister can hold at least three days of
diurnal emissions without vapor breaking-through the canister. For vessels using carbon
canisters as an evaporative emission control strategy, we established a multiple day cycle here as
well so that the passive purging can be observed. In the automotive test, the canister is loaded,
then purged during an engine test prior to the first day of testing. Because we are anticipating
canisters on marine applications to be passively purged we are using a different approach. Prior
to the first day of testing, the canister is loaded to full working capacity, then run over the
diurnal test temperature cycle to allow one day of passive purging. The test result is then based
on the highest recorded value in the following three days.
For fuel systems using a sealed system (or sealed-system with pressure relief), we do not
believe that a three day test is necessary. Prior to the first day of testing, the fuel is stabilized at
the initial test temperature. Following this stabilization, the SHED is purged and a single diurnal
temperature cycle run. Because this technology does not depend on purging or storage capacity
of a canister, multiple days of testing should not be necessary. Therefore, we established a one-
day test for the following technologies: sealed system without pressure relief, sealed system with
5-104
-------
Feasibility of Evaporative Emission Control
a pressure relief valve, sealed bladder fuel tanks, sealed fuel tanks with a volume compensating
air bag.
5.6.3.2 Test Fuel
Consistent with the automotive test procedures, the test must take place using
certification gasoline with a vapor pressure of 9.0 RVP. We do not require ethanol to be blended
into the test fuel. Although ethanol has a significant effect on permeation, it would not be
expected to affect diurnal emissions except in that it may affect fuel vapor pressure.
Diurnal emissions are not only a function of temperature and fuel volatility, but of the
size of the vapor space in the fuel tank as well. Consistent with the automotive procedures, the
fill level at the start of the test must be 40 percent of the nominal capacity of the fuel tank.
Nominal capacity of the fuel tank is defined as the volume of fuel, specified by the manufacturer,
to which the fuel tank can be filled when sitting in its intended position. The vapor space that
normally occurs in a fuel tank, even when "full," is not considered in the nominal capacity of the
fuel tank.
5.6.3.3 Tank Configuration
Installed marine fuel tanks are typically equipped with a vent line. As shown above, this
vent line can impact the emissions determined over the test procedure because it largely restricts
diffusion losses. Therefore, open vent marine fuel tanks that are designed with a connection for
a vent line must be equipped with a one meter fuel line to more accurately reflect real world
emissions. This should only be necessary for baseline configurations.
The majority of marine fuel tanks are made of plastic. Even plastic fuel tanks designed to
meet our standards are expected to have some amount of permeation. However, over the length
of the diurnal test, if it were performed on a new tank that had not been previously exposed to
fuel, the effect of permeation on the test results should be insignificant. For fuel tanks that have
reached their stabilized permeation rate (such as testing on in-use tanks), we believe that it is
appropriate to correct for permeation. In such a case, the permeation rate could be measured
from the fuel tank and subtracted from the final diurnal test result. The fuel tank permeation rate
has to be stabilized on the 9 RVP test fuel used for the diurnal test and measured either over the
diurnal temperature cycle or at a constant temperature (28 ± 2°C). This test measurement is
made just prior (within 24 hours) to the diurnal emission test to ensure that the permeation rate
does not change prior to the diurnal test. In addition, the test fuel needs to remain in the fuel
tank between the permeation and diurnal tests to ensure a stable permeation rate. The fuel tank
could be emptied to change test fuels and test set ups; however, this period will not be allowed to
exceed one hour. As an alternative to stabilizing the permeation rate prior to testing, the
permeation could be measured immediately before and after the diurnal test, and the lower
permeation rate used to correct the diurnal test results. In this case, the test fuel is not removed
after the diurnal test, and the second permeation test begins within 8 hours of the end of the
diurnal test.
5-105
-------
Final Regulatory Impact Analysis
5.6.3.4 Carbon Canister Engineering Design
Design-based certification may be used as an option to performing the above test. For
vessels using a carbon canister to control diurnal emissions, it is important to ensure that the
canister design is sufficient to achieve the standards. The following discussion outlines the
requirements that are necessary to ensure adequate canister design. These design parameters and
their associated test procedures are largely based on our understanding of current industry
practices for marine grade carbon.215
5.6.3.4.1 Carbon canister capacity
In a passive purge system, the storage capacity of the carbon canister must be properly
matched to the fuel system. Ideally, the canister is large enough to take full advantage of the
passive purge caused by cooling of the fuel tank. By creating more open sites in the canister,
greater vapor collection is possible during the next heating event. If a canister is undersized,
then the vessel would not likely meet the standards. On the other hand, after a certain point,
increasing the size of the canister offers little additional emission control. Once the system
reaches a stabilized purge/load condition, the emission reduction potential is based on the
portion of the canister that purges and loads rather than the full volume of the canister.
The storage capacity of a carbon canister is based both on the volume of the canister and
the working capacity of the carbon. Butane working capacity (BWC) is a measure of the vapor
storage capacity of the carbon and is expressed in units of mass of butane per unit of volume.
The BWC of the carbon must be at least 9 g/dL based on the test procedures specified in ASTM
D5228-92.216 Under this test procedure, butane vapor is fed through a carbon sample at a
specified rate, until the mass of the carbon sample reaches equilibrium. The butane is then
purged off with dry air. BWC of the carbon sample is calculated from the difference in the
measured mass of the carbon sample before and after the purge.
Using the ASTM test procedure, the BWC represents the full saturated capacity of the
canister and not the amount of vapor that the canister will hold before breakthrough occurs.
Under the EPA automotive test procedure in 40 CFR 86.134-96, the canister capacity is based on
the amount of butane loaded in the canister until 2 grams of breakthrough is measured.
However, the ASTM procedure gives a repeatable measure that is currently used by industry.
The design standard of 9 g/dL is based on this test procedure and therefore accounts for the
differences in the ASTM and existing EPA automotive procedure.
Based on the data presented earlier in this chapter, we are requiring that the volume of
the carbon canister must be a minimum of 0.04 liters of carbon per gallon of fuel tank capacity
for fuel tanks installed in trailerable boats (<26 feet in length and <8.5 feet in width). For larger
boats, the fuel temperature may be less affected by diurnal temperature swings for two reasons.
First, these fuel tanks are in larger vessels which are more likely to be stored in the water and
therefore, subject to smaller temperature fluctuations. Second, these fuel tanks are generally
larger and have larger thermal inertia in the fuel which may lead to lower temperature
fluctuation. Therefore, for fuel tanks installed on non-trailerable boats (>26 feet in length or
5-106
-------
Feasibility of Evaporative Emission Control
>8.5 feet in width), the design minimum volume is 0.016 liters of carbon every gallon per gallon
of fuel tank capacity.
5.6.3.4.2 Carbon humidity resistance
In a marine environment, the carbon may be exposed to more humid air, on average, than
in land-based applications such as cars and trucks. Traditional carbons used in automotive
applications can adsorb water, thereby closing sites off to hydrocarbons. With active purge and
carbon heating during refueling vapor collection, the water vapor is easily purged off the carbon.
Under this rule, we are basing the design specification on a passive purge canister design and are
not requiring onboard refueling vapor recovery. Therefore, we believe that the carbon should be
resistant to moisture in the air. In the in-use program discussed above, marine grade carbon was
used that was developed specifically for high humidity applications.217
Design-based certification requirements for humidity resistance are based on the
specifications of the humidity-resistant carbon used in the in-use demonstration program. This
carbon meets a moisture adsorption capacity maximum of 0.5 grams of water per gram of carbon
at 90 percent relative humidity and a temperature of 25±5°C. This limit is based on a test
procedure where dried carbon is exposed to water vapor and the pressure in the sample chamber
is controlled to achieve the correct partial pressure of the water to achieve the desired relative
humidity. The adsorption of water in the carbon is calculated based on the reduction in pressure
in the sample chamber. More detail on this test procedure is available in the docket.218
5.6.3.4.3 Carbon durability
Another issue that has been raised with regard to canister use in marine applications is
the durability of the canister under the shocks that can be observed on a marine vessel.
Automotive applications see shocks and vibration as well and the carbon is protected by packing
it under pressure in the canister. To address the concern of carbon durability, however, we are
including a carbon strength requirement. This strength requirement is consistent with the
specifications for the carbon used in the in-use test program described above, which was
designed to have a higher hardness value and lower dust attrition rates than typical automotive
carbons.
The industry procedure for carbon pellet strength is to determine the average pellet size
in a sample of carbon before and after a pan hardness test. Pellet size is determined by
separating the carbon by size using sieves. The pan hardness test involves shaking the carbon in
a pan with steel balls over a fixed period of time. The pellet strength is determined by taking the
ratio of the average pellet size of the carbon before and after the pan and ball attrition test. Pellet
strength must be at least 85 percent. The test procedure is ASTM D3802-79219 with two
variations. First, as discussed above, hardness is defined as the ratio of mean particle diameter
before and after the attrition test. Second, the attrition test uses twenty 1/2M steel balls and ten 3A"
steel balls rather than fifteen of each as specified in ASTM D3802-79. These variations on the
ASTM procedure reflect common industry practice for pelletized carbons in contrast to the
original test procedure which were intended for granular carbons.220
5-107
-------
Final Regulatory Impact Analysis
5.6.3.4.4 Canister design
The design of the canister itself is important in building an effective and durable carbon
canister system. The canister should be made of a material that is compatible with the
application. For instance, the material should be fuel resistant and durable. Where a flame test
is required by the Coast Guard, the material should be able to pass this test on its own or with a
protective cover. In addition, the canister material must have good structural integrity at
temperatures that it would be exposed to in a boat. If the material changes in dimension at
temperature, that flexing may loosen the carbon packing, allowing the carbon to move and
eventually deteriorate. The canister should be installed in the boat in such a way that undue
stress is not placed on the canister. It should also be properly constructed so that there are no
leaks in the canister.
The canister must be packed in such a way that the carbon does not move inside the
canister in-use. If the carbon were able to move, it would eventually break down under
vibration. Over time the carbon could deteriorate into dust which could eventually escape from
the canister. This is not an issue with a carbon canister that uses a properly designed and
installed volume compensator. The basic design of a volume compensator is that compression is
held on the carbon bed with a spring. A mesh or foam cover is used on the volume compensator
that will allow air to pass through, but will hold the carbon pellets in place.
The carbon should be packed into the canister in such a way that there is a consistent size
of carbon pellets throughout the canister. If the carbon settles in the storage hopper, it would be
possible for some canisters to be filled largely with the smallest diameter carbon pellets (or dust)
which would increase the pressure restriction of the canister. Also, if the carbon is not packed
properly when placed into the canister, it could later settle leading to a volume reduction of the
carbon that is too large for the volume compensator to address.
The carbon canister design must allow for a proper flow path of vapor and air through the
carbon bed. In current carbon canister designs, an air gap is typically installed upstream of the
carbon bed. Flow directors may be molded into this air gap. The purpose of the air gap is to
allow the vapor or purge air to disperse and flow through the entire carbon bed. Even with a
small air gap, the vapor will disperse because it will attempt to follow the path of least resistance
through the canister. Without the air gap, the flow could be predominately in the center of the
carbon (or wherever the intake hose connection is located). In addition, to prevent flow
restriction, the carbon granules must have a minimum mean diameter of 3.1 mm based on the
procedures in ASTM D2862.221
The geometry of a carbon canister can affect the effectiveness of the control system. For
instance, a long, narrow canister will have higher efficiency than a short wide canister. This is
because some breakthrough can occur if the pathway is too short for the flow of vapor. Based on
one study, the effectiveness of the carbon canister increases notably until a length to diameter
ratio of about 3.5 is achieved.222 At higher ratios, less of an impact on efficiency was observed.
At too high of a length to diameter ratio, significant back pressure may occur in the system.
5-108
-------
Feasibility of Evaporative Emission Control
5.6.3.4.5 Integration with Fuel System
It is important that a carbon canister system be appropriately integrated into the fuel
system. For instance, the canister must be positioned in the vent line, and potentially a liquid
separation valve added, to ensure that liquid fuel does not reach the canister during refueling.
We also expect the fuel system design to minimize spit-back out of the fill neck during refueling.
A design that caused fuel to stream out the fill neck during refueling, even with a fuel nozzle
shut-off mechanism, is not acceptable.
5.7 Impacts on Noise, Energy, and Safety
The Clean Air Act requires EPA to consider potential impacts on noise, energy, and
safety when establishing the feasibility of new emission standards for marine vessels.
5.7.1 Noise
In this case, we do not expect evaporative emission controls to have any impact on noise
from Small SI equipment or marine vessels because noise from the affected parts of the fuel
system is insignificant.
5.7.2 Energy
We anticipate that the evaporative emission standards will have a positive impact on
energy. By capturing or preventing the loss of fuel through evaporation, we estimate that the
lifetime average fuel savings will be about 1.4 gallons for an average piece of Small SI
equipment and 28 gallons for an average boat. This translates to a fuel savings of about 45
million gallons for Small SI equipment and 26 million gallons for Marine SI vessels in 2030
when most of the affected equipment used in the U.S. is expected to have evaporative emission
control.
5.7.3 Safety
As part of the development of this rule, EPA performed a technical study on the safety of
emission control technology for Small SI equipment and Marine SI vessels.223 The conclusions
of this study are presented below. Although the study focuses on equipment with engines less
than 37 kilowatts, the conclusions drawn for marine apply to boats with larger engines as well as
ABYC, USCG, UL, and SAE requirements do not distinguish between engine sizes.
EPA has reviewed the fuel hose and fuel tank characteristics for NHH and HH equipment
and evaluated control technology which could be used to reduce evaporative emissions from
these two subcategories. This technology is capable of achieving reductions in fuel tank and fuel
hose permeation without an adverse incremental impact on safety. For fuel hoses and fuel tanks,
the applicable consensus standards, manufacturer specific test procedures and EPA requirements
are sufficient to ensure that there will be no increase in the types of fuel leaks that lead to fire
and burn risk in use. Instead, these standards will reduce vapor emissions both during operation
5-109
-------
Final Regulatory Impact Analysis
and in storage. That reduction, coupled with some expected equipment redesign, is expected to
lead to reductions in the risk of fire or burn without affecting component durability.
We also conducted a design and process Failure Mode and Effects Analyses (FMEA)
comparing current Phase 2 and Phase 3 compliant engines and equipment to evaluate
incremental changes in risk probability as a way of evaluating the incremental risk of upgrading
Phase 2 engines to meet Phase 3 emission standards.224 This is an engineering analysis tool to
help engineers and other professional staff on the FMEA team to identify and manage risk. In a
FMEA, potential failure modes, causes of failure, and failure effects are identified and a
resulting risk probability is calculated from these results. This risk probability is used by the
FMEA team to rank problems for potential action to reduce or eliminate the causal factors.
Identifying these causal factors is important because they are the elements that a manufacturer
can consider reducing the adverse effects that might result from a particular failure mode.
Our FEMA evaluated permeation and running loss controls on nonhandheld engines. We
found that these controls do not increase the probability of fire and burn risk from those expected
with current fuel systems, but could in fact lead to directionally improved systems from a safety
perspective. Finally, the running loss control program for nonhandheld equipment will lead to
changes that are expected to reduce risk of fire during in-use operation. Moving fuel tanks away
from heat sources, improving cap designs to limit leakage on tip over, and requiring a tethered
cap will all help to eliminate conditions which lead to in-use problems related to fuel leaks and
spillage. Therefore, we believe that the application of emission control technology to reduce
evaporative emissions from these fuel hoses and fuel tanks will not lead to an increase in
incremental risk of fires or burns and in some cases is likely to at least directionally reduce such
risks.
EPA has reviewed the fuel hose and fuel tank characteristics for marine vessels and
evaluated control technology which could be used to reduce evaporative emissions from boats.
With regard to fuel hoses, fuel tanks, and diurnal controls, there are rigorous USCG, ABYC, UL,
and SAE standards which manufacturers will continue to meet for fuel system components. All
of these standards are designed to address the in-use performance of fuel systems, with the goal
of eliminating fuel leaks. The low permeation fuel hoses and tanks needed to meet the Phase 3
requirements need to pass these standards and every indication is that they will pass.
Furthermore, the EPA permeation certification requirements related to emissions
durability will add an additional layer of assurance. Low permeation fuel hoses are used safely
today in many marine vessels. Low permeation fuel tanks and diurnal emission controls have
been demonstrated in various applications for many years without an increase in safety risk.
Furthermore, a properly designed fuel system with fuel tank and fuel hose permeation controls
and diurnal emission controls will reduce the fuel vapor in the boat, thereby reducing the
opportunities for fuel related fires. In addition, using improved low permeation materials
coupled with designs meeting USCG and ABYC requirements should reduce the risk of fuel
leaks into the vessel. EPA believes that the application of emission control technologies on
marine engines and vessels for meeting the evaporative emissions standards will not lead to an
5-110
-------
Feasibility of Evaporative Emission Control
increase in incremental risk of fires or burns, and in many cases may incrementally decrease
safely risks in certain situations.
5-111
-------
Final Regulatory Impact Analysis
APPENDIX 5A: Diurnal Temperature Traces
Figure 5A-1: Temperature Trace for Personal Watercraft on Trailer
O)
0)
•o
2
0)
Q.
0)
12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM
Figure 5A-2: Temperature Trace for Jet Boat on Trailer
95.00
90.00
85.00
80.00
O)
0)
75.00
I
0)
Q.
0)
70.00
65.00
60.00
55.00
50.00
45.00
12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM
5-112
-------
Feasibility of Evaporative Emission Control
Figure 5A-3: Temperature Trace for Runabout on Trailer
105.00
12:00 AM
12:OOPM 12:00 AM
12:OOPM 12:00 AM
12:OOPM 12:00 AM
12:00 PM
Figure 5A-4: Temperature Trace for Jet Boat on Trailer
::OOAM 12:OOPM
12:00 AM 12:OOPM 12:00 AM 12:OOPM 12:00 AM 12:OOPM
5-113
-------
Final Regulatory Impact Analysis
Figure 5A-5: Temperature Trace for Runabout in Water
a
2
o>
a.
E
a>
Fuel Temp
Ambient Temp
Water Temp
12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM
12:00 AM
Figure 5A-6: Temperature Trace of Deckboat in Water
Fuel Temp
Ambient Temp
Water Temp
12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM 12:OOAM
5-114
-------
Feasibility of Evaporative Emission Control
APPENDIX 5B: Emission Results for Small SI Equipment Fuel Tanks
Showing Effect of Venting on Diffusion
5B.1 Diffusion Effects from Variable Temperature Diurnal Testing
Figure 5B-1: Diurnal/Diffusion Test Results for BM Metal Fuel Tank (2 Labs)
•190
-inn
ft n
"w1
S. 6 o
&
o
I
4 0
2n
o n s
^^^~ I U III pU I d LUI U
— •— Stock Cap Vent, Lab 1
Qtn^k O on \/ont I a hi O
OLUUr\ ^dfJ vclll, LdU Z
^^ Hose Vent, Lab 1
^*^ Hose Vent Lab 2
^^ ^>S
^^—
^^ ,-*~
-:--:'
**•_
_^ , -^ ^
^jr^ ^^^^^^^^^
K* ^i^^^1^^^^^
«^* ^^^==m^
j^j*^L-*-*
v^S-—* — «_
0 3 6 9 12 15 18 21 2
Test Hours
60.0
-^ 10 w ^ c
P o o o o c
O o O O O C
Temperature [deg C]
5-115
-------
Final Regulatory Impact Analysis
Figure 5B.1-2: Diurnal/Diffusion Test Results for BP Plastic Fuel Tank
8 0
7 0
6 0
5 0
(A
E
2 4 n
a
o
I 30
9 n
1 0
0 0 i
^^" ^^S^^ ^.
/ ^^-L
—^^ ^
4f To m n o TQ ti ir
* — * — M o a i TI e a u
^/ -•— Stock Cap
//* -^4^ Hose Vent
••••••••••
^i-^-*-^^*^~™^
0 3 6 9 12 15 18
Test Hours
4n n
*^*^ 35 0
30 0
o"
""" — ~^-\ 2.
9Q o
•e =
-ii~,\/m-it 150 J"
ap vent ' j-u jj
Vent 10.0 1"
a)
5 0
0 0
21 24
Figure 5B.1-3: Diurnal/Diffusion Test Results for HP Plastic Fuel Tank
re
O)
O
48.0
40.0
32.0
24.0
16.0
8.0
O
O)
0)
«
Q.
E
0.0
9 12 15
Test Hours
18
21
24
5-116
-------
Feasibility of Evaporative Emission Control
Figure 5B.1-4: Diurnal/Diffusion Test Results for TP Plastic Fuel Tank
E
2
S
0
X
1 . D
1 4
1 9
1 n
0 8
0 6
n 4
n 9
n n f
(
/^ ^V
y1^ ^^^
/
-^
A /»^*^
si^
~*^*_
3 3 6 9 12 1£
Test Hours
Temperature
^^ Stock Cap Vent
— ^ — noss vsm
^^^^
^— • ~_
18 21 2
35 0
30 0
oc n
i
20 0
L
15 0
10 0
5 0
On
4
O)
0)
2,
D
(0
Q)
Q.
0)
1-
5-117
-------
Final Regulatory Impact Analysis
5B.2 Isothermal Results for Small SI Equipment Fuel Tanks Showing Effect of Venting on
Diffusion
Figure 5B.2-1: Isothermal Diffusion Test Results for BM Metal Fuel Tank
7 nn
6.00
't/j'
EA nn
ss
1—1 "} nn
o 3-°°
I
? nn
1 nn
n nn •
c
•
M'
jm
/•_
" mT~
•^ Temp (cap)
_/ ^— Temp (hose)
f \ i
/• -m- Stock Cap Vent
^ —^- Hose Vent
M
*— ^ Jr-^*^^*^^^
V ^^^r^^
^z^^^^
• -•zzsL.
D 3 6 9 12 15 18 21 2
Test Hours
T^ n
OO.U
30 0
^O.U Q
O)
a)
20.0 ^.
£
15.0 ^
5
0)
m n Q.
> IU.U ^«
03
cn I-
o.O
n n
4
Figure 5B.2-2: Isothermal Diffusion Test Results for BP Plastic Fuel Tank
7 nn -,
6 00
5 on
"t/T
E4 nn
SS
1 — ' "} nn
o d-uu
X
p nn
1 nn
0.00 i
(
J9
J*/
^»
#^
#
*>
<*
*
<*
J*
*
Temp (mod cap)
Temp (cap)
Temp (hose)
-•- Stock Cap Vent
-m- Stock Cap Vent
—*— Hose Vent
•S
* i— • — •-^•=•=1 t=B=l=l=l=!
r— •— •— M^T^* -• * ^
J 3 6 9 12 15 18 21 2
Test Hours
35.0
•*n n
-1 -1 f0 ro c
en o en o en c
CD CD CD CD CD C
Temperature [deg C]
i 0.0
4
5-118
-------
Feasibility of Evaporative Emission Control
Figure 5B.2-3: Isothermal Diffusion Test Results for HP Plastic Fuel Tank
n /}E> -,
0 30
n 9*^
"to1
En °n
SS
1 — ' n 15
o °-1b
I
Om
n nn i
^^^
m^^^^^
^L9*^
^Z^^
*./
^^
^—Temperature
-m- Stock Cap Vent
-*^- Hose Vent
0 3 6 9 12 15 18 21 2
Test Hours
35.0
30.0
, 25'° 0
O)
0)
on n w
p en o y, c
CD CD o CD C
Temperature [c
Figure 5B.2-4: Isothermal Diffusion Test Results for TP Plastic Fuel Tank
0 35 oc n
0.30
n OEi
En pn
SS
- 015
o U-I!D
X
n m
Onn j
^ ^^^^^^ ^^^^^^^^0
^^^^^^" ^^^^^^ ^^^^^^
•MMMMMMMMMMMM^1*"111**^^^*^^^^^*
j^3
^^^
*^^
-^A^1
J*
Temp (cap)
Temp (hose)
^^ Stock Cap Vent
—^- Hose Vent
.uu • 1 1 1 1 1 1 1
0 3 6 9 12 15 18 21 2
Test Hours
OO.VJ
30.0
25.0 JJ
1 0)
a)
-^
p en o y, c
CD CD o CD C
Temperature [c
5-119
-------
Final Regulatory Impact Analysis
APPENDIX 5C: Diurnal Emission Results: Canister and Passive-Purge
Diurnal Emissions for a 30
50.00
45.00 H
40.00
.g 35.00
t/>
w 30.00
5
^ 25.00
'•£ 20.00
0
Q.
J 15.00
UJ
10.00
5.00
0.00
C
Fuel Tank with 2.1 L Can
Gallon Marine
ister, 72-96F
Baseline Emission Level
©
®
•
•
•
•
i
) 5 10 15
® ®
*4 *
* Diurnal Emissions
QMonday Outliers
20 25
•
•
30
Test Days
5-120
-------
Feasibility of Evaporative Emission Control
APPENDIX 5D: Material Properties of Common Fuel System Materials
This appendix presents data on permeation rates for a wide range of materials that can be
used in fuel tanks and hoses. The data also includes effects of temperature and fuel type on
permeation. Because the data was collected from several sources, there is not complete data on
each of the materials tested in terms of temperature and test fuel. Table D-l gives an overview
of the fuel systems materials included in the data set. Tables D-2 through D-3 present
permeation rates using Fuel C, a 10 percent ethanol blend (CE10), and a 15 percent methanol
blend (CE15) for the test temperatures of 23, 40, 50, and 60°C.
Table 5D-1: Fuel System Materials
Material Name
ACN NBR
Carilon
Celcon
CFM
CO
CR
CSM
E14659
E14944
ECO
ETER
ETFE
EVOH
FEE
FEP
FKM
FPA
FVMQ
GFLT
HOPE
HOPE
HNBR
LDPE
NBR
Nylon 12
PBT
PFA
Polyacetal
PTFE
PVDF
THV
Composition
acrylonitrile
aliphatic poly-ketone thermoplastic
acetal copolymer
fluoroelastomer
epichlorohydrin homopolymer
polychloroprene polymer
chlorosulfonated polyethylene
fluoropolymer film
fluoropolymer film
epichlorohydrin-ethylene oxide copolymer
epichlorohydrin-ethylene oxide terpolymer
ethylenetetrafluoroethylene, fluoroplastic
ethylene vinyl alcohol, thermoplastic
fluorothermoplastic
fluorothermoplastic
fluorocarbon elastomer
copolymer of tetrafluoroethylene and perfluoroalkoxy monomer
fluorovinyl methal silicone rubber (flourosilicone)
fluoroelastomer
high-density polyethylene
high density polyethylene
hydrogenated acrylonitrile-butadiene rubber
low density polyethylene
acrylonitrile-butadiene rubber
thermoplastic
polybutylene terephthalate, thermoplastic
fluorothermoplastic
thermoplastic
polytetrafluoroethylene, fluoroplastic
polyvinylidene fluoride, fluorothermoplastic
tetrafluoroethylene, hexafluoropropylene, vinyledene fluoride
5-121
-------
Final Regulatory Impact Analysis
Table 5D-2: Fuel System Material Permeation Rates at 23°C by Fuel Type
225,226,227,228,229,230
Material Name
HOPE
Nylon 12, rigid
EVOH
Polyacetal
PBT
PVDF
NBR (33% ACN)
HNBR (44%ACN)
FVMQ
FKM Viton A200 (66%F)
FKM Viton B70 (66%F)
FKM Viton GLT (65 %F)
FKM Viton B200 (68%F)
FKM Viton GF (70%F)
FKM Viton GFLT (67%F)
FKM -2 120
FKM - 5830
Teflon FEP 1000L
Teflon PTFE
Teflon PFA 1000LP
Tefzel ETFE 1000LZ
Nylon 12 (GM grade)
Nitrile
Silicone Rubber
Fluorosilicone
FKM
FE 5620Q (65.9% fluorine)
FE 5840Q (70.2% fluorine)
PTFE
ETFE
PFA
THV500
FuelC
g-mm/m2/day
35
0.2
-
-
-
-
669
230
455
0.80
0.80
2.60
0.70
0.70
1.80
8
1.1
0.03
-
0.18
0.03
6.0
130
-
-
-
-
-
0.05
0.02
0.01
0.03
Fuel CE10
g-mm/m2/day
_
-
-
-
-
-
1028
553
584
7.5
6.7
14
4.1
1.1
6.5
-
-
0.03
-
0.03
0.05
24
635
-
-
16
7
4
-
-
-
-
CM15
g-mm/m2/day
35
64
10
3.1
0.4
0.2
1188
828
635
36
32
60
12
3.0
14
44
8
0.03
0.05
0.13
0.20
83
1150
6500
635
-
-
-
0.08*
0.04*
0.05*
0.3
* tested on CM20.
5-122
-------
Feasibility of Evaporative Emission Control
Table 5D-3: Fuel System Material Permeation Rates at 40°C by Fuel Type
231,232
Material Name
Carilon
EVOH-F101
EVOH-XEP380
HOPE
LDPE
Nylon 12 (L2 10 IF)
Nylon 12 (L2140)
Celcon
Fortran PPSSKX-3 82
Celcon Acetal M90
Celanex PBT 3300 (30% GR)
Nylon 6
Dyneon El 465 9
Dyneon El 4944
ETFE Aflon COP
m-ETFE
ETFE Aflon LM730 AP
FKM-70 16286
GFLT 19797
Nitrile
FKM
FE 5620Q (65.9% fluorine)
FE 5840Q (70.2% fluorine)
THV-310X
THV-500
THV-610X
FuelC
g-mm/m2/day
0.06
<0.0001
<0.0001
90
420
2.0
1.8
0.38
-
-
-
-
0.25
0.14
0.24
0.27
0.41
11
13
-
-
-
-
-
0.31
-
Fuel CE10
g-mm/m2/day
1.5
0.013
-
69
350
28
44
2.7
0.12
0.35
3
26
-
-
0.67
-
0.79
35
38
1540
86
40
12
-
-
-
CM15
g-mm/m2/day
13
3.5
5.3
71
330
250
-
-
-
-
-
-
2.1
1.7
1.8
1.6
2.6
-
-
3500
120
180
45
5.0
3.0
2.1
Table 5D-4: Fuel System Material Permeation Rates at 50°C by Fuel Type
233
Material Name
Carilon
HOPE
Nylon 12(L2140)
Celcon
ETFE Afcon COP
FKM-70 16286
GFLT 19797
FuelC
g-mm/m2/day
0.2
190
4.9
0.76
-
25
28
Fuel CE10
g-mm/m2/day
3.6
150
83
5.8
1.7
79
77
CM15
g-mm/m2/day
_
-
-
-
-
-
-
5-123
-------
Final Regulatory Impact Analysis
Table 5D-5: Fuel System Material Permeation Rates at 60°C by Fuel Type 234-235-236-237
Material Name
Carilon
HOPE
Nylon 12 (L2140)
Celcon
ETFE Afcon COP
FKM-70 16286
GFLT 19797
polyeurethane (bladder)
THV-200
THV-310X
THV-510ESD
THV-500
THV-500 G
THV-610X
ETFE 6235 G
THV-800
FEP
FuelC
g-mm/m2/day
0.55
310
9.5
1.7
-
56
60
285
-
-
6.1
-
4.1
2.4
1.1
1.0
0.2
Fuel CE10
g-mm/m2/day
7.5
230
140
11
3.8
170
130
460
54
-
18
11
10
5.4
3.0
2.9
0.4
CM15
g-mm/m2/day
_
—
-
-
-
-
-
-
-
38
35
20
22
9.0
6.5
6.0
1.1
5-124
-------
Feasibility of Evaporative Emission Control
APPENDIX 5E: Diurnal Test Temperature Traces
Table 5E-1: Temperature vs. Time Sequence for Diurnal Testing
Test
Time*
[minutes]
0
60
120
180
240
300
360
420
480
540
600
660
720
780
840
900
960
1020
1080
1140
1200
1260
1320
1380
1440
Portable Fuel Tanks
SHED Air Temperature
Fahrenheit
72.0
72.5
75.5
80.3
85.2
89.4
93.1
95.1
95.8
96.0
95.5
94.1
91.7
88.6
85.5
82.8
80.9
79.0
77.2
75.8
74.7
73.9
73.3
72.6
72.0
Celsius
22.2
22.5
24.2
26.8
29.6
31.9
33.9
35.1
35.4
35.6
35.3
34.5
33.2
31.4
29.7
28.2
27.2
26.1
25.1
24.3
23.7
23.3
22.9
22.6
22.2
Installed Fuel Tanks
Trailerable Boat
Fuel Temperature
Fahrenheit
78.0
78.3
79.8
82.2
84.6
86.7
88.6
89.6
89.9
90.0
89.8
89.1
87.9
86.3
84.8
83.4
82.5
81.5
80.6
79.9
79.4
79.0
78.7
78.3
78.0
Celsius
25.6
25.7
26.5
27.9
29.2
30.4
31.4
32.0
32.2
32.2
32.1
31.7
31.0
30.2
29.3
28.6
28.0
27.5
27.0
26.6
26.3
26.1
25.9
25.7
25.6
Installed Fuel Tanks
Nontrailerable boat
Fuel Temperature
Fahrenheit
81.6
81.7
82.3
83.3
84.2
85.1
85.8
86.2
86.4
86.4
86.3
86.0
85.5
84.9
84.3
83.8
83.4
83.0
82.6
82.4
82.1
82.0
81.9
81.7
81.6
Celsius
27.6
27.6
27.9
28.5
29.0
29.5
29.9
30.1
30.2
30.2
30.2
30.0
29.7
29.4
29.1
28.8
28.5
28.3
28.1
28.0
27.9
27.8
27.7
27.6
27.6
* Repeat as necessary
5-125
-------
Final Regulatory Impact Analysis
Chapter 5 References
1. Reddy, S., "Prediction of Fuel Vapor Generation From a Vehicle Fuel Tank as a Function of Fuel RVP and
Temperature," SAE Paper 892089, 1989, Docket Identification EPA-HQ-OAR-2004-0008-0169.
2. Gushing, T., "Fuel Tank Hydrocarbon Emission Testing," prepared by Sterling Performance for the U.S. EPA,
Report #040107, September 30, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0137.
3. Gushing, T., "Fuel Tank Hydrocarbon Emission Testing," prepared by Sterling Performance for the U.S. EPA,
Report #040137, November 3, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0138.
4. "Evaporative Emissions from Offroad Equipment," California Air Resources Board, June 22, 2001, Docket
Identification EPA-HQ-OAR-2004-0008-0051.
5. "Docket No. A-2000-02; Notice or Proposed Rulemaking for the Control of Emissions from Spark-Ignition
Marine Vessels and Highway Motorcycles; 40 C.F.R. Parts 86, 90, 1045, 1051, and 1068," Comments from the
National Marine Manufacturers Association, Docket Identification EPA-HQ-OAR-2004-0008-0211.
6. "Before the United States Environmental Protection Agency, Comments by Brunswick Corporation, Notice of
Proposed Rulemaking, Part 1045 Control of Emissions from Spark-Ignition Marine Vessels," Comments from
Brunswick Corporation, Docket Identification EPA-HQ-OAR-2004-0008-0192.
7. "Diurnal Emissions Testing of Walk-Behind Mowers Configured with Fuel Tank Pressure Relief Valves
(September 2002)," California Air Resources Board, September 17, 2002, Docket Identification EPA-HQ-OAR-
2004-0008-0052.
8. "Aboveground Storage Tanks (AST); Enhanced Vapor Recovery Regulation," California Air Resources Board,
August 16, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0413.
9. Email from Pamela Gupta, California Air Resources Board, to Mike Samulski, U.S. EPA, "Aboveground Storage
Tanks (ASTs) Field Study Presentation," September 6, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0420.
10. Samulski, M., "Characterization and Control of Evaporative Emissions from Fuel Tanks in Nonroad
Equipment," SAE Paper 2006-32-0094, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0424.
11. Beverley, K., Clint, I, Fletcher, P., "Evaporation Rates of Pure Liquids Measured Using a Gravimetric
Technique," Phys. Chem. Chem. Phys., 1999, 1, 149-153, Received July 9, 1998, Accepted October 13, 1998,
Docket Identification EPA-HQ-OAR-2004-0008-0103.
12. Batterman, S., Yu, Y., Jia, C., Godwin, C., "Non-methane Hydrocarbon Emissions from Vehicle Fuel Caps,"
Atmospheric Environment 39 (2005), Received May 27, 2004, Accepted December 1, 2004, Docket Identification
EPA-HQ-OAR-2004-0008-0102.
13. Sterling Performance, "USEPA Fuel Tank Hydrocarbon Emission Testing," prepared for the U.S. EPA, Report
Number 040195, January 26, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0139.
14. "24 hr EPA Diurnal MicroSHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-1.1,
Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0140.
15. "24 hr EPA Diurnal Micro SHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-1.2,
Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0141.
5-126
-------
Feasibility of Evaporative Emission Control
16. "24 hr EPA Diurnal MicroSHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-1.3,
Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0142.
17. "24 hr EPA Isothermal MicroSHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-
2.1, Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0143.
18. "24 hr EPA Isothermal MicroSHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-
2.2, Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0144.
19. "24 hr EPA Isothermal MicroSHED of Lawn Mower Gas Tanks," Testing Services Group, Reports L504835-
2.3, Prepared for U.S. EPA, April 20, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0145.
20. Haskew, H., Cadman, W., "Evaporative Emissions Under Real Time Conditions," SAE Paper 891121, 1989,
Docket Identification EPA-HQ-OAR-2004-0008-0168.
21. Haskew, H., Cadman, W., Liberty, T., "The Development of a Real-Time Evaporative Emissions Test," SAE
Paper 901110, 1990, Docket Identification EPA-HQ-OAR-2004-0008-0170.
22. Tschantz, M., "Activated Carbon for Use in Marine Evaporative Control Applications," Presented by
MeadWestvaco Corporation at the 2004 International Boatbuilders Exposition, October 25, 2004, Docket
IdentificationEPA-HQ-OAR-2004-0008-0040.
23. Internal MeadWestvaco memorandum from M. Tschantz to C. Pierson, "Preliminary 3 mm SeaGuard™
Specifications," August 31, 2005, Docket IdentificationEPA-HQ-OAR-2004-0008-0136.
24. "A Picture Book Concerning the NMMA 2005 Summer Test Program: Carbon Canisters for Marine Diurnal
Control," Harold Haskew & Associates, Inc., July 12, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0127.
25. Dr. Michael Tschantz, "Summer Test Program Carbon Analysis," presented by Meadwestvaco Corporation at
the 2005 International Boatbuilders Exhibition and Conference, October 20, 2005, Docket Identification EPA-HQ-
OAR-2004-0008-0290.
26. Andreatta, D., Heydinger, G., Bixel, R., Park, I, Jorgensen, S., "Evaluation of the Ignition Hazard Posed by
Onboard Refueling Vapor Recovery Canisters," SAE Paper 2001-01-0731, 2001, Docket Identification EPA-HQ-
OAR-2004-0008-0494.
27. Giacomazzi, R., "The Evolution of Automotive Fuel Tanks," Presented at the International Boat Builders'
Exposition and Conference, October 11, 2007.
28. Engineered Composite Solutions, "Products; Carbon Canisters," retrieved from
www.engineeredcompositesolutions.com/Products.htm on April 24, 2008.
29. "Diurnal Testing of Off-Road Equipment Retrofitted with Carbon Canister Evaporative Emission Control
Systems (March 2003)," California Air Resources Board, March 25, 2003, Docket Identification EPA-HQ-OAR-
2004-0008-0053.
30. Haskew, H., Cadman, W., Liberty, T., "The Development of a Real-Time Evaporative Emissions Test," SAE
Paper 901110, 1990, Docket Identification EPA-HQ-OAR-2004-0008-0170.
31. Email from Reid Clontz, MeadWestvaco, to Mike Samulski, U.S. EPA, "Temperature and Ethanol Questions,"
July 7, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0423.
32. American Boat and Yacht Council, "H-24 Gasoline Fuel Systems," 2005, Docket Identification EPA-HQ-OAR-
2004-0008-0429.
5-127
-------
Final Regulatory Impact Analysis
33. Sutherland, S., Hayes, I, "Analysis of the 2007 California Survey of Outboard and Sailboat Owners Regarding
Use of Portable Outboard Marine Fuel Tanks," Prepared for the California Air Resources Board by the Institute for
Social Research at California State University, Sacramento, March 2007.
34. "Diurnal Emissions Testing of Walk-Behind Mowers Configured with Fuel Tank Pressure Relief Valves
(September 2002)," California Air Resources Board, September 17, 2002, Docket Identification EPA-HQ-OAR-
2004-0008-0052.
35. Letter from Bill Scott, Country Industries Technologies, to Mike Samulski U.S. EPA, regarding a gas tank vent
design, July 20, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0287.
36. Compact Membrane Systems, Inc., www.compactmembrane.com, Docket Identification EPA-HQ-O AR-2004-
0008-0100.
37. Koch, W., "Developing Technology for Enhanced Vapor Recovery: Part 1 - Vent Processors," Petroleum
Equipment & Technology, February/March 2001, Docket Identification EPA-HQ-OAR-2004-0008-0099.
38. Membrane Technology & Research, Inc., www.mtrinc.com, Docket Identification EPA-HQ-OAR-2004-0008-
0101.
39. "New Evaporative Control System for Gasoline Tanks," Memorandum from Chuck Moulis, EPA to Glenn
Passavant, Nonroad Center Director, EPA, March 1, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0261.
40. Letter from Kevin Madison, Top Dog Systems, to Mike Samulski, U.S. EPA, regarding the Top Dog vaporless
fuel system, April 23, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0258.
41. Facsimile from Bob Hazekamp, Top Dog Systems to Mike Samulski, U.S. EPA, "Permeation of Polyurethane
versus THV Materials @ 60°C," January 14, 2002, Docket A-2000-01, Docket Identification EPA-HQ-O AR-2004-
0008-0435.
42. Emails from Steve Vaitses, Seacurefill to Mike Samulski, U.S. EPA, "VCS diagrams 1-4 attachments," and
"VCS attachments 5-7," April 27, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0395.
43. Testing Services Group, "Marine Fuel Tank Vapor Trap Diurnal Emissions Testing; 48 hr EPA (custom)
Diurnal," TSG Report No. L605816-1.2, July 14, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0445.
44. Fox, J., "Fuel Temperature Measurements on Small Equipment," Fox Automotive, August 2004, Docket
Identification EPA-HQ-OAR-2004-0008-0181.
45. Gushing, T., "Fuel Tank Hydrocarbon Emission Testing," prepared by Sterling Performance for the U.S. EPA,
Report #040107, September 30, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0137.
46. Gushing, T., "Fuel Tank Hydrocarbon Emission Testing," prepared by Sterling Performance for the U.S. EPA,
Report #040137, November 3, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0138.
47. Wong, W., "Addition of Evaporative Emissions for Small Off-Road Engines," California Air Resources Board
OFFROAD Model Change Technical Memo, Revised April 21, 2003, Docket OAR-2004-0008, Document OAR-
2004-0008-0006.
48. Tuckner, P., Baker, J., "Fuel Permeation Testing using Gravimetric Methods," SAE Paper 2000-01-1096, 2000,
Docket Identification EPA-HQ-O AR-2004-0008-0160.
49. Allen, S.J., "Fuel Tank Permeability Test Procedure Development; Final Report," U.S. Coast Guard, December
1986, Docket Identification EPA-HQ-OAR-2004-0008-0231.
5-128
-------
Feasibility of Evaporative Emission Control
50. Nulman, M, Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
51. "Permeation Rates of Small Off Road Engine High-Density Polyethylene Fuel Tanks (April 2001 Testing), June
8, 2001, California Air Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0146.
52. "Permeation Rates of Small Off Road Engine High-Density Polyethylene Fuel Tanks (February 2001 Testing),
June 8, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0147.
53. "Permeation Rates of High-Density Polyethylene Fuel Tanks (June 2001), June 12, 2001, California Air
Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0148
54. "Permeation Rates of High-Density Polyethylene Fuel Tanks (May 2001), June 11, 2001, California Air
Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0149.
55. "Average Permeation Rates of High-Density Polyethylene Off-Road Equipment Fuel Tanks Using a Surface
Area Approach, (March 2002)," March 15, 2002, California Air Resources Board, Docket Identification EPA-HQ-
OAR-2004-0008-0256.
56. "Test Method 513; Determination of Permeation for Spill-Proof Systems," California Air Resources Board,
Adopted July 6, 2000, Docket Identification EPA-HQ-OAR-2004-0008-0150.
57. "OPEIHHPC Comments on EPA Proposed Phase 3 Rule for HH Fuel Tank Permeation," Outdoor Power
Equipment Institute," Presentation to EPA, February 5, 2008.
58. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data reports available in
Docket, Docket Identification EPA-HQ-OAR-2004-0008-0186.
59. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early Container Data,"
August 29, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0200.
60. "Test Method 513; Determination of Permeation for Spill-Proof Systems," California Air Resources Board,
Adopted July 6, 2000, Docket Identification EPA-HQ-OAR-2004-0008-0150.
61. Lockhart, M., Nulman, M., Rossi, G., "Estimating Real Time Diurnal Permeation from Constant Temperature
Measurements," SAE Paper 2001-01-0730, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0161.
62. Hopf, G., Ries, H., Gray, E., "Development of Multilayer Thermoplastic Fuel Lines with Improved Barrier
Properties," SAE Paper 940165, 1994, Docket Identification EPA-HQ-OAR-2004-0008-0174.
63. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
64. Haskew, H., McClement, D., "Fuel Permeation from Automotive Systems; Final Report," prepared for the
California Air Resources Board and the Coordinating Research Council, CRC Project No. E-65, September 2004,
Docket Identification EPA-HQ-OAR-2004-0008-0151.
65. Tuckner, P., Baker, J., "Fuel Permeation Testing using Gravimetric Methods," SAE Paper 2000-01-1096, 2000,
Docket Identification EPA-HQ-OAR-2004-0008-0160.
66. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
5-129
-------
Final Regulatory Impact Analysis
67. Stevens, M, Demorest, R., "Fuel Permeation Analysis Method Correction," SAE Paper 1999-01-0376, 1999,
Docket Identification EPA-HQ-OAR-2004-0008-0157.
68. Lockhart, M., Nulman, M., Rossi, G., "Estimating Real Time Diurnal Permeation from Constant Temperature
Measurements," SAE Paper 2001-01-0730, Docket Identification EPA-HQ-OAR-2004-0008-0161.
69. Testimony of and Presentation, H. Haskew, President, Harold Haskew & Associates, MI, October 7, 2002,
Docket Identification EPA-HQ-OAR-2004-0008-0202.
70. Haskew, H., McClement, D., "Fuel Permeation from Automotive Systems; Final Report," prepared for the
California Air Resources Board and the Coordinating Research Council, CRC Project No. E-65, September 2004,
Docket Identification EPA-HQ-OAR-2004-0008-0151.
71. Email from Jim Watson, California Air Resources Board, to Mike Samulski, U.S. EPA, "Vapor Loss for
Variable Volume Containers.xls," October 26, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0291.
72. Email from Harold Haskew, Harold Haskew & Associates, to Mike Samulski, U.S. EPA, "High-low permeation
data," April 9, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0238.
73. Kathios, D., Ziff, R., Petrulis, A., Bonczyk, J., "Permeation of Gasoline and Gasoline-alcohol Fuel Blends
Through High-Density Polyethylene Fuel Tanks with Different Barrier Technologies," SAE Paper 920164, 1992,
Docket Identification EPA-HQ-OAR-2004-0008-0172.
74. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data reports available in the
Docket, Docket Identification EPA-HQ-OAR-2004-0008-0186.
75. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early Container Data,"
August 29, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0200.
76. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine Fuel Tanks,"
California Air Resources Board, June 21, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0152.
77. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine Fuel Tanks,"
California Air Resources Board, March 7, 2003, Docket Identification EPA-HQ-OAR-2004-0008-0153.
78. Conversation between Mike Samulski, U.S. EPA and Tom Schmoyer, Sulfo Technologies, June 17, 2002.
79. "Sulfo Data", E-mail from Tom Schmoyer, Sulfotechnologies to Mike Samulski, U.S. EPA, March 17, 2003,
Docket Identification EPA-HQ-OAR-2004-0008-0223.
80. "ADDENDUM TO: Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine
Fuel Tanks," California Air Resources Board, March 27, 2003, Docket Identification EPA-HQ-OAR-2004-0008-
0154.
81. "Resin and Additives - SOS Compatible," Email from Tom Schmoyer, Sulfo Technologies to Mike Samulski
and Glenn Passavant, U.S. EPA, June 19, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0215.
82. Email from Jim Watson, California Air Resources Board, to Mike Samulski, U.S. EPA, "Attachment to Resin
List," August 30, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0234.
83. Walles, B., Nulford, L., "Five Year Durability Tests of Plastic Gas Tanks and Bottles with Sulfonation Barrier,"
Coalition Technologies, LTD, for Society of Plastics Industry, January 15, 1992, Docket Identification EPA-HQ-
OAR-2004-0008-0233.
5-130
-------
Feasibility of Evaporative Emission Control
84. Kathios, D., Ziff, R., "Permeation of Gasoline and Gasoline-Alcohol Fuel Blends Through High-Density
Polyethylene Fuel Tanks with Different Barrier Technologies," SAE Paper 920164, 1992, Docket Identification
EPA-HQ-OAR-2004-0008-0172.
85. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data reports available in
Docket, Docket Identification EPA-HQ-OAR-2004-0008-0186.
86. "Permeation Rates of Blitz Fluorinated High Density Polyethylene Portable Fuel Containers," California Air
Resources Board, April 5, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0155.
87. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early Container Data,"
August 29, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0200.
88. www.pensteel.co.uk/light/smp/fluorination.htm. A copy of this site, as downloaded on August 13, 2004, is
available in Docket, Docket Identification EPA-HQ-OAR-2004-0008-0214.
89. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine Fuel Tanks,"
California Air Resources Board, June 21, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0152.
90. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine Fuel Tanks,"
California Air Resources Board, March 7, 2003, Docket Identification EPA-HQ-OAR-2004-0008-0153.
91. "OPEIHHPC Comments on EPA Proposed Phase 3 Rule for HH Fuel Tank Permeation," Outdoor Power
Equipment Institute," Presentation to EPA, February 5, 2008.
92. Kathios, D., Ziff, R., "Permeation of Gasoline and Gasoline-Alcohol Fuel Blends Through High-Density
Polyethylene Fuel Tanks with Different Barrier Technologies," SAE Paper 920164, 1992, Docket Identification
EPA-HQ-OAR-2004-0008-0172.
93. "SORE Component EO Summary," email from Jim Watson, California Air Resources Board, to Mike Samulski,
U.S. EPA, August 18, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0405.
94. "Fluorination Information," e-mail from Doug McGregor, BMW, to Mike Samulski, US EPA, August 8, 2003,
Docket Identification EPA-HQ-OAR-2004-0008-0259.
95. "Final Report: Permeation Testing of Injection Molded Walk Behind Mower (WBM) Fuel Tanks using
California's TP-901 Procedures," Automotive Testing Laboratories, Inc., August 30, 2005, Docket Identification
EPA-HQ-OAR-2004-0008-0182.
96. "Baseline Test Results: Permeation Testing of Injection Molded Walk Behind Mower (WBM) Fuel Tanks using
Federal Test Procedures," Automotive Testing Laboratories, Inc., August 30, 2005, Docket Identification EPA-HQ-
OAR-2004-0008-0183.
97. "Final Results: Permeation Testing of Injection Molded Walk Behind Mower (WBM) Fuel Tanks using Federal
Test Procedures," Automotive Testing Laboratories, Inc., July 28, 2006, Docket Identification EPA-HQ-OAR-2004-
0008-0419.
98. "Selar RB Technical Information," Faxed from David Zang, Dupont, to Mike Samulski, U.S. EPA on May 14,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0221.
99. Lyn, K., "A Study on Problems with Aluminum Fuel Tanks in Recreational Boats," Underwriters Laboratories
Inc., sponsored by USCG, 1994, Docket Identification EPA-HQ-OAR-2004-0008-0098.
5-131
-------
Final Regulatory Impact Analysis
100. E-mail from Alan Dubin, Ticona, to Mike Samulski, U.S. EPA, "Fuel Permeation Chart and Aggressive Fuels
Brochure," July 31, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0196.
101. Samulski, M., "Permeation Emission Testing on Three Roto-Molded Acetal Copolymer Fuel Tanks," Memo to
Docket OAR-2004-0008, October 25, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0443.
102. E-mail from Jim Watson, California ARB to Mike Samulski, U.S. EPA, "fuel tank specs.xls," March 6, 2006,
Docket Identification EPA-HQ-OAR-2004-0008-0357.
103. Chambers, J., "Marine Fuel Containment... A Permanent Solution," Presentation by Engineered Composite
Structures Inc. at the 2004 International Boatbuilders Exposition, October 25, 2004, Docket Identification EPA-HQ-
OAR-2004-0008-0037.
104. "Fluorination Information," e-mail from Doug McGregor, BMW, to Mike Samulski, US EPA, August 8, 2003,
Docket Identification EPA-HQ-OAR-2004-0008-0259.
105. Fead, E., Vengadam, R., Rossi, G., Olejnik, A., Thorn, J., "Speciation of Evaporative Emissions from Plastic
Fuel Tanks," SAE Paper 981376, 1998, Docket Identification EPA-HQ-OAR-2004-0008-0175.
106. Email from Dennis Goodenow, California Air Resources Board, to Mike Samulski, U.S. EPA, "Coextruded
PFC Data, September 11, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0436.
107. "OPEIHHPC Comments on EPA Proposed Phase 3 Rule for HH Fuel Tank Permeation," Outdoor Power
Equipment Institute," Presentation to EPA, February 5, 2008.
108. Solvay Advanced Polymers, "Solvay Advanced Polymers Unveils Modified Ixef® Polyarylamide Barrier
Material for Automotive Fuel Systems," Press Release, March 2, 200, Docket Identification EPA-HQ-OAR-2004-
0008-0539.
109. "Test Report: EPA CCD-05-14 Fuel Tank Permeation Testing," Testing Services Group, TSG Report Numbers
L706619-1.1, 2.1, 3.1, and4.1A, October 19, 2007, Docket Identification EPA-HQ-OAR-2004-0008-XXXX.
110. O'Brien, Partridge, Clay, "New Materials and Multi-Layer Rotomolding Technology for Higher Barrier
Performance Rotomolded Tanks," Atofina Chemicals, Inc., Presented at the Topcon Rotational Molding by Design
Conference, June 6-8, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0044.
111. Samulski, M., "Permeation Emission Testing of Two Multi-Layer Roto-Molded Fuel Tanks," Memo to Docket
OAR-2004-0008, August 31, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0132.
112. Partridge, R., "Petro-SealTM for Ultra-Low Permeation," Arkema Inc., Presentation at 2004 International
Boatbuilders Exposition, October 25, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0038.
113. "Test Report 16514-1A: NMMA/USCG Tests of Prototype Fuel Tank for Atofina Chemicals, Inc.," IMANNA
Laboratory, July 16, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0416.
114. "Test Report 16544-1A: NMMA/USCG Tests of Prototype Fuel Tank for Atofina Chemicals, Inc.," IMANNA
Laboratory, July 16, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0417.
115. "Test Report 17474-1: NMMA/USCG Tests of 40 Gallon Prototype Fuel Tank for Arkema, Inc.," IMANNA
Laboratory, October 12, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0418.
116. "Final Report, Fuel Tank Testing, Procedure Number: SAE J288 DEC 02, Sections 4.2, 4.3," MGA Research
Corporation, October 19, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0406.
5-132
-------
Feasibility of Evaporative Emission Control
117. "SORE Component EO Summary," email from Jim Watson, California Air Resources Board, to Mike
Samulski, U.S. EPA, August 18, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0405.
118. Faler, Gary, "Cyclics CBT® Resin for Dual Layer Applications in Rotomolding," Cyclics Corporation, August
2006, Docket Identification EPA-HQ-OAR-2004-0008-0415.
119. Faler, Gary, ""Cyclics CBT® Resin for Rotomolding," Cyclics Corporation, Presented at ARMO 2006
Convention & Exposition, September, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0414.
120.Graham, B., Cook, D., "Measurement of Fuel Barrier Properties of Rotational Molded Materials," Article from
ANTEC 2006 Rotational Molding Division, May 12, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0404.
121. Fish, D., "Advanced Polymer Concepts," Presentation at the 2004 International Boatbuilders Exposition,
October 25, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0039.
122. Bauman, B., "Advances in Plastic Fuel Tanks," Presentation by Fluoro-Seal International at the 2004
International Boatbuilders Exposition, October 25, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0036.
123. Bauman, B., "Advances in Plastic Fuel Tanks," Presentation by Fluoro-Seal International at the 2004
International Boatbuilders Exposition, October 25, 2004, Docket Identification EPA-HQ-OAR-2000-0008,
Document OAR-2004-0008-0036.
124. SAE Surface Vehicle Standard, "Marine Fuel Hoses," Society of Automotive Engineers J 1527, Issued 1985-
12, Revised 1993-02, Docket Identification EPA-HQ-OAR-2004-0008-0177.
125. American Society for Testing and Materials, "Standard Test Method for Rubber Property-Effects of Liquids,"
ASTM D 471-06, November 2006, Docket Identification EPA-HQ-OAR-2004-0008-0511.
126. "Public Hearing on Proposed Evaporative Emission Standards for New Marine Vessels that Use Spark-Ignition
Engines," Held on Monday, October 7, 2002 at the U.S. Environmental Protection Agency, Office of Transportation
and Air Quality, in Ann Arbor, Michigan, De Scribe Reporting, Inc., Docket Identification EPA-HQ-OAR-2004-
0008-0265.
127. "Test Report: 24 hr MicroShed Test on 3 Standard Fuel Hoses," Testing Services Group, TSG Report Number
L201672-2.5, July 29, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0227.
128. Akron Rubber Development Laboratory, "Test Report PN# 50107" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0184.
129. Akron Rubber Development Laboratory, "Test Report PN# 50108" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0185.
130. "Test Report: Marine Fuel Hose Testing; 6-hour/28°C Constant Temperature MicroSHED after a cumulative
13-week soakinIE-10 Test Fuel at40°C," Testing Services Group, TSG Report Number L404396-1.1A, January 20,
2005, Docket Identification EPA-HQ-OAR-2004-0008-0407.
131. "Test Report: Permeation by Weight Loss," Testing Services Group, TSG Report Number L505371-6.1, May
15, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0409.
132. Harold Haskew, "NMMA/EPA Meeting," Harold Haskew & Associates, Inc, presented at NMMA/EPA
meeting on January 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0359.
133. SAE Vehicle Recommended Practice, "Personal Watercraft Fuel Systems," Society of Automotive Engineers J
2046, Issued January 19, 1993, Docket Identification EPA-HQ-OAR-2004-0008-0179.
5-133
-------
Final Regulatory Impact Analysis
134. Memorandum from Mike Samulski to Docket A-2000-02, "Hose Permeation Data from Saint-Gobain
Performance Plastics," October 17, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0210.
135. SAE Recommended Practice J30, "Fuel and Oil Hoses,"June 1998, Docket Identification EPA-HQ-OAR-
2004-0008-0176.
136. Baltz, G., "Effects of Alcohol Extended Fuels on the Rate of Fuel Hose Permeation," SAE Paper 880709,
Presented at the International Congress and Exposition February 29-March 4, 1988, Docket Identification EPA-HQ-
OAR-2004-0008-0167.
137. "OPEI Fuel Line Permeation Results," Walbro Engine Management, April 12, 2006, Docket Identification
EPA-HQ-OAR-2004-0008-0440.
138. "HHPC Evaluation of EPA Proposed Phase 3 Rule for Fuel Line Permeation," Outdoor Power Equipment
Institute, presentation to EPA, February 5, 2008.
139. Horvath, I, "Optimizing Permeation Resistance and Low Temperature Flexibility in Heat Resistant NBR Fuel
Hose," SAE Paper 790661, presented at the Passenger Car Meeting, June 11-15, 1979, Docket Identification EPA-
HQ-OAR-2004-0008-0166.
140. MacLachlan, I, "Automotive Fuel Permeation Resistance - A Comparison of Elastomeric Materials," SAE
Paper 790657, presented at the Passenger Car Meeting, June 11-15, 1979, Docket Identification EPA-HQ-OAR-
2004-0008-0165.
141. Testing Services Group, "Marine Fuel Hose MicroSHED Testing; 6-hour/28°C Constant Temperature
MicroSHED after a cumulative 13-week soak in IE-10 Test Fuel at 40°," TSG Report No. L404396-1.1A, January
24, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0407.
142. Testing Services Group, "Marine Fuel Hose MicroSHED Testing; 24-hourEPA MicroSHED after a cumulative
2-week soak in IE-10 Test Fuel at 40°," TSG Report No. L404396-2.1A, January 24, 2005, Docket Identification
EPA-HQ-OAR-2004-0008-0408.
143. Harold Haskew, "Marine Fuel System Hose Permeation Testing: 2nd Progress Report," Harold Haskew &
Associates, Inc., testing performed at Sterling Performance Testing Laboratory, October 3, 2005, Docket
IdentificationEPA-HQ-OAR-2004-0008-0358.
144. Harold Haskew, "Marine Fuel System Hose Permeation Testing: 2nd Progress Report," Harold Haskew &
Associates, Inc., testing performed at Sterling Performance Testing Laboratory, October 3, 2005, Docket
IdentificationEPA-HQ-OAR-2004-0008-0358.
145. Harold Haskew, "NMMA/EPA Meeting," Harold Haskew & Associates, Inc, presented at NMMA/EPA
meeting on January 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0359.
146. "Test Report: Permeation by Weight Loss," Testing Services Group, TSG Report Number L505371-6.1, May
15, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0409.
147. Sterling Performance, "U.S.E.P.A. Fuel Line Permeation 050085; Version 1.0," August 13, 2005, Docket
IdentificationEPA-HQ-OAR-2004-0008-0428.
148. Denbow, R., Browning, L., Coleman, D., "Report Submitted for WA 2-9, Evaluation of the Costs and
Capabilities of Vehicle Evaporative Emission Control Technologies," ICF, ARCADIS Geraghty & Miller, March
22, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0194.
5-134
-------
Feasibility of Evaporative Emission Control
149. Denbow, R., Browning, L., Coleman, D., "Report Submitted for WA 2-9, Evaluation of the Costs and
Capabilities of Vehicle Evaporative Emission Control Technologies," ICF, ARCADIS Geraghty & Miller, March
22, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0194.
150. SAE Recommended Practice J2260, "Nonmetallic Fuel System Tubing with One or More Layers," 1996,
Docket Identification EPA-HQ-OAR-2004-0008-0180.
151. 33CFR183.558
152. SAE Surface Vehicle Standard, "Marine Fuel Hoses," Society of Automotive Engineers J 1527, Issued 1985-
12, Revised 1993-02, Docket Identification EPA-HQ-OAR-2004-0008-0177.
153. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01, June
17, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0208.
154. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01,
August 6, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0207.
155. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01,
August 6, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0207.
156. Akron Rubber Development Laboratory, "Test Report; PN# 49503," Prepared for the U.S. EPA, September 3,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0187.
157. Akron Rubber Development Laboratory, "Test Report PN# 50107" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0184.
158. Akron Rubber Development Laboratory, "Test Report PN# 50108" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0185.
159. Akron Rubber Development Laboratory, "Test Report PN# 53609" Prepared for American Honda Motor
Company, September 26, 2003, Docket Identification EPA-HQ-OAR-2004-0008-0189.
160. Akron Rubber Development Laboratory, "Test Report PN# 53630" Prepared for the U.S. EPA, September 26,
2003, Docket Identification EPA-HQ-OAR-2004-0008-0190.
161. Akron Rubber Development Laboratory, "Test Report PN# 53908" Prepared for the U.S. EPA, October 20,
2003, Docket Identification EPA-HQ-OAR-2004-0008-0191.
162. Email from Mike Fauble, Avon Automotive to Mike Frederick Jr., Avon Automotive, "SAE J30 Test Results,"
April 26, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0394.
163. Fuller, R., "Unique Low Permeation Elastomeric Laminates for Fuel Hose; F200 Technology used in
Automotive Fuel Hose," Hearing Testimony for Dupont-Dow Elastomers on NPRM, October 7, 2002, Docket
Identification EPA-HQ-OAR-2004-0008-0197.
164. "Test Report: 24 hr MicroSHED Test of Fuel Line Assemblies After 8 Week Soak in Fuel C," Testing Services
Group, TSG Report Number L201672-3.1, June 25, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0229.
165. "Test Report: EPA MicroSHED Test on Fuel Line Assemblies 24-Hour EPA MicroSHED Test with Multipoint
@ 8-Weeks," Testing Services Group, TSG Report Number L201672-2.3, June 17, 2002, Docket Identification
EPA-HQ-OAR-2004-0008-0228.
-------
Final Regulatory Impact Analysis
166. "Test Report: EPA MicroSHED Test on Fuel Line Assemblies 24-Hour EPA MicroSHED Re-Test of Part
#01672-11," Testing Services Group, TSG Report Number L201672-3.1R, July 8, 2002, Docket Identification EPA-
HQ-OAR-2004-0008-0230.
167. Akron Rubber Development Laboratory, "Test Report PN# 50107" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0184.
168. Akron Rubber Development Laboratory, "Test Report PN# 50108" Prepared for the U.S. EPA, December 13,
2002, Docket Identification EPA-HQ-OAR-2004-0008-0185
169. Akron Rubber Development Laboratory, "Test Report PN# 52318" Prepared for the U.S. EPA, July 24,2003,
Docket Identification EPA-HQ-OAR-2004-0008-0188.
170. Harold Haskew, "Marine Fuel System Hose Permeation Testing: 2nd Progress Report," Harold Haskew &
Associates, Inc., testing performed at Sterling Performance Testing Laboratory, October 3, 2005, Docket
IdentificationEPA-HQ-OAR-2004-0008-0358.
171. Harold Haskew, "NMMA/EPA Meeting," Harold Haskew & Associates, Inc, presented at NMMA/EPA
meeting on January 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0359.
172. Akron Rubber Development Laboratory, "Test Report PN# 52318" Prepared for the U.S. EPA, July 24,2003,
Docket Identification EPA-HQ-OAR-2004-0008-0188.
173. "Meeting with DuPont and Teleflex on December 10, 2002," memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-02, December 10, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0260.
174. "Low perm marine fuel hose," e-mail from Rob Bentley, Teleflex to Alan Stout U.S. EPA, July 16, 2003,
Docket Identification EPA-HQ-OAR-2004-0008-0225.
175. Engineered Composite Solutions, "Press Release: Feb 2008 ECS develops low permeation primer bulb,"
retrieved from http://www.engineeredcompositesolutions.com/press_releases.htm on April 24, 2008.
176. BluSkies Marine Products, "Primer Bulb; A White Paper," retrieved from www.bluskies.us on June 25, 2008.
177. "SORE Component EO Summary," email from Jim Watson, California Air Resources Board, to Mike
Samulski, U.S. EPA, August 18, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0405.
178. Letter from William Guerry, Kelly Drye Collier Shannon, to Glenn Passavant, U.S. EPA, and Susan Bathalon,
U.S. Consumer Product Safety Commission, "EPA Proposal, Phase III, Small Engine Regulations - Fuel Lines for
Cold Weather Handheld Products," January 25, 2007, Docket Identification EPA-HQ-OAR-2004-0008-0524.
179. Outdoor Power Equipment Institute, "Fuel Hose Design in Relation to the Design and Operation of Handheld
Power Equipment," Presentation to EPA, June 21, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0399.
180. Outdoor Power Equipment Institute, "Comparison of NBR and FPM (Viton) Performance on Hand Held
Equipment," Presentation to EPA, Docket Identification EPA-HQ-OAR-2004-0008-0441.
181. "Alpha List of OEM Approvals," http://www.precixinc.com/oemlist.html, downloaded August 15, 2006,
Docket Identification EPA-HQ-OAR-2004-0008-0411.
182. "Viton: Volume Swell of Viton® Fluoroelastomers in Fuel C / Ethanol Blends," DuPont Performance
Elastomers, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0510.
5-136
-------
Feasibility of Evaporative Emission Control
183. "FKM / Genuine Viton®," http://www.o-ring.info/en/datasheets/fkm-genuine-viton/, downloaded August 15,
2006, Docket Identification EPA-HQ-OAR-2004-0008-0410.
184. Hatano, M, Otsuka, M, Ogata, C., Suetsugu, N, Amemiya, T., Kudo, M., "Development of New FKM O-
Rings with Superior Fuel-Oil Resistance and Low-Temperature Properties," SAE Paper 2005-01-1743, 2005, Docket
Identification EPA-HQ-OAR-2004-0008-043 3.
185. Yamaguchi, Y., Shojima, D., Ogata, C., Oya, K., Kawasaki, K., Kudo, M., "Development of the New FKM O-
Ring for a Cooling Device," 2005-01-1753, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0434.
186. "Technical Information; Dyneon Fluorelastomer LTFE 6400X," 3M, Issued February, 2003, Docket
Identification EPA-HQ-OAR-2004-0008-0521.
187. "Relative Permeation of Elastomers," Precix, downloaded from www.precixinc.com on February 14, 2007,
Docket Identification EPA-HQ-OAR-2004-0008-0525.
188. Balzer, J., "F-TPV: A New Horizon for TPE Technology," Presented by Daikin America to U.S. EPA on
February 14, 2007, Docket Identification EPA-HQ-OAR-2004-0008-0522.
189. Letter from Trevor J. Richards, President Spring Cove Marina, "Fuel Spills While Filling Pleasure Craft
Tanks," April 23, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0383.
190. Letter from Van DePiero, City of Pittsburg, California to Steve Burkholder, Enviro Fill, April 19, 2006, Docket
IdentificationEPA-HQ-OAR-2004-0008-0384.
191. American Boat and Yacht Council, "H-24 Gasoline Fuel Systems," 2005, Docket Identification EPA-HQ-
OAR-2004-0008-0429.
192. Email from Steve Burkholder, Enviro-Fill, to Mike Samulski and Alan Stout, U.S. EPA, "Overfill Prevention
System Developments," July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0447.
193. Boat U.S. Foundation, "Foundation Findings #40; Spill? What Spill?; Products to Keep Fuel Where it
Belongs-In Your Tank," www.boatus.com/foundation/Findings/fmdings40, March 2005, Docket Identification
EPA-HQ-OAR-2004-0008-0431.
194. Letter from Steve Burkholder, Enviro-Fill, to Mike Samulski, U.S. EPA, March 1, 2006, Docket Identification
EPA-HQ-OAR-2004-0008-0430.
195. Letter from Steve Burkholder, Enviro-Fill, to Mike Samulski, U.S. EPA, January 13, 2006, Docket
Identifications? A-HQ-OAR-2004-0008-0446.
196. Email from Steve Burkholder, Enviro-Fill, to Mike Samulski and Alan Stout, U.S. EPA, "Overfill Prevention
System Developments," July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0447.
197. SAE Recommended Practice J30, "Fuel and Oil Hoses," June 1998, Docket Identification EPA-HQ-OAR-
2004-0008-0176.
198. SAE Surface Vehicle Standard, "Marine Fuel Hoses," Society of Automotive Engineers J 1527, Issued 1985-
12, Revised 2007-01.
199. American Society for Testing and Materials, "Standard Test Method for Rubber Property-Effects of Liquids,"
ASTM D 471-06, November 2006, Docket Identification EPA-HQ-OAR-2004-0008-0511.
5-137
-------
Final Regulatory Impact Analysis
200. SAE Recommended Practice J1737, "Test Procedure to Determine the Hydrocarbon Losses from Fuel Tubes,
Hoses, Fittings, and Fuel Line Assemblies by Recirculation,"1997, Docket Identification EPA-HQ-OAR-2004-0008-
0178.
201. SAE Surface Vehicle Standard, "Nonmetallic Fuel System Tubing with One or More Layers," Society of
Automotive Engineers J 2260, Issued 1996-11, Docket Identification EPA-HQ-OAR-2004-0008-0180.
202. SAE Surface Vehicle Recommended Practice, "Test Procedure to Determine the Hydrocarbon Losses from
Fuel Tubes, Hoses, Fittings, and Fuel Line Assemblies by Recirculation," Society of Automotive Engineers J 1737,
Issued 1997-08, Docket Identification EPA-HQ-OAR-2004-0008-0178.
203. Draft SAE Information Report J1769, "Test Protocol for Evaluation of Long Term Permeation Barrier
Durability on Non-Metallic Fuel Tanks," Docket Identification EPA-HQ-OAR-2004-0008-0195.
204. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
205. "Permeation Rates of Small Off Road Engine High-Density Polyethylene Fuel Tanks (April 2001 Testing),
June 8, 2001, California Air Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0146.
206. "Permeation Rates of High-Density Polyethylene Fuel Tanks (June 2001), June 12, 2001, California Air
Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0148
207. "Permeation Rates of High-Density Polyethylene Fuel Tanks (May 2001), June 11, 2001, California Air
Resources Board, Docket Identification EPA-HQ-OAR-2004-0008-0149.
208. "Average Permeation Rates of High-Density Polyethylene Off-Road Equipment Fuel Tanks Using a Surface
Area Approach, (March 2002)," March 15, 2002, California Air Resources Board, Docket Identification EPA-HQ-
OAR-2004-0008-0256.
209. Haskew, H., McClement, D., "Fuel Permeation from Automotive Systems; Final Report," prepared for the
California Air Resources Board and the Coordinating Research Council, CRC Project No. E-65, September 2004,
Docket Identification EPA-HQ-OAR-2004-0008-0151.
210. Dickson, A., Goyet, C., "Handbook of Methods forthe Analysis of the Various Parameters of the Carbon
Dioxide System in Sea Water; Version 2," Prepared for the U.S. Department of Energy, SOP21 "Applying air
buoyancy corrections," September 29, 1997, Docket Identification EPA-HQ-OAR-2004-0008-0135.
211. Testing Services Group, "CARD TM-513: Portable Fuel Container Permeation Testing," Presented to U.S.
EPA on August 18, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0133.
212. Testing Services Group, "Technical Brief: Buoyancy Correction for Mass Measurement of Fuel Containers,"
TB 50819-1, August 19, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0134.
213. California Air Resources Board, "TP-502: Test Procedure for Determining Diurnal Emissions from Portable
Fuel Containers," proposed July 22, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0300.
214. "Test Method to Measure Fluid Permeation of Polymeric Materials by Speciation," SAE Surface Vehicle
Standard J2659, Issued December 2003, Docket Identification EPA-HQ-OAR-2004-0008-0512.
215. Internal MeadWestvaco memorandum from M. Tschantz to C. Pierson, "Preliminary 3 mm SeaGuard™
Specifications," August 31, 2005, Docket Identifications?A-HQ-OAR-2004-0008-0136.
5-138
-------
Feasibility of Evaporative Emission Control
216. "Standard Test Method for Determination of Butane Working Capacity of Activated Carbon," ASTM
International, Designation D5228-92 (Reapproved 2005), Docket Identification EPA-HQ-OAR-2004-0008-0347.
217. Tschantz, M, "Activated Carbon for Use in Marine Evaporative Control Applications," Presented by
MeadWestvaco Corporation at the 2004 International Boatbuilders Exposition, October 25, 2004, Docket
IdentificationEPA-HQ-OAR-2004-0008-0040.
218. "Carbon Specs," email fromM. Tschantz , MeadWestvaco to M. Samulski, U.S. EPA, November 18, 2005,
Docket Identification EPA-HQ-OAR-2004-0008-0421.
219. "Standard Test Method for Ball-Pan Hardness of Activated Carbon," ASTM International, Designation D3802-
79 (Reapproved 2005), Docket Identification EPA-HQ-OAR-2004-0008-0348.
220. "SOP 0960-610 vs. ASTM 3802," email from Mike Tschantz, MeadWestvaco to Mike Samulski, U.S. EPA,
January 13, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0412.
221. " Standard Test Method for Particle Size Distribution of Granular Activated Carbon," ASTM International,
Designation D2862-97 (Reapproved 2004).
222. Williams, R., Clontz, R., "Impact and Control of Canister Bleed Emissions," SAE Paper 2001-01-0733, 2001,
Docket Identification EPA-HQ-OAR-2004-0008-0432.
223. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
224. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
225. Hopf, G., Ries, H., Gray, E., "Development of Multilayer Thermoplastic Fuel Lines with Improved Barrier
Properties," SAE Paper 940165, 1994, Docket Identification EPA-HQ-OAR-2004-0008-0174.
226. Stahl, W., Stevens, R., "Fuel-Alcohol Permeation Rates of Fluoroelastomers, Fluoroplastics, and Other Fuel
Resistant Materials," SAE Paper 920163, 1992, Docket Identification EPA-HQ-OAR-2004-0008-0171.
227. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01, June
17, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0208.
228. Goldsberry, D., "Fuel Hose Permeation of Fluoropolymers," SAE Paper 930992, 1993, Docket Identification
EPA-HQ-OAR-2004-0008-0173.
229. Tuckner, P., Baker, J., "Fuel Permeation Testing Using Gravimetric Methods," SAE Paper 20001-01-1096,
2000, Docket Identification EPA-HQ-OAR-2004-0008-0160.
230. Fuller, R., "Unique Low Permeation Elastomeric Laminates for Fuel Hose; F200 Technology used in
Automotive Fuel Hose," Hearing Testimony for Dupont-Dow Elastomers on NPRM, October 7, 2002, Docket
Identification EPA-HQ-OAR-2004-0008-0197.
231. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
5-139
-------
Final Regulatory Impact Analysis
232. Duchesne, D., Hull, D., Molnar, A., "THV Fluorothermoplastics in Automotive Fuel Management Systems,"
SAE Paper 1999-01-0379, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0159.
233. Nulman, M, Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
234. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of Polymeric
Materials," SAE Paper 2001-01-1999, 2001, Docket Identification EPA-HQ-OAR-2004-0008-0162.
235. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01, June
17, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0208.
236. Facsimile from Bob Hazekamp, Top Dog Systems, to Mike Samulski, U.S. EPA, "Permeation of Polyurethane
versus THV Materials @ 60°C," January 14, 2002, Docket A-2000-01, Docket Identification EPA-HQ-OAR-2004-
0008-0435.
237. Duchesne, D., Hull, D., Molnar, A., "THV Fluorothermoplastics in Automotive Fuel Management Systems,"
SAE Paper 1999-01-0379, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0159.
5-140
-------
-------
Costs of Control
CHAPTER 6: Costs of Control
This chapter describes our approach to estimating the cost of complying with the new
emission standards. We start with a general description of the approach used to estimate costs,
then describe the technology changes we expect and assign costs to them. We also present an
analysis of the estimated aggregate cost to society.
6.1 Methodology
We developed the costs for individual technologies using estimates from ICF
Incorporated1, conversations with manufacturers, and other information as cited below. The
technology characterization reflects our current best judgment based on EPA's technology
demonstrations, engineering analysis, information from manufacturers, and the published
literature.
Costs of control include variable costs (for incremental hardware costs, assembly costs,
and associated markups) and fixed costs (for tooling, R&D, and certification). Variable costs are
marked up at a rate of 29 percent to account for the engine or equipment/vessel manufacturers'
overhead and profit.2 For technologies sold by a supplier to the engine manufacturers, an
additional 29 percent markup is included for the supplier's overhead and profit. Labor estimates
are marked up by 100 percent to reflect fringe and overhead charges including management,
supervision, general and administrative expenses, etc. All costs are in 2005 dollars.
The analysis presents an estimate of per-unit costs that will occur in the first year(s) of
new emission standards and the corresponding long-term costs. Long-term costs decrease due to
two principal factors. First, fixed costs are assessed for five years, after which they are fully
amortized and are then no longer part of the cost calculation. Second, manufacturers are
expected to learn over time to produce the engines with the new technologies or aftertreatment at
a lower cost. Consistent with analyses from other programs, we reduce estimated variable costs
by 20 percent beginning with the sixth year of production.3 The small spark ignited engine
industry and the marine industry have different reasons for the learning.
Learning for the Small SI industry is expected to occur in the catalyst muffler designs. It
will likely occur for two reasons: 1) over time the number of different muffler catalyst designs
may be reduced thereby decreasing substrate costs due to larger ordering volumes. 2) heat shield
manufacturing may become automated and/or designs more uniform. Learning will not occur for
other technologies such as electronic fuel injection systems for they currently exist on some
Small SI equipment and motorized vehicles such as scooters .
In the marine industry, manufacturers are less likely to put in the extra R&D effort for
low-cost manufacturing of engine families of relatively low sales volumes. Learning will occur
in two basic ways. As manufacturers produce more units, they will make improvements in
production methods to improve efficiency. The second way learning occurs is materials learning
where manufacturers reduce scrap. Scrap includes units that are produced but rejected due to
6-1
-------
Final Regulatory Impact Analysis
inadequate quality and material scrap left over from the manufacturing process. As production
starts, assemblers and production engineers will then be expected to find significant
improvements in fine-tuning the designs and production processes.
We believe it is appropriate to apply this learning factor here for the marine industries,
given that they are facing new emission regulations, some for the first time, and it is reasonable
to expect learning to occur with the experience of producing and improving emission-control
technologies. Manufacturers do not have significant experience with most of the emissions
controls that are anticipated for meeting the standards.
Many of the engine technologies available to Marine SI and Small SI engine
manufacturers to control emissions also have the potential to significantly improve engine
performance. This is clear from the improvements in automotive technologies. As cars have
continually improved emission controls, they have also greatly improved fuel economy,
reliability, power, and a reduced reliance on regular maintenance. Similarly, the fuel economy
improvements associated with converting from two-stroke to four-stroke engines is well
understood. We attempt to quantify these expected improvements for each type of engine below.
Even though the analysis does not reflect all the possible technology variations and
options that are available to engine manufacturers, we believe the projections presented here
provide a cost estimate representative of the different approaches manufacturers may ultimately
take. We expect manufacturers in many cases to find and develop approaches to achieve the
emission standards at a lower cost than we describe in this analysis.
6.2 Exhaust Emission Control Costs for Small SI Engines
This section presents our cost estimates for meeting the new exhaust emission standards
for Small land-based spark-ignition (Small SI) engines. EPA has relied upon model year 2008
certification data for this analysis to characterize the current Class I and Class II market and the
technology mix needed to comply with the Phase 3 standards. EPA chose not include data from
Chinese manufacturers in this analysis because we have no information on actual sales of their
engines in the United States. Manufacturers do submit sales estimates to EPA at the time of
certification. However, the sales estimates provided by Chinese manufacturers would suggest
that sales of Small SI nonhandheld engines have doubled over the last few years. Based on
discussions with nonhandheld engine manufacturers that have been certifying with EPA for over
ten years now, we do not believe this is the case and it is our understanding that sales of
nonhandheld engines from Chinese manufacturers are relatively small at this time. Therefore,
we believe it is appropriate to not include certification data from Chinese manufacturers in our
analysis.
In 1995, EPA finalized the first regulations for reducing emissions from small spark
ignited (SI) engines <19kW. Small spark ignited engine designs include side valve and overhead
valve engine configurations designated in two groups by engine displacement. Class I engines
are <225cc and Class II engines are >225cc and less than 19kW. The Phase 2 regulations for
these engines were set with the expectation that Class I side valve engines would be converted to
6-2
-------
Costs of Control
overhead valve design. Certification data from 2008 shows that engine manufacturers have been
able to achieve Phase 2 certification with the continued use of side valve engines in some cases.
A summary of the 2008 technology market mix is presented in Table 6.2-1.
For the final Phase 3 standards, the EPA 2008 certification database was referenced. It
was found that the majority of Class I engines were in need of some emission reduction and
therefore it is estimated that these engines would use catalysts and the related engine design
improvements required to use catalysts safely. For Class II engines, the 2008 certification
database revealed some engine families meet the Phase 3 emission levels and therefore
technologies are not required on all engine families. For those engine families needing emission
reduction technology, different technologies were assigned depending on whether the engine was
a one cylinder or a multiple cylinder engine. A number of one cylinder engine families were
estimated to use catalysts. For two or more cylinders, the largest engine family per engine
manufacturer needing emission reduction technology was assigned closed loop electronic fuel
injection. The remainder were assigned catalysts with the appropriate muffler setup. The
expected technology market mix is presented in Table 6.2-2.
Table 6.2-1: 2008 Technology Market Mix
sv
OHV
w/ Catalyst
w/ Other (EFI and/or watercooled)
Class I
66%
34%
0.003%
0
Class II
2%
98%
0.4%
1%
Table 6.2-2: Technology Market Mix Expectations for Phase 3 Engines
HC+NOx Emission Standards: 38% Reduction Class I, 34% Reduction Class II*
Exhaust Standard Implementation Date
SV
OHV
w/ Catalyst
w/ Other (EFI and/or watercooled)
2012
Class I
66%
34%
95%
0
2011
Class II
2%
98%
50%
6.6%
*EPA 2008 certification data
The following sections describe the technologies and related variable and fixed costs
followed by an analysis of aggregate costs. The costs are based on a report from ICF
International entitled "Small SI engine Technologies and Costs."4 Variable costs to the
manufacturers vary with the engine size and the emission technologies considered.
6-3
-------
Final Regulatory Impact Analysis
Manufacturers prices of all components were estimated from various sources including
information from engine and catalyst manufacturers and previous work performed by ICF
International on spark ignited engine technology. All hardware costs to the engine
manufacturers are subject to a 29 percent mark-up. This includes manufacturer overhead, profit,
dealer overhead and profit. A separate supplier markup of 29 percent is also applied to items
typically purchased from a suppliers such as fuel injection and catalysts. A 5 percent warranty
mark-up is added to hardware cost of specific technologies including electronics, to represent an
overhead charge covering warranty claims associated with new parts.
Fixed costs to the manufacturer include the cost of researching, developing and testing a
new technology. The cost of retooling the assembly line for the production of new parts as well
as engine certification including durability testing are also fixed costs. Design and
development fixed costs per month are listed in Table 6.2-3. Tooling and specific R&D costs are
listed in the following sections. Fixed costs for certification are listed in Section 6.2.3.
Table 6.2-3: Design and Development Costs
for use in Fixed Cost Estimates per Month 5
Hours
Rates
Costs
Design Costs Per Month
Engineer
160
$64.41
TOTAL Design Costs Per Month
$10,306
$10,306
Development Costs Per Month
Engineer
Technicians
Dynamometer Test
Time
160
320
20 tests
$64.41
$41.87
$250 ea
TOTAL Development Costs Per Month
$10,306
$13,398
$5,000
$28,704
6.2.1 Class I
Class I engines currently emitting at or below the Phase 2 emission standard of 16.1
g/kWh will need to reduce their engine out HC+NOx emissions by 30-50 percent to comply with
the Phase 3 emission standard of 10 g/kWh with an appropriate margin. A number of Class I
side valve (SV) engines have been redesigned for the Phase 1 and Phase 2 rulemakings, however
SV and overhead valve (OHV) engines will need a different approach to meet these emission
standards. One technology to reduce emissions to the Phase 3 levels is a three way catalyst with
appropriate precious metal loading for minimal CO conversion. EPA work has shown that
catalysts can function effectively through a dynamometer aging of 125 hours with a catalyst
conversion of about the same amount at high hours as low hours6. The amount of conversion is
6-4
-------
Costs of Control
only constrained by 1) the size of the catalyst to fit in the existing, or slightly larger, muffler, 2)
residence time of the exhaust gas along with 3) muffler surface and exhaust gas temperature
issues with respect to the amount of CO converted within a catalyst. EPA's work has been
shown to convert HC+NOx within a range of 3.8-6.7 g/kW-h (median approx 5.7g/kW-h) on
OHV engines and 3.8-10.3 g/kW-h on SV engines (median of 6.8 g/kW-h).
EPA's 2005 Phase 2 certification database lists OHV and SV engine HC+NOx emission
levels at low hours, a deterioration factor (df) and resultant certification levels. Engine
manufacturers with most regulated experience were considered for these df ranges for we are
most familiar with the performance of these engines. Engine families using credits to certify to
the emission standard with ABT were not included.
Table 6.2-4: 2005 EPA Certification Database with Catalyst Assumptions7
Technology
Type/UL
SV/125
OHV/ 125
OHV/250
OHV/500
Engine Out
"zero
hours"
(Min-Max)
10-11
6-15
7-15
8-14
DF
(Min-Max)
1-1.24
1-1.356
1-1.136
1-1.161
Certification
Level
(Min-Max)
13-14
9-16
8-12
8-15
Catalyst
conversion
(median
from EPA
work)
6.8
5.7
5.7
5.7
Engine with Catalyst
6.2-7.2
3.3-10.3
2.3-6.3
2.3-9.3
Table 6.2-4 is based on median HC+NOx catalyst conversion from EPA test work in the
Safety Study.8 The Safety Study also shows improvements in the cooling system design will
provide cooling to the engine and/or catalyst muffler system for reduced muffler skin
temperatures. Individual engine family applications will vary and engine improvements may be
required for durable and effective catalyst operation.
6.2.1.1 Engine Improvements for Class I
Improvements in engine combustion efficiency and engine cooling will assure the engine
systems support catalyst durability. Engine improvements for durable catalyst operation include
changes that are fixed costs and variable costs. Improvements in engine systems resulting in
fixed costs potentially include the following: 1) improved combustion chamber design for
optimized combustion, 2) improved piston design for reduced crevice volumes and reduced HC
emissions, 3) improved machining and casting tolerances for all combustion chamber
components, 4) improved cylinder head fin design for improved cooling, and 5) improved
carburetion for fuel delivery and system durability. Some engines would also benefit greatly
from 6) improved flywheel design in order to provide additional cooling to the engine and
muffler system. Clearly not all engines need these upgrades and many will implement few or
6-5
-------
Final Regulatory Impact Analysis
none.
Fixed costs per engine family for engine improvements are estimated at four months of design
work (one engineer) and six months of development work (one engineer, one technician and
dynamometer test time) along with tooling costs for the cylinder head, piston, connecting rod,
camshaft, carburetor, flywheel and setup changes. Tooling costs are estimated to be the same
across engine useful life categories with the exception of Class 1125 hour SV engines which
contains some engine families that are sold in much larger volumes and therefore would have
more tools to be modified. These fixed costs are presented in Table 6.2-5.
Table 6.2-5: Fixed Costs for Engine Improvements for Class I9
Engine Class
Useful life (his)
Valving
Class I
125
SV
125,250,500
OHV
R&D
Design (4 months)
Development (6 months)
TOTAL R&D per Engine Line
41,225
172,225
213,450
41,225
172,225
213,450
TOOLING COSTS
Cylinder Head
Piston
Connecting Rod
Camshaft
Carburetor
Flywheel
Setup Changes
TOTAL TOOLING per Engine Line
TOTAL FIXED
50,000
50,000
30,000
16,000
120,000
70,000
150,000
486,000
$699,450
25,000
25,000
15,000
8,000
60,000
35,000
75,000
243,000
$456,450
Variable cost items were identified from EPA field aging of engines from several engine
manufacturers. EPA performed several lawnmower in-use test programs in 2003 to 2005.
Several of the SV and OHV engines were equipped with catalysts. The process revealed that
potentially several engine design characteristics needed improvement in some cases in order for
catalysts to be successfully applied in-use. Items included: 1) fuel filter to screen out impurities
(assure do not encounter a stuck float and thereby excessive fuel flowed through the engine
6-6
-------
Costs of Control
coating the catalyst and rendering it inactive.), 2) incorporation of an intake gasket to assure
leaks do not develop in the intake system thereby resulting in hot engine operation and a number
of engine operational issues, 3) engine shroud screen over fan (avoid debris collecting in the
engine fan), and 4) improved engine cooling system for SV engines to assure the engine's piston
and combustion chamber walls stay in contact so oil does not seep past the rings and into the
combustion chamber (see Chapter 4) thereby potentially poisoning the catalyst. Lastly, the
incorporation of improved induction coils will reduce the opportunity for spark plug wire
failures and misfire events. Table 6.2-6 lists the variable costs for engine improvements for
Class I engines certified to various useful lives. Clearly not all engines need these upgrades to
succeed and many will implement few or none.
Table 6.2-6: Variable Costs for Engine Improvements for Class I
10
Engine Improvement
Fuel Filter Screens (80% of engine sales)
cost/engine: 0.02
Improved Intake Gaskets (75% of engine sales for
Class I 125 hour useful life)
cost/engine: 0.03
Screen over cooling fan (16% of 125 hr Class I)
cost/engine: 0.45
Larger Induction Coils (all)
Engine Manufacturer Cost
TOTAL w/Markup
29% OEM
Learning Curve w/ 29% Markup
(0.8*Total w/Markup)* 1.29
UL 125
SV
0.02
0.02
0.07
0.10
0.21
0.27
0.22
UL 125
OHV
0.02
~
0.07
0.10
0.19
0.24
0.19
UL250
0.02
~
~
0.10
0.12
0.15
0.12
UL500
~
~
~
0.10
0.10
0.13
0.10
6.2.1.2 Catalysts for Class I
The following paragraphs describe details on catalysts substrates, washcoat and precious
metal, and muffler shielding for Class I engines. Although commonly in use today, spark
arresters are discussed in the context of the overall design.
Based on catalyst/muffler development and emission testing by EPA (2004-2005), an
engine which has an HC+NOx exhaust ratio of 60/40 is best suited for the use of a catalyst in
Small SI engines for the catalyst can be designed for minimal CO oxidation and related heat
generation. This ratio can be found on OHV engines for they have efficient combustion
chambers. SV engines require slightly larger catalysts due to their less efficient combustion
chambers and less than optimum HC/NOx ratios. In addition, SV engines are more likely to
6-7
-------
Final Regulatory Impact Analysis
have oil seep past the piston rings due into the exhaust to cylinder distortion. A longer catalyst,
or the use of a pipe catalyst prior to the brick catalyst, allows it to survive for the full useful life
for the catalyst is poisoned from the front of the catalyst to the back. According to the EPA
Phase 2 certification database, Class I SV engine families are certified to the 125 hour useful life
and therefore the cost analysis includes two different catalyst costs for the 125 hour useful life.
The engines certified to the 250 and 500 useful life categories are all of OHV engine
design. As with the 125 hour category, catalyst substrate sizes are calculated as a percentage of
the engine displacement. The certification database was queried for this engine displacement
data and the displacements are sales weighted, as seen in Table 6.2-7. Catalyst volumes range
from 18 percent of the engine displacement for the 125 OHV useful life to 50 percent of the
engine displacement for the 500 hour useful life. Larger catalysts are needed for longer useful
life periods in order to provide the emission conversion durability. Specific costs for engines
within each useful life category will differ.
The substrate cost is based on an average cost of metallic and ceramic substrates as
presented in the ICF report11 due to the variety of Small SI equipment types and variety of
catalysts offered in the marketplace. This cost analysis estimates equal weighting of the substrate
types and therefore takes an average of the cost for both metallic and ceramic.
Due to the concern of oil sulfur poisoning in Class I engines, EPA envisions that a 5:1
ratio of Platinum/Rhodium precious metal would be used for these catalysts. The cost of
precious metals was taken from a 3 year average in price from 2003-2005. Washcoat material is
expected to be a 30/70 percent mixture of cerium and alumina oxide, respectively.
The design of the catalyst/muffler forms the basis for the degree of cooling needed at the
muffler and exhaust port. EPA's solution for muffler surface and exhaust gas cooling included
three steps 1) forcing the cooling air from the engine fan/cylinder head region to the muffler can
be achieved through a slight redesign of the engine's shroud, 2) a muffler shroud that is designed
to guide the cooling air around the entire muffler and exits at a specified location, and lastly 3)
and if when needed an ejector is added to the muffler at the exhaust gas outlet so the exhaust gas
can be combined with ambient air before being accessible to the user.
EPA's observation of a number of lawnmower engine designs revealed that the majority
of heat shields currently used on small engines need to be redesigned in order to allow the use of
air flow from the engine's fan to flow optimally around the muffler for cooling. The portion of
engines that do have such systems and will not incur this cost were removed from the cost
analysis and ICF's estimates for this technology were adjusted. EPA utilized the 2005
certification database to estimate sales and to calculate a percentage of engines that will be
estimated to redesign their muffler heat shield. Table 6.2-7 contains the variable costs for
catalysts, heat shields and spark arresters.
-------
Costs of Control
Table 6.2-7: Variable Catalyst Costs for Class I12
to Achieve the Phase 3 Standards
Useful Life
Engine Power (hp)
Engine Displacement (cc)
Catalyst Volume (cc)
Substrate Diameter (cm)
Substrate
Washcoat and Precious
Metal
Labor
Labor Overhead 40%
Supplier Markup 29%
Catalyst Manufacturer Price
Heat Shield*
Spark Arresters
Engine Manufacturer Cost
TOTAL w/Markup
29% OEM
UL125
SV
3.3
178
45
3.50
$1.97
$1.83
$1.40
$0.56
$1.67
$7.43
$0.50
$0.05
$7.98
$10.29
UL125
OHV
5.1
180
32
3.50
$1.53
$1.31
$1.40
$0.56
$1.39
$6.19
$0.29
$0.05
$6.53
$8.42
UL250
5.0
167
55
4.00
$2.32
$2.81
$1.40
$0.56
$2.06
$9.15
$0.18
$0.05
$9.38
$12.10
UL500
5.2
166
83
4.50
$3.22
$4.24
$1.40
$0.56
$2.73
$12.15
$0.14
$0.05
$12.34
$15.92
* Based on EPA's work with small engine equipment from 2003-2005, it has been observed that
some manufacturers have heat shielding that is sufficient or only needs slight modification.
These sales volumes have been removed and the resultant price recalculated.
The fixed costs related to catalyst development for Class I engine applications include
design (one engineer), of two months, and development (one engineer, one technician and
dynamometer time), for five months, of the muffler and heat shield. The inside of the muffler is
to be redesigned to house the catalyst, provide supplemental air when needed, and provide
baffling for the exhaust flow in order to maximize heat dissipation from the exhaust flow. The
muffler stamping will also need to be updated to account for the new design. A second critical
component of the catalyst/muffler system is the heat shield. The heat shield must be designed to
allow cooling air from the fan to flow around the muffler to maximize cooling of the muffler and
then exit at an optimum point. The muffler/heat shield system must be located at a
predetermined distance from the engine block in order to allow air to flow behind the muffler to
cool the backside. Setup changes also are incurred with these modified stampings. The total
tooling per engine line is estimated at $240,000 for Class I engines of 125 hour useful life and
6-9
-------
Final Regulatory Impact Analysis
$120,000 for Class I engines of other useful life periods. The difference is due to the additional
tooling for high volume SV engine families. Table 6.2-8 presents the fixed costs associated with
using catalysts on Class I engines.
Table 6.2-8: Fixed Costs for Catalysts for Class I Engines
13
Engine Class
Useful life (hrs)
Valving
Class I
125
SV
125, 250, 500
OHV
R&D
Design
(2 months)
Development (5 months)
TOTAL R&D per Engine Line
20,612
143,521
164,133
20,612
143,521
164,133
TOOLING COSTS
Modified Muffler Stamping
Heat Shield Stamping
Engine Shroud Modification
Setup Changes
TOTAL TOOLING per Engine Line
TOTAL FIXED COSTS
100,000
60,000
30,000
50,000
240,000
$404,133
50,000
30,000
15,000
25,000
120,000
$284,133
A learning curve of 20 percent is applied to costs for catalyst technology starting in the
sixth year after the standard is implemented. This somewhat conservative since the learning
normally occurs at 20 percent with a doubling of production which would thus be in the third or
fourth year. Optimized catalyst/muffler designs and manufacturing processes will likely be
developed as the industry becomes experienced in using mufflers with catalysts on Small SI
engines. The muffler washcoat will still be unique per engine family per engine manufacturer
for engine out emissions will differ. Table 6.2-9 presents the estimated learning curve impacts
on variable costs. The precious metal prices are determined in the marketplace and therefore
would not be affected by the learning curve.
6-10
-------
Costs of Control
Table 6.2-9: Learning Curve Variable Catalyst Costs for Class I
to Achieve the Phase 3 Standards
Useful Life
Engine Power (hp)
Engine Displacement (cc)
Catalyst Volume (cc)
Substrate Diameter (cm)
Substrate
Washcoat and Precious
Metal
Labor
Labor Overhead
Supplier Markup 29%
Manufacture Price
Heat Shield
(adjusted % for eng w/
sufficient heat shield)
Flame/Spark Arrester
Hardware Cost to
Manufacturer
w/Markup
29% OEM
UL 125
-SV
3.3
178
45
3.50
$1.57
$1.83
$1.40
$0.56
$1.55
$6.92
$0.40
$0.05
$7.37
$9.50
UL 125
-OHV
5.1
180
32
3.50
$1.22
$1.31
$1.40
$0.56
$1.30
$5.80
$0.23
$0.05
$6.08
$7.84
UL250
5.0
167
55
4.00
$1.86
$2.81
$1.40
$0.56
$1.92
$8.55
$0.14
$0.05
$8.74
$11.28
UL500
5.2
166
83
4.50
$2.58
$4.24
$1.40
$0.56
$2.55
$11.32
$0.11
$0.05
$11.49
$14.82
Table 6.2-10 contains the estimated total costs for Class I Phase 2 compliant engines to
meet the Phase 3 emission standards. Near term costs are those costs for the first five years.
Long term costs are those costs to which the learning curve has been applied.
6-11
-------
Final Regulatory Impact Analysis
Table 6.2-10: Class I Estimated Total Costs Per Engine (Variable) and
Per Engine Family (Fixed) to Achieve the Phase 3 Standards
Useful Life
Engine Displacement (cc)
Catalyst Volume (cc)
Substrate Diameter (cm)
UL 125 -
SV
178
45
3.50
UL 125 -
OHV
180
32
3.50
UL250
167
55
4.00
UL500
166
83
4.50
Variable Costs - Near Term
Engine Improvements
Catalyst
Total Variable Cost (Near)
$0.27
$10.29
$10.56
$0.24
$8.36
$8.60
$0.15
$12.10
$12.25
$0.13
$15.92
$16.05
Variable Costs - Long Term (with Learning)
Engine Improvements
Catalyst
Total Variable Cost (Long)
$0.22
$9.50
$9.72
$0.19
$7.84
$8.04
$0.12
$11.28
$11.39
$0.10
$14.82
$14.92
Fixed Costs
Engine Improvements
Catalyst
Total Fixed Costs
$699,450
$404,133
$1,103,583
$456,450
$284,133
$740,583
$456,450
$284,133
$740,583
$456,450
$284,133
$740,583
6.2.2 Class II
The Phase 3 HC+NOx emission standard for Class II is 8 g/k-Wh which is a 34 percent
emission reduction from the Phase 2 standards of 12.1 g/k-Wh. This standard is to be met at the
end of the regulatory useful life for each engine family. The EPA Phase 2 certification database
shows that the majority of engines in this Class are of OHV design however, approximately 2
percent of the engines are still side valve engine technology.
Class II side valve engines are currently certified to the Phase 2 standards with credits
from lower emitting OHV engines. The EPA 2005 certification database shows the majority of
overhead valve engines currently certifying HC+NOx at a range of 7-11 g/kW-h and side valve
engines certifying in the range of 13-14 g/kW-h. Lowering of the emission standard will reduce
the number of emission credits available for side valves to certify and therefore, it is assumed
that the remaining side valve engines will be phased out and replaced with currently produced
overhead valve engines or continue to be certified using ABT credits from a limited number of
6-12
-------
Costs of Control
lower emitting engine families.
Assuming a 2 g/kW-h compliance margin to 6 g/kW-h, emission reduction technologies
will need to be designed to reduce emissions 15-57 percent. Table 6.2-11 illustrates potential
engine out emissions with emission reduction technologies applied to Phase 2 engines. OHV
engines are expected to potentially include some engine improvements and/or catalysts or
electronic fuel injection.
Table 6.2-11: 2005 EPA Certification Database Summary With Catalyst Assumptions14
UL
OHV
250
500
1000
Engine Out
"zero hours"
(Min-Max)*
4.8-10.0
Median: 7.9
4.4-10.8
Median: 8.3
6.0-11.2
Median: 8.4
DF
(Min-Max)**
1-1.7
Median: 1.137
1-1.6
Median: 1.039
1-1.4
Median: 1.03
Certification
Level
(Min-Max)*
6.7-12.0
Median: 8.9
5.9-10.9
Median: 9.5
6.9-11.2
Median: 8.9
Catalyst
conversion
(non-EFI
engine)15
4.0
4.0
4.0
Engine with Catalyst
(Based on Median
values)
2.7-8.0
1.9-6.9
2.9-7.2
* Values of engines that meet the standard. 500 hr UL has a liquid cooled engine with catalyst that meets
a 2.6 g/kW-h HC+NOx and 1000 hr UL has the same that meets 1.8 g/kW-h HC+NOx.
**Some engines have catalysts and therefore claim a higher df
Class II contains several liquid cooled engines. These engines likely have the ability to
be enleaned to more of a degree due to the additional cooling assistance and therefore may not
need a catalyst to meet the Phase 3 emission standards.
6.2.2.1 Engine Improvements for Class II
Engine improvements include improved engine design and larger induction coils as
shown in Tables 6.2-12 and 6.2-13. Improvements in engine design will allow for more efficient
combustion and a more favorable HC:NOx ratio for the use of a reducing catalyst. A larger
induction coil will reduce the opportunity for spark plug wire failure and misfire events. It is
estimated that 1000 hour engines currently have sufficient induction coils and will not need this
improvement.
6-13
-------
Final Regulatory Impact Analysis
Table 6.2-12: Variable Costs for
Engine Improvements for Class II per Engine16
Larger Induction Coils
TOTAL
w/Markup 29% OEM
Learning w/29% OEM
(0.8*Total)*1.29
UL250
0.09
0.12
0.10
UL500
0.09
0.12
0.10
UL 1000
~
~
~
Improved engine design includes machining and casting tolerances, improved
combustion chamber configuration, reduced crevice volumes, better cooling (improved fin
design on cylinder head and oil control), improved flywheel design and improved carburetion.
Better carburetor performance is needed to assure floats do not stick and better cooling so
engines operate at cooler temperatures. Fixed costs include design (one engineer at 4 months),
development and tooling costs (one engineer, one technician and dynamometer time for 6
months) per engine family to achieve improved engine design. Projected fixed costs are
presented in Table 6.2-13. The fixed cost is estimated to be the same per engine family and is
estimated at $456,450.
6-14
-------
Costs of Control
Table 6.2-13: Fixed Costs for
Engine Improvements for Class II per Engine Family17
Engine Class
Useful life (hrs)
Valving
Class II
250,500,1000
OHV
R&D
Design (4 months)
Development (6 months)
TOTAL R&D per Engine Line
41,225
172,225
213,450
TOOLING COSTS
Cylinder Head
Piston
Connecting Rod
Camshaft
Carburetor
Flywheel
Setup Changes
TOTAL TOOLING per Engine Line
TOTAL FIXED
25,000
25,000
15,000
8,000
60,000
35,000
75,000
$243,000
$456,450
6.2.2.2 Catalysts for Class II
Further emission reduction can be achieved through the use of catalysts. The catalyst
must be designed for durability throughout the engine's regulatory useful life. A catalyst
efficiency of 25-45 percent is estimated for these engines. The catalyst technology that would
be utilized would be similar to that used for Class I engines. The exceptions include: 1) Class II
engines would not use supplemental air because the HC and NOx ratios are more favorable in
Class II OHV engines due to their more efficient combustion chamber and larger displacement
and horsepower, and 2) the precious metals in the catalysts range from
platinum/palladium/rhodium for 250 and 500 hour Class II engines to to palladium/rhodium
(5:1) for 1000 hour regulatory useful life engines.
Class II engine designs include engines 1 to 4 cylinders. Engines with two or more
cylinders have specific issues to be considered in terms of safety with regard to engine exhaust
and catalyst use and this will be addressed towards the end of this section. The variable costs for
6-15
-------
Final Regulatory Impact Analysis
catalysts of single cylinder engines are listed in Table 6.2-14. The catalyst substrate size is
calculated based on the engine displacement size. To utilize one value per regulatory useful life
category for this analysis, the engine horsepower and displacements were sales weighted with
values from the 2005 EPA certification database information. Catalyst volumes range from 33
percent of the engine displacement for the 250 useful life to 50 percent of the engine
displacement for the 1000 hour useful life. Larger catalysts are needed for longer useful life
periods in order to provide the emission conversion durability.
Catalyst substrate and heat shield variable costs will be decreased in the sixth year with a
learning curve of 20 percent. This somewhat conservative since the learning normally occurs at
20 percent with a doubling of production which would be in the third or fourth year. Optimized
catalyst/muffler designs and heat shield manufacturing processes will likely be developed as the
industry becomes experienced in application of the catalyst technology across their product line.
The muffler washcoat will likely still be unique per engine family per engine manufacturer and
therefore it is estimated there will likely not be a one size fits all catalyst/muffler design. The
precious metal prices are determined in the marketplace and therefore are not discounted over
time.
6-16
-------
Costs of Control
Table 6.2-14: Variable Catalyst Costs for Class II OHV Single Cylinder Engine
HC+NOx Emission Reduction to Phase 3 Standards
Useful Life
Engine Power (hp)
Engine Displacement (cc)
Catalyst Volume (cc)
Substrate Diameter (cm)
Substrate*
Washcoat and Precious Metal
Labor
Labor Overhead 40%
Supplier Markup 29%
Manufacture Price
Heat Shield
Spark Arrester
Hardware Cost to Manufacturer
w/Markup 29% OEM
Near Term Estimates
250
11.3
406
134
5.25
$4.78
$4.03
$1.40
$0.56
$3.12
$13.89
$4.23
$0.10
$18.22
$23.50
500
11.1
338
135
6.00
$4.81
$2.73
$1.40
$0.56
$2.75
$12.25
$3.96
$0.05
$16.26
$20.97
1000
9.5
329
165
7.00
$5.67
$4.10
$1.40
$0.56
$3.40
$15.13
$4.05
$0.05
$19.23
$24.80
Learning Curve Estimates
250
11.3
406
134
5.25
$3.82
$4.03
$1.40
$0.56
$2.84
$12.65
$3.38
$0.10
$16.14
$20.82
500
11.1
338
135
6.00
$3.84
$2.73
$1.40
$0.56
$2.47
$11.00
$3.17
$0.05
$14.23
$18.35
1000
9.5
329
165
7.00
$4.53
$4.10
$1.40
$0.56
$3.07
$13.66
$3.24
$0.05
$16.95
$21.87
* 50/50- split of metallic vs ceramic substrates
Fixed costs involve modification to the existing heat shield and cooling system. If the
muffler is in close proximity to the engine fan then cost for a heat shield can also be included
because in some cases the heat shields will need to be improved in order to direct cooling air
from the engine's flywheel over the muffler for muffler cooling. These fixed costs are presented
in Table 6.2-15.
6-17
-------
Final Regulatory Impact Analysis
Table 6.2-15: Fixed Costs for Class II OHV Single Cylinder Engine
Engine Class
Useful life (hrs)
Valving
II
125, 250, 500
OHV
R&D
Design
(2 months)
Development (5 months)
TOTAL R&D per Engine Line
20,612
143,521
164,133
TOOLING COSTS
Modified Muffler Stamping
Heat Shield Stamping
Engine Shroud Modification
Setup Changes
TOTAL TOOLING per Engine Line
TOTAL FIXED COSTS
50,000
30,000
15,000
25,000
120,000
$284,133
Carbureted V-Twins
Carbureted engines with more than one cylinder, ex: V-twins or more, have special
concerns when considering the use of catalyst application. Multi-cylinder engines may continue
to run if one cylinder misfires or does not fire at all. If this occurs, the results is raw unburned
fuel and air from one cylinder and hot exhaust gases from the other cylinder combining in the
muffler. In a catalyst muffler, this condition will likely result in continuous backfire which
would create high temperatures within the muffler and potentially destroy the catalyst. One
solution is to have separate catalyst mufflers for each cylinder. The two cylinders in the V-twins
currently share one muffler. If two mufflers are used, then the individual mufflers would likely
need to be slightly larger. Each individual muffler would need to be 25-30 percent larger than
one half the volume of the original. Since the two cylinders in the V-twins currently share one
muffler one option for consideration would be to package the two catalysts in separate chambers
within one larger muffler.
Costs for this new muffler design are listed in Tables 6.2-16 and 6.2-17. V-twin engines
from EPA's certification database were sales weighted for power and engine displacement per
regulatory useful life. ICF provided the estimates for existing muffler costs and new muffler
cost estimates.
18
6-18
-------
Costs of Control
Table 6.2-16: Variable Costs for Change to Two Mufflers for V-Twins
19
Engine Power (hp)
Engine Displacement - Total (cc)
Per Cylinder Displacement (cc)
Current Muffler Cost
New Muffler Cost (includes 2)
Hardware Cost to Manufacturer
OEM Markup @ 29%
Total Component Costs
250 OHV
16.3
605
393
($20.24)
$26.31
$6.07
$1.76
$7.83
500 OHV
20.1
632
411
($23.13)
$30.07
$6.94
$2.01
$8.95
1000 OHV
17.1
627
408
($22.57)
$29.34
$6.77
$1.96
$8.73
Fixed costs include modified muffler stamping, exhaust pipe changes and setup changes.
These costs are estimated at $100,000 per engine family. Special considerations were not
accounted for in the case where OEM's obtain their own muffler and assemble the muffler onto
the engine once the engine is received from the engine manufacturer. This analysis considers
that in most cases equipment manufacturers would buy their catalyst mufflers from the engine
manufacturer in order to avoid engine certification.
Table 6.2-17: Fixed Costs for Change to Two Mufflers for V-Twins2
Engine Power
Engine Displacement - Total (cc)
Per Cylinder Displacement
Modified Muffler Stamping
Exhaust Pipe Changes
Setup Changes
Total Tooling per Engine Line
250 OHV
16.3hp
605
393
$50,000
$25,000
$25,000
$100,000
500 OHV
20.1hp
632
411
$50,000
$25,000
$25,000
$100,000
1000 OHV
17.1hp
627
408
$50,000
$25,000
$25,000
$100,000
In this analysis, catalyst sizes are related to the engine cylinder size and therefore since
cylinders of V-twin engines are smaller than one cylinder Class II engines, costs are recalculated
from Table 6.2-14. Note that one catalyst is used in each muffler for a total of two catalysts.
Tables 6.2-18 and 6.2-19 present the projected variable and fixed catalyst costs for Class II OHV
V-twin engines.
6-19
-------
Final Regulatory Impact Analysis
Table 6.2-18: Variable Catalyst Costs for Class II OHV V-Twin Engine,
Near Term and Learning Curve Effect
Useful Life
Engine Power (hp)
Engine Displacement per Cylinder
Catalyst Volume (cc)
Substrate Diameter (cm)
Substrate*
Washcoat and Precious Metal
Labor
Labor Overhead 40%
Supplier Markup 29%
Manufacture Price per Catalyst
Two Catalysts ($x2)
Heat Shield (2)
Spark Arrester (2)
Hardware Cost to Manufacturer
Markup
29% OEM
New Muffler Differential
TOTAL COST
Near Term Costs
250
16.3
303
100
5.00
$3.74
$3.00
$1.40
$0.56
$2.52
$11.22
$22.45
$8.53
$0.20
$31.18
$9.04
$7.83
$48.05
500
21.0
316
126
5.00
$4.55
$2.55
$1.40
$0.56
$2.63
$11.68
$23.36
$9.76
$0.10
$33.22
$9.63
$8.95
$51.80
1000
17.1
314
157
5.50
$5.44
$3.91
$1.40
$0.56
$3.28
$14.59
$29.18
$10.50
$0.10
$39.79
$11.54
$8.73
$60.06
Learning Curve Effect
250
16.3
303
100
5.00
$2.99
$3.00
$1.40
$0.56
$2.31
$10.26
$20.52
$6.82
$0.20
$27.54
$7.99
$6.26
$41.97
500
21.0
316
126
5.00
$3.64
$2.55
$1.40
$0.56
$2.36
$10.51
$21.02
$7.81
$O.10
$28.92
$8.39
$7.16
$44.76
1000
17.1
314
157
5.50
$4.35
$3.91
$1.40
$0.56
$2.96
$13.19
$26.37
$8.4
$0.1
$34.87
$10.11
$6.98
$51.97
* 50/50- split of metallic vs ceramic substrates
6-20
-------
Costs of Control
Table 6.2-19: Fixed Costs for Class II OHV V-Twin Engine
Useful Lives
250, 500, 1000
R&D COSTS
Design (2 months)
Development (5 months)
TOTAL R&D
$20,612
$143,521
$164,133
TOOLING COSTS
Heat Shield Stamping
Engine Shroud Modification
Setup Changes
New Muffler Design
Total Tooling per Engine Line
TOTAL FIXED COSTS
$50,000
$25,000
$25,000
$100,000
$200,000
$364,133
Electronic Fuel Injection
Electronic fuel injection (EFI) is another solution for engines with two or more cylinders.
EFI will allow more equal fuel delivery between or among the engine cylinders. In addition, it
enables better atomization and more efficient fuel delivery during load pickup. If an engine
family is somewhat close to the Phase 3 standard currently then EFI may allow the engine to
meet the emission standards without a catalyst. If a small catalyst is needed, EFI allows the
engine to be setup for cylinder monitoring and can be shut down if all cylinders are not operating
properly. Due to the anticipated higher cost for EFI compared to catalyst, EPA estimates that
each engine manufacturer will initially apply EFI to the engine family, of two or more cylinders,
with the highest sales volume. Table 6.2-20 lists the estimated costs to apply electronic fuel
injection. The cost tables include subtracting the existing carburetor.
6-21
-------
Final Regulatory Impact Analysis
Table 6.2-20: Variable Costs for Electronic Fuel Injection - Open and Closed Loop
For Class II Engines and Applications with a Battery21
Injectors
Pressure Regulator
ECM/MAP Sensor
Throttle Body
Air Temperature Sensor
Fuel Pump
Oxygen Sensor
Wiring/Related Hardware
HARDWARE COST TO MANUFACTURE
OEM markup @ 29%
Warranty Markup @ 5%
Total Component Cost
Remove existing carburetor ($15) marked up 29%
EFI Technology Difference
Open Loop EFI
8.00
3.75
27.00
2.75
1.50
10.50
~
12.00
66.75
19.36
2.85
88.96
-19.35
$69.61
Closed Loop EFI
8.00
3.75
27.00
2.75
1.50
10.50
7.00
12.00
73.75
21.39
3.69
98.83
-19.35
$79.48
Fixed costs for electronic fuel injection are listed in Table 6.2-21. Open loop fuel
injection requires more research and development time due to the fact that it does not use an
oxygen sensor to keep the air/fuel ratio in check. This analysis estimatess all engines using
electronic fuel injection will be developed as closed loop fuel injection systems.
6-22
-------
Costs of Control
Table 6.2-21: Fixed Costs for Electronic Fuel Injection - Open and Closed Loop
For Class II Engines and Applications with a Battery
Design
Development
Modified Exhaust Manifold for O2 Sensor
Total Fixed Costs
Open Loop
$41,225
$229,633
—
$270,858
Closed Loop
$20,612
$57,408
$25,000
$103,020
6.2.2.3 Equipment Costs
The majority of Class I engines are sold as a unit and therefore the engine, fuel tank and
muffler are provided by the engine manufacturer to the equipment manufacturer. As shown in
EPA's Technical Study on the "Safety of Emission Controls for Nonroad Spark-Ignition Engines
<50 Horsepower", catalysts can be applied to Class I engines such that muffler temperatures are
equal to or less than those of the current Phase 2 product with minimal changes to the engine
package. Some engines may require larger mufflers to house a catalyst depending on current
muffler design. However the majority of equipment housing Class I engines are close coupled to
the engine with open access for air cooling and therefore it no equipment redesign costs are
applied to equipment manufacturers.
The majority of Class II engines are not sold as a unit. The current industry practice
includes equipment manufacturers purchasing the muffler separate from the engine. Based on
conversations with industry it is believed that for several reasons this practice will change to the
dominant practice being the equipment manufacturer purchasing the muffler from the engine
manufacturer. The offerings by the engine manufacturer will likely be influenced by the largest
customers and smaller equipment manufacturers will have a few set models from which to
choose. A limited amount of equipment redesign will be required on products.
EPA's work with catalysts in mufflers of two one-cylinder Class II lawn tractor engines
has revealed that the current muffler on this equipment type has plenty of room to accommodate
the catalyst and internal baffling to promote cooling of the exhaust gases. Smaller mufflers are
used in other applications in which engine noise is not of concern. EPA did not work with these
mufflers and therefore, it is uncertain if the catalyzed muffler will work in these mufflers. It is
possible that a larger muffler can may be required to accommodate the catalyst.
Changes that will be required on Class II engines with catalysts includes a heat shield for
the muffler (counted in catalyst costs), necessary sheet metal to direct cooling from the engine
flywheel to the muffler and any equipment design changes to accommodate a different engine
envelope.
Incorporating shrouding to direct the cooling air to and around the muffler is of most
6-23
-------
Final Regulatory Impact Analysis
importance. The shrouding added includes extending and rerouting some of the engine sheet
metal that is used to direct the air-flow out of the engine cylinder and blocking off the usual air
exit into the engine compartment. The air is routed out the bottom of the chassis instead. In
EPA's Class II one cylinder engine testing, the "touch-guard" was boxed in by closing off it's
slots, closing off one end, and reducing the size of the opening on the opposite end. The exhaust
exit was re-routed to a different location, and an ejector was added over the top of the exhaust.
The amount of additional metal is fairly minimal and relatively thin-gage. The best examples are
the Kohler CV490 on one of the Craftsman tractors and the Kohler SV590 on the Cub Cadet.
Detailed photos of the SV590 installation can be found in EPA's Safety Study.22
For equipment that use engines with catalysts and require heat shield or equipment
design changes, variable costs are estimated for the sheet metal and/or engine structure redesign
at $1.30 per piece of equipment. Since a portion of engines are assigned to EFI, or will likely
not require additional heat shield or equipment modifications due to current equipment design, it
is estimated that 60 percent of equipment will utilize increased sheet metal and/or engine
structure redesign. This yields a sales weighted average of $0.78 per equipment. Fixed costs
for R&D for the added sheet metal design and/or engine restructure are estimated at $30,000 per
equipment model and tooling changes are also estimated at $45,000 per model. These estimates
are based on the estimates for developing and applying heat shields in the catalyst cost estimates
for Class II and can be seen in Table 6.2-22.
Table 6.2-22: Average Equipment Costs Per Equipment Model
Heat Shield
Additional material for
equipment redesign or air
entrainment pathway
R&D
Tooling Changes
Variable Costs
-0- included in catalyst costs
1.30 per equipment
0.78 avg over all for 60% of
equipment
n/a
n/a
Fixed Costs
-0- included in catalyst costs
n/a
30,000
45,000
6.2.3 Compliance and Certification
The certification and compliance costs include engine dynamometer aging as well as
emission testing pre- and post-aging. Certification and compliance costs are included in this
analysis as fixed costs. After preliminary emission testing, engines are aged on the
dynamometer to the regulatory useful life. The aged engines are then emission tested. The
engine's emission levels must be below the new standards. If not, then the engine family cannot
be certified unless the excesses are offset with other engine families within a manufacturers
product line and the manufacturer must be involved in the averaging, banking and trading
program. Engine families will need to certify to the new emission standards using the updated
6-24
-------
Costs of Control
test procedure found in Chapter 4.
The Phase 2 certification database was used as the basis for the number of engine
families to be certified to these standards. The 2005 Certification database contains a number of
engine manufacturers that have certified to the Phase 1 emission standards (1997) as well as a
large number of additional engine manufacturers that have certified to the Phase 2 standards
(2002).
6.2.3.1 Measurement Protocol 1065 Compliance Costs
New to the small engine industry are the 1065 protocols for gaseous emission
measurement. These protocols are found in 40 CFR Part 1065. Depending on the analyzing
equipment used by the industry, the certification analyzers may have to be upgraded to the
estimated cost of $250,000. It is possible that less costly upgrades on some analyzers will be
available. A CVS system can be assembled for $50,000 given manufacturer ingenuity.
6.2.3.2 Certification Costs
Certification costs include emission testing after a short engine break-in period and aging
on a dynamometer to the full useful life and then repeat emission testing. Costs for
dynamometer aging of each Class and corresponding useful life are found in ICF's report "Small
SI Engine Technologies and Costs."23 The costs per dynamometer aged engines are estimated in
Table 6.2-3. are based on test setup, data analysis, engine aging operation, dyno costs, scheduled
maintenance, prototype engine cost and fuel.
Table 6.2-23: Dynamometer Aging Certification Costs Per Class and Useful Life
CLASS I
125
250
500
$9,532
$17,462
$33,353
CLASS II
250
500
1,000
$18,413
$34,658
$70,069
The costs for the emission compliance tests are found in Tables 6.2-24 and 6.2-25 and
they are the same for each engine regardless of useful life category. A total of two emission
tests after break-in and two at end of useful life are accounted for in this cost analysis. The
emission test costs are estimated at $2,012 each and are based on the costs for a private test
laboratory in 2005.24
Table 6.2-24: Emission Testing Costs Per Class
CLASS I
all useful lives
$8,048
CLASS II
all useful lives
$8,048
6-25
-------
Final Regulatory Impact Analysis
Table 6.2-25: Per Engine Family Emission Testing and Dynamometer
Aging Costs Per Class and Useful Life
CLASS I
125
250
500
$17,580
$25,510
$41,401
CLASS II
250
500
1,000
$26,461
$42,706
$78,117
6.2.4 LPG/CNG Engine Costs
Engine manufacturers and equipment manufacturers certify engines to run on LPG. The
number of engine families are obtained from EPA's 2008 Certification Database. Certification
costs found in Section 6.2.3.2 apply to these engines. Part 1065 compliance costs are not applied
since the engine manufacturers are the same as listed in the gasoline section (costs already
applied) and it is estimated that equipment manufacturers contract with a test lab due to the high
cost of maintaining an individual test lab.
For engine certification, all engine families will be required to be tested for baseline
emissions, see Table 6.2-26. Small volume engine manufacturers with a production of 10,000
engines or less can utilize an assigned deterioration factor and do not have to undergo
dynamometer aging or end of life emission testing. Those listed under dynamometer aging in
Table 6.2-26 will need to age the engines and perform end of life emission testing. The number
of engine families listed under catalyst development will need emission reduction technology
such as catalysts.
Table 6.2-26: Number of LPG Engine Families Per Class and
Useful Life Designation for Fixed Cost Analysis
CLASS I
UL
125
250
500
BaselineE
mission
Testing
~
~
2
Dynamo-meter
Aging + End of
Life Emission
Testing
~
~
2
Catalyst
Dev
~
~
1
CLASS II
UL
250
500
1000
Baseline
Emission
Testing
9
20
13
Dynamo-meter
Aging + End of
Life Emission
Testing
9
20
13
Catalyst
Dev
lcyl/2cyl
1
2
1
2
10
8
For Phase 3, companies with small volume production (<10,000) can use an assigned df.
6-26
-------
Costs of Control
Table 6.2-27 lists the certification costs.
Table 6.2-27: Certification Costs - LPG
Baseline Emission Testing
Dynamometer Aging
End of Life Emission Test
Total
Class I
$8,048
$66,706
$8,048
$82,802
Class II
$169,008
$1,769,774
$169,008
$2,107,790
As mentioned above, the technology to reduce emissions to the Phase 3 levels is
catalysts. Catalysts are currently being utilized on LPG engines as shown in EPA's 2008
Certification Database. Basic engine improvement design changes, accounted for in the gasoline
engine families, were not accounted for in these engines for they were already made in the base
engine before they were converted to run on LPG/CNG. Costs that will be applied to these
engines are R&D for catalyst formulation and variable parts costs which will need to be
formulated for the exhaust makeup from these engines. The majority of these engines are two
cylinder engines, however the concerns of the application of catalysts to these engine designs are
relieved in that some of the V-twin LPG engines are already certified with catalysts. Costs for
catalyst system redesign for some of the existing engine families are included in order for these
families to meet the Phase 3 standards. Table 6.2-28 lists the R&D and Tooling costs for
catalysts for LPG. Table 6.2-29 contains the totals for fixed cost for each class given the total
number of engine families listed in Table 6.2-26.
6-27
-------
Final Regulatory Impact Analysis
Table 6.2-28: Fixed Costs for Class II OHV Single Cylinder Engine - LPG
Engine Class
Useful life (his)
Valving
II
125, 250, 500
OHV
R&D
Design
Development (5 months)
TOTAL R&D per Engine Line
$20,612
$143,521
$164,133
TOOLING COSTS
TOTAL TOOLING per Engine Line
TOTAL FIXED COSTS
0*
$164,133
*LPG engines are modified from gasoline version engines. Tooling costs are not included for it
is estimated that catalyst volume for these engines will be determined based on a percentage of
engine displacement, as the gasoline version, and therefore the catalysts will fit into the same
muffler space.
Table 6.2-29: Total Fixed Costs for LPG
Engine Families 2005$
Catalyst R&D
Certification Cost
TOTAL
Class I
$492,399
$47,114
$539,413
Class II
$6,072,921
$1,146,438
$7,219,359
Certification data on gaseous fueled engines show that the HC:NOx ratio is higher in
NOx than in HC which is opposite from gasoline engines. Platinum will be used in the precious
metal mixture in order for the oxygen reduced from the NOx to be utilized to convert CO due to
the lack of HC. For Class I engines, the cost estimate presented in Table 6.2-7 is applicable
because it is calculated with a platinum/palladium/rhodium ratio of 5/0/1. For Class II engines,
the 500 and 1000 hour catalyst cost estimates will be modified in order to include more platinum
and all useful life periods will have resized catalysts based on the sales weighted engine
displacement in the certification listing of LPG engines. Table 6.2-30 lists the variable catalyst
costs for Class II OHV Engines, 250 and 500 hour useful life engines (no 1000 hour UL engines
are listed in the LPG certification). Two to three cylinder engines have higher displacement and
therefore costs are recalculated for those engine designs.
6-28
-------
Costs of Control
Table 6.2-30: Variable Catalyst Costs for Class II OHV Engines - LPG
HC+NOx Emission Reduction to Phase 3 Standards
Useful Life
Engine Power (hp)
Engine Displacement (cc)
Engine/Catalyst
Catalyst Volume (cc)*** (per
cylinder)
Substrate Diameter
Substrate* (per cylinder)
Washcoat and Precious Metal
Labor
Labor Overhead 40%
Supplier Markup 29%
Manufacture Price (per catalyst)
Total Catalyst Cost
Heat Shield (2 for v-twin)
Spark Arrester (2 for v-twin)
Hardware Cost to Manufacturer
w/Markup 29% OEM
Add'l Muffler for V-twin
Total Catalyst Cost for LPG
engines
Total Catalyst Cost for Gasoline
Engines
Cost Difference
1 cylinder
250
13.8
415
33%
137
5.25
5.55
4.24
$1.40
$0.56
$3.41
$15.16
$15.16
$4.23
$0.10
$19.49
$25.14
-
$24.14
$23.50
$1.64
500
17.8
389
40%
156
6.00
8.91
4.82
$1.40
$0.56
$4.55
$20.24
$20.24
$4.26
$0.05
$24.55
$31.67
-
$31.67
$20.97
$10.70
1000**
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
2 cylinders
250
18.2
597
33%
197
5.00
3.70
2.96
$1.40
$0.56
$2.50
$11.12
$22.24
$5.90
$0.20
$28.34
$8.22
$7.83
$44.40
$48.05
-$3.66
500
19.2
743
40%
297
5.00
5.20
4.46
$1.40
$0.56
$3.37
$14.99
$30.00
$6.92
$0.10
$37.00
$10.73
$8.95
$56.68
$51.80
$4.87
1000
23
751
50%
376
5.50
6.34
8.86
$1.40
$0.56
$4.97
$22.14
$44.24
$7.32
$0.10
$51.69
$14.99
$8.73
$75.41
$60.06
$15.37
* 50/50- split of metallic vs ceramic substrates
** No one cylinder LPG engines are certified to the 1000 hour useful life
*** these catalyst volumes were calculated from the engine disp in EPA's certification data for 2005
6-29
-------
Final Regulatory Impact Analysis
Calculations for the rulemaking have been completed using gasoline assumptions. To
account for the increase in costs due to some of the gasoline engines being used as LPG engines,
an increase in the total cost is added to the current gasoline engine variable cost total.
Table 6.2.-31 is an example of costs for 2012 in 2005$.
Table 6.2-31: Change in Variable Cost in 2012, 2005$ - LPG
Total Engine
Sales
Estimate per
Useful Life
2012
%of
LPG/CNG
Engines in
Useful Life
per Class
#ofCyl
Number
of
Engines
with
change in
Cost
Estimate
Variable
Cost
Change in
2012
Total Change in
costs in 2012
2005$
Class I
125 OHV
250
500
2,953,419
905,005
623,431
0%
0%
0.63%
1
1
1
0
0
3574
0
0
0
0
0
0
Class II
250
500
1000
3,334,488
724,231
821,463
0.58%
1.94%
3.19%
1
2
1
2
2
14,500
10,469
12,918
90,630
18,700
$1.38
-$2.95
$9.25
$5.17
$15.59
2012 Total Increase
$20,027
-$30,923
$119,523
$ 468,553
$ 291,517
$868,698
* use the same technology as gasoline counterpart
Table 6.2-32 contains the catalyst cost estimates for LPG engines including a learning
curve discount. This cost estimate is used in year six of the cost estimates.
6-30
-------
Costs of Control
Table 6.2-32: Variable Catalyst Costs with Learning Curve for Class II OHV Engines
LPG; HC+NOx Emission Reduction to Phase 3 Standards
Useful Life
Engine Power (hp)
Engine Displacement (cc)
Engine/Catalyst
Catalyst Volume (cc)*** (per
cylinder)
Substrate Diameter
Substrate* (per cylinder)
Washcoat and Precious Metal
Labor
Labor Overhead 40%
Supplier Markup 29%
Manufacture Price (per catalyst)
Total Catalyst Cost
Heat Shield (2 for v-twin)
Spark Arrester (2 for v-twin)
Hardware Cost to Manufacturer
w/Markup 29% OEM
Add'l Muffler for V-twin
Total Catalyst Cost for LPG
engines
Total Catalyst Cost for Gasoline
Engines
Cost Difference
1 cylinder
250
13.8
415
33%
137
5.25
4.44
4.24
$1.40
$0.56
$3.09
$13.73
$15.90
$3.38
$0.10
$17.21
$22.20
-
$22.20
$20.82
$1.38
500
17.8
389
40%
156
6.00
7.13
4.82
$1.40
$0.56
$4.03
$17.94
$24.88
$3.41
$0.05
$21.40
$27.61
-
$27.61
$18.35
$9.25
1000**
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
2 cylinders
250
18.2
597
33%
197
5.00
2.96
2.96
$1.40
$0.56
$2.29
$10.17
$20.33
$4.72
$0.20
$25.25
$7.32
$6.26
$38.84
$41.79
-$2.95
500
19.2
743
40%
297
5.00
4.16
4.46
$1.40
$0.56
$3.07
$13.65
$27.30
$5.54
$0.10
$32.93
$9.55
$7.16
$49.64
$44.47
$5.17
1000
23
751
50%
376
5.50
5.07
8.86
$1.40
$0.56
$4.61
$20.50
$41.00
$5.86
$0.10
$46.96
$13.62
$6.98
$67.56
$51.97
$15.59
* 50/50- split of metallic vs ceramic substrates
** No one cylinder LPG engines are certified to the 1000 hour useful life
*** these catalyst volumes were calculated from the engine disp in EPA's certification data for 2005
6-31
-------
Final Regulatory Impact Analysis
6.2.5 Small SI Aggregate Costs
Costs presented in the previous sections are combined here to present streams of costs.
The first, Section 6.2.5.1, presents variable costs (recurring costs) for meeting the Phase 3
exhaust standards. Section 6.2.5.2 presents a stream of fixed costs for meeting the Phase 3
exhaust standards. Costs are based on assuming all engines are gasoline engines. Additional
costs for LPG engines are included at the end of this section.
6.2.5.1 Variable Costs for Meeting Exhaust Standards
Variable costs for Class I are summarized in Table 6.2-10 for engine improvements and
catalysts in near term and long term (with learning) costs. Nearly all engines in Class I (96.9%
125 useful life (UL), 99.4% 250 UL and 72.65% 500 UL) are estimated to have both
technologies applied and therefore the costs are added according to useful life period and then
multiplied by the number of engines sold per useful life category, as will be discussed later. The
resultant variable costs per engine is presented in Table 6.2-33. Long term costs are six years
after the near term costs and include a 20 percent learning curve reduction for engine
improvement components, catalyst substrate and heat shield costs.
Variable costs for Class II are a combination of engine improvements and catalyst or
engine improvements and electronic fuel injection (EFI), see Section 6.2.2. Information on
engine designs and related certification emission results in the 2008 EPA Certification Database
were utilized to determine the percentage of technologies per useful life. A portion of the
engines, the largest multi-cylinder engine family per engine manufacturer needing emission
reduction, are assigned the use of electronic fuel injection. The remaining engine families are
assigned catalysts and related engine improvements. Some engines would not to require any
costs. Long term costs (learning) are six years after the near term costs and include a 20 percent
learning curve reduction for engine improvement components, catalyst substrate and heat shield
costs.
Table 6.2-32: Percentage Technologies Per Useful Life per Class II
Useful Life
250
500
1000
No changes
43.66%
59.84%
28.50%
2FI - Class II
V-twin
5.88%
5.62%
10.80%
V-twin
catalyst
0.61%
0.30%
18.24%
Catalvst-Sinale
Cylinder
49.85%
34.24%
42.46%
6-32
-------
Costs of Control
Table 6.2-33: Variable Costs Per Engine for Meeting
Exhaust Standards, Per Engine (2005$)
Useful Life (his)
125- SV
125 - OHV
250
500
1000
Class I
Near Term (20 12)
10.41
8.55
12.17
11.71
~
Long Term
(2017)*
9.43
7.80
11.33
10.87
~
Class II
Near Term (20 11)
~
~
16.8
11.93
30.07
Long Term
(2016)*
~
~
14.24
9.87
25.21
*Long term includes learning reduction
The total Small SI engine costs for the first 30 years (2008-2037) were estimated using
sales and growth estimates from the US EPA's NONROAD model. The percentage sales per
useful life category (Class I: 125, 250, 500, Class II: 250, 500, 1000) were calculated from the
manufacturer prescribed useful life period and yearly estimated sales per engine family in the
EPA 2008 Phase 2 certification database (confidential information). The percentages in
Table 6.2-34 were applied to US EPA's NONROAD model sales estimates and the results are
presented in Table 6.2-35 Note that snowblowers are not included for they only have to comply
with the evaporative standards since they are exempted from the exhaust emission standards.
Table 6.2-34: Small SI Engines
Sale Percentages per Useful Life
Useful Life
125- SV
125 - OHV
250
500
1000
Class I
66%
23%
6%
5%
—
Class II
—
—
74%
11%
15%
6-3
-------
Final Regulatory Impact Analysis
Table 6.2-35: Class I and Class II Projected Sales
per Useful Life Category (snowblowers excluded)
YEAR
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
2037
CLASS 1
125
SV
6,139,695
6,250,204
6,360,351
6,474,907
6,584,353
6,698,636
6,810,588
6,921,857
7,031,516
7,144,731
7,256,744
7,370,110
7,482,752
7,594,978
7,706,370
7,818,799
7,931,065
8,043,780
8,157,416
8,270,846
8,383,989
8,497,471
8,610,967
8,724,583
8,838,143
8,951,587
9,064,949
9,178,412
9,291,898
9,405,435
125
OHV
2,112,583
2,150,608
2,188,508
2,227,925
2,265,584
2,304,907
2,343,428
2,381,714
2,419,446
2,458,402
2,496,944
2,535,952
2,574,710
2,613,326
2,651 ,654
2,690,339
2,728,969
2,767,752
2,806,853
2,845,883
2,884,814
2,923,861
2,962,913
3,002,007
3,041 ,082
3,080,116
3,119,122
3,158,163
3,197,212
3,236,279
250
OHV
565,168
575,340
585,479
596,025
606,099
616,619
626,924
637,167
647,261
657,683
667,994
678,429
688,798
699,129
709,382
719,732
730,066
740,442
750,902
761 ,343
771 ,758
782,204
792,652
803,110
813,564
824,006
834,442
844,886
855,333
865,784
500
OHV
489,249
498,055
506,832
515,961
524,682
533,789
542,710
551 ,577
560,315
569,337
578,263
587,296
596,272
605,215
614,091
623,051
631 ,997
640,978
650,034
659,072
668,088
677,131
686,175
695,229
704,278
713,318
722,352
731 ,393
740,436
749,484
CLASS II
250
OHV
3,375,298
3,436,078
3,497,169
3,560,736
3,621 ,924
3,685,741
3,747,698
3,809,238
3,870,775
3,933,230
3,995,499
4,058,405
4,120,822
4,183,226
4,245,149
4,307,507
4,369,904
4,432,728
4,495,615
4,558,320
4,620,946
4,683,749
4,746,567
4,809,474
4,872,290
4,935,044
4,997,777
5,060,568
5,123,362
5,186,178
500
OHV
485,273
494,012
502,795
511,934
520,731
529,906
538,814
547,662
556,509
565,488
574,441
583,485
592,459
601,431
610,333
619,299
628,270
637,302
646,344
655,359
664,363
673,392
682,423
691 ,468
700,499
709,521
718,540
727,568
736,596
745,627
1,000
OHV
687,306
699,683
712,123
725,067
737,527
750,522
763,138
775,669
788,200
800,917
813,597
826,406
839,116
851 ,824
864,433
877,131
889,837
902,629
915,435
928,203
940,956
953,744
966,536
979,345
992,137
1,004,915
1,017,689
1 ,030,475
1 ,043,262
1 ,056,053
The Total Variable Costs were calculated using the sales information found in
Table 6.2-35 and applying the corresponding variable cost from Table 6.2-33. Results are
presented in Table 6.2-36. Engines used in snowblowers and handheld equipment will require
only evaporative control measures and these are presented in Section 6.5.
6-34
-------
Costs of Control
Table 6.2-36: Variable Costs for Meeting Phase 3 Exhaust Emission Standards, 2005$
Year
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
2037
Class 1: Engine only
125
-
_
-
-
86,472,768
87,973,661
89,443,931
90,905,230
92,345,393
86,511,329
87,867,634
89,240,307
90,604,228
91,963,103
93,311,878
94,673,217
96,032,587
97,397,383
98,773,331
100,146,794
101,516,776
102,890,863
104,265,113
105,640,825
107,015,858
108,389,476
109,762,116
111,135,969
112,510,105
113,884,858
250
-
-
-
-
7,376,417
7,504,448
7,629,867
7,754,521
7,877,372
7,449,638
7,566,432
7,684,635
7,802,085
7,919,100
8,035,245
8,152,473
8,269,530
8,387,055
8,505,540
8,623,812
8,741 ,783
8,860,108
8,978,447
9,096,912
9,215,318
9,333,603
9,451 ,803
9,570,108
9,688,438
9,806,820
500
-
-
-
-
6,141,854
6,248,457
6,352,885
6,456,676
6,558,965
6,191,089
6,288,151
6,386,385
6,483,993
6,581 ,239
6,677,763
6,775,185
6,872,467
6,970,137
7,068,606
7,166,896
7,264,937
7,363,272
7,461,619
7,560,070
7,658,473
7,756,775
7,855,006
7,953,324
8,051 ,663
8,150,046
Class II: Engine & Equipment
250
-
-
-
59,835,023
60,863,232
61,935,618
62,976,746
64,010,872
55,130,893
56,020,430
56,907,330
57,803,283
58,692,279
59,581 ,099
60,463,047
61 ,351 ,207
62,239,924
63,134,711
64,030,407
64,923,505
65,815,481
66,709,969
67,604,683
68,500,650
69,395,337
70,289,133
71,182,631
72,076,956
72,971 ,320
73,865,999
500
-
-
-
6,105,598
6,210,517
6,319,944
6,426,181
6,531 ,704
5,493,685
5,582,326
5,670,704
5,759,984
5,848,571
5,937,140
6,025,024
6,113,527
6,202,086
6,291 ,250
6,380,504
6,469,500
6,558,384
6,647,518
6,736,674
6,825,955
6,915,109
7,004,174
7,093,210
7,182,327
7,271 ,449
7,360,602
1,000
-
-
-
21,799,867
22,174,477
22,565,182
22,944,499
23,321 ,265
19,873,736
20,194,399
20,514,111
20,837,087
21,157,554
21,477,959
21,795,886
22,116,052
22,436,419
22,758,975
23,081 ,857
23,403,804
23,725,346
24,047,793
24,370,323
24,693,303
25,015,822
25,338,020
25,660,111
25,982,500
26,304,902
26,627,419
6.2.5.2 Fixed Costs
Fixed costs for the small spark ignition engines include test cell modifications for 1065
compliance, emission certification of engine families as well as R&D and tooling expenditures
for engine design changes and equipment modifications. This section presents the aggregate
fixed costs for small spark ignition engines.
The test procedure for small spark ignited engines for this rulemaking is governed by Part
1065 with regulation specific details in Part 1054. Evaluation of Part 1065 reveals that there are
some differences in calibration procedures with existing Part 90 and programming must be
changed to calculate via 1065. As industry begins to apply 1065 requirements in its test cells
6-35
-------
Final Regulatory Impact Analysis
there may be some other modifications that will be revealed. To cover these costs, EPA is
allocating $600,000 per engine manufacturer for 1065 compliance. The number of engine
manufacturers is taken from the the certification database which lists 16 different engine
manufacturers of nonhandheld engines and 15 engine manufacturers of handheld engines. The
certification database also lists a number of new offshore manufacturers. These companies
typically certify through independent test laboratories within the United States and therefore
only encounter costs for these upgrades through increased service fees. For this cost analysis,
one additional manufacturer for nonhandheld and handheld is added to the certification database
totals to cover test labs. Therefore, for nonhandheld engine manufacturers, a total of 17 test
facilities at 600,000 per test facility yields a total estimated cost of $10,200,000. For the purpose
of this cost analysis, this cost is spread evenly across all useful lives per class for a total of
1,700,000 for each. Upgrades for test cells for handheld engines are calculated at $9,600,000.
Engine manufacturers are to begin using compliant test cells for new engine family certifications
beginning in 2013. This cost analysis estimates engine manufacturers will invest in their test
cells from 2008-2011.
Table 6.2-37: Fixed Costs for Compliance
with 1065, 2005$ (thousands)
TOTAL
CLASS I
125
1,700
250
1,700
500
1,700
CLASS II
250
1,700
500
1,700
1000
1,700
HANDHELD
9,600
Investment Per Year
2008
2009
2010
7011
425
425
425
475
425
425
425
475
425
425
425
475
425
425
425
475
425
425
425
475
425
425
425
475
2,400
2,400
2,400
7400
Each engine family must certify each year to the emissions standards applicable in that
year. This cost analysis assumes no carryover data, but that all engine families will undergo
durability aging and emission testing. The number of engine families per Class and per useful
life category were taken from EPA's 2008 Certification Database. For Class I, the 2008 database
lists 66 engine families from traditionally regulated companies. For Class II, the 2008 database
lists 121 engine families. The engine families are designated per useful life class by the claim in
the certification application. The estimates in Table 6.2-38 represent the number of engine
families per useful life designation used in this cost analysis to calculate fixed costs. Costs for
certifiers of LPG engines are covered in Section 6.2.4.
6-36
-------
Costs of Control
Table 6.2-38: Number of Engine Families Per Class
and Useful Life Designation for Certification
CLASS I
125
250
500
31
15
20
CLASS II
250
500
1000
39
18
64
It should be noted that the certification database does contain certifications from a large
number of companies that have a short history of compliance and claim large sales numbers.
These companies were not used in this analysis for we are not yet convinced they are actually
selling in this country nor in the numbers they claim. Engine families still certified to Phase 1
(either through credits, small engine family flexibilities or averaging) were also not included.
For Class II, there are a number of small volume engine families which have not yet been
certified to Phase 2 due to flexibilities in that rulemaking. Due to the low volume sales, these
engine families were estimated to be certified to the 250 hour useful life. For Class I-A, engine
families are being moved to the <80cc category where they already meet the handheld emission
standard. Class I-B engines are traditionally low volume sales engine families; we believe that
they will likely be incorporated into the engine manufacturers ABT programs and certification of
these low volume sales engine families will be covered without engine improvement.
The total engine exhaust emission certification costs are calculated by taking the number
of engine families from Table 6.2-38 and multiplying them by the emission test and
dynamometer aging costs from Table 6.2-23. This analysis estimates that engine certification
costs are expended over two years prior to standard implementation as shown in Table 6.2-39.
The combined 1065 compliance and engine certification costs are presented in Table 6.2-40.
Table 6.2-39: Engine Certification Costs to Exhaust Standards
2008
2009
2010
7011
CLASS I
125
272,490
777 490
250
191325
191375
500
455,411
455 411
CLASS II
250
635,064
635,064
500
811,414
811,414
1000
3,007,505
3,007,505
Handheld
$0
Table 6.2-40: Total Stream of Costs for
Engine Certification by Year, (thousands)
2008
2009
2010
7011
CLASS I
125
425
425
697
697
250
425
425
616
616
500
425
425
880
880
CLASS II
250
425
1,060
1,060
475
500
425
1,236
1,236
475
1000
425
3,432
3,432
475
Handheld
2,400
2,400
2,400
2400
6-37
-------
Final Regulatory Impact Analysis
Fixed costs for engine research and development and tooling changes to meet exhaust
emission standards are presented throughout sections 6.2.1 Class I and 6.2.2. Class II. The fixed
costs include engine improvements, engine improvements with catalyst development or EFI
development and application. Class I engine families are assigned engine improvements and/or
engine improvements and catalyst development costs. The number of engine families per Class
are from the 2008 EPA Certification Database. Table 6.2-41 presents the number of engine
families estimated per technology package for Class I and Table 6.2-42 presents the number of
engine families estimated per technology for Class II.
Table 6.2-41: Estimated Number of Engine Families per Technology Package, Class I
Technology/Useful Life
- Engine Improvements (all)
- Catalysts
125
31
26
250
15
13
500
20
14
Table 6.2-42: Estimated Number of Engine Families per Technology Package, Class II
Technology/Useful Life
- Engine Improvements (all)
- One Cylinder Engine Catalyst
- Two or More Cylinders per Engine for Catalyst
- Electronic Fuel Injection on Two or More Cylinder Engines
250
39
16
7
4
500
18
4
3
3
1000
64
14
27
5
Table 6.2-43: Total Fixed Costs (thousands)
for Engines to Meet Phase 3 Exhaust Emission Standards
l&D
Pooling
TOTAL
CLASS I
125
SV
5,286
10,164
15,450
125
OHV
5,598
5,571
11,169
250
OHV
5,335
5,205
10,540
500
OHV
6,567
6,540
13,107
CLASS II
250
OHV
12,412
12,412
15,450
500
OHV
5,225
12,412
15,450
1000
OHV
20,780
12,412
15,450
Table 6.2-44: Total Fixed Costs Investment (thousands)
for Engines to Meet Phase 3 Exhaust Emission Standards
2008
2009
2010
2011
CLASS 1
125-sv
1,322
1,322
6,404
6,404
125-ohv
1,400
1,400
4,185
4,185
250
1,334
1,334
3,936
3,936
500
1,642
1,642
4,912
4,912
CLASS 11
250
4,301
11,150
11,150
500
2,398
5,263
5,263
1000
7,419
18,498
18,498
6-38
-------
Costs of Control
Total fixed costs for Small SI exhaust emissions are shown in Table 6.2-45.
Table 6.2-45: Certification and Technology Fixed Costs
for Engines to Meet Phase 3 Exhaust Standards
2008
2009
2010
2011
Class 1
125
3,146
3,146
11,286
11.286
250
1,759
1,759
4,553
4.553
500
2,067
2,067
5,792
5.792
Class 11
250
4,562
12,046
12,046
425
500
2,167
5,843
5,843
425
1.000
7,352
21,438
21,438
425
Handheld
2,400
2,400
2,400
2.400
Equipment companies using Class II engines are also estimated to incur fixed costs in
redesigning equipment models to incorporate Phase 3 Class II engines. The PSR database shows
there are 413 businesses using Class II engines.25 Assuming each business on average produces
two unique models requiring clearly different redesign yields a number of 826 redesigns.
Table 6.2-22 contains equipment costs per equipment model and Table 6.2-46 contains the total
equipment costs.
Table 6.2-46: Total Class II Equipment Cost
2009
2010
250
10325000
10325000
500
10325000
10325000
1000
10325000
10325000
6.2.5.3 Operating Cost Savings
The application of electronic fuel injection to an estimated 6.6% of the Class II engines is
expected to result in fuel savings. Fuel savings from the use of fuel injection on Class II engines
is estimated at 10 percent. Kohler has been offering a fuel injected Class II engine for nearly 10
years and two articles (1996 OEM Off-Highway and 1998 Diesel Progress)26'27 claim 15-20
percent fuel savings over carbureted engines. We elected to conservatively use a figure often
percent. In calculating the fuel savings, we use a gasoline price of $1.81 per gallon without
taxes.28 Table 6.2-47 presents estimated fuel savings for Class II engines with electronic fuel
injection. The improvements and catalyst application to Class I engines are estimated to result in
no operating or fuel savings. Fuel savings that are obtained from evaporative reduction
technologies are presented later in the evaporative portion of this chapter.
6-39
-------
Final Regulatory Impact Analysis
Table 6.2-46: Fuel Savings from the Increased
Use of Electronic Fuel Injection on Class II Engines
Year
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
7037
Gallons
0
0
0
3,916,719
7,074,990
10,071,145
11,966,500
13,835,431
15,252,178
16,221,164
16,966,562
17,576,831
18,104,416
18,532,855
18,916,185
19,267,974
19,607,579
19,935,933
20,259,628
20,579,305
20,896,206
21,210,286
21,521,549
21,830,758
22,138,949
22,446,266
22,752,947
23,059,059
23,363,883
23 667 468
Fuel Savings $
0
0
0
$7,104,929
$12,834,033
$18,269,057
$21,707,230
$25,097,472
$27,667,451
$29,425,191
$30,777,343
$31,884,371
$32,841,410
$33,618,599
$34,313,960
$34,952,106
$35,568,148
$36,163,782
$36,750,965
$37,330,860
$37,905,717
$38,475,459
$39,040,090
$39,600,995
$40,160,054
$40,717,527
$41,273,846
$41,829,132
$42,382,085
«47 937 787
6-40
-------
Costs of Control
6.2.5.4 Total Aggregate Costs
The aggregate costs for meeting the exhaust emission standards are presented in
Table 6.2-47. Aggregate costs include variable costs and fixed costs for engine manufacturers
(technology, certification, 1065 compliance), equipment manufacturers and LPG engine families
and converters. An average cost per engine is presented in Table 6.2-48 and the aggregate costs
with fuel savings is presented in Table 6.2-49.
Table 6.2-47: Total Aggregate for 30 year Cost Analysis
for Exhaust Emission Standard Compliance without Fuel Savings, 2005$
Year
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
2037
Exhaust Only
CLASS I
7,012,721
7,012,721
21,671,947
21,671,947
99,991,039
101,726,567
103,426,684
105,116,426
106,781,731
100,152,057
101,722,217
103,311,327
104,890,306
106,463,442
108,024,886
109,600,875
111,174,585
112,754,576
114,347,477
115,937,502
117,523,496
119,114,243
120,705,179
122,297,808
123,889,649
125,479,854
127,068,925
128,659,401
130,250,205
131,841,723
CLASS II
15,393,779
71,614,262
71,614,262
93,177,678
93,481,939
95,129,055
96,728,158
98,316,508
85,010,277
86,381,917
87,749,492
89,131,026
90,501,833
91,872,369
93,232,307
94,601,825
95,972,201
97,351,938
98,733,075
100,110,208
101,485,609
102,864,885
104,244,509
105,626,063
107,005,647
108,383,854
109,761,603
111,140,627
112,519,710
113,899,279
1065
Certification
Upgrades
Handheld
2,400,OOC
2,400,OOC
2,400,OOC
2,400,OOC
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
Table 6.2-48 presents a sales weighted average per-equipment cost estimate for engines
6-41
-------
Final Regulatory Impact Analysis
meeting the Phase 3 exhaust standards. Note that the fixed costs are invested prior to the
implementation date for Class I and Class II engines and therefore the variable costs are what
remain in these years. Near term costs for Class I are from 2012-2016 and for Class II are from
2011-2015. Long term costs include a learning curve deduction in manufacturing/production for
these engines in the 6th year and the Class I costs start in 2017 and Class II starts in 2016.
Table 6.2-48: Sales Weighted Average Per-Equipment
Cost Estimates (Without Fuel Savings) for Exhaust Standards, 2005$
Short Term Costs
^years 1-5) per Class per
Useful Life
Near Term
Lone Term*
Class I
125
10.48
9.01
250
7.51
11.33
500
11.12
10.87
Class 11
250
17.70
15.02
500
13.66
10.76
1000
31.93
26.49
Handheld
(no Exhaust)
0.00
0.00
* Long term is with learning, if applicable
The aggregate costs with fuel savings is presented in Table 6.2-49. Fuel savings are
available from Class II engines using electronic fuel injection and start in 2011 which is the first
year of standard implementation.
6-42
-------
Costs of Control
Table 6.2-49: Total Aggregate for 30 year Cost Analysis
for Exhaust Emission Standard Compliance with Fuel Savings, 2005$
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
2037
Class l
7,012,721
7,012,721
21,671,947
21,671,947
99,991,039
101,726,567
103,426,684
105,116,426
106,781,731
100,152,057
101,722,217
103,311,327
104,890,306
106,463,442
108,024,886
109,600,875
111,174,585
112,754,576
114,347,477
115,937,502
117,523,496
119,114,243
120,705,179
122,297,808
123,889,649
125,479,854
127,068,925
128,659,401
130,250,205
131.841.723
LJiass II
15,393,779
71,614,262
71,614,262
89,260,959
86,406,949
85,057,910
84,761,659
84,481,077
69,758,099
70,160,753
70,782,931
71,554,195
72,397,418
73,339,514
74,316,122
75,333,851
76,364,623
77,416,005
78,473,447
79,530,903
80,589,404
81,654,598
82,722,959
83,795,305
84,866,697
85,937,588
87,008,656
88,081,568
89,155,827
90.231.811
Handheld
2,400,00(
2,400,00(
2,400,00(
2,400,00(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
C
At a 7 percent discount rate, over 30 years, the estimated annualized cost to
manufacturers for Small SI exhaust emission control, without fuel savings, is $182 million. The
corresponding estimated annualized fuel savings due to the use of electronic fuel injection on
Class II engines is $24 million. At a 3 percent discount rate, the estimated annualized cost to
manufacturers for Small SI exhaust emission control, without fuel savings, is $189 million. The
corresponding estimated annualized fuel savings due to the use of electronic fuel injection on
Class II engines is $27 million.
6-43
-------
Final Regulatory Impact Analysis
6.3 Exhaust Emission Control Costs for Outboard and Personal Watercraft
Marine Engines
This section presents our cost estimates for meeting the new exhaust emission standards
for outboard and personal watercraft marine engines.
As of about a decade ago, outboard and personal watercraft (OB/PWC) engines were
primarily two-stroke carbureted engines. There were no emission control requirements. Since
then, manufacturers have used two primary strategies to meet exhaust emission standards. The
first is two-stroke direct injection. By injecting the fuel directly into the combustion chamber
after the exhaust port closes, the short-circuiting fuel losses with traditional two-strokes can be
largely eliminated. The second approach is to convert to using four-stroke engines, either
carbureted or fuel-injected. One other approach that has been used by one PWC manufacturer
has been the use of a two-way catalyst in the exhaust of a two-stroke engine. Today, engine
sales are a mix of old and new technology. We anticipate that the standards will largely be met
by phasing out the old-technology engines and using technology already available in the
marketplace.
Since California ARB has adopted standards similar to the new national standards,
manufacturers have already started with design and testing efforts to meet our standards. To
reflect this in the cost analysis, we include no estimated costs for R&D to introduce the various
emission-control technologies. This reflects the expectation that manufacturers will not need to
conduct additional R&D for EPA's requirements, since they are introducing those technologies
for sale in California. As noted below, we are including estimated R&D expenditures as part a
compliance cost, because EPA's NTE standards represent an incremental requirement beyond
what California ARB has adopted.
For the purpose of this analysis, we divide outboards into five power categories and PWC
into three power categories. We present cost estimates of various emission-control technologies
for each of these power categories. Additional detail on the per-engine costs presented in this
section is available in the docket.29 Table 6.3-1 presents these power categories and the engine
size we use to represent each category.
Table 6.3-1: Engine Sizes Used for Cost Analysis
Outboard Engines
Personal
Watercraft
Engines
Power Range
0-25 hp
25-50 hp
50-100 hp
100-175 hp
>175 hp
50-100 hp
100-175 hp
>175 hp
Engine Power
9.9 hp
40 hp
75 hp
125 hp
225 hp
85 hp
130 hp
175 hp
Displacement
0.25 L
0.76 L
1.60 L
1.80L
3.00 L
1.65 L
1.85L
2.50 L
Cylinders
2
3
3
4
6
2
3
4
6-44
-------
Costs of Control
6.3.1 Two-Stroke Direct Injection
Traditional outboards use carbureted two-stroke engine designs where the fuel and air are
mixed in the carburetor then pumped into the combustion chamber through the crankcase. The
piston itself acts to open and close the intake and exhaust ports. As a result, fuel may be lost out
the exhaust port. Better control of the fuel can be achieved using indirect injection in place of
the carburetor; however, this does not prevent short-circuiting losses. Indirect injection is
primarily used on the largest two-stroke engines. Direct-injection has been used by
manufacturers to reduce emissions from two-stroke outboards. By injecting the fuel directly into
the cylinder after the exhaust port is closed, short-circuiting losses can be minimized.
Table 6.3-2 and 6.3-3 present incremental costs of applying direct injection to outboards and
PWC, respectively. For the largest power category, costs are presented incremental to indirect
injection. For the remaining categories, costs are presented incremental to carbureted engines.
For 135 hp PWC engine, incremental costs are presented for both IDI and carbureted engines
because baseline engines in this power category use both approaches.
Table 6.3-2: Outboard—Projected Incremental Costs for 2-Stroke Direct Injection
Hardware Cost to Manufacturer
carburetor(s)
fuel metering solenoids
IDI injectors
fuel distributer
pressure regulator
air compressor
air regulator
throttle body position sensor
intake manifold
fuel pump
electronic control module
air intake temperature sensor
manifold air pressure sensor
injection timing sensor/timing wheel
wiring/related hardware
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
9.9 hp
carb.
($28)
$36
~
~
~
$80
$15
$30
$5
$3
$85
$5
$10
$5
$20
$266
$13
$5
$82
$13
$380
40 hp
carb.
($114)
$60
~
~
~
$100
$15
$35
$5
$0
$90
$5
$10
$8
$30
$244
$15
$6
$77
$12
$354
75 hp
carb.
($135)
$66
~
~
~
$120
$17
$35
$9
($5)
$95
$5
$11
$9
$30
$257
$19
$8
$82
$13
$379
125 hp
carb.
($165)
$96
~
~
~
$140
$20
$40
$10
($6)
$100
$5
$11
$10
$50
$311
$22
$9
$99
$16
$456
225 hp
IDI
~
$156
($102)
($25)
($35)
$165
$22
$10
($5)
($35)
$0
$0
$0
$0
$0
$151
$14
$6
$49
$8
$228
6-45
-------
Final Regulatory Impact Analysis
Table 6.3-3: PWC—Projected Incremental Costs for 2-Stroke Direct Injection
Hardware Cost to Manufacturer
carburetor(s)
fuel metering solenoids
IDI injectors
fuel distributer
pressure regulator
air compressor
air regulator
throttle body position sensor
intake manifold
fuel pump
electronic control module
air intake temperature sensor
manifold air pressure sensor
injection timing sensor/timing wheel
wiring/related hardware
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
85 hp
carb.
($114)
$44
~
~
~
$120
$17
$35
$9
($5)
$95
$5
$11
$9
$20
$246
$19
$8
$79
$12
$364
130 hp
carb.
($165)
$72
~
~
~
$140
$20
$40
$10
($6)
$100
$5
$11
$10
$30
$267
$22
$9
$86
$13
$398
130 hp
IDI
~
$72
($51)
($20)
($30)
$140
$20
$0
($10)
($30)
$0
$0
$0
$0
$0
$91
$12
$5
$31
$5
$144
175 hp
IDI
~
$104
($68)
($25)
($35)
$165
$22
$0
($5)
($35)
$0
$0
$0
$0
$0
$123
$12
$5
$41
$6
$186
6-46
-------
Costs of Control
6.3.2 Migration from Two-Stroke to Four-Stroke Engines
The primary technology that manufacturers are using to meet exhaust emissions
standards has been to convert their product offering more to four-stroke engines. Because four-
stroke engines are common in the market today, we do not include costs for research and
development or warranty. Rather, we anticipate that manufacturers will sell more of the four-
stroke engines and phase out the carbureted two-stroke designs as a result of the new standards.
Tables 6.3-4 and 6.3-5 below present a comparison between costs for two-stroke and four-stroke
outboard and PWC engines, respectively. These costs are based on prices for current product
offerings.
Table 6.3-4: Outboard—Projected Incremental Costs for 4-Stroke
2-stroke baseline technology
4-stroke control technology
2-stroke cost
4-stroke cost
Markup at 29%
Total Incremental Cost
9.9 hp
carb
carb
$900
$1,124
$65
$289
40 hp
carb
carb
$2,101
$2,633
$154
$686
75 hp
carb
carb
$3,076
$3,861
$228
$1,013
125 hp
carb
EFI
$4,195
$5,504
$380
$1,689
225 hp
DPI
EFI
$6,339
$7,761
$412
$1,834
Table 6.3-5: PWC—Projected Incremental Costs for 4-Stroke
2-stroke baseline technology
4-stroke control technology
2-stroke cost
4-stroke cost
Markup at 29%
Total Incremental Cost
85 hp
carb
EFI
$3,319
$4,350
$299
$1,330
130 hp
DPI
EFI
$4,578
$5,587
$293
$1,302
175 hp
DPI
EFI
$5,862
$7,207
$390
$1,735
6-47
-------
Final Regulatory Impact Analysis
6.3.3 Four-Stroke Electronic Fuel Injection
Manufacturers can gain better control of their fuel and air management through the use of
electronic fuel injection. This is often used in larger OB/PWC engines today. For this analysis,
we consider the use of a port fuel-injection system, which refers to individual injectors located at
each intake port in the engine. In addition to the injectors, this system includes a fuel rail,
pressure regulator, electronic control module, manifold air pressure and temperature sensors, a
high pressure fuel pump, a throttle assembly, a throttle position sensor, and a magnetic
crankshaft pickup for engine speed. Tables 6.3-6 and 6.3-7 present the incremental costs of a
port fuel-injection system compared to a carburetor-based fuel system for outboards and personal
watercraft, respectively.
Table 6.3-6: Outboard—Projected Incremental Costs for 4-Stroke EFI
Hardware Costs
carburetor(s)
injectors
fuel rail
pressure regulator
intake manifold
throttle body position sensor
fuel pump
electronic control module
air intake temperature sensor
manifold air pressure sensor
injection timing sensor
wiring/related hardware
Hardware Cost to Manufacturer
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
9.9 hp
($28)
$34
$40
$15
$5
$30
$13
$95
$5
$10
$5
$20
$244
$3
$1
$72
$12
$332
40 hp
($114)
$51
$55
$15
$5
$35
$10
$100
$5
$10
$8
$30
$210
$4
$2
$63
$11
$289
75 hp
($135)
$51
$65
$20
$6
$35
$10
$105
$5
$11
$9
$30
$212
$4
$2
$63
$11
$291
125 hp
($165)
$68
$70
$30
$10
$40
$14
$110
$5
$11
$10
$40
$243
$4
$2
$72
$12
$333
225 hp
($240)
$102
$80
$35
$15
$50
$17
$115
$5
$11
$10
$60
$260
$4
$2
$77
$13
$356
6-48
-------
Costs of Control
Table 6.3-7: PWC—Projected Incremental Costs for 4-Stroke EFI
Hardware Costs
carburetor(s)
injectors
fuel rail
pressure regulator
intake manifold
throttle body position sensor
fuel pump
electronic control module
air intake temperature sensor
manifold air pressure sensor
injection timing sensor
wiring/related hardware
Hardware Cost to Manufacturer
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
85 hp
($135)
$34
$65
$20
$6
$35
$10
$105
$5
$11
$9
$20
$185
$4
$2
$55
$9
$255
130 hp
($165)
$51
$70
$30
$10
$40
$14
$110
$5
$11
$10
$30
$216
$4
$2
$64
$11
$297
175 hp
($240)
$68
$80
$35
$15
$50
$17
$115
$5
$11
$10
$40
$206
$4
$2
$61
$10
$283
6-49
-------
Final Regulatory Impact Analysis
6.3.4 Catalysts
We believe the OB/PWC exhaust emission standards can be achieved without the use of
catalysts. At this time, three-way catalysts have not been demonstrated on OB/PWC engines.
However, one manufacturer has been using a two-way catalyst on PWCs with 2-stroke engines
for several years. We include research and development costs for this technology because it is
not currently used in the marine industry, but is an alternative we assess in Chapter 11. Catalyst
sizes and formulations are based on the analysis discussed below for SD/I engines. Tables 6.3-8
and 6.3-9 present the incremental cost of adding catalysts to four-stroke, electronic fuel-injection
OB and PWC engines, respectively.
Table 6.3-8: Outboard—Projected Incremental Costs for Catalytic Control
Catalyst Unit Price
catalyst volume (L)
substrate diameter (cm)
substrate
ceria/alumina
Pt/Pd/Rd
can (18 gauge SS)
Total Material Cost
Labor
labor overhead at 40%
supplier markup at 29%
Manufacturer Price per Unit
Hardware Cost to Manufacturer
catalyst
exhaust manifold modifications
oxygen sensor
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
Fixed Cost to Manufacturer
research & development
tooling
units/year
years to recover
Fixed Cost/Unit
Total Incremental Cost
9.9 hp
0.09
4.5
$2
$1
$2
$0.4
$6
$14
$6
$8
$33
$33
$15
$25
$73
$1
$1
$22
$2
$99
$342,788
$75,000
5,000
5
$23
$122
40 hp
0.27
6.0
$4
$3
$7
$0.8
$15
$14
$6
$10
$45
$45
$17
$25
$87
$1
$1
$26
$2
$116
$352,938
$75,000
5,600
5
$21
$137
75 hp
0.56
8.5
$5
$6
$16
$1
$29
$14
$6
$14
$62
$62
$20
$25
$107
$1
$1
$32
$2
$143
$362,068
$75,000
6,400
5
$19
$162
125 hp
0.63
9.0
$6
$7
$18
$1
$32
$14
$6
$15
$67
$67
$25
$25
$117
$1
$1
$34
$3
$156
$372,980
$75,000
5,900
5
$21
$177
225 hp
1.05
10.0
$8
$12
$29
$2
$52
$14
$6
$21
$92
$92
$30
$25
$147
$1
$1
$43
$3
$195
$388,643
$75,000
4,700
5
$27
$222
6-50
-------
Costs of Control
Table 6.3-9: PWC—Projected Incremental Costs for Catalytic Control
Catalyst Unit Price
catalyst volume (L)
substrate diameter (cm)
substrate
ceria/alumina
Pt/Pd/Rd
can (18 gauge SS)
Total Material Cost
Labor
labor overhead at 40%
supplier markup at 29%
Manufacturer Price per Unit
Hardware Cost to Manufacturer
catalyst
exhaust manifold modifications
oxygen sensor
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hour
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
Fixed Cost to Manufacturer
research & development
tooling
units/year
years to recover
Fixed Cost/Unit
Total Incremental Cost
85 hp
0.58
9.0
$5
$7
$16
$1
$30
$14
$6
$14
$63
$63
$35
$25
$123
$1
$1
$36
$3
$165
$363,502
$75,000
1,700
5
$71
$236
130 hp
0.65
9.0
$6
$7
$18
$1
$33
$14
$6
$15
$68
$68
$40
$25
$133
$1
$1
$39
$3
$177
$371,332
$75,000
5,300
5
$23
$200
175 hp
0.88
9.0
$7
$10
$25
$2
$44
$14
$6
$18
$82
$82
$45
$25
$152
$1
$1
$45
$4
$202
$381,016
$75,000
1,000
5
$126
$328
6.3.5 Certification and Compliance
Outboard and PWC engines must already be certified to meet the current EPA HC+NOx
exhaust emission standards. We therefore do not anticipate any increase in clerical work
associated with these standards. In addition, manufacturers are likely to meet the new standards
by selling more of their lower-emission engines, which are certified today. However,
manufacturers may need to adjust engine calibrations to meet the new standard and collect
further data to demonstrate compliance with the not-to-exceed zone. We therefore allow on
average two months of R&D for each engine family as part of the certification process.
Considering two engineers and three technicians and the corresponding testing costs for the two-
6-51
-------
Final Regulatory Impact Analysis
month period, we estimate a total cost of $130,000 per engine family. Unless engine designs
were significantly changed, manufacturers could recertify engine families each year using
carryover of this original test data. If this cost is amortized over five years of engine sales with
an average volume of 5,500 engines per family for outboards and 4,200 engines per family for
PWC, the resulting cost is $5 per engine for outboards and $6 for PWC.
6.3.6 Operating Cost Savings
We anticipate that the standards will largely be met on average by phasing out old, high-
emitting technologies, such as carbureted two-stroke engines and replacing them with currently
available clean technologies such as four-stroke engines and direct-injection two-stroke engines.
In addition to having lower emissions, these newer-technology engines have significantly lower
fuel consumption. Over the life of an engine, these fuel savings result in significant operating
cost savings. In calculating the fuel savings, we use a gasoline price of $1.81 per gallon without
taxes.
30
The largest portion of the fuel savings would come from phasing out carbureted
crankcase-scavenged two-stroke engines. As discussed in Chapter 4, scavenging losses from
these engines can result in more than 25 percent of the fuel passing through the engine unburned.
In addition, we model incremental fuel-consumption benefits between fuel-injected two-stroke
engines, carbureted four-stroke engines, and fuel-injected four strokes. These fuel consumption
rates and their derivation are described in more detail in the docket.31
Table 6.3-10: Projected Fuel Savings for OB/PWC Engines
Annual Per-Engine Gallons Consumed
Average Life (years)
Anticipated Reduction in Fuel Consumption
Lifetime Gallons Saved
Lifetime Cost Savings
Discounted Cost Savings (7%)
Outboard
72
19
5.2%
72
$130
$75
PWC
225
9.9
4.7%
103
$187
$142
6.3.7 Total OB/PWC Engine Costs
As discussed above, we anticipate that manufacturers would meet the standards largely
by changing their technology mix from older to newer technologies. For this reason, our
estimated per-engine costs for the average OB/PWC engine reflect a mix of technology changes.
Table 6.3-11 presents the baseline technology mix by power class. This technology mix is based
on an analysis of sales projections submitted to EPA by OB/PWC manufacturers at time of
certification. These sales projections are confidential, but a general description of this analysis is
6-52
-------
Costs of Control
available in the docket.32
Table 6.3-11: Baseline Technology Mix for OB/PWC Engines
Outboards
9.9 hp
40 hp
75 hp
125 hp
225 hp
PWC
85 hp
130 hp
175 hp
2-Stroke
Carbureted
24%
32%
20%
20%
0%
30%
5%
0%
2-Stroke
Indirect Injection
0%
0%
0%
0%
25%
60%
0%
70%
2-Stroke
Direct Injection
0%
2%
10%
30%
60%
10%
5%
30%
4-Stroke
Carbureted
76%
35%
0%
0%
0%
0%
0%
0%
4-Stroke
Fuel Injection
0%
32%
70%
50%
15%
0%
90%
0%
To develop the control technology mix, we made three adjustments to the baseline
technology mix. First, we considered that all the 2-stroke carbureted and indirect injection
engines would be replaced by either 2-stroke direct injection or 4-stroke engines. Second, we
included calibration costs for the for the 2-stroke direct injection and 4-stroke engines for better
emission performance. These engines are well below the existing HC+NOx standards; however,
there is currently wide variability in certified emission levels. We believe the standards will
require engine manufacturers to pay closer attention to emissions calibrations for their higher-
emitting new technology engines. Third, we included the conversion of a small number of 2-
stroke direct injection engines to 4-stroke based on product plans conveyed to us in private
conversations with manufacturers. While there is no way of knowing exactly what the actual
technology mix will be, we believe our analysis represents a reasonable scenario. Table 6.3-12
presents the projected technology mix for this control scenario.
6-53
-------
Final Regulatory Impact Analysis
Table 6.3-12: Projected Control Technology Mix for OB/PWC Engines
Outboards
9.9 hp
40 hp
75 hp
125 hp
225 hp
PWC
85 hp
130 hp
175 hp
2-Stroke
Carbureted
0%
0%
0%
0%
0%
0%
0%
0%
2-Stroke
Indirect Injection
0%
0%
0%
0%
0%
0%
0%
0%
2-Stroke
Direct Injection
0%
2%
10%
30%
50%
0%
5%
30%
4-Stroke
Carbureted
100%
66%
20%
0%
0%
100%
0%
0%
4-Stroke
Fuel Injection
0%
32%
70%
70%
50%
0%
95%
70%
We developed the per-engine costs based on the technology mix and technology cost
tables presented above. As discussed above, our cost estimates include both variable and fixed,
and we distinguish between near-term and long-term costs. Because our analysis amortizes fixed
costs over 5 years, the long-term costs are made up of variable costs only. Variable costs are
lower in the long term due to the learning effect discussed above. Table 6.3-13 presents these
average per-engine cost estimates.
Table 6.3-13: OB/PWC Per-Engine Cost Estimates (Without Fuel Savings)
OB aggregate
9.9 hp
40 hp
75 hp
125 hp
225 hp
PWC aggregate
85 hp
130 hp
175 hp
Short Term (years 1-5)
Fixed
$11
$5
$5
$8
$15
$27
$19
$29
$14
$45
Variable
$280
$69
$216
$203
$338
$690
$340
$870
$85
$1,290
Total
$291
$74
$222
$210
$353
$717
$359
$899
$98
$1,336
Long Term (years
6-10)
$224
$55
$173
$162
$270
$552
$272
$696
$68
$1,032
6.3.8 OB/PWC Aggregate Costs
Aggregate costs are calculated by multiplying the per-engine variable cost estimates
described above by projected engine sales. These variable costs are then added to the fixed costs
6-54
-------
Costs of Control
as incurred. Engine sales are based on estimates supplied by the National Marine Manufacturers
Association (www.nmma.org) and projections for future years are based on the growth rates in
the NONROAD model. Fuel-consumption reductions are calculated using the NONROAD
based on population estimates. These population estimates in the NONROAD model are similar
to those estimated by NMMA. A description of the sales and population data and our analysis of
the data are available in the docket.33 Table 6.3-14 presents the projected costs of meeting the
exhaust emission standards over a 30-year time period, with and without the fuel savings. Fuel
savings from the evaporative emission standards are not included in this table, but they are
presented separately below.
The population and sales data reported by NMMA, suggest that the NONROAD model
may somewhat underestimate the useful life of outboard and personal watercraft marine engines.
If useful life were back-calculated—dividing NMMA population by sales and adjusted for
growth—we would get a longer average life estimate. As a result, the per-engine fuel savings
described above may be understated. Because the current approach gives us a conservative
benefits estimate, and because we do not have new data on average lives for marine engines to
update the estimates in the NONROAD model, we are not updating the model at this time. For
this reason, the 30-year stream may give a better view of the impact of the fuel savings than the
per-engine analysis.
At a 7 percent discount rate, over 30 years, the estimated annualized cost to
manufacturers for OB/PWC exhaust emission control is $95 million. The corresponding
estimated annualized fuel savings due to more efficient engines is $48 million. At a 3 percent
discount rate, the estimated annualized cost to manufacturers for OB/PWC exhaust emission
control is $95 million. The corresponding estimated annualized fuel savings due to more
efficient engines is $57 million.
6-55
-------
Final Regulatory Impact Analysis
Table 6.3-14: Projected 30-Year Aggregate Cost Stream for OB/PWC Engines
Year
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
2037
Without Fuel Savings
OB
$8,347,493
$8,347,493
$83,474,059
$84,087,276
$84,700,492
$85,313,708
$85,926,924
$69,232,112
$69,716,553
$70,200,994
$70,685,435
$71,169,876
$71,654,317
$72,138,757
$72,623,198
$73,107,639
$73,592,080
$74,076,521
$74,564,028
$75,051,534
$75,539,041
$76,026,548
$76,514,055
$77,001,562
$77,489,069
$77,976,576
$78,464,083
$78,951,590
$79,439,097
$79,926,603
PWC
$3,773,815
$3,773,815
$26,775,058
$26,971,752
$27,168,447
$27,365,142
$27,561,836
$22,206,825
$22,362,213
$22,517,602
$22,672,991
$22,828,380
$22,983,769
$23,139,157
$23,294,546
$23,449,935
$23,605,324
$23,760,712
$23,917,085
$24,073,457
$24,229,829
$24,386,201
$24,542,574
$24,698,946
$24,855,318
$25,011,690
$25,168,062
$25,324,435
$25,480,807
$25,637,179
With Fuel Savings
OB
$8,347,493
$8,347,493
$79,497,283
$76,139,850
$72,816,127
$69,526,261
$66,259,388
$45,684,931
$42,316,350
$38,991,304
$35,708,089
$32,473,153
$29,314,449
$26,232,921
$23,265,756
$20,500,653
$18,356,730
$16,648,131
$15,128,141
$13,767,757
$12,656,743
$11,676,039
$10,870,140
$10,196,651
$9,613,547
$9,148,071
$8,757,202
$8,450,580
$8,228,217
$8,059,025
PWC
$3,773,815
$3,773,815
$24,485,325
$22,376,416
$20,292,925
$18,241,565
$16,224,620
$8,714,463
$6,764,827
$4,885,794
$3,100,364
$1,454,031
$587,570
$(9,566)
$(471,718)
$(844,125)
$(1,149,312)
$(1,387,033)
$(1,568,545)
$(1,705,040)
$(1,796,384)
$(1,842,344)
$(1,854,155)
$(1,865,959)
$(1,877,784)
$(1,889,601)
$(1,901,412)
$(1,913,224)
$(1,925,042)
$(1,936,853)
6.4 Exhaust Emission Control Costs for Sterndrive/Inboard Marine
Engines
This section presents our cost estimates for meeting the new exhaust emission standards
for sterndrive and inboard marine engines.
Sterndrive and inboard (SD/I) marine engines are typically "marinized" using automotive
engine blocks. There are a few exceptions where unique engine blocks are used, but these
applications represent a very small portion of the sales volume. Typical automotive blocks are
3.0 liter in-line 4-cylinder engines, 4.3 liter V-6 engines, and V-8 engines ranging from 5.0 to 8.2
liters total displacement. For purposes of this analysis, we present costs for an in-line 4 cylinder
engine, a V-6 engine, and three V-8 engine configurations. In addition, this analysis considers
costs to the original engine manufacturer and to the engine "marinizer." Additional detail on the
6-56
-------
Costs of Control
projected costs may be found in the docket.34
Because California ARB has adopted standards similar to the new national standards,
manufacturers have already started with design and testing efforts to meet our standards. To
reflect this in the cost analysis, we include no estimated costs for R&D to introduce the various
emission-control technologies. This reflects the expectation that manufacturers will not need to
conduct additional R&D for EPA's requirements, since they are introducing those technologies
for sale in California. As noted below, we are including estimated R&D expenditures as part a
compliance cost, because EPA's NTE standards represent an incremental requirement beyond
what California ARB has adopted.
6.4.1 Fuel Injection
Current SD/I engines are sold with carburetors or with fuel-injection systems. The
smaller 3.0 L 14 engines are typically carbureted while the larger 8.1 and 8.2 L V8 engines are
typically fuel injected. Our estimate is that about 25-30 percent of V6 engines and 70-80 percent
of the 5.0 - 6.2L V8 engines are currently sold with fuel injection. For the purpose of this
analysis we anticipate that all SD/I engines will need to be fuel injected to meet the new
emission standards. Fuel injection allows better control of the air-to-fuel ratio in the engine and
exhaust for better emission design control and catalyst efficiency.
We consider the use of a port fuel-injection system for this analysis, which refers to
individual injectors located at each intake port in the engine. In addition to the injectors, this
system includes a fuel rail, pressure regulator, electronic control module, manifold air pressure
and temperature sensors, a high pressure fuel pump, a throttle assembly, a throttle position
sensor, and a magnetic crankshaft pickup for engine speed. We also consider a cool fuel system
to prevent the occurrence of vapor lock in the fuel lines. Table 6.4-1 presents the incremental
costs of a port fuel-injection system compared to a carburetor-based fuel system. Because this
technology is widely used today, we include fixed costs for final calibrations as part of the cost
of certification and compliance in Section 6.4.4.
6-57
-------
Final Regulatory Impact Analysis
Table 6.4-1: Projected Incremental Hardware Costs for Controlled Port Fuel Injection
Hardware Cost to Manufacturer
carburetor
injectors
pressure regulator
fuel filter
intake manifold
fuel rail
throttle assembly (w/ position sensor)
cool fuel system (w/ pump)
electronic control module
air intake temperature sensor
manifold air pressure sensor
crank position sensor
wiring/related hardware
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hr
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
3.0L 14
($140)
$68
$15
$1
$14
$80
$150
$115
$70
$5
$14
$16
$80
$488
$3
$1
$143
$24
$659
4.3L V6
($145)
$102
$15
$1
$25
$80
$150
$120
$65
$5
$14
$16
$80
$528
$4
$2
$155
$26
$715
5.0L V8
($145)
$136
$15
$1
$25
$80
$150
$120
$65
$5
$14
$16
$80
$562
$4
$2
$165
$28
$760
5.7L V8
($145)
$136
$15
$1
$30
$80
$150
$120
$65
$5
$14
$16
$80
$567
$4
$2
$166
$28
$767
8.1L V8
($145)
$160
$15
$1
$40
$80
$60
$120
$60
$5
$14
$16
$80
$506
$4
$2
$148
$25
$685
6-58
-------
Costs of Control
6.4.2 Exhaust Gas Recirculation
We do not anticipate that manufacturers will use exhaust gas recirculation (EGR) to meet
the exhaust emission standards. However, in developing this rule, we considered the option of a
standard based on emission reductions possible through the use of EGR. This analysis is
reflected in our alternatives discussion in Chapter 11. For this analysis, we consider an EGR
system with a valve, plumbing, and modification to the intake manifold. Table 6.4-2 presents
incremental variable costs of a controlled engine with EGR compared to an uncontrolled engine
with port fuel injection and no EGR.
Table 6.4-2: Projected Incremental Hardware Costs for Exhaust Gas Recirculation
Hardware Cost to Manufacturer
intake manifold
exhaust gas recirculation
exhaust manifold
oxygen sensors
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hr
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
3.0L 14
$5
$25
$2
$17
$49
$1
$0
$15
$2
$67
4.3L V6
$5
$25
$5
$34
$69
$1
$0
$20
$3
$94
5.0L V8
$10
$25
$5
$34
$74
$1
$0
$22
$4
$101
5.7L V8
$10
$25
$5
$34
$74
$1
$0
$22
$4
$101
8.1L V8
$10
$25
$5
$34
$74
$1
$0
$22
$4
$101
6-59
-------
Final Regulatory Impact Analysis
6.4.3 Catalysts
We anticipate that manufacturers will use small three-way catalysts to meet the SD/I
exhaust emission standards. A catalyst will likely be placed in the riser of each exhaust manifold
upstream of where the water and exhaust gases mix. Catalyst sizes and configurations are based
on the developmental catalyst efforts on SD/I engines discussed in Chapter 4. Costs are included
to modify the exhaust manifolds for packaging of the catalyst. We believe these catalysts will be
used in conjunction with port fuel injection and closed-loop electronic control. Therefore, we
include the cost of an oxygen sensor upstream of each catalyst. The costs in Table 6.4-3 are
presented incremental to an open-loop port fuel injection.
Table 6.4-3: Projected Incremental Hardware Costs for Catalytic Control
Catalyst Unit Price
catalyst volume (L) (each)
number of catalysts
substrate diameter (cm)
substrate
ceria/alumina
Pt/Pd/Rd
can (18 gauge SS)
Total Material Cost
labor at $28/hr
labor overhead at 40%
supplier markup at 29%
Manufacturer Price per Unit
Hardware Cost to Manufacturer
catalysts
oxygen sensors
exhaust manifold
Total Incremental Hardware Cost
Engine Manufacturer Markup
labor at $28/hr
labor overhead at 40%
markup at 29%
warranty markup at 5%
Total Incremental Component Cost
3.0L 14
1.00
1
9.5
$8
$11
$28
$3
$51
$5
$2
$17
$74
$74
$17
$10
$101
$2
$1
$30
$5
$139
4.3L V6
0.75
2
8.3
$7
$9
$21
$3
$39
$5
$2
$13
$59
$119
$34
$20
$173
$1
$0
$50
$9
$233
5.0L V8
0.88
2
9.0
$7
$10
$25
$3
$45
$5
$2
$15
$66
$132
$34
$20
$186
$1
$0
$54
$9
$251
5.7L V8
1.00
2
9.5
$8
$11
$28
$3
$51
$5
$2
$17
$74
$148
$34
$25
$207
$1
$0
$60
$10
$279
8.1L V8
1.40
2
11.0
$10
$16
$39
$4
$69
$5
$2
$22
$98
$195
$34
$30
$259
$1
$0
$76
$13
$349
As discussed above, we do not include research and development costs in our fixed costs
for SD/I engines. However, we do include tooling costs that would be associated with ramping
up production of California engines for the entire United States. These tooling costs are
presented in Table 6.4-4.
6-60
-------
Costs of Control
Table 6.4-4: Projected Incremental Tooling Costs for Catalytic Control
Fixed Costs to Engine Manufacturer
tooling
units/year
years to recover
fixed costs/unit
Fixed Costs to Engine Marinizer
tooling
units/year
years to recover
fixed costs/unit
Total Incremental Fixed Costs
3.0L 14
$30,000
15,000
5
$1
$35,000
2,000
5
$5
$5
4.3L V6
$35,000
15,000
5
$1
$45,000
2,000
5
$6
$6
5.0L V8
$40,000
15,000
5
$1
$50,000
2,000
5
$7
$7
5.7L V8
$40,000
15,000
5
$1
$55,000
2,000
5
$7
$8
8.1L V8
$45,000
15,000
5
$1
$55,000
1,000
5
$14
$15
6.4.4 Certification and Compliance
We estimate that certification costs for SD/I engines would come to about $130,000 per
engine family. We expect that manufacturers would combine similar engines into the same
family. The above certification cost estimate allows for two months of R&D for each engine
family as part of the certification process. This would include two engineers and three
technicians and the corresponding testing costs for the two-month period. Unless engine designs
were significantly changed, engine families could be recertified each year using carryover of this
original test data. If this cost is amortized over five years of engine sales with an average
volume of 2,000 engines per family, the resulting cost is $13 per engine.
6.4.5 Operating Cost Savings
We anticipate that manufacturers will convert their remaining carbureted engines to fuel
injection to meet the new standards. We believe this will result in fuel savings because of the
better fuel control offered by fuel injection compared to carburetion. The fuel consumption rates
we use for carbureted and fuel injected SD/I engines and their derivation are described in more
detail in the docket.35 We use the price of gasoline discussed earlier in this chapter.
6-61
-------
Final Regulatory Impact Analysis
Table 6.4-5: Projected Fuel Savings for SD/I Engines
Annual Per-Engine Gallons Consumed
Average Life (years)
Anticipated Reduction in Fuel Consumption
Lifetime Gallons Saved
Lifetime Cost Savings
Discounted Cost Savings (7%)
228
19.7
2.3%
101
$185
$105
6.4.6 Total SD/I Engine Costs
We expect that SD/I engine manufacturers would use catalytic converters and electronic
fuel injection to meet the standards. In 2003, about 60 percent of SD/I engines were sold with
electronic fuel injection. This estimate is based on confidential sales information submitted to
the California Air Resources Board by SD/I manufacturers certifying to the 2003 California
exhaust emission standards. The manufacturers who certified in California represent more than
90 percent of U.S. sales of SD/I engines. Manufacturers have indicated to us that they are
moving in the direction of selling more fuel-injected engines and using carburetors only on their
low-cost "introductory" engines. For this cost analysis, we use the projected technology mix for
2009 from the NONROAD model which projects that about 85 percent of SD/I engines sold will
be fuel-injected. Table 6.4-6 presents our estimates of the sales mix between carbureted and fuel-
injected SD/I engines.
Table 6.4-6: Baseline Technology Mix for SD/I Engines
3.0LI-4
4.3LV-6
5.0LV-8
5.7LV-8
8.1LV-8
high performance
2003 MY California Certification
Carbureted
100%
75%
40%
10%
100%
~
Fuel Injection
0%
25%
60%
90%
0%
~
Projected 2009 Baseline
Carbureted
50%
20%
5%
0%
0%
50%
Fuel Injection
50%
80%
95%
100%
100%
50%
We developed the per-engine costs by assigning costs for electronic fuel injection for
engine models that are projected to be carbureted in 2009. Except for high-performance engines,
we also apply costs for catalysts. As discussed above, our cost estimates include both variable
and fixed costs, and we distinguish between near-term and long-term costs. Because our
analysis amortizes fixed costs over 5 years, the long-term costs are made up of variable costs
only. These variable costs are lower in the long term due to the learning effect discussed above.
6-62
-------
Costs of Control
Table 6.4-7 presents these average per-engine cost estimates. Fixed costs for high-performance
engines were based on an engine family size of 50 engines, compared to 2,000 engines for
traditional SD/I engines.
Table 6.4-7: SD/I Per-Engine Cost Estimates (Without Fuel Savings)
SD/I Aggregate
3.0L
4.3L
5.0L
5.7L
8.1L
high performance
Short Term (years 1-5)
Fixed
$21
$18
$19
$20
$21
$28
$280
Variable
$334
$465
$377
$297
$279
$349
$257
Total
$355
$483
$396
$317
$300
$377
$537
Long Term (years
6-10)
$266
$372
$301
$238
$223
$279
$216
6.4.7 SD/I Aggregate Costs
Aggregate costs are calculated by multiplying the per-engine variable cost estimates
described above by projected engine sales. These variable costs are then added to the fixed costs
as incurred. Engine sales are based on estimates supplied by the National Marine Manufacturers
Association (www.nmma.org) and projections for future years are based on the growth rates in
the NONROAD model. Fuel consumption reductions are calculated using the NONROAD
based on population estimates. These population estimates in the NONROAD model are similar
to those estimated by NMMA. A description of the sales and population data and our analysis of
the data is available in the docket.36 Table 6.4-8 presents the projected costs of the rule over a
30-year time period with and without the fuel savings that would be expected from meeting the
exhaust emission standards. Fuel savings from the evaporative emission standards are not
included in this table, but they are presented separately below.
At a 7 percent discount rate, over 30 years, the estimated annualized cost to
manufacturers for SD/I exhaust emission control is $28 million. The corresponding estimated
annualized fuel savings due to more efficient engine controls is $8 million. At a 3 percent
discount rate, over 30 years, the estimated annualized cost to manufacturers for SD/I exhaust
emission control is $28 million. The corresponding estimated annualized fuel savings due to
more efficient engine controls is $10 million.
6-63
-------
Final Regulatory Impact Analysis
Table 6.4-8: Projected 30-Year Aggregate Cost Stream for SD/I Engines
Year
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
2037
Without Fuel Savings
$4,971,723
$4,971,723
$31,909,535
$32,143,949
$32,378,362
$32,612,775
$32,847,189
$26,361,975
$26,546,438
$26,730,902
$26,915,366
$27,099,830
$27,284,293
$27,468,757
$27,653,221
$27,837,685
$28,022,149
$28,206,612
$28,392,244
$28,577,875
$28,763,506
$28,949,137
$29,134,769
$29,320,400
$29,506,031
$29,691,662
$29,877,294
$30,062,925
$30,248,556
$30,434,187
With Fuel Savings
$4,971,723
$4,971,723
$31,207,005
$30,512,049
$29,822,674
$29,128,682
$28,440,869
$21,040,258
$20,317,643
$19,601,606
$18,896,409
$18,205,484
$17,523,304
$16,828,978
$16,141,908
$15,467,722
$14,804,334
$14,154,769
$13,524,906
$12,917,862
$12,338,401
$11,804,572
$11,474,294
$11,259,905
$11,088,594
$10,948,483
$10,831,791
$10,732,954
$10,650,884
$10,582,255
6.5 Evaporative Emission Control Costs for Small SI Equipment
This section presents our cost estimates for meeting the new evaporative emission
standards for land-based equipment using small spark-ignition engines.
In our analysis of the costs of the evaporative emission standards for Small SI equipment,
we consider the approximately 250 equipment types used in the NONROAD model to determine
emission inventories. These equipment types are then aggregated into the five engine classes,
with each class divided by general equipment types and between residential and commercial
applications. For each of these aggregate categories, we determine weighted average hose
lengths and tank sizes which we use as inputs to our cost calculations. These inputs are
presented in more detail in the evaporative emission inventory discussion in Chapter 3. This
discussion presents our cost estimates as a function of hose length and tank size. In addition, we
present examples of costs for four typical Small SI equipment configurations which include a
handheld (HH) configuration, a walk-behind mower (WBM), and two other non-handheld
6-64
-------
Costs of Control
(NHH) configurations. These configurations, which are presented in Table 6.5-1, are based on
average tank sizes and hose lengths used in our inventory model (see Chapter 3). Although these
typical configurations do not, by any means, represent all of the equipment types included in our
cost calculations, they should give a good indication of how we performed our analysis.
Table 6.5-1: Typical Small SI Equipment Configurations
Fuel Tank Capacity (gallons)
Fuel Tank Material*
Fuel Tank Molding Process
Fuel Tank Weight (Ibs.)
Fuel Hose Length (in.)
Fuel Hose Inner Diameter (in.)
HH
0.25
HOPE
IM/BM
0.6
4
0.125
WBM
0.5
HOPE
IM/BM
0.8
8
0.25
NHH#1
2
HOPE
IM/BM
1.8
24
0.25
NHH #2
5
XLPE
RM
5.9
36
0.25
* HDPE = high-density polyethylene, XLPE = cross-link polyethylene
* IM = injection-molded, BM = blow-molded, RM = rotational-molded
The fuel tank weights are based on measurements made in our lab on many of the fuel
tanks that were included in our evaporative emission test programs. The higher weight to
capacity ratio of the smaller fuel tank is due to the smaller surface to volume ratio and due to
extra structural components often molded as part of the fuel tanks. We use the fuel tank weight
to determine costs of material changes. The method used to mold the fuel tank and material used
affect the permeation control strategies that may be used. This effect is discussed below.
Note that some handheld equipment has structurally-integrated constructions where the
fuel tank is part of the structure of the equipment. These fuel tanks are typically made out of
nylon 6 with up to 30 percent fiberglass reinforcement. Data in Chapter 5 suggest that these fuel
tanks would be able to meet the tank permeation standards without changing the fuel tank
material.
6.5.1 Hose Permeation
Barrier fuel hose incremental costs estimates are based on costs shared confidentially by
component manufacturers. These costs are supported by the costs of existing products used in
other nonroad and automotive applications.3738'39 For baseline fuel lines, we consider nitrile
rubber hose such as that used to meet SAE J30 R7 recommendations. For handheld equipment,
we consider the baseline fuel lines to be injected-molded rubber hose for structurally-integrated
constructions and clear elastomeric tubing for other equipment.
For this analysis, we considered three primary approaches to reducing permeation from
fuel hoses. The first was the use of thermoplastic fuel lines such as those used in automotive
applications. The incremental cost of these fuel lines is about $0-0.10/ft compared to typical
hose used on Small SI equipment. However, there have been concerns expressed in the past by
manufacturers that this fuel line is not flexible or durable enough for small nonroad applications.
6-65
-------
Final Regulatory Impact Analysis
Two other approaches are using thermoplastic or thermoelastomer barrier materials in the
fuel hose construction. Our estimate is that thermoplastic fuel lines, such as Teflon or THV800,
would result in an incremental cost to the manufacturer of about $0.75-0.85 per foot.
Manufacturers have expressed in the past that they would have to upgrade their fuel clamps for
the use of thermoplastic barrier hose. Therefore, we include an incremental cost for the two
clamps totaling $0.10. Manufacturers have recently shared with us that they believe the
standards can be met through the use of a lower cost approach. In this approach, the barrier layer
is made of a thermoelastomer such as FKM. Our estimate of the incremental cost for this
approach is $0.20-0.30 per foot. Although the high flexibility of thermoelastomers such as FKM
may allow manufacturers to use existing hose clamps, we also include the hose clamp cost due to
the uncertainty of how manufacturers will construct their equipment with the new hose.
In some handheld applications, the fuel lines are molded in intricate custom shapes rather
than extruded like traditional hoses. In these designs, a section of the fuel line is inside the fuel
tank while the remainder is external to the fuel tank. In addition, a vent line may be molded into
the same part. Because the tanks are typically sealed with a one way valve on the vent, the vent
lines are exposed to saturated vapor. The fuel lines may be formed from molded cured rubber
such as NBR or injection-molded out of a rubberized plastic such as Alcryn. A low permeation
approach would be to mold the fuel lines out of FKM which is a thermoelastomer used in other
fuel line applications. Based on a sample of six fuel lines (two of which included vent lines) we
got an average weight of 11 grams (0.025 Ibs.). Based on cost estimates of $1.00/lb. for NBR
and $10-15/lb. for FKM, we get a cost estimate of $0.25 to $0.35 per fuel line. Manufacturers
have raised the concern that if a new material is used, that they may need to modify their hose
connectors to make sure that the hose does not pull off the barbs. To account for this, we include
a $0.10 cost for the addition of clamps or hose connector modifications.
Table 6.5-2 presents the estimated incremental costs of low permeation hose for four
typical equipment configurations. These costs include the markup discussed above for overhead
and profit. Because these hose constructions are established technology, we consider the short
and long-term costs to be the same. We believe the standards can be achieved using a
thermoelastic barrier and therefore use these costs in our analysis.
Table 6.5-2: Fuel Line Permeation Cost Estimates for Typical Small SI Equipment
thermoplastic barrier hose
thermoelastic barrier hose
thermoelastic molded fuel line
HH
4", 1/8" ID
$0.54
$0.28
$0.48
WBM
8", 1/4" ID
$0.86
$0.34
NA
NHH#1
2 ft, 1/4" I.D.
$2.32
$0.77
NA
NHH#2
3 ft, 1/4" I.D.
$3.42
$1.10
NA
6.5.2 Tank Permeation
As discussed in earlier chapters, plastic fuel tanks for Small SI equipment are constructed
in one of three primary molding processes: blow-molding, injection-molding, and rotational
6-66
-------
Costs of Control
molding. Blow-molded tanks are primarily made of high-density polyethylene (HDPE),
injection-molded tanks are primarily HDPE or nylon, and rotational molded tanks are primarily
cross-link polyethylene (XLPE). Because the molding process can affect the permeation control
approaches available, we discuss the technologies for each approach individually.
6.5.2.1 All HDPE fuel tanks
Surface treatments can be used to reduce permeation from HDPE fuel tanks, whether
they are blow-molded, injection-molded, or rotational-molded. Our surface treatment cost
estimates are based on price quotes from a companies that specialize in fluorination40 and
sulfonation.41 In the fluorination process, costs are based on the number of fuel tanks that will fit
into the fluorination treatment chamber. Therefore, costs are higher for larger fuel tanks,
because less tanks will fit in the chamber. The price sheet referenced for our fluorination prices
assumes rectangular shaped containers. These fuel tanks would stack easily in the fluorination
treatment chamber with little wasted space. However, tor irregular shaped fuel tanks, less fuel
tanks would fit in the treatment chamber due to dead space between the tanks when they are
placed in the support baskets in the chamber. To account for this inefficiency with typical
shaped fuel tanks, we consider a void space equal to about 25 percent of the volume of the fuel
tank. For handheld equipment, we consider a void space of 100 percent because of the
structurally-integrated nature of many tanks.
For sulfonation, the shape of the fuel tanks is less of an issue because the treatment
process is limited only by the spacing on the production line which is roughly the same for the
range of fuel tank sizes used in Small SI equipment. These prices do not include the cost of
transporting the tanks; we estimated that shipping, handling and overhead costs would be an
additional $0.03 to $0.76 per fuel tank depending on tank size (using the same void space
estimates as above).42
Manufacturers, with high enough production volumes, could reduce the costs of
sulfonating fuel tanks by constructing an in-house treatment facility. The cost of a sulfonation
production line facility that could treat 150-500 thousand fuel tanks per year (depending on tank
size) would be approximately $800,000.43 This facility, which is designed to last at least 10
years, is made up of a SO3 generator, a scrubber to clean up used gas, a conveyor belt, and
injection systems for the SO3 gas and for the neutralizing agent (ammonia solution). The
manufacturer of this equipment estimates that the operating costs, which includes electricity and
chemicals, would be about 3 cents per tank. We based our costs on a production capacity of
300,000 units per year for handheld tanks and 150,000 units per year for non-handheld tanks. In
the long term, the costs would be based on the full life of the equipment which we estimate to be
10 years for this analysis. Finally, we use a labor rate of $28/hr with a 40 percent markup for
overhead which is consistent with our engine costs above and apply one full time employee to
operation of the sulfonation machine. A manufacturer that sulfonates its fuel tanks in-house
would not need to pay shipping costs. In the long run, we calculate that this approach will be
less expensive than shipping tanks to an outside facility.
6-67
-------
Final Regulatory Impact Analysis
6.5.2.2 Blow-molded fuel tanks
Manufacturers may reduce permeation from blow-molded fuel tanks by blending in a low
permeation material such as ethylene vinyl alcohol (EVOH) with the HDPE. This is typically
known by its trade name, Selar. The EVOH in the plastic forms non-continuous barrier platelets
in the tank during blow-molding that make it harder for fuel to permeation through the walls of
the tank. Using this approach, no changes should be necessary in the blow-molding equipment,
so the costs are based on increased material costs. We used 10 percent EVOH which costs about
$3-4 per pound and 90 percent HDPE which costs about $0.65-0.75 per pound.44 This equates to
a price increase of about $0.35 per pound. We then applied the material weights shown in
Table 6.5-1 to estimate costs per tank for this technology.
For higher production volumes, manufacturers may consider blow molding multi-layer
fuel tanks with continuous barriers. Practically, a new blow-molding machine would be required
because four or five additional injection screws would be necessary for the barrier layer, two
adhesion layers, an additional HDPE layer, and potentially a regrind layer. A machine that could
blow-mold multi-layer tanks would approximately double the price of the blow-molding
machine. For this analysis, we use a mono-layer machine cost of $1,000,000 and a multi-layer
machine cost of $3,000,00045, resulting in an increase in machine cost of $2,000,000. In
addition, tooling costs for each new tank design would be about $50,000. For this analysis we
considered a fuel tank with a material composition of 3 percent EVOH at $3.50/lb, 4 percent
adhesive layer at $l/lb, 45 percent regrind, and the remainder HDPE. Our analysis uses a total
annual production of 80,000-160,000 blow-molded tanks per year, depending on tank size
(smaller sizes would allow more tanks per mold), with 5 different molds. Capital costs are
amortized over 5 years in the short term and 10 years in the long-term (reflecting a 10 year life
of the machine).
6.5.2.3 Injection-molded fuel tanks
The technologies discussed above for blow-molded fuel tanks do not appear to be
feasible for injection-molded fuel tanks. The non-continuous barrier platelet approach does not
work well in this process because of the high shear stresses associated with injection molding.
Multi-layer rotomolded tanks would have to be formed by making separate molds, then fusing
the layers when the tank sides are welded together. While this may be possible, it would be
cumbersome. Barrier treatments would work for fuel tanks injected out of HDPE, but many
handheld tanks are injection molded out of nylon for better thermal resistance. At this time, it
appears that fluorination and sulfonation would not work effectively on nylon tanks. However,
nylon has low permeation on gasoline, and some nylon formulations are capable of meeting the
standards which are based on test fuel with 10 percent ethanol.
The advantages of injection molding are that it has lower tooling costs than blow-
molding and it is a faster molding process than rotational-molding. Although injection-molding
does not lend itself well to multi-layer construction, there is another process with similar costs
and production rates called thermoforming which does. Thermoforming entails using sheets of
plastic that are heated and pulled into a mold using vacuum suction. As with injection molding,
6-68
-------
Costs of Control
two halves are then joined together. In thermoforming, however, the sides are combined while
the plastic is still molten rather than by welding as is used in injection-molding. By using sheets
of extruded multi-layer plastic, thermoforming can be used to produce low-permeation, multi-
layer fuel tanks.
Because the thermoforming process requires extruded sheets, this process requires the
addition of an extruder. A small extruder, which would support several thermoforming machines
considered in this analysis would cost $2-3 million. The thermoforming machine itself would
cost about two-thirds that of an injection molding machine because it has less moving parts (such
as the injection screw). However, we estimate that two thermoforming machines would be
necessary to maintain the cycle time possible with an injection molding machine. At the same
time, hot plate welding machines would not be necessary because the tanks halves are assembled
in the thermoforming machine. We use an incremental cost savings of $100,000 for the molding
machine. Mold costs are somewhat lower for thermoforming as well because they are made of
aluminum rather than hardened steel. We estimate that a four-cavity injection mold would cost
about $60-80,000 while a four-cavity thermoforming mold would cost $20-30,000. For this
analysis we use a production of 300,000 tanks per year using 5 different molds. In the short
term, we amortize the fixed costs over 5 years, while in the long term we use 10 years to
represent the full life of the machines. Incremental material costs are based on 3 percent EVOH
and 4 percent adhesion material to create the barrier layer.
Another option would be to mold the entire fuel tank of a low permeation material such
as an acetal copolymer, or a thermoplastic polyester. These materials have list prices in the
range of about $1-2 per pound which is about double the material cost of HDPE, but comparable
to the cost of nylon.46 In addition, these fuel tanks could be made out of metal, which does not
permeate. For larger marine fuel tanks, metal tanks are available that cost about 25-30 percent
more than plastic fuel tanks (made under low volume construction). Private conversations with
Small SI equipment manufacturers suggest that making small fuel tanks out of metal could
increase the cost of the tanks for Small SI equipment by 200-300 percent and would limit the
possibility of constructing complex designs.
6.5.2.4 Rotational-molded fuel tanks
Many larger fuel tanks are rotationally molded. This process is more cost-effective for
smaller production volumes than blow-molding or injection-molding because of the lower
tooling costs for new tank designs. However, this process is slower which limits its usefulness
for large production volumes. Typically, rotational-molded fuel tanks manufactured for Small SI
equipment are made of cross-link polyethylene (XLPE). Although XLPE is more expensive than
HDPE which may also be used in the rotational-molding process, it is considered to be more
impact resistant than HDPE. This is important because the rotational molded fuel tanks are often
larger fuel tanks mounted on the outside of the equipment where it could be exposed to impacts
such as stepping, thrown rocks, branches, etc.
As discussed in Chapter 5, neither sulfonation or fluorination has been demonstrated to
be successful in creating a barrier on XLPE that would meet the new standards. Therefore, we
6-69
-------
Final Regulatory Impact Analysis
look to multi-layer approaches for our cost estimates. In the rotational-molding process, fuel
tanks may be formed with two layers. The traditional method is to add the first material to the
mold prior to entering the oven, and once that shell forms to add a second material through the
use of a drop box in the mold. Depending on the complexity and size of a drop box, it can add
from $1,000 to nearly $9,000 to the cost of the mold.47'48'49 One manufacturer is currently
making multi -layer rotational-molded fuel tanks for use Small SI equipment without the use of a
drop box. Their approach is proprietary, but the material manufacturer is making efforts to
develop an alternative to using a drop box as well.50 For this analysis, we include a $5,000 cost
for a drop box in the short term, but not in the long term. In addition, we do not project that this
process will have an increase on the cycle processing time because the increased heating time is
offset by decreased cooling time. The inner layer could be molded out of an acetal copolymer,
nylon, or even HDPE which could then be surface treated. Typical acetal copolymers cost about
the same as XLPE, although the rotational-molding grade may cost a little more.51 We use a cost
of $1.50/lb. for this acetal copolymer compared to XLPE which is approximately $1.20/lb.
Nylon, which can range in cost from $2 to $6 depending on the grade may also be used in
conjunction with XLPE to provide a permeation barrier. The advantage of nylon is that it bonds
to XLPE better than acetal copolymers. For this analysis, we consider the use of nylon at
$4.00/lb in a fuel tank with a 1 mm barrier and 4-5mm average total wall thickness. We
amortize the fixed cost of the drop boxes over 5 years of production of 1000 tanks per year for
each mold.
Another material is also available for molding an inner layer in rotomolded XLPE fuel
tanks. This material is poly butylene terephthalate cyclic oligimor and is known by the trade
name CBT®. With this material, no drop box is necessary. The CBT is added in the mold with
the XLPE resin. During the molding process, the XLPE shell forms in the mold. Due to
differences in viscosity and temperature properties, the CBT goes to the inside of the fuel tank.
It then polymerizes to form an inner liner. We use a cost of $5/lb. for CBT in this analysis and
use the same barrier thickness as discussed above.
Another technology that has been demonstrated for reducing permeation from XLPE fuel
tanks is a low permeation epoxy barrier. To apply this barrier, an adhesion treatment must first
be performed to increase the fuel tank surface energy so that the epoxy will adhere to the XLPE.
This can be done through a low level fluorination treatment. For this analysis, we use the cost of
level 1 fluorination.52 We use the same void space and shipping costs discussed above for our
fluorination cost analysis. The epoxy could be applied by dipping the fuel tank or spraying it on
like paint and then must be cured using UV light. We include a fixed cost of $10,000 for a
volume of 100,000 fuel tanks per year to account for coating and curing equipment. In addition,
we apply the cost of one full time employee to apply the coating and use a labor rate of $28/hr
with a 40 percent markup for overhead which is consistent with our engine costs above. For
traditional epoxies, we estimate that the cost would be $6-7/lb. Manufacturers have commented
that UV-curable epoxy, which could be processed much faster, would cost $12-15/lb.53'54 We
use a cost of $12/lb. for this analysis. Because only a thin coating needed (we use 0.125 mm),
the epoxy layer makes up only about 3 percent of the material of the fuel tank. Because there are
benefits to the epoxy coating such as allowing the fuel tank to be painted, there may be an
incentive to use this technology even on HDPE fuel tanks. For that reason, we estimated the cost
6-70
-------
Costs of Control
for smaller HDPE tanks as well using the same general assumptions except for a larger
production volume of 150,000 tanks per year due to their smaller size.
6.5.2.5 Summary of Fuel Tank Costs per Equipment
Table 6.5-3 summarizes the incremental costs of the fuel tank permeation emission-
control strategies discussed above. For technologies sold by a supplier to the engine
manufacturers, an additional 29 percent markup is included for the supplier's overhead and
profit. Both long-term and short-term costs are presented. The long-term costs account for the
stabilization of the capital investments and the learning curve effect discussed above. We use
the same material and shipping costs for our short-term and long-term estimates because these
cost components are well established with a wide range of applications. As discussed above, for
the multilayer fuel tank constructions, we consider an EVOH barrier for hand-held and Class I
equipment and nylon barrier for Class II equipment.
Table 6.5-3: Tank Permeation Control Cost Estimates for Typical Small SI Equipment
fluorination3'b: short term
long term
sulfonation3'b: short term
long term
non-continuous platelets3
multi-layer3: short term
EVOH long term
multi-layer0: short term
PA 11 long term
multi-layer0: CBT
thermo-formingb: short term
long term
acetal-copolymer3'b'°
metal construction3'13'0
epoxy coating3'b'°: short term
long term
HH
0.25 gallons
IM/BM
$0.62
$0.50
$0.64
$0.52
$0.17
$4.13
$2.01
NA
NA
$0.36
$0.20
$0.62
$1.94
$1.26
$1.01
WBM
0.5 gallons
IM/BM
$0.77
$0.63
$1.25
$1.01
$0.22
$4.08
$1.98
NA
NA
$0.53
$0.29
$0.79
$3.87
$1.32
$1.06
NHH#1
2 gallons
IM/BM
$3.10
$2.52
$1.40
$1.16
$0.51
$3.80
$1.75
NA
NA
$1.50
$0.82
$1.82
$5.16
$2.56
$2.08
NHH#2
5 gallons
RM
NA
NA
NA
NA
$5.54
$3.40
$5.77
NA
$2.28
$9.68
$5.69
$4.64
a incremental to traditional
b incremental to traditional
c incremental to traditional
blow-molding
inj ection-molding
rotational-molding
6-71
-------
Final Regulatory Impact Analysis
6.5.3 Venting Losses
Venting losses are made up of diurnal breathing losses and running losses which are
similar to diurnal emissions except that the heating event is caused by the engine. We are
requiring that equipment manufacturers install systems to capture their running losses by sealing
the fuel tank and venting vapor to the engine intake. For the purpose of our cost analysis, we
consider a system with a purge hose running from the fuel tank to the engine intake (with 2 hose
clamps) that is the same length of the fuel hose. We use a cost of $0.25/ft for the hose and $0.10
each for the two hose clamps. This is consistent with the above cost analysis for low permeation
hose. We also consider a fuel cap redesign to meet the sealing requirements with a one way
valve to prevent a vacuum from occurring in the fuel tank as fuel is drawn out to the engine. We
use a cost of $1 for the valve and cap redesign. Also, we include a cost of $0.10 to account for a
limiting flow orifice in the purge line. Finally, using the labor costs discussed above, we
calculate an incremental assembly labor cost of about $0.20 per engine.
Diurnal emissions could be captured through the use of a carbon canister. The carbon
then could be purged by air drawn into the fuel tank as the fuel cools. This is known as passive
purge. This system would be similar to the running loss control system except that venting
would occur through a canister and the valving would be modified to provide liquid/vapor
separation. This valve would prevent fuel from entering the canister if the equipment were
tipped over. We estimate the cost of a canister to vary based on size ranging from about $2 for a
1 quart tank to about $4 for a five gallon tank. The majority of these canister costs for small fuel
tanks are for the canister, connections, and mounting hardware. As the fuel tank size increases,
the carbon becomes a more significant fraction of the cost. For this analysis, we add the cost of
the canister to the cost of running loss control and include another $0.20 for assembly costs.
Diurnal emissions could be controlled further through an active purge canister system. In
an active purge system, the canister would also be purged by the engine during operation. The
added components of this system compared to the passive purge system would include a line to
the air filter (or separate air filter for the canister breathing line) and a purge valve. This
amounts to an additional cost of $0.15/ft for the air line, $0.20 for two clamps, $1 for the purge
valve, and another $0.20 for assembly.
6-72
-------
Costs of Control
Table 6.5-4: Venting Control Cost Estimates for Typical Small SI Equipment
running loss: short term
long term
passive purge canister* : short term
long term
active purge canister* * : short term
long term
WBM
0.5 gallons
8", 1/4" ID
$2.06
$1.65
$3.07
$2.45
$1.93
$1.54
NHH#1
2 gallons
2 ft, 1/4" ID
$2.32
$1.85
$3.82
$3.06
$2.19
$1.75
NHH#1
5 gallons
3 ft, 1/4" I.D.
$2.51
$2.01
$4.38
$3.51
$2.38
$1.91
* incremental to running loss control
** incremental to passive purge canister
6.5.4 Certification and Compliance
The running loss standards call for manufacturers to certify their running loss systems
based on design rather than requiring emission testing. However, they will still need to integrate
the emission-control technology into their designs and there will be some engineering and
clerical effort need to submit the required information for certification. We expect that in the
early years, plastic fuel tank manufacturers will perform durability and permeation testing on
their fuel tanks for certification. They will be able to carry over this data in future years and will
be able to carry across this data to other fuel tanks made of similar materials and using the same
permeation control strategy regardless of tank size or shape. Typical certification costs may be
spread between the tank manufacturer, hose manufacturer, and equipment manufacturer. For the
sake of this analysis, we combine the tank, hose, and boat certification costs to calculate the total
certification of an average fuel system. We estimate that 90 percent of fuel tank sales in Small
SI equipment are plastic and the remainder are metal.
For the first year we estimate fuel tank durability and certification testing to cost about
$15,000 per tank manufacturer on the assumption that the manufacturer will use the same
materials and permeation control strategy for all of their fuel tanks to reduce costs. Low
permeation fuel lines are largely an established technology. However, we include a cost of
$1,000 to perform certification testing on fuel lines. In addition, we estimate about $10,000 for
engineering and clerical work for the equipment manufacturers.
For handheld equipment manufacturers, we spread these costs over sales of 500,000 units
per year. For handheld and Class I equipment manufacturers, which are integrated
manufacturers, we base the costs on average annual sales per manufacturer. We estimate the
average annual sales to be about 500,000 units for handheld equipment and 100,000 units for
Class I equipment. Generally for Class II equipment, a large number equipment manufacturers
purchase their engines from a smaller number of engine manufacturers. We estimate average
annual sales per year to be 50,000 units for Class II.
6-73
-------
Final Regulatory Impact Analysis
As with other fixed costs, we amortized the cost over 5 years of sales to calculate per unit
certification costs. Combining these costs, we get average fuel system integration and
compliance costs of about $0.01 for handheld equipment, $0.05 for Class I equipment, and $0.10
for Class II equipment.
6.5.5 Operating Cost Savings
Evaporative emissions are essentially fuel that is lost to the atmosphere. Over the
lifetime of a piece of Small SI equipment, this can result in a significant loss of fuel. The
reduction in evaporative emissions would therefore result in meaningful fuel savings which can
be directly related to operating cost savings based on an average density of 6 Ibs/gallon for
gasoline (based on lighter hydrocarbons which evaporate first) and the price of gasoline
described above. Table 6.5-5 presents the estimated fuel savings for Small SI equipment
associated with the evaporative emission standards.
Table 6.5-5: Projected Evaporative Fuel Savings for Small SI Equipment
Evaporative HC Reduced [Ibs/life]
Lifetime Gallons Saved
Lifetime Cost Savings
Average Equipment Life [years]
Discounted Cost Savings (7%)
Handheld
1.4
0.2
$0.41
4.2
$0.40
Class I
4.8
0.8
$1.45
5.3
$1.31
Class II
28.6
4.7
$8.55
5.9
$5.96
6.5.6 Total Small SI Equipment Costs
We expect that Small SI manufacturers will use a variety of technologies to meet the fuel
tank permeation standards. As discussed above, many options are available so the technologies
chosen will depend on the baseline fuel tank construction, the equipment application, and the
manufacturers' particular design philosophies. Hose permeation standards will likely be met
through the use of barrier hose constructions.
For the purpose of this analysis, we divided Small SI equipment into 23 categories to
better quantify differences in costs that may be associated with different equipment applications.
Earlier in this chapter, engine costs are presented as a function of design life. However, we
believe evaporative emission costs are more a function of the application than the design life due
to the differences in hose lengths and tank sizes and constructions. Manufacturers would not
likely design a less robust fuel system for equipment used with lower hour engines. Table 6.5-6
presents our assessment of the mix of the fuel system constructions used today. This assessment
is based on the NONROAD 2005 model and on confidential information supplied by Small SI
equipment manufacturers.
6-74
-------
Costs of Control
Table 6.5-6: Baseline Technology Mix for Small SI Equipment
Equipment Class
Fuel Line Description
Length ft*
construction
Fuel Tank Construction
gallons
material/process* *
Handheld Equipment
Class III commercial
Class III residential
Class IV commercial
Class IV residential
Class V
0.25
0.25
0.33
0.33
0.50
rubber hose
rubber hose
6% molded line
24% molded line
52% molded line
0.9
0.3
0.4
0.3
0.5
HOPE
HOPE
6% Nylon/94% HOPE
24% Nylon/76% HOPE
52% Nylon/48% HOPE
Class I Equipment
ag/const/gen ind/mat hand
commercial mowers
residential mowers
com. other L&G
res. other L&G
pumps/comp/press. wash
snow equipment
utility /rec. vehicles
welders/generators
0.72
0.72
0.62
0.72
0.62
0.72
0.63
0.72
0.72
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
1.6
0.8
0.4
1.1
0.6
0.8
0.3
3.6
0.8
100%IM
90%IM/10%BM
100%IM
90%IM/10%BM
100%IM
100%IM
100%IM
100%IM
100%IM
Class II Equipment
ag/const/gen ind/mat hand
commercial mowers
residential mowers
com. other L&G
res. other L&G
pumps/comp/press. wash
snow equipment
utility /rec. vehicles
welders/generators
3.6
6.5
3.2
1.5
1.1
2.6
1.2
2.7
3.8
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
rubber hose
5.4
4.7
2.6
1.2
5.0
4.7
0.7
3.9
6.0
60% EM/40% RM
60% EM/40% RM
70/18/12%IM/BM/RM
60% EM/40% RM
70/18/12%IM/BM/RM
60% EM/40% RM
60% EM/40% RM
60% EM/40% RM
60% EM/40% RM
* we use 1/8" I.D. for handheld and 1/4" I.D. for non-handheld hose
** IM = injection molded HOPE, BM = blow-molded HOPE, RM = rotational-molded XLPE
We base our fuel tank costs on several technologies. In our cost analysis for handheld
engines, we model costs based on fluorination for HDPE tanks, but we do not apply costs to
tanks that are molded out of nylon as these tanks would likely meet the standards today. For
non-handheld equipment, we split the costs of permeation control of injected molded FfDPE fuel
tanks 50/50 between fluorination and converting to multi-layer thermoformed constructions with
an EVOH barrier. For blow-molded fuel tanks, we base our costs on using a multi-layer
blowmolded construction with an EVOH barrier. For rotational-molded XLPE fuel tanks, we
base our costs on rotational-molding a nylon layer in the tank.
6-75
-------
Final Regulatory Impact Analysis
For fuel line permeation, we distinguish between the costs for traditional hose versus
molded fuel lines. Fuel hose costs are based on using a fluoroelastomer barrier within the
traditional construction. For molded fuel lines, we base the costs on molding the parts
completely out of a high-grade fluoroelastomer. We do not apply costs to fuel lines used in cold-
weather equipment.
As discussed above, our cost estimates include both variable and fixed costs, and we
distinguish between near-term and long-term costs. Because our analysis amortizes fixed costs
over 5 years, the long-term costs are generally made up of variable costs only. The exception to
this is fuel tank permeation control strategies where more expensive molding equipment is used.
We assume an equipment life of 10 years, so in the long term, the amortized additional cost of
the molding equipment is half, on average, of the short-term amortized cost over 5 years (5 years
of amortized payments/10 years of equipment life = l/2). In addition, variable costs are lower in
the long term due to the learning effect discussed in Section 6.1. Table 6.5-7 presents these
average per-engine cost estimates.
Table 6.5-7: Small SI per Equipment Cost Estimates (Without Fuel Savings)
Handheld aggregate
tank permeation
hose permeation
Class I aggregate
tank permeation
hose permeation
running loss
Class II aggregate
tank permeation
hose permeation
running loss
Short Term (years 1-5)
Fixed
$0.01
$0.01
$0
$0.47
$0.45
$0.02
$0
$1.26
$1.21
$0.04
$0
Variable
$0.81
$0.62
$0.19
$2.58
$0.33
$0.33
$1.92
$5.47
$2.15
$1.09
$2.23
Total
$0.82
$0.63
$0.19
$3.05
$0.78
$0.35
$1.92
$6.73
$3.37
$1.13
$2.23
Long Term (years 6-10)
Fixed
$0
$0
$0
$0.19
$0.19
$0
$0
$0.69
$0.69
$0
$0
Variable
$0.69
$0.50
$0.19
$2.01
$0.27
$0.20
$1.53
$4.48
$1.73
$0.96
$1.78
Total
$0.69
$0.50
$0.19
$2.20
$0.46
$0.20
$1.53
$5.16
$2.42
$0.96
$1.78
6.5.7 Small SI Equipment Aggregate Costs
Aggregate costs are calculated by multiplying the per-equipment variable cost estimates
described above by projected equipment sales. Fixed costs are added as incurred. Fuel savings
are calculated directly from the projected HC reductions due to the evaporative emission
standards. Table 6.5-8 presents the projected costs of the rule over a 30-year time period with
and without the fuel savings associated with reducing evaporative emissions.
At a 7 percent discount rate, over 30 years, the estimated annualized cost to
manufacturers for Small SI evaporative emission control is $65 million. The estimated
corresponding annualized fuel savings due to control of evaporative emissions from Small SI
6-76
-------
Costs of Control
equipment is $53 million. At a 3 percent discount rate, the estimated annualized cost to
manufacturers for Small SI evaporative emission control is $68 million. The estimated
corresponding annualized fuel savings due to control of evaporative emissions from Small SI
equipment is $59 million.
Table 6.5-8: Projected 30-Year Aggregate Cost Stream for Small SI Evap
Year
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
2037
Without Fuel Savings
Handheld
$224,312
$5,942,463
$5,816,345
$5,923,313
$7,752,522
$7,883,770
$6,810,314
$6,922,632
$7,034,855
$7,147,090
$7,259,067
$7,371,143
$7,483,470
$7,595,660
$7,707,763
$7,819,853
$7,931,999
$8,044,212
$8,156,448
$8,268,656
$8,380,840
$8,493,060
$8,605,303
$8,717,528
$8,829,741
$8,941,949
$9,054,168
$9,166,396
$9,278,617
$9,390,834
Class I
$4,648,719
$4,732,285
$9,215,094
$9,277,641
$30,008,965
$29,122,813
$29,609,400
$30,093,095
$30,569,878
$25,830,285
$26,235,145
$26,644,831
$27,051,954
$27,457,582
$27,860,231
$28,266,579
$28,672,351
$29,079,725
$29,490,381
$29,900,304
$30,309,200
$30,719,307
$31,129,464
$31,540,051
$31,950,436
$32,360,405
$32,770,086
$33,180,125
$33,590,247
$34,000,550
Class II
$7,105,928
$13,749,794
$12,455,089
$35,396,135
$36,003,944
$35,892,762
$36,495,479
$37,094,450
$32,298,515
$32,819,027
$33,338,022
$33,862,125
$34,382,368
$34,902,482
$35,418,720
$35,938,441
$36,458,489
$36,981,966
$37,505,944
$38,028,453
$38,550,330
$39,073,626
$39,597,054
$40,121,181
$40,644,586
$41,167,490
$41,690,229
$42,213,436
$42,736,659
$43,260,062
With Fuel Savings
Handheld
$224,312
$5,707,932
$5,135,637
$4,845,509
$6,079,623
$5,677,010
$4,191,701
$4,019,074
$3,902,020
$3,897,310
$3,924,431
$3,968,150
$4,021,523
$4,080,576
$4,139,728
$4,198,990
$4,258,379
$4,317,834
$4,377,298
$4,436,735
$4,496,146
$4,555,595
$4,615,066
$4,674,519
$4,733,961
$4,793,398
$4,852,844
$4,912,301
$4,971,750
$5,031,196
Class I
$3,697,028
$2,634,260
$6,081,748
$5,276,627
$21,187,318
$16,186,469
$13,744,050
$12,389,454
$11,344,204
$5,337,945
$4,963,819
$4,768,572
$4,660,933
$4,604,761
$4,586,374
$4,607,753
$4,648,613
$4,691,017
$4,733,531
$4,775,266
$4,815,980
$4,857,899
$4,899,855
$4,942,267
$4,984,464
$5,026,247
$5,067,730
$5,109,583
$5,151,512
$5,193,635
Class II
$4,886,033
$8,715,782
$4,833,539
$20,401,655
$14,937,731
$9,525,916
$6,260,075
$3,382,349
$(4,288,346)
$(5,765,846)
$(6,881,442)
$(7,709,855)
$(8,357,363)
$(8,820,939)
$(9,198,114)
$(9,511,148)
$(9,799,648)
$(10,064,140)
$(10,315,949)
$(10,558,303)
$(10,793,367)
$(11,021,049)
$(11,243,155)
$(11,460,730)
$(11,676,805)
$(11,890,232)
$(12,101,049)
$(12,309,918)
$(12,516,613)
$(12,721,142)
6.6 Costs of Evaporative Emission Controls for Marine Vessels
This section presents our cost estimates for meeting the new evaporative emission
standards for marine vessels.
To determine the cost impacts of the evaporative emission standards on marine fuel
systems, we considered three primary marine applications. The first is a portable fuel tank with
a detachable fuel line and a primer bulb. The second is a personal watercraft vessel. The third is
6-77
-------
Final Regulatory Impact Analysis
a larger vessel with an installed fuel tank and fuel lines meeting SAE J1527 specifications. In
our cost analysis, we consider a wide range of vessel sizes for each of these categories.
However, to simplify this discussion we only present our cost estimates for the three typical
applications shown in Table 6.6-1. For this illustration, costs are based on vessels with one fuel
tank and one engine. Although these typical configurations do not, by any means, represent all
of the vessel types included in our cost calculations, they should give a good indication of how
we performed our analysis.
Table 6.6-1: Typical Marine Vessel Fuel System Configurations
Fuel Tank Capacity (gallons)
Fuel Tank Material*
Fuel Tank Molding Process
Fuel Tank Weight (Ibs.)
Fuel Hose: Length (ft.)
Inner Diameter (in.)
Vent Hose: Length (ft.)
Inner Diameter (in.)
Fill Neck: Length (ft.)
Inner Diameter (in.)
Portable Tank
6
HOPE
blow-molded
4.4
6, primer bulb
1/4
-
-
-
-
PWC
17
HOPE
blow-molded
12
5.7
1/4
2
1/4
1.9
1.5
Installed Tank
57
XLPE
rotational-molded
55
9.9
3/8
8.0
5/8
10.1
1.5
* HDPE = high-density polyethylene, XLPE = cross-link polyethylene
Fuel tank weights are based on measurements of fuel tanks used in our permeation testing
and are used to determine material costs. XLPE fuel tanks are typically thicker walled; thus they
typically weigh more per gallon of capacity. Fuel hose lengths are based on conversations with
(and confidential business information from) boat builders and fuel system suppliers. This data
is within the range of hose lengths included in the written comments made by one boat builder
on our earlier proposal.55
6.6.1 Hose Permeation
There are several grades of fuel system hose used in marine applications. For sterndrive
and inboard (SD/I) applications, Title 33 of the Code of Federal Regulations, Part 183 defines
fuel system requirements. These requirements reference SAE J1527 for fuel hose specifications.
For personal watercraft (PWC), fuel line specifications are defined in SAE J2046. For
outboards, no fuel hose specifications exist. Typically, larger vessels, with installed fuel tanks
use SAE J1527 Class I hose for lines filled with fuel and Class II hose for lines containing fuel
vapor. Inner diameters (ID) of these fuel system lines are typically 3/8" for fuel lines, 5/8" for
vent lines, and 1.5" for fill necks. PWC typically have fuel supply/return hose with a 1/4" ID.
Portable marine fuel tanks for outboards typically have fuel lines with a 1/4" ID and a primer
bulb. Fill neck hose is made by wrapping several layers of materials over a mandrill and
vulcanizing the rubber in an oven. The remaining fuel lines are typically extruded. Fuel hose
meeting the CFR requirements typically has several layers for durability and flame resistance.
6-78
-------
Costs of Control
Barrier fuel hose incremental costs estimates are based on costs of existing products used
in marine and automotive applications.56'57'58'59'60 Because the manufacturing process is not
fundamentally changed in adding a barrier layer, this cost is mostly the result of more expensive
materials. For 1/4" hose such as used in some small outboards and personal watercraft, we
estimate a cost increase of $0.25/ft for a thermoelastic barrier and $0.85/ft for a thermoplastic
barrier. These costs are consistent with the costs described above for Small SI equipment.
SD/I vessels are required to use marine fuel hose meeting Coast Guard requirements
specified in 33 CFR part 183. This hose is recommended by the American Boat and Yacht
Council for outboard boats not using portable fuel tanks as well. Marine hose with a nylon
barrier is available today that meets these requirements. The cost differential of traditional
versus marine barrier hose for fuel and vent lines in the market today varies from no cost at all to
more than $1 per foot. One hose distributer stated that they sell both non-barrier and barrier hose
at the same price. They stated that the fuel resistance provided by the barrier layer allows the
hose construction to use a thinner wall and therefore use less rubber. Another hose distributor,
lists about a $1 cost markup for Al barrier hose compared to their B1 marine hose. Note that B1
hose does not meet the Coast Guard fire requirements for fuel lines and this may be part of the
reason for the cost differential. For this analysis, we use a cost increase of $0.50/ft for fuel hose
and $1.00 for vent hose for vessels with installed fuel tanks. We use a higher incremental cost
for vent hose because this hose typically has a larger diameter, requiring more material.
For IVa" fill neck hose, we estimate a cost increase of $2.00/ft. This cost increase is
based on our estimates of material and labor costs. The fill neck hose would be constructed in
the same manner as today except that a thin barrier layer would be included in the multi-layer
construction. One hose distributer advertises barrier fill-neck hose with a price markup of $9 per
foot. However, this cost markup likely represents the high costs typical of special orders where
setup costs must be spread over low hose production. Currently, little or none of this hose is
purchased by boat builders. Our price estimate is more consistent with differences in cost for
barrier versus non-barrier chemical hose manufactured in the same manner.
We do not expect the addition of a barrier layer to affect the flexibility of the hose
because marine hose is already fairly stiff and because the barrier layer is very thin and flexible.
In fact, the barrier hose samples we tested appeared a little more flexible than the baseline hose
because less wall thickness was needed for permeation control. Therefore, we believe special
hose clamps or fittings will not typically be required.
Primer bulbs are typically formed from molded cured rubber such as NBR or injection-
molded out of a rubberized plastic such as Alcryn. Primer bulbs could also be molded from
FKM which is a fluoroelastomer used in fuel line applications. Primer bulbs typically weigh
between 0.1 and 0.2 Ibs, nitrile costs about $1.00/lb and FKM costs about $10-15/lb depending
on the level of fluorine in the material. If the whole primer bulb was molded out of FKM, it
would increase the material cost by about $1.50-2.00 per primer bulb. Alternatively,
manufacturers could save on material costs by injection molding an inner layer of Alcryn and
curing a coating of FKM over this shell. Using a higher grade of FKM ($15/lb) could help
minimize the amount of the fluoroelastomer needed. For the multi-layer design, we assume
6-79
-------
Final Regulatory Impact Analysis
about 30-50 percent of the material would be FKM which results in a material cost increase of
about $0.90 per primer bulb.
Table 6.6-2 presents our estimates of incremental costs for low permeation marine fuel
system hose. Primer bulb costs are presented both for 100 percent FKM and multi-layer
constructions. The incremental cost for the 1/4" fuel lines are presented for the thermoelastic
barrier and the costs for the heavier fuel hose are based on costs of existing nylon barrier marine
hose. These costs include a markup, and no long-term cost savings are applied to these costs
because they are primarily material costs.
Table 6.6-2: Hose Permeation Control Cost Estimates for Typical Marine Vessels
primer bulb
100% FKM
multi-layer
fuel supply/return
fill neck
vent hose
Portable Tank
6', 1/4" ID fuel hose
primer bulb
$2.13
$1.16
$1.94
-
-
PWC
5.7', 1/4" ID fuel hose
1.9', 1.5 "ID fill neck
2.0', 1/4" ID vent hose
-
-
$1.84
$5.16
$0.65
Installed Tank
9.9', 3/8" ID fuel hose
10. 1', 1.5 "ID fill neck
8.0', 5/8 "ID vent hose
-
-
$6.58
$26.12
$10.29
6.6.2 Tank Permeation
Portable fuel tanks and fuel tanks used in personal watercraft are typically blow-molded
out of HDPE and have a capacity ranging from 4 to!8 gallons. Because of the manufacturing
process and material used, some permeation control technologies are available that are different
from what would be feasible for larger rotational-molded fuel tanks. Larger, low-production
volume marine fuel tanks are typically rotational-molded out of XLPE. Rotational-molding is
used for smaller production runs because of the much lower relative tooling costs compared to
blow-molding. For fuel tanks in vessels that are subject to the 33 CFR 183 fuel system
requirements, manufacturers have found that fuel tanks molded out of HDPE will not pass the
fire test, while XLPE fuel tanks will. Therefore, XLPE is used in rotational-molded marine fuel
tanks.
6.6.2.1 Blow-Molded Fuel Tanks
Our surface treatment cost estimates are based on price quotes from companies that
specialize in this fluorination61 and sulfonation.62 The fluorination costs are a function of the
geometry of the fuel tanks because they are based on how many fuel tanks can be fit in a
treatment chamber. The price sheet referenced for fluorination assumes rectangular shaped
containers. For irregular shaped fuel tanks, the costs would be higher because they could not
efficiently utilize the chamber volume. There would be significant void space. We consider a
6-80
-------
Costs of Control
void space equal to about 25 percent of the volume of the fuel tank. For sulfonation, the shape of
the fuel tanks is less of an issue because the treatment process is limited only by the spacing on
the production line which is roughly the same for the range of fuel tank sizes used for
portable and personal watercraft fuel tanks. These prices do not include the cost of transporting
the tanks; we estimated that shipping, handling and overhead costs would be an additional $0.40-
$1.40 per fuel tank, for tanks ranging from 4-18 gallons.63
As discussed above for Small SI fuel tanks, manufacturers, with high enough production
volumes, could reduce the costs of sulfonating fuel tanks by constructing an in-house treatment
facility. We base our costs for marine fuel tanks on 150,000 tanks per year and use this approach
for our long-term cost determination for sulfonation.
Our estimate of the cost for non-continuous barrier platelets (generally known as Selar) is
based on increased material costs. No changes should be necessary to the blow-molding
equipment. We used 10 percent ethylene vinyl alcohol (EVOH) which is about $3-4 per pound
and 90 percent HDPE which is about $0.65-0.75 per pound.64 This equates to a price increase of
about $0.35 per pound. We then applied the material weights shown in Table 6.5-1 to estimate
costs per tank for this technology.
For higher production volumes, manufacturers may consider blow molding multi-layer
fuel tanks with continuous barriers. Practically, a new blow-molding machine would be required
because four or five additional injection screws would be necessary for the barrier layer, two
adhesion layers, an additional HDPE layer, and potentially a regrind layer. A machine that could
blow-mold multi-layer tanks would approximately double the price of the blow-molding
machine For this analysis, we use a mono-layer machine cost of $1,000,000 and a multi-layer
machine cost of $3,000,000 for smaller tanks and $4,000,000 for larger tanks (>6 gallons)65,
resulting in an increase in machine cost of $2,000,000-$3,000,000. In addition, tooling costs for
each new tank design would be about $50,000. For this analysis we considered a fuel tank with
a material composition of 3 percent EVOH at $3.50/lb, 4 percent adhesive layer at $l/lb, 45
percent regrind, and the remainder HDPE. Our analysis uses a total annual production of
60,000-80,000 blow-molded tanks per year, depending on tank size, with 5 different molds.
Capital costs are amortized over 5 years in the short term and 10 years in the long-term
(reflecting a 10 year life of the machine).
6.6.2.2 Rotational-Molded Fuel Tanks
Most installed fuel tanks are rotational-molded out of XLPE for the reasons discussed
above. As discussed above, barrier treatments have not been demonstrated to provide effective
permeation control for XLPE. In addition, Selar and traditional multi-layer blow-molding
approaches do not work for rotational-molded cross-link polyethylene fuel tanks.
Two approaches were discussed above in the Small SI section for rotational-molded
XLPE fuel tanks: 1) dual-layer molding with a barrier layer and 2) epoxy coating of fuel tanks.
These approaches could also be applied to marine fuel tanks. For the dual layer approach,
marine fuel tank manufacturers have expressed concern that the acetal copolymer will not adhere
6-81
-------
Final Regulatory Impact Analysis
well to the XLPE. For large fuel tanks, this could be an issue because the layers could pull apart
and cause leaks at the fittings. As an alternative, one company has developed an approach using
a high grade, non-hygroscopic nylon known a polyamide 11 as a barrier layer. This material
costs about $5-7/lb compared to XLPE which costs about $1.20/lb. The barrier layer would
likely be about 20 percent of the total material. Using a nylon cost of $6/lb. and a barrier fraction
of 30 percent, we get an average material cost of $2.64/lb. For the short term, we add a $5,000
cost to the mold or a drop box which we amortize over 100 tanks per year for 5 years.
Consistent with the analysis for Small SI equipment, we do not include the cost of a drop box in
the long term because of the ongoing development of a process that does not require a drop
box.66 In fact, one manufacture is already using a proprietary process to mold multi-layer
rotational-molded fuel tanks without a drop box.
Another material is also available for molding an inner layer in rotomolded XLPE fuel
tanks. This material is poly butylene terephthalate cyclic oligimor and is known by the trade
name CBT®. With this material, no drop box is necessary. The CBT is added in the mold with
the XLPE resin. During the molding process, the XLPE shell forms in the mold. Due to
differences in viscosity and temperature properties, the CBT goes to the inside of the fuel tank.
It then polymerizes to form an inner liner. We use a cost of $5/lb. for CBT in this analysis and
use the same barrier thickness as discussed above.
Another technology that has been demonstrated for reducing permeation from XLPE fuel
tanks is a low permeation epoxy barrier. To apply this barrier, an adhesion treatment must first
be performed to increase the fuel tank surface energy so that the epoxy will adhere to the XLPE.
This can be done through a low level fluorination treatment. For this analysis we use the cost of
level 1 fluorination.67 We use the same void space and shipping costs discussed above for our
fluorination cost analysis. Shipping costs are estimated to range from $4-$ 10 per tank for
20-130 gallon tanks. The epoxy could be applied by dipping the fuel tank or spraying it on like
paint and then the epoxy must be allowed to cure. We include a fixed cost of $10,000 for a
volume of 15,000 fuel tanks per year to account for coating and curing equipment. In addition,
we apply the cost of part of one employee's time (using a labor standard of 15,000 tanks
annually per employee) time to apply the coating and use a labor rate of $28/hr with a 40 percent
markup for overhead which is consistent with our engine costs above. We estimate that the
epoxy cost would be $6-7/lb. Manufacturers have commented that UV-curable epoxy, which
could be processed much faster, would cost $12-15/lb.68'69 We use a cost of $12/lb. for this
analysis. However with only a thin coating needed (we use 0.125 mm), the epoxy layer makes
up only about 2.0-2.5 percent of the material of the fuel tank. Because there are benefits to the
epoxy coating such as allowing the fuel tank to be painted, there may be an incentive to use this
technology even on HOPE fuel tanks. For that reason, we estimated the cost for portable fuel
tanks as well using the same general assumptions except for a larger production volume of
100,000 tanks per year with a increased labor standard due to the smaller tank sizes.
6.6.2.3 Other Marine Fuel Tank Constructions
We do not anticipate that the permeation standard would affect the cost of metal fuel
tanks. Although some permeation can occur at rubber seals (such as for the sending unit), this
6-82
-------
Costs of Control
would be small due to the small exposed surface area of the seals.
Another type of fuel tank construction that is used in some applications, such as offshore
racing boats, is fiberglass fuel tanks. This fiberglass is commonly made of vinyl ester or epoxy
which have high permeation rates. One manufacturer has developed a fiberglass composite that
uses treated volcanic ash in a carrier matrix to create a non-continuous permeation barrier. This
composite is known as an unsaturated polyester nanocomposite (UPE). In addition to being a
low permeation technology for fiberglass tanks, this construction could also be used as an
alternative for metal or plastic fuel tanks. These low permeation fiberglass constructions can be
fabricated or molded. We estimate that fabricated fiberglass composite fuel tanks would cost at
least as much as metal fuel tanks because of the labor involved in hand constructing the tanks.
However, these fuel tanks may also be molded with an average mold cost of $2,500.™ For the
purposes of this analysis we use a cost increase of 20 percent when comparing this technology to
rotational-molded fuel tanks which is a somewhat lower than the cost of a metal fuel tank.
6.6.2.4 Summary of Fuel Tank Costs per Vessel
Table 6.6-3 summarizes the incremental costs of the fuel tank permeation emission-
control strategies discussed above. For technologies sold by a supplier to the engine
manufacturers, an additional 29 percent markup is included for the supplier's overhead and
profit. Both long-term and short-term costs are presented. The long-term costs account for the
stabilization of the capital investments and the learning curve effect discussed above. We use
the same material, shipping, and fluorination costs for our short-term and long-term estimates
because these cost components are well established with a wide range of applications. As
discussed above, for the multilayer fuel tank constructions, we consider an EVOH barrier for
portable and PWC fuel tanks and a polyamide 11 barrier for rotational-molded fuel tanks. UPE
fiberglass nanocomposite costs presented here are incremental to rotational-molded XLPE tanks.
6-83
-------
Final Regulatory Impact Analysis
Table 6.6-3: Tank Permeation Control Cost Estimates for Typical Marine Vessels
fluorination: short term
long term
sulfonation: short term
long term
non-continuous platelets
multi-layer: short term
EVOH long term
multi-layer: short term
PA11 long term
multi-layer: CBT
UPE fiberglass short term
nanocomposite long term
epoxy coating: short term
long term
Portable Tank
6 gallons
$9.30
$7.44
$1.67
$1.26
$1.27
$7.74
$4.22
NA
NA
NA
$5.47
$4.85
PWC
17 gallons
$26
$21
$3.27
$1.29
$3.37
$15
$8.58
NA
NA
NA
$12
$11
Installed Tank
57 gallons
NA
NA
NA
$81
$68
$81
$68
$54
$48
$39
$43
$39
6.6.3 Venting Losses
For portable fuel tanks, the standards would require the fuel cap to be modified to remove
the user-controlled screw and add a one-way valve. We estimate that the cost of a vacuum relief
valve would be about $0.50 more than the manual valve used on portable fuel tanks today. We
double this cost to account for upgrading the valve for marine applications. For personal
watercraft, we are not claiming any costs or benefits because these vessels already seal their fuel
tanks with a pressure relief valve.71
Larger fuel tanks are currently vented to atmosphere. One emission-control technology
that could be used to meet our standards would be to seal the fuel tank and use a 1 psi pressure
relief valve to prevent over-pressure. However, manufacturers have commented that their fuel
tanks are not designed to withstand pressure and that the current molding process does not lend
itself to making the fuel tanks more pressure resistant. Their fuel tanks currently deflect
significantly at pressures as low as 1 psi. However, for some fuel tank constructions, a sealed
system may be a viable option. For our cost analysis of this approach, we estimate the cost of a
pressure relief valve to be about $1 based on products available in automotive applications. We
double this cost to account for either upgrading the valve for marine applications or adding a
redundant valve for safety reasons. For this case, we consider in the costs, changes in the fuel
tank design to make it more able to withstand 1 psi of pressure. We estimate that if
manufacturers were to make changes to the geometry of the fuel tank to help withstand 1 psi of
pressure without significant deflection, it could increase the material needed by 10 to 30 percent.
We include a cost estimate of $2,500 for the development of each new mold and amortize it over
6-84
-------
Costs of Control
100 tanks per year for 5 years. If the pressure relief valve is placed in the fill-neck cap, no vent
hose would be needed, which would reduce the cost of the fuel system. For the long-term cost
estimate, we consider the cost savings of removing the vent line. For this analysis, based on
conversations with boat builders, we divide the aftermarket hose price72 by four to represent the
cost of the hose to the boat builder.
Diurnal emissions may also be controlled through the use of a carbon canister in the vent
line. The carbon would be purged by air drawn into the fuel tank as the fuel cools. This is
known as passive purge. With a canister system, no significant pressure would build up in the
fuel tank. The canister would be packaged in the existing vent line and a float valve or other
liquid/vapor separation device would be added to the fuel system to ensure that liquid fuel would
not enter the vent line during refueling. We include a cost of $2 for this valve and $0.40 for two
additional hose clamps. In our cost estimates, we consider a canister using marine grade carbon
which is harder and more moisture resistant than typical carbon used in automotive applications.
Data shows that about 2 liters of carbon would be necessary for a 50 gallon fuel tank.73 We
estimate the cost of a canister to vary based on size ranging from about $12 for a 20 gallon tank
to about $38 for a 100 gallon tank.
Pressure could be completely eliminated using a bladder fuel tank because there would
be no vapor space. Based on conversations with a manufacturer of bladder fuel tanks, the
incremental cost of adding a bladder to a fuel tank would increase the fuel tank cost by 30-100
percent, depending on the size and shape of the fuel tank. As with a control strategy using a
pressure relief valve in the fill neck, no vent hose would be needed with a bladder fuel tank.
Pressure in the fuel tank can be minimized by reducing the vapor space in the fuel tank.
A volume compensating air bag can be used to minimize pressure. This air bag would need to be
about 1/4 to 1/3 the volume of the fuel tank. For this analysis we use 1/3 the cost of the bladder
fuel tank to account for the smaller bag size. We also include the cost of a low pressure psi
valve which could be used in conjunction with this technology as a safety backup.
Table 6.6-4: Venting Control Cost Estimates for Typical Marine Vessels
pressure relief valve:
passive purge canister:
bladder fuel tank:
volume compensating
air bag:
short term
long term
short term
long term
short term
long term
short term
long term
Portable Fuel Tank
6 gallons
$1.29
$1.03
NA
NA
NA
NA
NA
NA
Installed Fuel Tank
57 gallons
$26
$21
$32
$25
$259
$207
$91
$73
6-85
-------
Final Regulatory Impact Analysis
6.6.4 Certification and Compliance
We anticipate that manufacturers will use design based certification to as an alternative to
emission testing to meet the diurnal emission requirements. However, they will still need to
integrate the emission-control technology into their designs and there will be some engineering
and clerical effort need to submit the required information for certification. We expect that in
the early years, plastic fuel tank manufacturers will perform durability and permeation testing on
their fuel tanks for certification. They will be able to carry over this data in future years and will
be able to carry across this data to other fuel tanks made of similar materials and using the same
permeation control strategy regardless of tank size or shape. Typical certification costs may be
spread between the tank manufacturer, hose manufacturer, and boat builder. For the sake of this
analysis we combine the tank, hose, and boat certification costs to calculate the total certification
of an average fuel system. We estimate that 80 percent of fuel tank sales are plastic and about 25
percent of fuel tanks sold are portable fuel tanks.
For the first year we estimate fuel tank durability and certification testing to cost about
$15,000 per tank manufacturer on the assumption that the manufacturer will use the same
materials and permeation control strategy for all of their fuel tanks to reduce costs. Low
permeation fuel lines are largely established technology. However, we include a cost of $1,000
to perform certification testing on marine hose. In addition, we estimate about $10,000 for
engineering and clerical work for the tank and hose manufacturers. Boat builder certification
should be a simple letter referencing the tank and hose certificates and design requirements. We
consider a cost of $500 for this effort.
For portable fuel tank manufacturers we spread these costs over sales of 25,000 tanks per
year. For PWC manufacturers, which are integrated manufacturers, we base the costs on average
annual PWC sales which we estimate to be about 15,000 units per year. For vessels with
installed fuel tanks, the same tank manufacturer will often sell to many boat builders. Therefore,
we base the cost on average sales per tank manufacturer which we estimate to be about 40,000
per year. Although there is currently a limited offering of marine fuel hose products today, we
conservatively use the same lower unit volumes as for fuel tanks when applying hose testing
costs. This represents the scenario where portable fuel tank manufacturers and PWC
manufacturers perform their own hose testing, while smaller boat builders rely on data from the
hose manufacturers. For non-integrated boat builders using installed fuel tanks, we estimate that
the average sales per year is approximately 250 vessels.
As with other fixed costs, we amortized the cost over 5 years of sales to calculate per unit
certification costs. Combining these costs, we get average fuel system integration and
compliance costs of about $0.22 for portable fuel tanks, $0.35 for PWC, and $0.53 for fuel
systems on other vessels.
6.6.5 Operating Cost Savings
Evaporative emissions are essentially fuel that is lost to the atmosphere. Over the
lifetime of a marine vessel, this can result in a significant loss of fuel. The reduction in
6-86
-------
Costs of Control
evaporative emissions would therefore result in meaningful fuel savings which can be directly
related to operating cost savings based on an average density of 6 Ibs/gallon for gasoline (based
on lighter hydrocarbons which evaporate first) and the price of gasoline described above.
Table 6.6-5 presents the estimated fuel savings for marine vessels associated with the
evaporative emission standards.
Table 6.6-5: Projected Evaporative Fuel Savings for Marine Vessels
Evaporative HC Reduced [Ibs/life]
Lifetime Gallons Saved
Lifetime Cost Savings
Average Equipment Life [years]
Discounted Cost Savings (7%)
Portable
80
13
$24
12.7
$17
PWC
53
9
$16
9.9
$12
Installed
228
38
$68
17
$42
6.6.6 Total Marine Vessel Costs
We expect that marine vessel manufactures will make use of a variety of technologies to
meet the fuel tank permeation and diurnal emission standards. As discussed above, many
options are available so the technologies chosen will depend on the baseline fuel tank
construction, the vessel type, and the manufacturer's particular preferences. The hose
permeation standards will likely be met through the use of barrier hose constructions.
In calculating the costs of this rule, we consider the marine vessel categories in the
NONROAD model. NONROAD divides marine vessels into outboard, personal watercraft, and
SD/I applications and further subdivides these applications into several engine power categories.
This analysis uses the unique hose and tank sizes for each subcategory in the NONROAD model
and described in Chapter 3. For this analysis, we treat all vessels with outboard engines up to 25
hp as having portable fuel tanks made of plastic. This analysis considers all PWC to have plastic
fuel tanks as well. Based on our understanding of the market share of plastic versus aluminum
tanks, we use a split of 30 percent metal and 70 percent plastic for installed fuel tanks.
We base our cost analysis on likely technologies that manufactures may use. For
portable and PWC fuel tanks and, we base our tank permeation control costs on multi-layer
coextrusion with an EVOH barrier. For larger installed fuel tanks, we split the costs 50/50
between dual-layer rotational-molded tanks with a nylon barrier and the use of a low-permeation
epoxy coating over the tanks in a post molding process. Diurnal control costs are based on
sealed systems for portable marine tanks, current technology for PWC, and passive canister
systems for vessels with installed fuel tanks. Fuel supply line costs are based on thermoelastic
barrier technology. No costs or benefits are claimed for vent hose or fill neck hose.
As discussed above, our cost estimates include both variable and fixed costs, and we
6-87
-------
Final Regulatory Impact Analysis
distinguish between near-term and long-term costs. Because our analysis amortizes fixed costs
over 5 years, the long-term costs are generally made up of variable costs only. The exception to
this is fuel tank permeation control strategies where more expensive molding equipment is used.
We assume an equipment life of 10 years, so in the long term, the amortized additional cost of
the molding equipment is half, on average, of the short-term amortized cost over 5 years (5 years
of amortized payments/10 years of equipment life = l/2). In addition, variable costs are lower in
the long term due to the learning effect discussed in Section 6.1. Table 6.6-6 presents these
average per-engine cost estimates.
Table 6.6-6: Per Vessel Evaporative Emission Cost Estimates (Without Fuel Savings)
Portable aggregate
tank permeation
hose permeation
diurnal venting
PWC aggregate
tank permeation
hose permeation
diurnal venting
Installed aggregate
tank permeation
hose permeation
diurnal venting
Short Term (years 1-5)
Fixed
$6.65
$6.64
$0.01
$0
$12.95
$12.93
$0.01
$0
$0.63
$0.23
$0.01
$0.40
Variable
$5.39
$1.00
$3.10
$1.29
$4.49
$2.64
$1.84
$0
$73.55
$35.31
$6.54
$31.69
Total
$12.04
$7.65
$3.10
$1.29
$17.43
$15.58
$1.86
$0
$74.18
$35.54
$6.54
$32.09
Long Term (years 6-10)
Fixed
$3.21
$3.21
$0
$0
$6.30
$6.30
$0
$0
$0
$0
$0
$0
Variable
$5.13
$1.00
$3.10
$1.03
$4.49
$2.64
$1.84
$0
$61.53
$29.63
$6.54
$25.35
Total
$8.34
$4.22
$3.10
$1.03
$10.79
$8.94
$1.84
$0
$61.53
$29.63
$6.54
$25.35
6.6.7 Marine Vessel Aggregate Costs
Aggregate costs are calculated by multiplying the per-vessel variable cost estimates
described above by projected vessel sales and adding in fixed costs as incurred. Vessel sales are
based on estimates from the National Marine Manufacturers Association (www.nmma.org) and
projections for future years are based on the growth rates in the NONROAD model. A
description of the sales and population data and our analysis of the data are available in the
docket.74 Fuel savings are calculated directly from the projected HC reductions due to the
evaporative emission standards. Table 6.6-7 presents the projected costs of the rule over a 30-
year time period with and without the fuel savings associated with reducing evaporative
emissions. For the purposes of combining these costs with the exhaust emission costs described
above, we also present the projected costs by engine type in Table 6.6-8.
The population and sales data reported by NMMA, suggest that the NONROAD model
may somewhat underestimate the useful life of outboard and personal watercraft marine vessels.
If useful life were back-calculated—dividing NMMA population by sales and adjusted for
growth—we would get a longer average life estimate. As a result, the per-vessel fuel savings
6-8
-------
Costs of Control
described above may be understated. Because the current approach gives us a conservative
benefits estimate, and because we do not have new data on average lives for marine vessels to
update the estimates in the NONROAD model, we are not updating the model at this time. For
this reason, the 30-year stream may give a better view of the impact of the fuel savings than the
per-vessel analysis.
At a 7 percent discount rate, over 30 years, the estimated annualized cost to
manufacturers for marine evaporative emission control is $21 million. The estimated
corresponding annualized fuel savings due to control of evaporative emissions from boats is $22
million. At a 3 percent discount rate, the estimated annualized cost to manufacturers for marine
evaporative emission control is $23 million. The estimated corresponding annualized fuel
savings due to control of evaporative emissions from boats is $27 million.
Table 6.6-7: Projected 30-Year Aggregate Cost Stream for Marine Vessels
Year
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
2037
Without Fuel Savings
Portable
$36,474
$803,917
$857,095
$594,826
$599,164
$603,502
$607,839
$593,848
$598,004
$602,159
$606,314
$610,470
$614,625
$618,781
$622,936
$627,091
$631,247
$635,402
$639,584
$643,765
$647,947
$652,129
$656,310
$660,492
$664,674
$668,855
$673,037
$677,219
$681,400
$685,582
PWC
$67,297
$1,461,840
$1,395,094
$855,964
$862,207
$868,449
$874,691
$880,933
$887,097
$893,262
$899,426
$905,590
$911,754
$917,918
$924,083
$930,247
$936,411
$942,575
$948,778
$954,982
$961,185
$967,388
$973,591
$979,794
$985,998
$992,201
$998,404
$1,004,607
$1,010,810
$1,017,014
Installed
$425,096
$2,437,783
$2,118,774
$13,113,533
$25,655,601
$25,841,342
$26,027,084
$26,212,826
$24,120,397
$22,380,137
$22,534,578
$22,689,018
$22,843,458
$22,997,898
$23,152,338
$23,306,778
$23,461,218
$23,615,659
$23,771,076
$23,926,494
$24,081,911
$24,237,329
$24,392,747
$24,548,164
$24,703,582
$24,859,000
$25,014,417
$25,169,835
$25,325,252
$25,480,670
With Fuel Savings
Portable
$36,474
$550,181
$299,968
$(352,563)
$(787,374)
$(1,194,787)
$(1,593,995)
$(2,011,013)
$(2,403,445)
$(2,791,322)
$(3,170,164)
$(3,541,264)
$(3,893,751)
$(4,217,582)
$(4,502,321)
$(4,715,380)
$(4,892,031)
$(5,047,877)
$(5,189,699)
$(5,318,214)
$(5,440,156)
$(5,551,536)
$(5,655,549)
$(5,745,992)
$(5,828,655)
$(5,903,314)
$(5,963,979)
$(6,018,898)
$(6,068,556)
$(6,116,187)
PWC
$67,297
$1,401,963
$1,271,709
$516,639
$308,552
$103,461
$(97,802)
$(295,433)
$(487,095)
$(673,068)
$(850,586)
$(1,000,808)
$(1,134,586)
$(1,212,981)
$(1,270,967)
$(1,318,185)
$(1,357,627)
$(1,391,089)
$(1,418,920)
$(1,441,800)
$(1,460,469)
$(1,475,218)
$(1,486,959)
$(1,496,432)
$(1,505,907)
$(1,515,381)
$(1,524,855)
$(1,534,329)
$(1,543,803)
$(1,553,278)
Installed
$425,096
$1,913,103
$1,028,968
$10,845,446
$21,179,415
$19,143,903
$17,118,227
$15,098,006
$10,806,108
$6,882,907
$4,872,062
$2,878,316
$901,070
$(1,069,567)
$(3,020,477)
$(4,922,205)
$(6,725,979)
$(8,386,163)
$(9,827,107)
$(11,086,083)
$(12,248,543)
$(13,295,437)
$(14,225,862)
$(15,039,376)
$(15,694,795)
$(16,270,514)
$(16,778,147)
$(17,215,402)
$(17,601,830)
$(17,945,566)
6-89
-------
Final Regulatory Impact Analysis
Table 6.6-8: Projected 30-Year Aggregate Cost Stream
for Marine Vessels by Engine Type
Year
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
2037
Without Fuel Savings
OB
$319,871
$2,656,653
$2,502,577
$8,678,125
$15,269,926
$15,380,478
$15,491,029
$15,583,252
$14,360,302
$13,168,410
$13,259,282
$13,350,154
$13,441,026
$13,531,898
$13,622,771
$13,713,643
$13,804,515
$13,895,387
$13,986,834
$14,078,282
$14,169,729
$14,261,176
$14,352,623
$14,444,071
$14,535,518
$14,626,965
$14,718,413
$14,809,860
$14,901,307
$14,992,754
PWC
$67,297
$1,461,840
$1,395,094
$855,964
$862,207
$868,449
$874,691
$880,933
$887,097
$893,262
$899,426
$905,590
$911,754
$917,918
$924,083
$930,247
$936,411
$942,575
$948,778
$954,982
$961,185
$967,388
$973,591
$979,794
$985,998
$992,201
$998,404
$1,004,607
$1,010,810
$1,017,014
SD/I
$141,699
$585,047
$473,292
$5,030,234
$10,984,838
$11,064,366
$11,143,894
$11,223,423
$10,358,098
$9,813,887
$9,881,610
$9,949,333
$10,017,057
$10,084,780
$10,152,503
$10,220,227
$10,287,950
$10,355,674
$10,423,826
$10,491,978
$10,560,130
$10,628,281
$10,696,433
$10,764,585
$10,832,737
$10,900,889
$10,969,041
$11,037,193
$11,105,345
$11,173,497
With Fuel Savings
OB
$319,871
$1,906,143
$912,594
$5,726,980
$10,170,561
$8,145,273
$6,134,930
$4,108,869
$776,555
$(2,506,681)
$(4,483,567)
$(6,438,915)
$(8,363,328)
$(10,247,532)
$(12,075,309)
$(13,785,575)
$(15,367,391)
$(16,791,879)
$(17,990,346)
$(19,002,879)
$(19,924,799)
$(20,741,603)
$(21,455,586)
$(22,105,569)
$(22,684,596)
$(23,205,543)
$(23,661,696)
$(24,055,883)
$(24,404,310)
$(24,716,953)
PWC
$67,297
$1,401,963
$1,271,709
$516,639
$308,552
$103,461
$(97,802)
$(295,433)
$(487,095)
$(673,068)
$(850,586)
$(1,000,808)
$(1,134,586)
$(1,212,981)
$(1,270,967)
$(1,318,185)
$(1,357,627)
$(1,391,089)
$(1,418,920)
$(1,441,800)
$(1,460,469)
$(1,475,218)
$(1,486,959)
$(1,496,432)
$(1,505,907)
$(1,515,381)
$(1,524,855)
$(1,534,329)
$(1,543,803)
$(1,553,278)
SD/I
$141,699
$557,141
$416,341
$4,765,903
$10,221,480
$9,803,842
$9,389,302
$8,978,125
$7,626,108
$6,598,266
$6,185,465
$5,775,967
$5,370,647
$4,960,383
$4,552,511
$4,147,990
$3,749,381
$3,357,839
$2,973,540
$2,598,583
$2,236,099
$1,894,631
$1,574,175
$1,320,201
$1,161,147
$1,031,715
$919,570
$821,583
$733,924
$655,200
6.7 Cost Sensitivity Analysis
In developing the cost estimates described above, EPA used data from a wide variety of
sources. These sources included conversations with manufacturers and vendors, published
material costs, government cost tracking, and sales literature. In addition, we discussed many of
our cost estimates with industry experts. Through this process we have received information
suggesting that there is the potential for variability in some of the cost estimates used as inputs to
this analysis. For instance, fuel prices have been rising over the past few years which affects the
dollar value of our fuel savings estimates.
In this section, we perform an analysis of the sensitivity of our cost estimates to the
6-90
-------
Costs of Control
observed variation in costs for several input components of the cost analysis. The input
components that we are focusing on for the sensitivity analysis are those that would be expected
to have a significant effect on the final cost results. These are components that we either
observed high variability when collecting the data, or industry has raised issues about the
uncertainty of the technology which may lead to cost uncertainty.
We are focusing on five elements of the cost analysis for this sensitivity analysis. These
five elements are:
1. gasoline prices
2. precious metal costs
3. fraction of Small SI equipment manufacturers that design their own mufflers
4. electronic fuel injection on all Class II engines with multiple cylinders
5. costs of rotational-molded tank technologies
6.7.1 Gasoline Price Sensitivity
To estimate fuel savings in the above analysis, we used fuel price information obtained
from the U.S. Department of Energy, Energy Information Administration (EIA) which posts
gasoline price samples throughout the year on-line.75 For 2004 and 2005, national fuel prices are
based on an analysis of fuel prices by PADD as reported by EIA in 2006.76 For years later than
2005, we use the estimates reported in the 2008 Annual Energy Outlook report also developed
by EIA.77 Based on this information, the national average fuel price, with taxes, grew from
$2.20 in 2005 to $2.99 in 2008. This price estimate includes both a $0.184/gallon federal excise
tax and approximately a $0.2 I/gallon average state excise tax.78 Subtracting these taxes, we get
a fuel cost of $2.60/gallon for 2008.
To investigate the sensitivity of the cost analysis in this chapter to gasoline fuel price, we
looked at the U.S. average fuel prices for 2004 and 2007. These price estimates were calculated
in the same manner as the 2005 estimate. Table 6.7-1 presents these estimates. Fuel savings are
directly related to the gasoline price used in the cost analysis. Therefore, if the 2004 average
gasoline price were used in the cost analysis, the estimated fuel savings would have been about
22 percent lower. If the 2008 projected price were used, the estimated fuel savings would have
been about 33 percent higher. Because of the recent trend of increasing gasoline prices, we may
be understating the fuel savings in our cost analysis. However, using the 2005 fuel price is
consistent with our use of 2005 dollars for the costs in this chapter.
Table 6.7-1 U.S. Average Gasoline Prices [S/Gallon]
Year
2004
2005
2006
2007
2008 (projected)
with taxes
$1.80
$2.20
$2.63
$2.77
$2.99
without taxes
$1.41
$1.81
$2.44
$2.37
$2.60
6-91
-------
Final Regulatory Impact Analysis
As discussed above, our analysis of fuel savings uses a constant fuel price for all future
years. To the extent that fuel prices were to fluctuate in future years, this could have an impact
on realized fuel savings. To investigate the sensitivity of our analysis to future fuel prices, we
considered fuel price projections from the EIA's 2008 Annual Energy Outlook.79 EIA
projections include primary estimates of fuel prices, known as the "reference case," as well as
"high price" estimates. These projections, which include taxes on motor gasoline, are shown in
Figure 6.7-1. EIA projections from AEO 2007 and AEO 2006 are also presented for
comparison. Note that the EIA reference cases show relatively flat fuel price projections beyond
2010, when the fuel savings associated with this rule would be realized.
Figure 6.7-1: EIA Motor Gasoline Projections [Include Taxes]
$4.00
$3.50
$3.00
! $2.50
S
.c
+j
!i $2.00
2
|) $1.50
$1.00
$0.50
$-
EIA Motor Gasoline Projections
(Sales weighted-average price for all grades. Includes Federal, State, and local taxes)
FRM
AEO 2006
AEO 2007
AEO 2008
- - 'AEO 2008 HIGH PRICE
2008
2013
2018
2023
2028
Table 6.7-2 presents our fuel savings estimates for this rule and presents a comparison
with modified estimates using the EIA fuel price projections. Consistent with the above
discussion, we adjusted the fuel price estimates to remove taxes on motor gasoline. Compared to
the primary EPA estimate, using the reference case projections would result in about a 9 percent
increase in estimated fuel savings. Using the high price estimate would result in about a 60
percent increase in estimated fuel savings.
6-92
-------
Costs of Control
Table 6.7-2: Sensitivity of Fuel Savings Estimates to Gasoline Price Projections
Year
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
2037
Price without Taxes Per Gallon
Primary
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
$1.81
AEO 2008
Reference
$2.60
$2.36
$2.16
$2.09
$2.02
$1.95
$1.94
$1.86
$1.80
$1.82
$1.84
$1.91
$1.97
$1.93
$1.95
$1.96
$1.95
$1.97
$1.98
$1.99
$2.02
$2.04
$2.06
$2.06a
$2.06a
$2.06a
$2.06a
$2.06a
$2.06a
$2.06a
Annualized Savings
[million dollars]
AEO 2008
High Price
$2.60
$2.39
$2.45
$2.48
$2.51
$2.55
$2.61
$2.62
$2.67
$2.73
$2.77
$2.82
$2.89
$2.97
$3.01
$3.01
$3.03
$3.00
$3.01
$3.04
$3.06
$3.09
$3.13
$3.13a
$3.13a
$3.13a
$3.13a
$3.13a
$o i o a
3.13
$3.13a
3%
7%
Fuel Savings [million dollars]
Primary
$3
$8
$20
$45
$72
$97
$118
$137
$154
$168
$181
$194
$204
$214
$224
$233
$241
$248
$255
$261
$267
$272
$277
$282
$286
$290
$294
$297
$301
$304
$180
$156
AEO 2008
Reference
$5
$11
$24
$52
$80
$105
$126
$140
$153
$168
$184
$204
$221
$228
$241
$251
$259
$269
$278
$287
$297
$306
$314
$319
$324
$329
$333
$337
$341
$345
$197
$169
AEO 2008 High
Price
$5
$11
$27
$61
$100
$137
$169
$198
$226
$253
$277
$301
$325
$351
$371
$386
$402
$411
$423
$437
$450
$464
$479
$487
$494
$501
$507
$514
$520
$525
$293
$249
a Based on estimate for 2030. AEO 2008 does not project fuel prices beyond 2030.
6.7.2 Variation in Precious Metal Prices
Precious metal prices for Platinum and Rhodium have increased over the past 5 years.80
Prices for palladium are currently at their 1998 levels. However, a large spike in palladium
prices was seen in 2000 and 2001. Due to the high variability of this market, we get higher
precious metal cost estimates if we based the price estimates on a recent single month average
(September 2006). If we look at an average over a longer time period (10 years) we calculate
lower platinum costs, but higher rhodium and palladium costs. These precious metal price
estimates are presented in Table 6.7-3.
6-93
-------
Final Regulatory Impact Analysis
Table 6.7-3: Precious Metal Prices [per troy oz]
Rhodium
Palladium
Platinum
ICF 3 year Average
$1,121
$210
$811
September 2006
$4,835
$316
$1,134
10 Year Average
$1,356
$341
$623
6.7.2.1 Sensitivity of Small SI Catalyst Costs to Precious Metal Costs
To look at the sensitivity of our cost analysis for Small SI exhaust emission control, we
considered the precious metal cost variability described above. Based on the amount of each of
these precious metals in our projected catalyst designs, Table 6.7-4 presents the impact on per-
engine costs of using the spot price and 10 year average price in our analysis. These costs,
which are broken down by class and useful life, are presented for the near term without fuel
savings.
Table 6.7-4: Sensitivity of Small SI Total Per Engine Cost Estimates
to Precious Metal Costs
CLASS
UL
TECH
RULE Cost/Equip
(3 yr avg precious
metal price)
I
125
OHV/SV
14.12
I
250
OHV
19.82
I
500
OHV
26.07
II
250
OHV
46.21
II
500
OHV
50.83
II
1000
OHV
92.17
SEPTEMBER 2006 PRICE
Cost/Equip
Increase
% Increase
$15.69
$1.57
10%
$22.60
$2.78
12%
$30.25
$4.18
14%
$47.48
$1.27
3%
$52.67
$1.84
4%
$96.11
$3.94
4%
10 YEAR AVERAGE
Cost/Equip
Increase
%Increase
$13.91
-$0.21
-1.5%
$19.45
-$0.37
-1.9%
$25.51
-$0.56
-2.2%
45.84
$-0.37
-1%
$51.39
$0.56
1%
$93.80
$1.63
2%
6.7.2.1 Sensitivity of SD/I Catalyst Costs to Precious Metal Costs
To look at the sensitivity of our cost analysis for SD/I exhaust emission control, we
considered the precious metal cost variability described above. Based on the amount of each of
6-94
-------
Costs of Control
these precious metals in our projected catalyst designs, Table 6.7-5 presents the impact on per-
engine costs of using the spot price and 10 year average price in our analysis. These costs,
which are presented for each of the engine sizes used above for the primary cost analysis, are
near term costs without fuel savings.
6-95
-------
Final Regulatory Impact Analysis
Table 6.7-5: Sensitivity of SD/I Cost Estimates to Precious Metal Costs
Primary Analysis
3.0L 14
$483
4.3L V6
$396
5.0L V8
$317
5.7L V8
$300
8.1L V8
$377
September 2006 Precious Metal Prices
Cost
Increase
% Increase
$511
$28
5%
$417
$21
5%
$342
$24
7%
$328
$28
8%
$416
$39
9%
10 Year Average Precious Metal Prices
Cost
Increase
% Increase
$479
-$4
-1%
$393
-$3
-1%
$314
-$4
-1%
$296
-$4
-1%
$371
-$6
-2%
Catalyst manufacturers usually buy precious metals on contract, not at the market spot
price. Our primary analysis values appear reasonable.
6.7.3 Portion of Equipment Manufacturers Designing Own Muffler System and
Recertifying the Engine
This analysis considers that equipment manufacturers will purchase the muffler design
provided by the engine manufacturer in the engine's certified engine configuration. However,
due to the fact that engine manufacturers will likely not be able to provide catalysts in all of the
muffler designs used by equipment manufacturers, the smaller volume equipment manufacturer
will need to pick their muffler from the limited offerings of the engine manufacturer.
The muffler designs may or may not fit into the equipment produced by the equipment
manufacturer. If it does not, then the equipment manufacturer may choose to utilize the catalyst
brick from their engine manufacturer and work with a muffler manufacturer to redesign their
existing muffler. If they choose this option, then they must undergo expenses to redesign the
muffler and heat shield to apply the catalyst safely. The equipment manufacturer must also pay
for emission test of the new engine/muffler configuration as well as pay the certification fee to
EPA for engine certification.
Applications which may find issues using a predetermined muffler design include those
that have close coupled equipment shrouding or a closed equipment structure. EPA estimates
that 10 percent of equipment companies will find themselves in this situation with at least one
piece of equipment in their product line. Given there are an estimated 413 companies, 41
companies with three differently designed models each yields 123 models. Given that there are
at times more than one engine used in an equipment design, we can assume two engine types per
model - this yields a total of 246 redesigns and certifications. The fixed costs for this work are
listed in Table 6.7-6.
6-96
-------
Costs of Control
Table 6.7-6: Costs for Equipment Manufacturers
6-97
-------
Final Regulatory Impact Analysis
to Perform Engine Certification, Class II OHV
Muffler/Heat Shield Design
Emission Test per Certified Engine Configuration
Estimated EPA Certification Fee
TOTAL Per Equipment Model Per Engine Type
10% of Equipment Manufacturers = 41 (x41)
Three equipment models per equipment mfr.
Two engine types per Equipment Model (x2)
TOTAL ESTIMATED COST
Fixed Costs
$75,000
$2012
$800
$77,812
41
123
246
$19,141,
752
If this occurred it would add about $19 million dollars to the total compliance cost or
about 0.86 percent of the total 30 year cost net present value.
6.7.4 Electronic Fuel Injection on Class II Engines with Multiple Cylinders
The current analysis states that only a portion of an engine manufacturers Class II engine
families of two or more cylinders per engine will incorporate electronic fuel injection. In the
event that success with the technology results in all Class II engines of two or more cylinders
using the technology, then the cost stream of this rulemaking will change. Table 6.7-7 compares
the estimated costs of catalysts and fuel injection.
Table 6.7-7: Cost Comparison Between Catalyst and EFI
Technology
Class II V-twin
250
500
1000
Variable Costs
V-Twin Catalyst
Electronic Fuel
Injection
Difference
$49.59
$78.99
$28.40
$53.47
$78.99
$25.52
$62.32
$78.99
$16.67
Fixed Costs
V-Twin Catalyst
Electronic Fuel
$364,133
$103,020
$364,133
$103,020
$364,133
$103,020
6-98
-------
Costs of Control
Difference
-$261,113
-$261,113
-$261,113
The resultant change in cost/equipment for this is shown in Table 6.7-8. The costs
presented here are for the near term and long term without fuel savings. The reason that costs do
not change very much overall is due to the fact that there is still a significant portion of Class II
engines that are single cylinder whose costs estimates are not changing.
Table 6.7-8
Sales Weighted Average Cost Per Class II Equipment
250
500
1000
Short Term (first year - includes fixed costs)
Primary analysis
All Class II V-Twin to EFI
Difference
$46.21
$46.80
$0.59
1.3%
$50.83
$49.71
-$1.12
2.2%
$92.17
$91.55
-$0.62
0.67%
Long Term (6th year and beyond)
Primary analysis
All Class II V-Twin to EFI
Difference
$32.56
$33.16
$0.60
1.8%
$27.13
$27.15
$0.02
0.07%
$49.80
$50.62
$0.82
1.6%
The estimated fuel savings for a residential riding mower is $39.00 net present value over
its lifetime. EFI is estimated to cost $79.00 after consideration of the savings from removal of
the existing carburetor. Therefore, the increase in the overall hardware cost with fuel savings is
$40.00.
6.7.5 Costs of Rotational-Molded Tank Technologies
Many of the fuel tank permeation control technologies discussed in Chapter 5 are used
widely today. One exception is multi-layer rotationally-molded fuel tanks. One tank
manufacturer is currently producing fuel tanks for Small SI equipment with a nylon inner layer.
This manufacturer has stated that they are able to produce these fuel tanks using the normal
molding process without additional equipment. However, other manufacturers who sell tanks
into Small SI and marine applications have expressed concern that they do not know how to
mold tanks with nylon inner liners without the use of a drop box. As described above, a drop
box is an added component on a mold that opens during the molding process to add a second
layer of material into the mold. These manufacturers have indicated that they are working with
another material, CBT (discussed above and in Chapter 5), that would not require a drop box.
6-99
-------
Final Regulatory Impact Analysis
However, they have not finished their evaluation of this technology. Marine fuel tank
manufacturers have expressed the concern that if the cost of plastic fuel tanks were too high, that
more boat builders may begin using aluminum fuel tanks.
To examine the uncertainty in what technologies will be used to reduce permeation from
rotationally molded fuel tanks, we considered three factors listed below. As with the analysis
above, we present costs for typical fuel tank sizes rather than trying to present every fuel tank
size considered in the cost model. The two fuel tank sizes used here are a 5 gallon tank for
Small SI equipment and a 57 gallon fuel tank for boats.
1. Cost of using a drop box in the rotational-molding process
2. Sensitivity to variations in material costs
3. Consideration of replacing plastic with metal fuel tanks in marine industry
In the analysis described above, we include a $5,000 cost per mold in the near term to
account for the cost using drop boxes. This cost was based on a range of cost estimates supplied
by tank manufacturers ranging from $1,000 to nearly $9,000 per mold for adding drop boxes. In
the long term we projected that tank manufacturers would all be able to mold fuel tanks without
the use of a drop box. This projection was based on the current practices of one manufacturer
and on alternative processes that other manufacturers are investigating today. To look at the
sensitivity of tank permeation control costs for rotationally-molded fuel tanks, we consider costs
without drop boxes and with $9,000 drop boxes.
Table 6.7-9: Sensitivity of Rotomolded Tank Cost Estimates to Drop Box Cost
Primary Analysis ($5,000 drop box)
5 Gallon Small SI Tank
$5.54
57 Gallon Boat Tank
$81
Without Drop Box
Cost
Increase
% Increase
With $9,000 Drop Box
Cost
Increase
% Increase
$4.25
($1.29)
-23%
$6.58
1.04
19%
$68
($13)
-16%
$92
$10
13%
The analysis above considers three multi-layer approaches to rotationally-molded fuel
tanks. These approaches are molding with a nylon inner layer using a drop box, molding with a
slightly more expensive CBT layer without a drop box, and a post processing epoxy coating. All
three of these approaches would be sensitive to changes in barrier material prices. Because these
are new materials for fuel tank applications, it would be possible that material costs would
decrease over time with increased production volumes. At the same time, increases in material
6-100
-------
Costs of Control
costs could occur, especially for materials with prices tied closely to petroleum prices (such as
polyethylene). To consider the sensitivity of fuel tank cost to material costs, we consider the fuel
tank construction with a nylon barrier. Here we consider both a 20 percent decrease and a 20
percent increase in material costs, both for the nylon and the cross-link polyethylene. This
translates a cross-link polyethylene cost ranging from $0.96 to $1.44/lb. and nylon costs ranging
from to a nylon cost ranging from $3.20 to $4.80/lb. for Small SI and $4.8 to $7.2/lb. for marine
fuel tanks.
Table 6.7-10: Sensitivity of Rotomolded Tank Cost Estimates to Material Cost
Primary Analysis
5 Gallon Small SI Tank
$5.54
57 Gallon Boat Tank
$81
20% Decrease in Material Costs
Cost
Increase
% Increase
$5.18
($0.85)
-15%
$68
($14)
-17%
20% Increase in Material Costs
Cost
Increase
% Increase
$6.40
$0.86
15%
$95
($14)
17%
Marine fuel tanks that are installed in marine vessels are primarily rotationally-molded
out of cross-link polyethylene. However, many fuel tank are also made of aluminum. Very
large fuel tanks (typically greater in size than rotationally-molded fuel tanks) are often made out
of fiberglass. Marine fuel tank manufacturers making rotationally-molded fuel tanks have
expressed the concern that if the costs were to increase too high, that many boat builders would
switch to using aluminum fuel tanks. Based on conversations with industry, plastic fuel tanks
sell for about 2/3 to 3/4 the price of aluminum fuel tanks.
One manufacturer of multi-layer rotationally-molded fuel tanks with a nylon inner layer
has stated that they sell these fuel tanks at a price about 50 percent higher than traditional mono-
layer fuel tanks. Although this puts the plastic tanks into the price range of metal fuel tanks,
there are other downstream costs that would also need to be considered. Boat builders have
indicated that it is common for aluminum fuel tanks to corrode when exposed to water. For this
reason, they typically include a large access panel to the fuel tank when metal fuel tanks are
used. The use of an access panel greatly reduces the cost of replacing a fuel tank if necessary.
This access panel adds cost and complexity to the boat and may affect where the fuel tank can be
positioned in the boat. Boat manufacturers have indicated that, when plastic fuel tanks are used,
the only access required is to the hose connections on one end of the fuel tank.
In addition to the cost of an access panel for removing corroded tanks, the cost of
replacing the fuel tank must be considered. This would essentially double the price of the metal
tank, even without considering labor costs. In addition, fuel spills could create other damage in
6-101
-------
Final Regulatory Impact Analysis
the boat or even a safety hazard. Repeated problems with fuel tank corrosion could hurt the
reputation of the boat builder and leave them open to litigation. For these reasons, many boat
builders that have already chosen to use plastic fuel tanks would be expected to continue to use
these fuel tanks, even if they were roughly the same cost as metal fuel tanks.
We analyzed at two effects that could have an impact on our estimate of the price of low
permeation plastic fuel tanks. It seems unlikely that a high cost drop box would be necessary
given that one manufacturer is already producing multi-layer tanks without using a drop box. In
addition, the CBT technology is designed to not require the use of a drop box. While material
costs may fluctuate, it is not likely that a 20 percent increase in nylon would be observed. The
volume of this material sold is large and this rule would not be expected to limit availability of
the material. In addition, manufacturers have indicated that nylon prices have not risen greatly
with increased petroleum costs. Even with a 20 percent material price increase it seems unlikely
that boat builders would switch to using metal tanks. Manufacturers using plastic tanks have
indicated that they do so more for durability advantages with respect to corrosion than for a price
savings. In addition, the life time cost savings of plastic fuel tanks would outweigh the material
price increase. These lifetime cost savings include the installation of access ports to allow
replacement of the tanks, actual replacement of corroded tanks, and customer perception of poor
quality if tanks were to corrode.
6-102
-------
Costs of Control
Chapter 6 References
1. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
2. "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, Docket Identification EPA-
HQ-OAR-2004-0008-0204.
3. For further information on learning curves, see Chapter 5 of the Economic Impact, from "Regulatory Impact
Analysis—Control if Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and
Gasoline Sulfur Control Requirements," U.S. EPA, December 1999, EPA420-R-99-023. A copy of this document is
included in Air Docket A-2000-01, at Document No. II-A-83. The interested reader should also refer to previous
final rules for Tier 2 highway vehicles (65 FR 6698, February 10, 2000), marine diesel engines (64 FR 73300,
December 29, 1999), nonroad diesel engines (63 FR 56968, October 23, 1998), and highway diesel engines (62 FR
54694, October 21, 1997).
4. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
5."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
6. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
7. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
8. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
9."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
10."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
11. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
12. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
13. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
6-103
-------
Final Regulatory Impact Analysis
14. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
15. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
16. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
17."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
18."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
19."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
20."Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
21. "Small SI Engine Technologies and Costs, Draft Report" ICF International, prepared for U.S. Environmental
Protection Agency, August 2006, Docket Identification EPA-HQ-OAR-2004-008-0253.
22. "EPA Technical Study on the Safety of Emission Controls for Nonroad Spark-Ignition Engines < 50
Horsepower," U.S. Environmental Protection Agency, EPA420-R-06-006, March 2006, Docket Identification EPA-
HQ-OAR-2004-0008-0329.
23. "Small SI Engine Technologies and Costs, Draft Report," ICF International, August 2006, Docket Identification
EPA-HQ-OAR-2004-008-0253.
24. "Small SI Engine Technologies and Costs, Draft Report," ICF International, August 2006, Docket Identification
EPA-HQ-OAR-2004-008-0253.
25. "Summary of 2002 Nonhandheld Equipment Market by Manufacturer Size", EPA Memo by Phil Carlson to
Docket EPA_HQ_OAR_2004_0008_0542.
26. "Build it, and they will come," OEM Off-Highway, page 74-78, November 1996.
27. "EFI Hits the Small Engine Market," Diesel Progress, pages 52-54, July 1998.
28. Memorandum from Alex Rogozhin, RTI, to Chi Li, U.S. EPA, "Calculation of Motor Gasoline Prices in Small SI
Rule EIA," November 10, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0449.
29. "Marine Outboard and Personal Watercraft SI Engine Technologies and Costs," ICF Consulting, prepared for
the U.S. Environmental Protection Agency, July 2006, Docket Identification EPA-HQ-OAR-2004-0008-0452.
30. Memorandum from Alex Rogozhin, RTI, to Chi Li, U.S. EPA, "Calculation of Motor Gasoline Prices in Small SI
Rule EIA," November 10, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0449.
6-104
-------
Costs of Control
31. "Updates to Technology Mix, Emissions Factors, Deterioration Rates, Power Distribution, and Fuel
Consumption Estimates for SI Marine Engines in the NONROAD Emissions Inventory Model," Memorandum from
Mike Samulski, U.S. EPA to Docket, November 30, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0361.
32. "Updates to Technology Mix, Emissions Factors, Deterioration Rates, Power Distribution, and Fuel
Consumption Estimates for SI Marine Engines in the NONROAD Emissions Inventory Model," Memorandum from
Mike Samulski, U.S. EPA to Docket, November 30, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0361.
33. Samulski, M., "Marine SI Sales and Price Estimates," U.S. EPA, August 30, 2004, Docket Identification EPA-
HQ-OAR-2004-0008-0448.
34. "Sterndrive and Inboard Marine SI Engine Technologies and Costs," ICF Consulting, prepared for the U.S.
Environmental Protection Agency, July 2006, Docket Identification EPA-HQ-OAR-2004-0008-0453.
35. "Updates to Technology Mix, Emissions Factors, Deterioration Rates, Power Distribution, and Fuel
Consumption Estimates for SI Marine Engines in the NONROAD Emissions Inventory Model," Memorandum from
Mike Samulski, U.S. EPA to Docket, November 30, 2005, Docket Identification EPA-HQ-OAR-2004-0008-0361.
36. Samulski, M., "Marine SI Sales and Price Estimates," U.S. EPA, August 30, 2004, Docket Identification EPA-
HQ-OAR-2004-0008-0448.
37. Trident Marine Hose, "Retail Price List 2001," Docket Identification EPA-HQ-OAR-2004-0008-0226.
38. Denbow, R., Browning, L., Coleman, D., "Report Submitted for WA 2-9, Evaluation of the Costs and
Capabilities of Vehicle Evaporative Emission Control Technologies," ICF, ARCADIS Geraghty & Miller, March
22, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0194.
39. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01,
August 6, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0207.
40. "Information on Costs and Effectiveness of Fluorination Received from Fluoroseal," Memorandum from Mike
Samulski to Docket A-2000-1, March 27, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0201.
41. "Visit to Sulfo Technologies LLC on April 18, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket
A-2000-01, April 22, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0232.
42. "Shipping Costs," Memorandum from Glenn Passavant, U.S. EPA to Docket A-2000-01, March 27, 2002,
Docket Identification EPA-HQ-OAR-2004-0008-0222.
43. "Visit to Sulfo Technologies LLC on April 18, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket
A-2000-01, April 22, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0232.
44. "Plastic News," Resin Pricing for November 8, 2004, www.plasticnews.com, Docket Identification EPA-HQ-
OAR-2004-0008-0442.
45. Ward, Geoff, "Potential Small Entity Representative (SER) Response to Initial Outreach Document: Exhaust
and Evaporative Emissions Control from Small SI Engines and Equipment and Marine SI Engines and Vessels,"
Agri-Industrial Plastics, July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0427.
46. "Plastic News," Resin Pricing for November 8, 2004, www.plasticnews.com, Docket Identification EPA-HQ-
OAR-2004-0008-0442.
47. "Visit to Kracor, Inc.," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-02, March 31, 2003,
Docket Identification EPA-HQ-OAR-2004-0008-0249.
6-105
-------
Final Regulatory Impact Analysis
48. Written comments from Robert Porter, Inca Molded Products to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0426.
49. Written comments from Jim Harden, Grady-White Boats to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0437.
50. "Petro-Seal for Ultra-Low Fuel Permeation," Ron Partridge, Arkema Inc., Presentation at the International Boat
Exposition, October 25, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0252.
51. "Plastic News," Resin Pricing for November 8, 2004, www.plasticnews.com, Docket Identification EPA-HQ-
OAR-2004-0008-0442.
52. "Information on Costs and Effectiveness of Fluorination Received from Fluoroseal," Memorandum from Mike
Samulski to Docket A-2000-1, March 27, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0201.
53. Written comments from Robert Porter, Inca Molded Products to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0426.
54. Written comments from Jim Harden, Grady-White Boats to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0437.
55. "Comments by Brunswick Corporation; Notice of Proposed Rulemaking; Part 1045 Control of Emissions from
Spark-Ignition Marine Vessels," Rolf Lichtner, Mercury Marine, January 7, 2003, Docket Identification EPA-HQ-
OAR-2004-0008-0192.
56. Trident Rubber Inc., "Trident Marine Hose, Retail Price List 2001," Docket Identification EPA-HQ-
OAR-2004-0008-0226.
57. Denbow, R., Browning, L., Coleman, D., "Report Submitted for WA 2-9, Evaluation of the Costs and
Capabilities of Vehicle Evaporative Emission Control Technologies," ICF, ARCADIS Geraghty & Miller, March
22, 1999, Docket Identification EPA-HQ-OAR-2004-0008-0194.
58. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket A-2000-01,
August 6, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0207.
59. Lawrence Marine; NovaFlex Group, "Industrial Price List: Marine Products," October 15, 2004, Docket
IdentificationEPA-HQ-OAR-2004-0008-0439.
60. Trident Rubber Inc., "Trident Marine Hose, Retail Price List, October 2005," Docket Identification EPA-HQ-
OAR-2004-0008-0444.
61. "Information on Costs and Effectiveness of Fluorination Received from Fluoroseal," Memorandum from Mike
Samulski to Docket A-2000-1, March 27, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0201.
62. "Visit to Sulfo Technologies LLC on April 18, 2002," Memorandum from Mike Samulski, U.S. EPA to Docket
A-2000-01, April 22, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0232.
63. "Shipping Costs," Memorandum from Glenn Passavant, U.S. EPA to Docket A-2000-01, March 27, 2002,
Docket Identification EPA-HQ-OAR-2004-0008-0222.
64. "Plastic News," Resin Pricing for November 8, 2004, www.plasticnews.com, Docket Identification EPA-HQ-
OAR-2004-0008-0442.
6-106
-------
Costs of Control
65. Ward, Geoff, "Potential Small Entity Representative (SER) Response to Initial Outreach Document: Exhaust
and Evaporative Emissions Control from Small SI Engines and Equipment and Marine SI Engines and Vessels,"
Agri-Industrial Plastics, July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0427.
66. "Petro-Seal for Ultra-Low Fuel Permeation," Ron Partridge, Arkema Inc., Presentation at the International Boat
Exposition, October 25, 2004, Docket Identification EPA-HQ-OAR-2004-0008-0252.
67. "Information on Costs and Effectiveness of Fluorination Received from Fluoroseal," Memorandum from Mike
Samulski to Docket A-2000-1, March 27, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0201.
68. Written comments from Robert Porter, Inca Molded Products to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 24, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0426.
69. Written comments from Jim Harden, Grady-White Boats to Phil Carlson, U.S.EPA regarding impacts of
regulation on small business, July 25, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0437.
70. "Marine Fuel Containment; A Permanent Solution," Jason Chambers, Engineered Composite Structures Inc.,
Presentation at the 2004 International Boat Exposition, October 25, 2004, Docket Identification EPA-HQ-
OAR-2004-0008-0247.
71. Meetings with NMMA on Marine Emission Standards," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-02, March 4, 2002, Docket Identification EPA-HQ-OAR-2004-0008-0250.
72. Trident Marine Hose, "Retail Price List 2001," Docket Identification EPA-HQ-OAR-2004-0008-0226.
73. "Activated Carbon for Use in Marine Evaporative Control Applications," MeadWestvaco Corporation,
Presentation at the International Boat Exposition, October 25, 2004, Docket Identification EPA-HQ-
OAR-2004-0008-0248.
74. Samulski, M., "Marine SI Sales and Price Estimates," U.S. EPA, August 30, 2004, Docket Identification EPA-
HQ-OAR-2004-0008-0448.
75. Energy Information Administration, "Retail Gasoline Historical Prices," downloaded from
http://www.eia.doe.gov/oil_gas/petroleum/data_publications/wrgp/mogas_history.html, on February 28, 2008.
76. Memorandum from Alex Rogozhin, RTI, to Chi Li, U.S. EPA, "Calculation of Motor Gasoline Prices in Small SI
Rule EIA," November 10, 2006, Docket Identification EPA-HQ-OAR-2004-0008-0449.
77. Energy Information Administration, "Annual Energy Outlook 2008; with Projections to 2030,"
DOE/EIA-0383(2008), June 2008.
78. Federal Highway Administration, "Tax Rates on Motor Fuel - 2003," downloaded from
http://www.eia.doe.gov/oil_gas/petroleum/data_publications/wrgp/gasoline_taxes.html on October 25, 2006, Docket
Identification EPA-HQ-OAR-2004-0008-0451.
79. Energy Information Administration, "Annual Energy Outlook 2008; with Projections to 2030,"
DOE/EIA-0383(2008), June 2008.
80. Johnson Matthey, "Platinum Today," www.platinum.matthey.com, downloaded October 25, 2006, Docket
IdentificationEPA-HQ-OAR-2004-0008-0422.
6-107
-------
Cost Per Ton
CHAPTER 7: Cost Per Ton
This Chapter will present the cost effectiveness analysis we completed for our proposed
small spark ignition engine (<19 kW) and recreational marine (personal water craft,
sterndrive/inboard and outboard) emission standards. Under Clean Air Act section 213, 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
setting process is not necessarily premised on setting the most cost effective standards, even
though this is a significant factor. Cost-effectiveness is a useful tool in evaluating the
appropriateness of our standards.
The cost-effectiveness analysis described in this chapter relies in part on cost information
from Chapter 6 and emissions information from Chapter 3 to estimate the dollars per ton of
emission reductions produced from our proposed standards. We have calculated the cost
effectiveness using a 30-year net present value approach that accounts for all costs and emission
reductions over a 30-year period. 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.
Section 7.1 describes the calculation behind the 30 year net present value cost
effectiveness and Section 7.2 lists the results of the calculations for our combined small spark
ignition standards (exhaust and evaporative) and marine engines (exhaust and evaporative).
Table 7.2-.5 lists the results for the 30-year net present value cost effectiveness analysis for
Small SI and Marine. The results of the cost-effectiveness of comparative programs are listed in
Table 7.2-6.
7.1 30-Year Net Present Value Cost Effectiveness (Cost per Ton)
We have 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 only the new
equipment/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 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 means by which to assess the cost effectiveness of
these programs.
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 Chapters 3 and 6,
respectively. The calculation of 30-year net present value cost-effectiveness follows the pattern
described above for the per-engine analysis:
7-1
-------
Final Regulatory Impact Analysis
DNAE =
Where:
DNAE = Reduction in nationwide 30-year net present value emissions in tons
= Reduction in nationwide emissions in tons for year i of the program
= Year of the program, counting from year 1 to year 30
and
i-2008
Where:
DNAC = Nationwide 30-year net present value costs in dollars
(NC); = Nationwide costs in dollars for year i of the program
i = Year of the program from year 1 to year 30
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 Chapter 3. The results are given in
Tables within the following section.
7.2 Results
We calculated the cost-effectiveness of our program on a 30-year net present value basis
separately for our proposed Small SI standards <19kW and recreational marine standards. To do
this, we summed net present value of total costs from Chapter 6, and divided by the sum of the
net present value of tons reduced from Chapter 3. These costs and emission reductions are
repeated in Appendices 7-A and 7-B. The results are given in Table 7.2-1 to 7.2-2 for Small SI
engines and equipment and 7.2-3 and 7.2-4 for recreational marine engines and vessels.
Table 7.2-1: 30-year Net Present Value Cost-effectiveness of the Standards
for Small SI Engines <19kW Without Fuel Savings (7 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$2,257
$809
$3067
NPVReduction
(tons)
1,785,000
1,098,000
2,883,000
Cost per Ton
$1,264
$736
$1,063
7-2
-------
Cost Per Ton
Table 7.2-2: 30-year Net Present Value Cost-effectiveness of the Standards
for Small SI Engines <19kW With Fuel Savings (7 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$1,959
$151
$2,110
NPVReduction
(tons)
1,785,000
1,098,000
2,883,000
Cost per Ton
$1,097
$137
$856
Table 7.2-3: 30-year Net Present Value Cost-effectiveness of the Standards
for Marine Engines Without Fuel Savings (7 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$1,521
$270
$1,790
NPVReduction
(tons)
1,826,000
461,000
2,287,000
Cost per Ton
$833
$585
$783
Table 7.2-4: 30-year Net Present Value Cost-effectiveness of the Standards
for Marine Engines With Fuel Savings (7 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$823
($6)
$817
NPVReduction
(tons)
1,826,000
461,000
2,287,000
Cost per Ton
$451
—
$357
Because 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, the cost and cost-effectiveness estimates for future proposed mobile source
programs could also reflect a 3 percent discount rate. The 3 percent rate is in the 2 to 3 percent
7-3
-------
Final Regulatory Impact Analysis
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). 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 7.2-5 through 7.2-8.
Table 7.2-5: 30-year Net Present Value Cost-effectiveness of the Standards
for Small SI Engines <19kW Without Fuel Savings (3 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$3,718
$1,327
$5,044
NPVReduction
(tons)
3,227,000
1,932,000
5,159,000
Cost per Ton
$1,152
$687
$978
Table 7.2-6: 30-year Net Present Value Cost-effectiveness of the Standards
for Small SI Engines <19kW With Fuel Savings (3 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$3,181
$170
$3,351
NPVReduction
(tons)
3,227,000
1,932,000
5,159,000
Cost per Ton
$986
$88
$650
Table 7.2-7: 30-year Net Present Value Cost-effectiveness of the Standards
for Marine Engines Without Fuel Savings (3 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$2,407
$444
$2,852
NPVReduction
(tons)
3,425,000
885,000
4,310,000
Cost per Ton
$703
$502
$662
7-4
-------
Cost Per Ton
Table 7.2-8: 30-year Net Present Value Cost-effectiveness of the Standards
for Marine Engines With Fuel Savings (3 percent discount rate)
Pollutants
HC+NOx
Exhaust
Evaporative
Exhaust + Evap
NPV Costs
(million $)
$1,100
($86)
$1,014
NPVReduction
(tons)
3,425,000
885,000
4,310,000
Cost per Ton
$321
—
$235
Because one primary purpose of cost-effectiveness is to compare our program to
alternative programs, we listed the cost effectiveness of several previous EPA actions for
controlled emissions from mobile sources for NOx and NMHC in Table 7.2-9. The programs
shown in these tables are those for which cost-effectiveness was calculated in a similar manner
allowing for a comparison. (Note: costs adjusted to 2005 dollars.)
Table 7.2-9: Cost-effectiveness of
Recent Mobile Source Exhaust Emission Programs for HC+NOx, 2005$
(7 percent discount with fuel savings)
Program
2002 HH engines Phase 2
2001 NHH Engines Phase 2
1998 Marine SI engines
2004 Comm Marine CI
2007 Large SI exhaust
2006 ATV exhaust
2006 off-highway motorcycle
2006 recreational marine CI
2010 snowmobile
2006 <50cc highway motorcycle
2010 Class 3 highway motorcycle
$/ton
840
neg*
1900
200
80
300
290
700
1430
1860
1650
* fuel savings outweigh engineering/hardware costs
Costs adjusted to 2005$ using http://wwwl.jsc.msa.gov/bu2/inflateGDP.html
Permeation and other evaporative emission control measures we have implemented for
highway and off-highway motorcycles, large SI engines, ATVs, and snowmobiles have all had
cost effectiveness values of less than $0/ton due to the fuel savings.
The analyses supporting the values in Table 7.2-6 were conducted over the past ten years
and thus not all were done on a purely identical basis in terms of their analytical approach (e.g.,
factors such as cost streams and cost recovery). By comparing values in Table 7.2-6 for
NOx+HC to those presented above we can see that the cost-effectiveness of our proposed Small
SI and recreational Marine SI standards fall within the range of these other programs. Some
7-5
-------
Final Regulatory Impact Analysis
previous programs have been more cost effective (lower $/ton) 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, because earlier reductions are usually easier and less expensive to
achieve and the least costly means for reducing emissions is generally pursued first.
This proposed rule also will bring environmental benefits related to reductions in carbon
monoxide (CO) emissions and emissions of direct particulate matter (PM). We have elected to
base our cost effectiveness analysis solely on HC+NOx for two reasons. First, with regard to
PM and CO, no new or additional technology beyond that needed to achieve the proposed
HC+NOx standards is expected to be required. These reductions will occur as part of the
technology and related efforts to meet the HC+NOx standards. Second, in the case of PM, we
are not setting standards but do expect reductions to occur as a result of engine changes and in
some cases the use of aftertreatment. In neither case is significant additional effort needed.
7-6
-------
Final Regulatory Impact Analysis
CHAPTERS: Cost-Benefit Analysis
8.1 Overview
This chapter presents our analysis of the health and environmental benefits that are
estimated to occur as a result of the final Small SI and Marine SI engine standards throughout
the period from initial implementation through 2030. Nationwide, the engines subject to the
final emission standards in this rule are a significant source of nonroad mobile source air
pollution. The final standards will reduce exposure to direct PM2.5, NOx, VOCs and air
toxics emissions and help avoid a range of adverse health effects associated with ambient
ozone and PM2.5 levels.
EPA is required by Executive Order (E.O.) 12866 to estimate the benefits and costs of
regulations with estimated annual impacts of over 100 million dollars. Such regulations tend
to include major new pollution control regulations. To estimate these benefits and costs, the
analysis presented here attempts to answer three questions: (1) what are the physical health
and welfare effects projected to result from particulate matter (PM) and ozone precursors
(direct PM, VOCs and NOx)? (2) what is the monetary value of the projected changes in
health and welfare attributable to the final rule? and (3) how do the projected monetized
benefits compare to the projected costs? This analysis constitutes one part of EPA's thorough
examination of the relative merits of this regulation.
The benefits analysis relies on three major components to answer these questions:
• Calculation of the projected impact of the final rule on the national nonroad emissions
inventory of precursors to ozone and PM2.5, specifically NOx, VOCs and direct PM, for
two future years (2020 and 2030).
• Air quality modeling for 2020 and 2030 to determine projected changes in ambient
concentrations of ozone and PM2.5, reflecting baseline and post-control emissions
inventories.
• A benefits analysis to determine the projected changes in human health and welfare, both
in terms of physical effects and monetary value, that result from the projected changes in
ambient concentrations of ozone and PM2.5 for the modeled standards.
A wide range of human health and welfare effects are linked with exposure to PM,
VOCs and NOx. Recent studies have linked short-term ozone exposures with premature
mortality. Exposure to ozone has also been linked to a variety of respiratory effects including
hospital admissions and illnesses resulting in school absences. Potential human health effects
associated with PM2 5 range from premature mortality to morbidity effects linked to long-term
(chronic) and shorter-term (acute) exposures (e.g., respiratory and cardiovascular symptoms
resulting in hospital admissions, asthma exacerbations, and acute and chronic bronchitis).
Welfare effects potentially linked to PM include materials damage and visibility impacts,
8-2
-------
Cost-Benefit Analysis
while ozone can adversely affect the agricultural and forestry sectors by decreasing yields of
crops and forests.
The benefits modeling is based on peer-reviewed studies of air quality and health and
welfare effects associated with improvements in air quality and peer-reviewed studies of the
dollar values of those public health and welfare effects. All of the benefit estimates for the
control options in this analysis are based on an analytical structure and sequence consistent
with benefits analyses performed for the recent analysis of the final Ozone NAAQS and the
final PM NAAQS analysis.1'2 For a more detailed discussion of the principles of benefits
analysis used here, we refer the reader to those documents, as well as to the EPA Guidelines
for Economic Analysis.
Table 8.1-1 summarizes the annual monetized health and welfare benefits associated
with the final standards for two years, 2020 and 2030. The estimates in Table 8.1-1, and all
monetized benefits presented in this chapter, are in year 2005 dollars. There are a few items
to note about these benefits:
• Using a conservative benefits estimate, the 2020 benefits outweigh the costs by a factor of
5. Using the upper end of the benefits range, the benefits could outweigh the costs by a
factor of 19. Likewise, in 2030 benefits outweigh the costs by at least a factor of 8 and
could be as much as a factor of 34. Thus, even taking the most conservative benefits
assumptions, benefits of the final standards clearly outweigh the costs.
• Emissions and air quality modeling decisions are made early in the analytical process. For
this reason, the emission control scenarios used in the air quality and benefits modeling
are slightly different than the final emission control program. The differences reflect
further refinements of the regulatory program since we performed the air quality modeling
for this rule. Chapter 3 of the RIA describes the changes in the inputs and resulting
emission inventories between the preliminary assumptions used for the air quality
modeling and the final regulatory scenario.
• The RIA for the proposal for this rulemaking only quantified benefits from PM; in the
current RIA we quantify and monetize the ozone-related health impacts associated with
the final rule. The science underlying the analysis is based on the current ozone criteria
document.3 The analytic approach to characterizing uncertainty is consistent with the
analysis used in the RIA for the final Ozone NAAQS.
• In a recent report on the estimation of ozone-related premature mortality published by the
National Research Council (NRC),4 a panel of experts and reviewers concluded that
ozone-related mortality should be included in estimates of the health benefits of reducing
ozone exposure. The report also recommended that the estimation of ozone-related
premature mortality be accompanied by broad uncertainty analyses while giving little or
no weight to the assumption that there is no causal association between ozone exposure
and premature mortality. Because EPA has yet to develop a coordinated response to the
NRC report's findings and recommendations, however, we have retained the approach to
8-3
-------
Final Regulatory Impact Analysis
estimating ozone-related premature mortality used in RIA for the final Ozone NAAQS.
EPA will specifically address the report's findings and recommendations in future
rulemakings.
Table 8.1-1. Estimated Monetized PM- and Ozone-Related Health Benefits of the Small
SI and Marine SI Engine Standards
2030 Total Ozone and PM Benefits - PM Mortality Derived from American Cancer Society Analysis3
Premature Ozone
Mortality Function or
Assumption
NMMAPS
Meta-analysis
Reference
Bell etal., 2004
Bell etal, 2005
Ito et al., 2005
Levy etal., 2005
Assumption that association is not causal6
Mean Total Benefits
(Billions, 2005$, 3%
Discount Rate)c'd
$2.4
$3.7
$4.4
$4.4
$1.8
Mean Total Benefits
(Billions, 2005$, 7%
Discount Rate) c'd
$2.2
$3.5
$4.2
$4.3
$1.6
2030 Total Ozone and PM Benefits - PM Mortality Derived from Expert Elicitationb
Premature Ozone
Mortality Function or
Assumption
NMMAPS
Meta-analysis
Reference
Bell et al., 2004
Bell et al., 2005
Ito et al., 2005
Levy etal., 2005
Assumption that association is not causal6
Mean Total Benefits
(Billions, 2005$, 3%
Discount Rate) c'd
$1.7 -$9.7
$3.0 -$11
$3.7 -$12
$3.7 -$12
$1.1 to $9.1
Mean Total Benefits
(Billions, 2005$, 7%
Discount Rate) c'd
$1.6 -$8.8
$2.9 -$10
$3.6 -$11
$3.7 -$11
$1.0 -$8.2
a Total includes ozone and PM2 5 benefits. Range was developed by adding the estimate from the ozone
premature mortality function to the estimate of PM25-related premature mortality derived from the American
Cancer Society analysis (Pope et al., 2002).
b Total includes ozone and PM25 benefits. Range was developed by adding the estimate from the ozone
premature mortality function to both the lower and upper ends of the range of the PM2 5 premature mortality
functions characterized in the expert elicitation. The effect estimates of five of the twelve experts included in the
elicitation panel fall within the empirically-derived range provided by the ACS and Six-Cities studies. One of
the experts fall below this range and six of the experts are above this range. Although the overall range across
experts is summarized in this table, the full uncertainty in the estimates is reflected by the results for the full set
of 12 experts. The twelve experts' judgments as to the likely mean effect estimate are not evenly distributed
across the range illustrated by arraying the highest and lowest expert means.
0 Note that total benefits presented here do not include a number of unqualified benefits categories. A
detailed listing of unqualified health and welfare effects is provided in Table 8.4-1.
d Results reflect the use of both a 3 and 7 percent discount rate, as recommended by EPA's Guidelines
for Preparing Economic Analyses and OMB Circular A-4. Results are rounded to two significant digits for ease
of presentation and computation.
e A recent report published by the National Research Council (NRC, 2008) recommended that EPA
"give little or no weight to the assumption that there is no causal association between estimated reductions in
premature mortality and reduced ozone exposure."
Table 8.1-1 reflects those human health and welfare effects we are able to quantify and
monetize. However, the full complement of known or suspected human health and welfare
effects associated with PM, ozone and air toxics remain unquantified because of current
limitations in methods or available data. We have not quantified potential health and welfare
effects of ozone and PM because impact functions are not available or do not provide easily
interpretable outcomes (e.g., changes in heart rate variability, acid and particulate deposition
8-4
-------
Cost-Benefit Analysis
damage to cultural monuments and other materials, and reductions in acidification of lakes
and streams and eutrophication in coastal areas). As a result, we may underestimate the total
benefits attributable to the implementation of the final standards.
This chapter is organized as follows. In Section 8.2, we provide an overview of the air
quality impacts modeled for the final standards that are used as inputs to the benefits analysis.
In Section 8.3, we discuss how uncertainty is characterized in this analysis. Section 8.4
discusses the literature on ozone- and PM-related health effects and describes the specific set
of health impact functions we used in the benefits analysis. Section 8.5 describes the
economic values selected to estimate the dollar value of ozone- and PM-related health
impacts. In Section 8.6, we report the results of the analysis for human health and welfare
effects. Finally, Section 8.7 presents a comparison of the costs and benefits associated with
the final standards. There are also two appendices associated with this chapter. The first,
Appendix 8 A, presents the results of the health-based cost effectiveness analysis. The second,
Appendix 8 A, presents the results of sensitivity analyses of key parameters in the benefits
analysis.
8.2 Air Quality Impacts for Benefits Analysis
In Chapter 2, we summarize the methods for and results of estimating air quality for
the 2020 and 2030 base case and final control scenario. These air quality results are in turn
associated with human populations and ecosystems to estimate changes in health and welfare
effects. For the purposes of the benefits analysis, we focus on the health effects that have
been linked to ambient changes in ozone and PM2.5 related to emission reductions estimated
to occur due to the final standards. We estimate ambient PM2.5 and ozone concentrations
using the Community Multiscale Air Quality model (CMAQ). The air quality modeling
Technical Support Document (TSD), which can be found in the docket for this rule, contains
detailed information about the modeling conducted for this rule. In this section, we describe
how the modeled air quality results were used for the benefits analysis.
We remind the reader that the emission control scenarios used in the air quality and
benefits modeling are slightly different than the final emission control program. The
differences reflect further refinements of the regulatory program since we performed the air
quality modeling for this rule. Emissions and air quality modeling decisions are made early in
the analytical process. Chapter 3 of the RIA describes the changes in the inputs and resulting
emission inventories between the preliminary assumptions used for the air quality modeling
and the final regulatory scenario.
8.2.1 Converting CMAQ Outputs to Full-Season Profiles for Benefits Analysis
This analysis extracted hourly, surface-layer PM and ozone concentrations for each
grid cell from the standard CMAQ output files. For ozone, these model predictions are used
8-5
-------
Final Regulatory Impact Analysis
in conjunction with the observed concentrations obtained from the Aerometric Information
Retrieval System (AIRS) to generate ozone concentrations for the entire ozone season.A!B
The predicted changes in ozone concentrations from the future-year base case to future-year
control scenario serve as inputs to the health and welfare impact functions of the benefits
analysis (i.e., the Environmental Benefits Mapping and Analysis Program [BenMAP]).
To estimate ozone-related health and welfare effects for the contiguous United States,
full-season ozone data are required for every BenMAP grid-cell. Given available ozone
monitoring data, we generated full-season ozone profiles for each location in two steps: (1)
we combined monitored observations and modeled ozone predictions to interpolate hourly
ozone concentrations to a grid of 12-km by 12-km population grid cells for the contiguous 48
states, and (2) we converted these full-season hourly ozone profiles to an ozone measure of
interest, such as the daily 8-hour maximum.C'D
For PM2.5, we also use the model predictions in conjunction with observed monitor
data. CMAQ generates predictions of hourly PM species concentrations for every grid. The
species include a primary coarse fraction (corresponding to PM in the 2.5 to 10 micron size
range), a primary fine fraction (corresponding to PM less than 2.5 microns in diameter), and
several secondary particles (e.g., sulfates, nitrates, and organics). PM2.5 is calculated as the
sum of the primary fine fraction and all of the secondarily formed particles. Future-year
estimates of PM2.5 were calculated using relative reduction factors (RRFs) applied to 2002
ambient PM2.5 and PM2.5 species concentrations. A gridded field of PM2.5 concentrations was
created by interpolating Federal Reference Monitor ambient data and IMPROVE ambient
data. Gridded fields of PM2 5 species concentrations were created by interpolating EPA
speciation network (ESPN) ambient data and IMPROVE data. The ambient data were
interpolated to the CMAQ 12 km grid.
The procedures for determining the RRFs are similar to those in EPA's draft guidance
for modeling the PM2 5 standard (EPA, 1999). The guidance recommends that model
predictions be used in a relative sense to estimate changes expected to occur in each major
PM2.5 species. The procedure for calculating future-year PM2.5 design values is called the
"Speciated Modeled Attainment Test (SMAT)." EPA used this procedure to estimate the
ambient impacts of the final emissions controls. Full documentation of the revised SMAT
methodology is contained in the Air Quality Modeling TSD.
A The ozone season for this analysis is defined as the 5-month period from May to September.
B Based on AIRS, there were 961 ozone monitors with sufficient data (i.e., 50 percent or more days reporting at
least nine hourly observations per day [8 am to 8 pm] during the ozone season).
c The 12-km grid squares contain the population data used in the health benefits analysis model, BenMAP.
D This approach is a generalization of planar interpolation that is technically referred to as enhanced Voronoi
Neighbor Averaging (EVNA) spatial interpolation. See the BenMAP manual for technical details, available for
download at http://www.epa.gov/ai^enmap.
8-6
-------
Cost-Benefit Analysis
8.2.2 Ozone and PMZ5 Air Quality Results
This section provides a summary of the predicted ambient PM2.5 and ozone
concentrations from the CMAQ model for the 2020 and 2030 base cases and changes
associated with the final rule. Table 8.2-1 provides those ozone and PM2.5 metrics for grid
cells in the modeled domain that enter the health impact functions for health benefits
endpoints. The population-weighted average reflects the baseline levels and predicted
changes for more populated areas of the nation. This measure better reflects the potential
benefits through exposure changes to these populations.
Table 8.2-1. Summary of CMAQ-Derived Population-Weighted Ozone and PM2.5 Air
Quality Metrics for Health Benefits Endpoints Due to the Final Small SI and Marine SI
Engine Standards
Statistic3
2020
Baseline
Change"
2030
Baseline
Change"
Ozone Metrics: National Population- Weighted Average (ppb)°
Daily 1 -Hour Maximum Concentration
Daily 8-Hour Maximum Concentration
Daily 8-Hour Average Concentration
Daily 24-Hour Average Concentration
47.60
44.07
42.63
35.39
0.078
0.066
0.062
0.047
46.91
43.47
42.06
35.02
0.108
0.093
0.088
0.068
PM2 5 Metrics: National Population- Weighted Average (ug/m3)
Annual Average Concentration
9.41
0.015
9.38
0.021
a Ozone and PM2 5 metrics are calculated at the CMAQ grid-cell level for use in health effects estimates
based on the results of spatial and temporal Voronoi Neighbor Averaging. Ozone metrics are calculated over
relevant time periods during the daylight hours of the "ozone season" (i.e., May through September). For the 8-
hour average, for example, the relevant time period is 9 am to 5 pm.
b The change is defined as the base-case value minus the control-case value.
0 Calculated by summing the product of the projected CMAQ grid-cell population and the estimated
CMAQ grid cell seasonal ozone concentration and then dividing by the total population.
8.3 Characterizing Uncertainty: Moving Toward a Probabilistic
Framework for Benefits Assessment
The National Research Council (NRC)5 highlighted the need for EPA to conduct
rigorous quantitative analysis of uncertainty in its benefits estimates and to present these
estimates to decision makers in ways that foster an appropriate appreciation of their inherent
uncertainty. In response to these comments, EPA's Office of Air and Radiation (OAR) is
developing a comprehensive strategy for characterizing the aggregate impact of uncertainty in
key modeling elements on both health incidence and benefits estimates. Components of that
process include emissions modeling, air quality modeling, health effects incidence estimation,
and valuation.
In benefit analyses of air pollution regulations conducted to date, the estimated impact
of reductions in premature mortality has accounted for 85% to 95% of total benefits.
8-7
-------
Final Regulatory Impact Analysis
Therefore, it is particularly important to characterize the uncertainties associated with
reductions in premature mortality. The health impact functions used to estimate avoided
premature deaths associated with reductions in ozone have associated standard errors that
represent the statistical errors around the effect estimates in the underlying epidemiological
studies.E In our results, we report credible intervals based on these standard errors, reflecting
the uncertainty in the estimated change in incidence of avoided premature deaths. We also
provide multiple estimates, to reflect model uncertainty between alternative study designs. In
addition, we characterize the uncertainty introduced by the inability of existing empirical
studies to discern whether the relationship between ozone and pre-mature mortality is causal
by providing an effect estimate preconditioned on an assumption that the effect estimate for
pre-mature mortality from ozone is zero.
For premature mortality associated with exposure to PM, we follow the same approach
that has been used in several recent RIAs.F'G'H First, we use Monte Carlo methods for
estimating random sampling error associated with the concentration response functions from
epidemiological studies and economic valuation functions. Monte Carlo simulation uses
random sampling from distributions of parameters to characterize the effects of uncertainty on
output variables, such as incidence of premature mortality. Specifically, we used Monte
Carlo methods to generate confidence intervals around the estimated health impact and dollar
benefits. Distributions for individual effect estimates are based on the reported standard
errors in the epidemiological studies. Distributions for unit values are described in Table 8.5-
1.
Second, we use the results of our expert elicitation of the concentration response
function describing the relationship between premature mortality and ambient PM2.5
concentration.1^ Incorporating only the uncertainty from random sampling error omits
important sources of uncertainty (e.g., in the functional form of the model; whether or not a
E Health impact functions measure the change in a health endpoint of interest, such as hospital admissions, for a
given change in ambient ozone or PM concentration.
F U.S. Environmental Protection Agency, 2004a. Final Regulatory Analysis: Control of Emissions from
Nonroad Diesel Engines. EPA420-R-04-007. Prepared by Office of Air and Radiation. Available at
http://www.epa.gov/nonroad-diesel/2004fr/420r04007.pdf
G U.S. Environmental Protection Agency, 2005. Regulatory Impact Analysis for the Clean Air Interstate Rule.
EPA452/-03-001. Prepared by Office of Air and Radiation. Available at:
http://www.epa.gov/interstateairqualitv/tsd0175.pdf
H U.S. Environmental Protection Agency, 2006. Regulatory Impact Analysis for the PM NAAQS. EPA Prepared
by Office of Air and Radiation. Available at: http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205-
Benefits.pdf
1 Expert elicitation is a formal, highly structured and well documented process whereby expert judgments,
usually of multiple experts, are obtained (Ayyb, 2002).
1 Industrial Economics, Inc. 2006. Expanded Expert Judgment Assessment of the Concentration-Response
Relationship Between PM2.5 Exposure and Mortality. Prepared for EPA Office of Air Quality Planning and
Standards, September. Available at: http://www.epa.gov/ttn/ecas/regdata/Uncertaintv/pm ee report.pdf
-------
Cost-Benefit Analysis
threshold may exist). This second approach attempts to incorporate these other sources of
uncertainty.
Use of the expert elicitation and incorporation of the standard errors approaches
provide insights into the likelihood of different outcomes and about the state of knowledge
regarding the benefits estimates. Both approaches have different strengths and weaknesses,
which are fully described in Chapter 5 of the PM NAAQS RIA.
These multiple characterizations, including confidence intervals, omit the contribution
to overall uncertainty of uncertainty in air quality changes, baseline incidence rates,
populations exposed and transferability of the effect estimate to diverse locations.
Furthermore, the approach presented here does not yet include methods for addressing
correlation between input parameters and the identification of reasonable upper and lower
bounds for input distributions characterizing uncertainty in additional model elements. As a
result, the reported confidence intervals and range of estimates give an incomplete picture
about the overall uncertainty in the estimates. This information should be interpreted within
the context of the larger uncertainty surrounding the entire analysis.
8.4 Health Impact Functions
Health impact functions measure the change in a health endpoint of interest, such as
hospital admissions, for a given change in ambient ozone or PM concentration. Health impact
functions are derived from primary epidemiology studies, meta-analyses of multiple
epidemiology studies, or expert elicitations. A standard health impact function has four
components: 1) an effect estimate from a particular study; 2) a baseline incidence rate for the
health effect (obtained from either the epidemiology study or a source of public health
statistics such as the Centers for Disease Control); 3) the size of the potentially affected
population; and 4) the estimated change in the relevant ozone or PM summary measures.
A typical health impact function might look like:
Ay = jv (*>•**-l),
where yo is the baseline incidence (the product of the baseline incidence rate times the
potentially affected population), P is the effect estimate, and Ax is the estimated change in the
summary pollutant measure. There are other functional forms, but the basic elements remain
the same. Section 6.2 described the ozone and PM air quality inputs to the health impact
functions. The following subsections describe the sources for each of the other elements:
size of potentially affected populations; effect estimates; and baseline incidence rates.
8.4.1 Potentially Affected Populations
The starting point for estimating the size of potentially affected populations is the
2000 U.S. Census block level dataset.6 Benefits Modeling and Analysis Program (BenMAP)
incorporates 250 age/gender/race categories to match specific populations potentially affected
8-9
-------
Final Regulatory Impact Analysis
by ozone and other air pollutants. The software constructs specific populations matching the
populations in each epidemiological study by accessing the appropriate age-specific
populations from the overall population database. BenMAP projects populations to 2020
using growth factors based on economic projections.7
8.4.2 Effect Estimate Sources
The most significant monetized benefits of reducing ambient concentrations of ozone
and PM are attributable to reductions in human health risks. EPA's Ozone and PM Criteria
Documents8'9 and the World Health Organization's 2003 and 200410'11 reports outline
numerous health effects known or suspected to be linked to exposure to ambient ozone and
PM. EPA recently evaluated the PM literature for use in the benefits analysis for the 2006
PM NAAQS RIA. Because we used the same literature for the PM benefits analysis in this
RIA, and also in the RIA for the proposed rule, we do not provide a detailed discussion of
individual effect estimates for PM in this section. Instead, we refer the reader to the 2006 PM
NAAQS RIA and the proposed Small SI and Marine SI RIA for details.K
The RIA for the proposal for this rulemaking only quantified benefits from PM; in the
current RIA we quantify and monetize the ozone-related health and environmental impacts
associated with the final rule using an approach consistent with the final ozone NAAQS RIA.
More than one thousand new ozone health and welfare studies have been published since EPA
issued the 8-hour ozone standard in 1997. Many of these studies investigated the impact of
ozone exposure on health effects such as: changes in lung structure and biochemistry; lung
inflammation; asthma exacerbation and causation; respiratory illness-related school absence;
hospital and emergency room visits for asthma and other respiratory causes; and premature
death. We provide a discussion of those ozone-related impacts in this section. For a more
detailed discussion of the health effects of ozone exposure, we point the reader to EPA's
ozone Criteria Document.12
It is important to note that we were not able to separately quantify all of the PM and
ozone health effects that have been reported in the ozone and PM criteria documents in this
analysis for four reasons: (1) the possibility of double counting (such as hospital admissions
for specific respiratory diseases); (2) uncertainties in applying effect relationships that are
based on clinical studies to the potentially affected population; (3) the lack of an established
concentration-response relationship; or 4) the inability to appropriately value the effect (for
example, changes in forced expiratory volume) in economic terms. Table 8.4-1 lists the
human health and welfare effects of pollutants affected by the final standards. Table 8.4-2
lists the health endpoints included in this analysis.
K U.S. Environmental Protection Agency, 2005. Regulatory Impact Analysis for the PM NAAQS. EPA Prepared
by Office of Air and Radiation. Available at: http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205--
Benefits.pdf pp. 5-29.
8-10
-------
Cost-Benefit Analysis
Table 8.4-1 Human Health and Welfare Effects of Pollutants Affected by the
Final Standards
Pollutant/Effect
Quantified and Monetized in Base
Estimates
Unqualified Effects - Changes in:
PM/Health
b
PM/Welfare
Ozone/Health
Ozone/Welfare
Premature mortality based on both
cohort study estimates and on expert
elicitation0'
Bronchitis: chronic and acute
Hospital admissions: respiratory
and cardiovascular
Emergency room visits for asthma
Nonfatal heart attacks (myocardial
infarction)
Lower and upper respiratory illness
Minor restricted-activity days
Work loss days
Asthma exacerbations (asthmatic
population)
Respiratory symptoms (asthmatic
population)
Subchronic bronchitis cases
Low birth weight
Pulmonary function
Chronic respiratory diseases other than chronic bronchitis
Nonasthma respiratory emergency room visits
UVb exposure (+/-)e
Premature mortality: short-term
exposures
Hospital admissions: respiratory
Emergency room visits for asthma
Minor restricted-activity days
School loss days
Asthma attacks
Acute respiratory symptoms
Decreased outdoor worker
productivity
Visibility in Southeastern Class I areas
Visibility in northeastern and Midwestern Class I areas
Household soiling
Visibility in we stern U.S. Class I areas
Visibility in residential and non-Class I areas
UVb exposure (+/-)e
Cardiovascular emergency room visits
Chronic respiratory damage8
Premature aging of the lungs8
Nonasthma respiratory emergency room visits
UVb exposure (+/-)e
Yields for commercial crops
Yields for commercial forests and noncommercial crops
Damage to urban ornamental plants
Recreational demand from damaged forest aesthetics
Ecosystem functions
UVb exposure (+/-)e
8-11
-------
Final Regulatory Impact Analysis
Pollutant/Effect
Quantified and Monetized in Base
Estimates
Unqualified Effects - Changes in:
Nitrogen
Deposition/
Welfare
NOx/Health
HC/Toxics
Health11
HC/Toxics
Welfare11
Commercial forests due to acidic sulfate and nitrate
deposition
Commercial freshwater fishing due to acidic deposition
Recreation in terrestrial ecosystems due to acidic
deposition
Commercial fishing, agriculture, and forests due to
nitrogen deposition
Recreation in estuarine ecosystems due to nitrogen
deposition
Ecosystem functions
Passive fertilization
Lung irritation
Lowered resistance to respiratory infection
Hospital admissions for respiratory and cardiac diseases
Cancer, including lung (benzene, 1,3-butadiene,
formaldehyde, acetaldehyde, naphthalene)
Anemia (benzene)
Disruption of production of blood components (benzene)
Reduction in the number of blood platelets (benzene)
Excessive bone marrow formation (benzene)
Depression of lymphocyte counts (benzene)
Reproductive and developmental effects (1,3-butadiene)
Irritation of eyes and mucus membranes (formaldehyde)
Respiratory irritation (formaldehyde)
Asthma attacks in asthmatics (formaldehyde)
Asthma-like symptoms in non-asthmatics (formaldehyde)
Irritation of the eyes, skin, and respiratory tract
(acetaldehyde)
Upper respiratory tract irritation and congestion (acrolein)
Neurotoxicity (n-hexane, toluene, xylenes)
Direct toxic effects to animals
Bioaccumulation in the food chain
Damage to ecosystem function
Odor
Primary quantified and monetized effects are those included when determining the primary estimate of
total monetized benefits of the final standards.
In addition to primary economic endpoints, there are a number of biological responses that have been
associated with PM health effects including morphological changes and altered host defense mechanisms. The
public health impact of these biological responses may be partly represented by our quantified endpoints.
c Cohort estimates are designed to examine the effects of long term exposures to ambient pollution, but
relative risk estimates may also incorporate some effects due to shorter term exposures (see Kunzli, 2001 for a
discussion of this issue).
While some of the effects of short-term exposure are likely to be captured by the cohort estimates,
there may be additional premature mortality from short-term PM exposure not captured in the cohort estimates
included in the primary analysis.
e May result in benefits or disbenefits.
The public health impact of biological responses such as increased airway responsiveness to stimuli,
inflammation in the lung, acute inflammation and respiratory cell damage, and increased susceptibility to
respiratory infection are likely partially represented by our quantified endpoints.
8-12
-------
Cost-Benefit Analysis
8 The public health impact of effects such as chronic respiratory damage and premature aging of the
lungs may be partially represented by quantified endpoints such as hospital admissions or premature mortality,
but a number of other related health impacts, such as doctor visits and decreased athletic performance, remain
unqualified.
h The categorization of unqualified toxic health and welfare effects is not exhaustive.
8-13
-------
Final Regulatory Impact Analysis
Table 8.4-2. Ozone- and PM-Related Health Endpoints
Endpoint
Pollutant
Study
Study Population
Premature Mortality
Premature mortality -
daily time series, non-
accidental
Premature mortality
— cohort study, all-
cause
Premature mortality,
total exposures
Premature mortality
— all-cause
ozone
PM25
PM25
PM25
Bell et al (2004) (NMMAPS study)13
Meta-analyses:
Belletal(2005)14
Ito et al (2005)15
Levyetal(2005)16
Pope et al. (2002)17
Laden et al. (2006)18
Expert Elicitation (lEc, 2006)19
Woodruff etal. (1997)20
All ages
>29 years
>25 years
>24 years
Infant (<1 year)
Chronic Illness
Chronic bronchitis
Nonfatal heart attacks
PM25
PM25
Abbey etal. (1995)21
Peters etal. (200 1)22
>26 years
Adults (>18 years)
Hospital Admissions
Respiratory
Cardiovascular
Asthma-related ER
visits
ozone
PM25
PM25
PM25
PM25
PM25
PM25
ozone
Pooled estimate:
Schwartz (1995) - ICD 460-519 (all resp)23
Schwartz (1994a; 1994b) - ICD 480-486
(pneumonia)24'25
Moolgavkar et al. (1997) - ICD 480-487
(pneumonia)26
Schwartz (1994b) - ICD 491-492, 494-496
(COPD)
Moolgavkar et al. (1997) - ICD 490-496
(COPD)
Burnett etal. (200 1)27
Pooled estimate:
Moolgavkar (2003)— ICD 490-496 (COPD)28
Ito (2003)— ICD 490-496 (COPD)29
Moolgavkar (2000)— ICD 490-496 (COPD)30
Ito (2003)— ICD 480-486 (pneumonia)
Sheppard (2003)— ICD 493 (asthma)31
Pooled estimate:
Moolgavkar (2003)— ICD 390-429 (all
cardiovascular)
Ito (2003)— ICD 410-414, 427-428 (ischemic
heart disease, dysrhythmia, heart failure)
Moolgavkar (2000)— ICD 390-429 (all
cardiovascular)
Pooled estimate:
Jaffe et al (2003)32
Peel et al (2005)33
Wilson etal(2005)34
>64 years
<2 years
>64 years
20-64 years
>64 years
<65 years
>64 years
20-64 years
5-34 years
All ages
All ages
8-14
-------
Cost-Benefit Analysis
Endpoint
Asthma-related ER
visits (con't)
Pollutant
PM25
Study
Norrisetal. (1999)35
Study Population
0-18 years
Other Health Endpoints
Acute bronchitis
Upper respiratory
symptoms
Lower respiratory
symptoms
Asthma exacerbations
Work loss days
School absence days
Minor Restricted
Activity Days
(MRADs)
PM25
PM25
PM25
PM25
PM25
ozone
ozone
PM25
Dockeryetal. (1996)36
Popeetal. (1991)37
Schwartz and Neas (2000)38
Pooled estimate:
Ostro et al. (200 1)39 (cough, wheeze and
shortness of breath)
Vedal et al. (1998)40 (cough)
Ostro (1987)41
Pooled estimate:
Gilliland et al. (200 1)42
Chenetal. (2000)43
Ostro and Rothschild (1989)44
Ostro and Rothschild (1989)
8-12 years
Asthmatics, 9-11
years
7-14 years
6-18 years3
18-65 years
5-17 years'3
18-65 years
18-65 years
a The original study populations were 8 to 13 for the Ostro et al. (2001) study and 6 to 13 for the Vedal et al.
(1998) study. Based on advice from the Science Advisory Board Health Effects Subcommittee (SAB-HES),
we extended the applied population to 6 to 18, reflecting the common biological basis for the effect in
children in the broader age group. See: U.S. Science Advisory Board. 2004. Advisory Plans for Health
Effects Analysis in the Analytical Plan for EPA's Second Prospective Analysis -Benefits and Costs of the
Clean Air Act, 1990—2020. EPA-SAB-COUNCIL-ADV-04-004. See also National Research Council
(NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution Regulations. Washington,
DC: The National Academies Press.
b Gilliland et al. (2001) studied children aged 9 and 10. Chen et al. (2000) studied children 6 to 11. Based on
recent advice from the National Research Council and the EPA SAB-HES, we have calculated reductions in
school absences for all school-aged children based on the biological similarity between children aged 5 to 17.
In selecting epidemiological studies as sources of effect estimates, we applied several
criteria to develop a set of studies that is likely to provide the best estimates of impacts in the
U.S. To account for the potential impacts of different health care systems or underlying
health status of populations, we give preference to U.S. studies over non-U.S. studies. In
addition, due to the potential for confounding by co-pollutants, we give preference to effect
estimates from models including both ozone and PM over effect estimates from single-
pollutant models.45'46
A number of endpoints that are not health-related also may significantly contribute to
monetized benefits. Potential welfare benefits associated with ozone exposure include:
increased outdoor worker productivity; increased yields for commercial and non-commercial
crops; increased commercial forest productivity; reduced damage to urban ornamental plants;
increased recreational demand for undamaged forest aesthetics; and reduced damage to
ecosystem functions.47'48 While we include estimates of the value of increased outdoor
worker productivity, estimation of other welfare impacts is beyond the scope of this analysis.
8-15
-------
Final Regulatory Impact Analysis
8.4.2.1 Ozone Exposure Metric
Both the NMMAPS analysis and the individual time series studies upon which the
meta-analyses were based use the 24-hour average or 1-hour maximum ozone levels as
exposure metrics. L The 24-hour average is not the most relevant ozone exposure metric to
characterize population-level exposure. Given that the majority of the people tend to be
outdoors during the daylight hours and concentrations are highest during the daylight hours,
the 24-hour average metric is not appropriate. The maximum 1-hour average metric uses an
exposure window different than that that used for the current ozone NAAQS. Together, this
means that the most biologically relevant metric is the maximum 8-hour average, which has
also been the metric for ozone NAAQS since 1997. Thus, for the final rule analysis, we have
converted ozone mortality health impact functions that use a 24-hour average or 1-hour
maximum ozone metric to maximum 8-hour average ozone concentration using standard
conversion functions.
This practice is consistent both with the available exposure modeling and with the
form of the current ozone standard. This conversion also does not affect the relative
magnitude of the health impact function. An equivalent change in the 24-hour average,
maximum 1-hour average, and maximum 8-hour average will provide the same overall change
in incidence of a health effect. The conversion ratios are based on observed relationships
between the 24-hour average and maximum 8-hour average ozone values. For example, in the
Bell et al., 2004 analysis of ozone-related premature mortality, the authors found that the
relationship between the 24-hour average, the maximum 8-hour average, and the maximum 1-
hour average was 2:1.5:1, so that the derived health impact effect estimate based on the
maximum 1-hour average should be half that of the effect estimate based on the 24-hour
values (and the maximum 8-hour average three-quarters of the 24-hour effect estimate).
8.4.2.2 Premature Mortality Effect Estimates
While particulate matter is the criteria pollutant most clearly associated with
premature mortality, recent research suggests that short-term repeated ozone exposure likely
contributes to premature death. The 2006 Ozone Criteria Document states: "Consistent with
observed ozone-related increases in respiratory- and cardiovascular-related morbidity, several
newer multi-city studies, single-city studies, and several meta-analyses of these studies have
provided relatively strong epidemiologic evidence for associations between short-term ozone
exposure and all-cause mortality, even after adjustment for the influence of season and PM"
(EPA, 2006: E-17).49 The epidemiologic data are also supported by newly available
experimental data from both animal and human studies which provide evidence suggestive of
L An exposure metric is a measure of air quality calculated as the average or maximum of modeled ambient
concentrations over a relevant time period, such as during the daylight hours of the "ozone season" (which is
May through September for this analysis). The 24-hour average is therefore calculated as the average of all
hourly ozone concentrations throughout the day (from 12am to ll:59pm). The 8-hour maximum is the
maximum hourly value observed between 9am and 5pm each day. The 1-hour maximum is the maximum hourly
value observed throughout an entire day.
8-16
-------
Cost-Benefit Analysis
plausible pathways by which risk of respiratory or cardiovascular morbidity and mortality
could be increased by ambient ozone. With respect to short-term exposure, the ozone Criteria
Document concludes: "This overall body of evidence is highly suggestive that ozone directly
or indirectly contributes to non-accidental and cardiopulmonary-related mortality, but
additional research is needed to more fully establish underlying mechanisms by which such
effects occur" (pg. E-18).
With respect to the time-series studies, the conclusion regarding the relationship
between short-term exposure and premature mortality is based, in part, upon recent city-
specific time-series studies such as the Schwartz (2004) analysis in Houston and the Huang et
al. (2004) analysis in Los Angeles.M This conclusion is also based on recent meta-analyses by
Bell et al. (2005), Ito et al. (2005), and Levy et al. (2005), and a new analysis of the National
Morbidity, Mortality, and Air Pollution Study (NMMAPS) data set by Bell et al. (2004),
which specifically sought to disentangle the roles of ozone, PM, weather-related variables,
and seasonality. The 2006 Criteria Document states that "the results from these meta-
analyses, as well as several single- and multiple-city studies, indicate that co-pollutants
generally do not appear to substantially confound the association between ozone and
mortality" (p. 7-103). However, CASAC raised questions about the implications of these
time-series results in a policy context. Specifically, CASAC emphasized that".. .while the
time-series study design is a powerful tool to detect very small effects that could not be
detected using other designs, it is also a blunt tool" (Henderson, 2006: 3). They point to
findings (e.g., Stieb et al., 2002, 2003) that indicated associations between premature
mortality and all of the criteria pollutants, indicating that "findings of time-series studies do
not seem to allow us to confidently attribute observed effects to individual pollutants" (id.).
They note that "not only is the interpretation of these associations complicated by the fact that
the day-to-day variation in concentrations of these pollutants is, to a varying degree,
determined by meteorology, the pollutants are often part of a large and highly correlated mix
of pollutants, only a very few of which are measured" (id.). Even with these uncertainties, the
CASAC Ozone Panel, in its review of EPA's Staff Paper, found "... premature total non-
accidental and cardiorespiratory mortality for inclusion in the quantitative risk assessment to
be appropriate."
Consistent with the methodology used in the ozone risk assessment found in the
Characterization of Health Risks found in the Review of the National Ambient Air Quality
Standards for Ozone: Policy Assessment of Scientific and Technical Information, we included
ozone mortality in the primary health effects analysis, with the recognition that the exact
magnitude of the effects estimate is subject to continuing uncertainty. We used effect
estimates from the Bell et al. (2004) NMMAPS analysis, as well as effect estimates from the
three meta-analyses.
M For an exhaustive review of the city-specific time-series studies considered in the ozone staff paper, see: U.S.
Environmental Protection Agency, 2007. Review of the National Ambient Air Quality Standards for Ozone:
Policy Assessment of Scientific and Technical Information. Prepared by the Office of Air and Radiation.
Available at http://www.epa.gov/ttn/naaqs/standards/ozone/data/2007_0l_ozone_staff_paper.pdf. pp. 5-36.
8-17
-------
Final Regulatory Impact Analysis
In a recent report on the estimation of ozone-related premature mortality published by
the National Research Council (NRC),50 a panel of experts and reviewers concluded that
ozone-related mortality should be included in estimates of the health benefits of reducing
ozone exposure. The report also recommended that the estimation of ozone-related premature
mortality be accompanied by broad uncertainty analyses while giving little or no weight to the
assumption that there is no causal association between ozone exposure and premature
mortality. Because EPA has yet to develop a coordinated response to the NRC report's
findings and recommendations, however, we have retained the approach to estimating ozone-
related premature mortality used in RIA for the final Ozone NAAQS. EPA will specifically
address the report's findings and recommendations in future rulemakings.
We estimate the change in mortality incidence and estimated credible intervalN
resulting from application of the effect estimate from each study and present them separately
to reflect differences in the study designs and assumptions about causality. However, it is
important to note that this procedure only captures the uncertainty in the underlying
epidemiological work, and does not capture other sources of uncertainty, such as uncertainty
in the estimation of changes in air pollution exposure (Levy et al., 2000).
8.4.2.3 Respiratory Hospital Admissions Effect Estimates
Detailed hospital admission and discharge records provide data for an extensive body
of literature examining the relationship between hospital admissions and air pollution. This is
especially true for the portion of the population aged 65 and older, because of the availability
of detailed Medicare records. In addition, there is one study (Burnett et al., 2001) providing
an effect estimate for respiratory hospital admissions in children under two.
Because the number of hospital admission studies we considered is so large, we used
results from a number of studies to pool some hospital admission endpoints. Pooling is the
process by which multiple study results may be combined in order to produce better estimates
of the effect estimate, or p. For a complete discussion of the pooling process, see Abt (2005).°
To estimate total respiratory hospital admissions associated with changes in ambient ozone
concentrations for adults over 65, we first estimated the change in hospital admissions for
each of the different effects categories that each study provided for each city. These cities
included Minneapolis, Detroit, Tacoma and New Haven. To estimate total respiratory
hospital admissions for Detroit, we added the pneumonia and COPD estimates, based on the
effect estimates in the Schwartz study (1994). Similarly, we summed the estimated hospital
admissions based on the effect estimates the Moolgavkar study reported for Minneapolis
(Moolgavkar et al., 1997). To estimate total respiratory hospital admissions for Minneapolis
using the Schwartz study (1994), we simply estimated pneumonia hospital admissions based
on the effect estimate. Making this assumption that pneumonia admissions represent the total
N A credible interval is a posterior probability interval used in Bayesian statistics, which is similar to a
confidence interval used in frequentist statistics.
0 Abt Associates, Incorporated. Environmental Benefits Mapping and Analysis Program, Technical Appendices.
May 2005. pp. 1-3
8-18
-------
Cost-Benefit Analysis
impact of ozone on hospital admissions in this city will give some weight to the possibility
that there is no relationship between ozone and COPD, reflecting the equivocal evidence
represented by the different studies. We then used a fixed-effects pooling procedure to
combine the two total respiratory hospital admission estimates for Minneapolis. Finally, we
used random effects pooling to combine the results for Minneapolis and Detroit with results
from studies in Tacoma and New Haven from Schwartz (1995). As noted above, this pooling
approach incorporates both the precision of the individual effect estimates and between-study
variability characterizing differences across study locations.
8.4.2.4 Asthma-Related Emergency Room Visits Effect Estimates
We used three studies as the source of the concentration-response functions we used to
estimate the effects of ozone exposure on asthma-related emergency room (ER) visits: Peel et
al. (2005); Wilson et al. (2005); and Jaffe et al. (2003). We estimated the change in ER visits
using the effect estimate(s) from each study and then pooled the results using the random
effects pooling technique (see Abt, 2005). The study by Jaffe et al. (2003) examined the
relationship between ER visits and air pollution for populations aged five to 34 in the Ohio
cities of Cleveland, Columbus and Cincinnati from 1991 through 1996. In single-pollutant
Poisson regression models, ozone was linked to asthma visits. We use the pooled estimate
across all three cities as reported in the study. The Peel et al. study (2005) estimated asthma-
related ER visits for all ages in Atlanta, using air quality data from 1993 to 2000. Using
Poisson generalized estimating equations, the authors found a marginal association between
the maximum daily 8-hour average ozone level and ER visits for asthma over a 3-day moving
average (lags of 0, 1, and 2 days) in a single pollutant model. Wilson et al. (2005) examined
the relationship between ER visits for respiratory illnesses and asthma and air pollution for all
people residing in Portland, Maine from 1998-2000 and Manchester, New Hampshire from
1996-2000. For all models used in the analysis, the authors restricted the ozone data
incorporated into the model to the months ozone levels are usually measured, the spring-
summer months (April through September). Using the generalized additive model, Wilson et
al. (2005) found a significant association between the maximum daily 8-hour average ozone
level and ER visits for asthma in Portland, but found no significant association for
Manchester. Similar to the approach used to generate effect estimates for hospital
admissions, we used random effects pooling to combine the results across the individual study
estimates for ER visits for asthma. The Peel et al. (2005) and Wilson et al. (2005) Manchester
estimates were not significant at the 95 percent level, and thus, the confidence interval for the
pooled incidence estimate based on these studies includes negative values. This is an artifact
of the statistical power of the studies, and the negative values in the tails of the estimated
effect distributions do not represent improvements in health as ozone concentrations are
increased. Instead these should be viewed as a measure of uncertainty due to limitations in
the statistical power of the study. Note that we included both hospital admissions and ER
visits as separate endpoints associated with ozone exposure, because our estimates of hospital
admission costs do not include the costs of ER visits, and because most asthma ER visits do
not result in a hospital admission.
8-19
-------
Final Regulatory Impact Analysis
8.4.2.5 Minor Restricted Activity Days Effects Estimate
Minor restricted activity days (MRADs) occur when individuals reduce most usual
daily activities and replace them with less-strenuous activities or rest, but do not miss work or
school. We estimated the effect of ozone exposure on MRADs using a concentration-
response function derived from Ostro and Rothschild (1989). These researchers estimated
the impact of ozone and PM2.5 on MRAD incidence in a national sample of the adult working
population (ages 18 to 65) living in metropolitan areas. We developed separate coefficients
for each year of the Ostro and Rothschild analysis (1976-1981), which we then combined for
use in EPA's analysis. The effect estimate used in the impact function is a weighted average
of the coefficients in Ostro and Rothschild (1989, Table 4), using the inverse of the variance
as the weight.
8.4.2.6 School Absences Effect Estimate
Children may be absent from school due to respiratory or other acute diseases caused,
or aggravated by, exposure to air pollution. Several studies have found a significant
association between ozone levels and school absence rates. We use two studies (Gilliland et
al., 2001; Chen et al., 2000) to estimate changes in school absences resulting from changes in
ozone levels. The Gilliland et al. study estimated the incidence of new periods of absence,
while the Chen et al. study examined daily absence rates. We converted the Gilliland et al.
estimate to days of absence by multiplying the absence periods by the average duration of an
absence. We estimated 1.6 days as the average duration of a school absence, the result of
dividing the average daily school absence rate from Chen et al. (2000) and Ransom and Pope
(1992) by the episodic absence duration from Gilliland et al. (2001). Thus, each Gilliland et
al. period of absence is converted into 1.6 absence days.
Following recent advice from the National Research Council (2002), we calculated
reductions in school absences for the full population of school age children, ages five to 17.
This is consistent with recent peer-reviewed literature on estimating the impact of ozone
exposure on school absences (Hall et al. 2003). We estimated the change in school absences
using both Chen et al. (2000) and Gilliland et al. (2001) and then, similar to hospital
admissions and ER visits, pooled the results using the random effects pooling procedure.
8.4.2.7 Worker Productivity
To monetize benefits associated with increased worker productivity resulting from
improved ozone air quality, we used information reported in Crocker and Horst (1981).
Crocker and Horst examined the impacts of ozone exposure on the productivity of outdoor
citrus workers. The study measured productivity impacts. Worker productivity is measuring
the value of the loss in productivity for a worker who is at work on a particular day, but due to
ozone, cannot work as hard. It only applies to outdoor workers, like fruit and vegetable
pickers, or construction workers. Here, productivity impacts are measured as the change in
income associated with a change in ozone exposure, given as the elasticity of income with
respect to ozone concentration. The reported elasticity translates a ten percent reduction in
ozone to a 1.4 percent increase in income. Given the national median daily income for
8-20
-------
Cost-Benefit Analysis
outdoor workers engaged in strenuous activity reported by the U.S. Census Bureau (2002),
$68 per day (2000$), a ten percent reduction in ozone yields about $0.97 in increased daily
wages. We adjust the national median daily income estimate to reflect regional variations in
income using a factor based on the ratio of county median household income to national
median household income. No information was available for quantifying the uncertainty
associated with the central valuation estimate. Therefore, no uncertainty analysis was
conducted for this endpoint.
8.4.2.8 Unqualified Effects
8.4.2.8.1 Direct Ozone Effects on Vegetation
The Ozone Criteria Document notes that "current ambient concentrations in many
areas of the country are sufficient to impair growth of numerous common and economically
valuable plant and tree species." (U.S. EPA, 2006, page 9-1). Changes in ground-level ozone
resulting from the implementation of alternative ozone standards are expected to affect crop
and forest yields throughout the affected area. Recent scientific studies have also found the
ozone negatively impacts the quality or nutritive value of crops (U.S. EPA, 2006, page 9-16).
Well-developed techniques exist to provide monetary estimates of these benefits to
agricultural producers and to consumers. These techniques use models of planting decisions,
yield response functions, and the supply of and demand for agricultural products. The
resulting welfare measures are based on predicted changes in market prices and production
costs. Models also exist to measure benefits to silvicultural producers and consumers.
However, these models have not been adapted for use in analyzing ozone-related forest
impacts. Because of resource limitations, we are unable to provide agricultural or benefits
estimates for the final rule.
An additional welfare benefit expected to accrue as a result of reductions in ambient
ozone concentrations in the United States is the economic value the public receives from
reduced aesthetic injury to forests. There is sufficient scientific information available to
reliably establish that ambient ozone levels cause visible injury to foliage and impair the
growth of some sensitive plant species (U.S. EPA, 2006, page 9-19). However, present
analytic tools and resources preclude EPA from quantifying the benefits of improved forest
aesthetics.
Urban ornamentals (floriculture and nursery crops) represent an additional vegetation
category likely to experience some degree of negative effects associated with exposure to
ambient ozone levels and likely to affect 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. The farm production value of ornamental crops was estimated at over $14
billion in 2003 (USDA, 2004). This is therefore a potentially important welfare effects
category. However, information and valuation methods are not available to allow for
8-21
-------
Final Regulatory Impact Analysis
plausible estimates of the percentage of these expenditures that may be related to impacts
associated with ozone exposure.
8.4.2.8.2 Nitrogen Deposition
Deposition to Estuarine and Coastal Waters
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). A recent study found that for the period 1990-2002,
atmospheric deposition accounted for 17 percent of nitrate loadings in the Gulf of Mexico,
where severe hypoxic zones have been existed over the last two decades (Booth and
Campbell, 2007)p.
Reductions in atmospheric deposition of NOx are expected to reduce the adverse
impacts associated with nitrogen deposition to estuarine and coastal waters. However, direct
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
functions and on estimates of the value of environmental responses. Because neither
appropriate 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.
Deposition to Agricultural and Forested Land
Implementation strategies for alternative standards which reduce 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.
p Booth, M.S., and C. Campbell. 2007. Spring Nitrate Flux in the Mississippi River Basin: A Landscape Model
with Conservation Applications. Environ. Sci. Technol.; 2007; ASAP Web Release Date: 20-Jun-2007; (Article)
DOI: 10.1021/es070179e
8-22
-------
Cost-Benefit Analysis
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). Moreover, any positive effect that nitrogen deposition has on forest productivity would
enhance the level of carbon dioxide sequestration as well.Q'R's
On the other hand, there is evidence that forest ecosystems in some areas of the United
States (such as the western U.S.) 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.
8.4.2.8.3 Ultraviolet Radiation
Atmospheric ozone absorbs a harmful band of ultraviolet radiation from the sun called
UV-B, providing a protective shield to the Earth's surface. The majority of this protection
occurs in the stratosphere where 90% of atmospheric ozone is located. The remaining 10% of
the Earth's ozone is present at ground level (referred to as tropospheric ozone) (NAS, 1991;
NASA). Only a portion of the tropospheric fraction of UV-B shielding is from anthropogenic
sources (e.g., power plants, byproducts of combustion). The portion of ground level ozone
associated with anthropogenic sources varies by locality and over time. Even so, it is
reasonable to assume that reductions in ground level ozone would lead to increases in the
same health effects linked to in UV-B exposures. These effects include fatal and nonfatal
melanoma and non-melanoma skin cancers and cataracts. The values of $15,000 per case for
non-fatal melanoma skin cancer, $5,000 per case for non-fatal non-melanoma skin cancer, and
$15,000 per case of cataracts have been used in analyses of stratospheric ozone depletion
(U.S. EPA, 1999). Fatal cancers are valued using the standard VSL estimate, which for 2020
is $6.6 million (1999$). UV-B has also been linked to ecological effects including damage to
crops and forest. For a more complete listing of quantified and unquantified UV-B radiation
effects, see Table G-4 and G-7 in the Benefits and Costs of the Clean Air Act, 1990-2010
(U.S. EPA, 1999. UV-B related health effects are also discussed in the context of
stratospheric ozone in a 2006 report by ICF Consulting, prepared for the U.S. EPA.
Q Peter M. Vitousek et. al, "Human Alteration of the Global Nitrogen Cycle: Causes and Consequences" Issues
in Ecology No. 1 (Spring) 1997.
R Knute J. Nadelhoffer et. al., "Nitrogen deposition makes a minor contribution to carbon sequestration in
temperate forests" Nature 398, 145-148 (11 March 1999)
s
Martin Kochy and Scott D. Wilson, "Nitrogen deposition and forest expansion in the northern Great Plains
Journal of Ecology Journal of Ecology 89 (5), 807-817
8-23
-------
Final Regulatory Impact Analysis
There are many factors that influence UV-B radiation penetration to the earth's
surface, including latitude, altitude, cloud cover, surface albedo, PM concentration and
composition, and gas phase pollution. Of these, only latitude and altitude can be defined with
small uncertainty in any effort to assess the changes in UV-B flux that may be attributable to
any changes in tropospheric ozone as a result of any revision to the ozone NAAQS. Such an
assessment of UV-B related health effects would also need to take into account human habits,
such as outdoor activities (including age- and occupation-related exposure patterns), dress and
skin care to adequately estimate UV-B exposure levels. However, little is known about the
impact of these factors on individual exposure to UV-B.
Moreover, detailed information does not exist regarding other factors that are relevant
to assessing changes in disease incidence, including: type (e.g., peak or cumulative) and time
period (e.g., childhood, lifetime, current) of exposures related to various adverse health
outcomes (e.g., damage to the skin, including skin cancer; damage to the eye, such as
cataracts; and immune system suppression); wavelength dependency of biological responses;
and interindividual variability in UV-B resistance to such health outcomes. Beyond these well
recognized adverse health effects associated with various wavelengths of UV radiation, the
Criteria Document (section 10.2.3.6) also discusses protective effects of UV-B radiation.
Recent reports indicate the necessity of UV-B in producing vitamin D, and that vitamin D
deficiency can cause metabolic bone disease among children and adults, and may also
increase the risk of many common chronic diseases (e.g., type I diabetes and rheumatoid
arthritis) as well as the risk of various types of cancers. Thus, the Criteria Document
concludes that any assessment that attempts to quantify the consequences of increased UV-B
exposure on humans due to reduced ground-level ozone must include consideration of both
negative and positive effects. However, as with other impacts of UVB on human health, this
beneficial effect of UVB radiation has not previously been studied in sufficient detail.
The Agency is currently evaluating the feasibility of estimating the effects of
increased UVB exposures resulting from reductions in tropospheric ozone. Please refer to the
final Ozone NAAQS RIA for a sensitivity analysis that explores the quantification of UV-B-
related health effects.51
8.4.2.8.4 Climate Implications of Tropospheric Ozone
Although climate and air quality are generally treated as separate issues, they are
closely coupled through atmospheric processes. Ozone, itself, is a major greenhouse gas and
climate directly influences ambient concentrations of ozone.
The concentration of tropospheric ozone has increased substantially since the pre-
industrial era and has contributed to warming. Tropospheric ozone is (after CO2 and CH4)
the third most important contributor to greenhouse gas warming. The National Academy of
Sciences recently stated1 that regulations targeting ozone precursors would have combined
T National Academy of Sciences, "Radiative Forcing of Climate Change: Expanding the Concept and Addressing
Uncertainties," October 2005.
8-24
-------
Cost-Benefit Analysis
benefits for public health and climate. As noted in the OAQPS Staff Paper, the overall body
of scientific evidence suggests that high concentrations of ozone on a regional scale could
have a discernible influence on climate. However, the Staff Paper concludes that insufficient
information is available at this time to quantitatively inform the secondary NAAQS process
with regard to this aspect of the ozone-climate interaction.
Climate change can affect tropospheric ozone by modifying emissions of precursors,
chemistry, transport and removal.u Climate change affects the sources of ozone precursors
through physical response (lightning), biological response (soils, vegetation, and biomass
burning) and human response (energy generation, land use, and agriculture). Increases in
regional ozone pollution are expected due to higher temperatures and weaker circulation.
Simulations with global climate models for the 21st century indicate a decrease in the lifetime
of tropospheric ozone due to increasing water vapor which could decrease global background
ozone concentrations.
The Intergovernmental Panel on Climate Change (IPCC) recently released a reportv
which projects, with "virtual certainty," declining air quality in cities due to warmer and
fewer cold days and nights and/or warmer/more frequent hot days and nights over most land
areas. The report states that projected climate change-related exposures are likely to affect the
health status of millions of people, in part, due to higher concentrations of ground level ozone
related to climate change.
The IPCC also reports* that the current generation of tropospheric ozone models is
generally successful in describing the principal features of the present-day global ozone
distribution. However, there is much less confidence in the ability to reproduce the changes
in ozone associated with perturbations of emissions or climate. There are major discrepancies
with observed long-term trends in ozone concentrations over the 20th century, including after
1970 when the reliability of observed ozone trends is high. Resolving these discrepancies is
needed to establish confidence in the models.
uDenman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C. Heinze,
E. Holland, D. Jacob, U. Lohmann, S Ramachandran, P.L. da Silva Bias, S.C. Wofsy and X. Zhang, 2007:
Couplings Between Changes in the Climate System and Biogeochemistry. In: Climate Change 2007: The
Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M.
Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA.
v IPCC, Climate Change 2007: Climate Change Impacts, Adaptation and Vulnerability,
Summary for Policymakers
w
Denman, et al, 2007: Couplings Between Changes in the Climate System and Biogeochemistry. In: Climate
Change 2007: The Physical Science Basis.
8-25
-------
Final Regulatory Impact Analysis
The EPA is currently leading a research effort with the goal of identifying changes in
regional US air quality that may occur in a future (2050) climate, focusing on fine particles
and ozone. The research builds first on an assessment of changes in US air quality due to
climate change, which includes direct meteorological impacts on atmospheric chemistry and
transport and the effect of temperature changes on air pollution emissions. Further research
will result in an assessment that adds the emission impacts from technology, land use,
demographic changes, and air quality regulations to construct plausible scenarios of US air
quality 50 years into the future. As noted in the Staff Paper, results from these efforts are
expected to be available for consideration in the next review of the ozone NAAQS.
8.4.3 Baseline Incidence Rates
Epidemiological studies of the association between pollution levels and adverse health
effects generally provide a direct estimate of the relationship of air quality changes to the
relative risk of a health effect, rather than estimating the absolute number of avoided cases.
For example, a typical result might be that a 100 ppb decrease in daily ozone levels might, in
turn, decrease hospital admissions by 3 percent. The baseline incidence of the health effect is
necessary to convert this relative change into a number of cases. A baseline incidence rate is
the estimate of the number of cases of the health effect per year in the assessment location, as
it corresponds to baseline pollutant levels in that location. To derive the total baseline
incidence per year, this rate must be multiplied by the corresponding population number. For
example, if the baseline incidence rate is the number of cases per year per 100,000 people,
that number must be multiplied by the number of 100,000s in the population.
Table 8.4-3 summarizes the sources of baseline incidence rates and provides average
incidence rates for the endpoints included in the analysis. For both baseline incidence and
prevalence data, we used age-specific rates where available. We applied concentration-
response functions to individual age groups and then summed over the relevant age range to
provide an estimate of total population benefits. In most cases, we used a single national
incidence rate, due to a lack of more spatially disaggregated data. Whenever possible, the
national rates used are national averages, because these data are most applicable to a national
assessment of benefits. For some studies, however, the only available incidence information
comes from the studies themselves; in these cases, incidence in the study population is
assumed to represent typical incidence at the national level. Regional incidence rates are
available for hospital admissions, and county-level data are available for premature mortality.
We have projected mortality rates such that future mortality rates are consistent with our
projections of population growth (Abt Associates, 2005).
8-26
-------
Cost-Benefit Analysis
Table 8.4-3. National Average Baseline Incidence Rates3
Endpoint
Mortality
Respiratory
Hospital
Admissions.
Asthma ER visits
Minor Restricted
Activity Days
(MRADs)
School Loss
Days
Endpoint
Asthma
Exacerbations
Source
CDC Compressed Mortality File,
accessed through CDC Wonder
(1996-1998)
1999 NHDS public use data files'3
2000 NHAMCS public use data
files'; 1999 NHDS public use data
filesb
Ostro and Rothschild
(1989, p. 243)
National Center for Education
Statistics (1996) and 1996 HIS
(Adams et al., 1999, Table 47);
estimate of 180 school days per
year
Source
Ostro etal. (2001)
Vedaletal. (1998)
Notes
non-
accidental
incidence
incidence
incidence
all-cause
Rate per 100 people per yeard by Age Group
<18
0.025
0.043
1.011
990.0
18-24
0.022
0.084
1.087
780
25-34
0.057
0.206
0.751
780
Notes
Incidence (and
prevalence) among
asthmatic African-
American children
Incidence among
asthmatic children
Daily wheeze
Daily cough
Daily dyspnea
Daily wheeze
Daily cough
Daily dyspnea
35-44
0.150
0.678
0.438
780
45-54
0.383
1.926
0.352
780
55-64
1.006
4.389
0.425
780
65+
4.937
11.62
9
0.232
Rate per 100 people per year
0.076(0.173)
0.067(0.145)
0.037 (0.074)
0.038
0.086
0.045
The following abbreviations are used to describe the national surveys conducted by the National Center for
Health Statistics: HIS refers to the National Health Interview Survey; NHDS - National Hospital Discharge Survey;
NHAMCS - National Hospital Ambulatory Medical Care Survey.
b See ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NHDS/
0 See ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NHAMCS/
All of the rates reported here are population-weighted incidence rates per 100 people per year. Additional details
on the incidence and prevalence rates, as well as the sources for these rates are available upon request.
8.5 Economic Values for Health Outcomes
Reductions in ambient concentrations of air pollution generally lower the risk of future
adverse health effects for a large population. Therefore, the appropriate economic measure is
willingness-to-pay (WTP) for changes in risk of a health effect rather than WTP for a health
effect that would occur with certainty (Freeman, 1993). Epidemiological studies generally
provide estimates of the relative risks of a particular health effect that is avoided because of a
reduction in air pollution. We converted those to units of avoided statistical incidence for ease
of presentation. We calculated the value of avoided statistical incidences by dividing
8-27
-------
Final Regulatory Impact Analysis
individual WTP for a risk reduction by the related observed change in risk. For example,
suppose a pollution-reduction regulation 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 death is $1 million
($100/0.0001 change in risk).
WTP estimates generally are not available for some health effects, such as hospital
admissions. In these cases, we used the cost of treating or mitigating the effect as a primary
estimate. These cost-of-illness (COI) estimates generally understate the true value of
reducing the risk of a health effect, because they reflect the direct expenditures related to
treatment, but not the value of avoided pain and suffering (Harrington and Portney, 1987;
Berger, 1987). We provide unit values for health endpoints (along with information on the
distribution of the unit value) in Table 8.5-1. All values are in constant year 2000 dollars,
adjusted for growth in real income out to 2020 using projections provided by Standard and
Poor's. Economic theory argues that WTP for most goods (such as environmental protection)
will increase if real income increases. Many of the valuation studies used in this analysis
were conducted in the late 1980s and early 1990s. Because real income has grown since the
studies were conducted, people's willingness to pay for reductions in the risk of premature
death and disease likely has grown as well. We did not adjust cost of illness-based values
because they are based on current costs. Similarly, we did not adjust the value of school
absences, because that value is based on current wage rates. Table 8.5-1 presents the values
for individual endpoints adjusted to year 2020 income levels. The discussion below provides
additional details on ozone related endpoints not previously included in the proposal for this
rule. For details on valuation estimates for PM-related endpoints, see the 2006 PM NAAQS
RIA and the proposed Small SI and Marine SI RIA.
8.5.1 Mortality Valuation
To estimate the monetary benefit of reducing the risk of premature death, we used the
"value of statistical lives" saved (VSL) approach, which is a summary measure for the value
of small changes in mortality risk for a large number of people. The VSL approach applies
information from several published value-of-life studies to determine a reasonable monetary
value of preventing premature mortality. The mean value of avoiding one statistical death is
estimated to be roughly $6.2 million at 1990 income levels (2005$), and $7.5 million at 2020
income levels. This represents an intermediate value from a variety of estimates in the
economics literature (see the 2006 PM NAAQS RIA for more details on the calculation of
VSL).
8.5.2 Hospital Admissions Valuation
In the absence of estimates of societal WTP to avoid hospital visits/admissions for
specific illnesses, estimates of total cost of illness (total medical costs plus the value of lost
productivity) typically are used as conservative, or lower bound, estimates. These estimates
are biased downward, because they do not include the willingness-to-pay value of avoiding
pain and suffering.
8-28
-------
Cost-Benefit Analysis
The International Classification of Diseases (ICD-9, 1979) code-specific COI
estimates used in this analysis consist of estimated hospital charges and the estimated
opportunity cost of time spent in the hospital (based on the average length of a hospital stay
for the illness). We based all estimates of hospital charges and length of stays on statistics
provided by the Agency for Healthcare Research and Quality (AHRQ 2000). We estimated
the opportunity cost of a day spent in the hospital as the value of the lost daily wage,
regardless of whether the hospitalized individual is in the workforce. To estimate the lost
daily wage, we divided the 1990 median weekly wage by five and inflated the result to year
2005$ using the CPI-U "all items." The resulting estimate is $135.59. The total cost-of-
illness estimate for an ICD code-specific hospital stay lasting n days, then, was the mean
hospital charge plus $136 • n.
8.5.3 Asthma-Related Emergency Room Visits Valuation
To value asthma emergency room visits, we used a simple average of two estimates
from the health economics literature. The first estimate comes from Smith et al. (1997), who
reported approximately 1.2 million asthma-related emergency room visits in 1987, at a total
cost of $186.5 million (1987$). The average cost per visit that year was $155; in 2005$, that
cost was $386.32 (using the CPI-U for medical care to adjust to 2005$). The second estimate
comes from Stanford et al. (1999), who reported the cost of an average asthma-related
emergency room visit at $323.23 (in 2005$), based on 1996-1997 data. A simple average of
the two estimates yields a (rounded) unit value of $355.
8.5.4 Minor Restricted Activity Days Valuation
No studies are reported to have estimated WTP to avoid a minor restricted activity
day. However, one of EPA's contractors, lEc (1993) has derived an estimate of willingness to
pay to avoid a minor respiratory restricted activity day, using estimates from Tolley et al.
(1986) of WTP for avoiding a combination of coughing, throat congestion and sinusitis. The
lEc estimate of WTP to avoid a minor respiratory restricted activity day is $38.37 (1990$), or
about $59 (2005$).
Although Ostro and Rothschild (1989) statistically linked ozone and minor restricted
activity days, it is likely that most MRADs associated with ozone exposure are, in fact, minor
respiratory restricted activity days. For the purpose of valuing this health endpoint, we used
the estimate of mean WTP to avoid a minor respiratory restricted activity day.
8.5.5 School Absences
To value a school absence, we: (1) estimated the probability that if a school child
stays home from school, a parent will have to stay home from work to care for the child; and
(2) valued the lost productivity at the parent's wage. To do this, we estimated the number of
families with school-age children in which both parents work, and we valued a school-loss
day as the probability that such a day also would result in a work-loss day. We calculated this
value by multiplying the proportion of households with school-age children by a measure of
8-29
-------
Final Regulatory Impact Analysis
lost wages.
We used this method in the absence of a preferable WTP method. However, this
approach suffers from several uncertainties. First, it omits willingness to pay to avoid the
symptoms/illness that resulted in the school absence; second, it effectively gives zero value to
school absences that do not result in work-loss days; and third, it uses conservative
assumptions about the wages of the parent staying home with the child. Finally, this method
assumes that parents are unable to work from home. If this is not a valid assumption, then
there would be no lost wages.
For this valuation approach, we assumed that in a household with two working
parents, the female parent will stay home with a sick child. From the Statistical Abstract of
the United States (U.S. Census Bureau, 2001), we obtained: (1) the numbers of single,
married and "other" (widowed, divorced or separated) working women with children; and (2)
the rates of participation in the workforce of single, married and "other" women with
children. From these two sets of statistics, we calculated a weighted average participation rate
of 72.85 percent.
Our estimate of daily lost wage (wages lost if a mother must stay at home with a sick
child) is based on the year 2000 median weekly wage among women ages 25 and older (U.S.
Census Bureau, 2001). This median weekly wage is $551. Dividing by five gives an estimated
median daily wage of $103. To estimate the expected lost wages on a day when a mother has
to stay home with a school-age child, we first estimated the probability that the mother is in
the workforce then multiplied that estimate by the daily wage she would lose by missing a
work day: 72.85 percent times $103, for a total loss of $75. Using the CPI-U for all items to
adjust to 2005$, the value equals approximately $85. This valuation approach is similar to
that used by Hall et al. (2003).
8-30
-------
Cost-Benefit Analysis
Table 8.5-1. Unit Values Used for Economic Valuation of Health Endpoints (2005$)a
Health Endpoint
Premature Mortality (Value of a
Statistical Life): PM25- and
Ozone-related
Chronic Bronchitis (CB)
Nonfatal Myocardial Infarction
(heart attack)
3% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
7% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
$6,200,000
$380,000
$82,958
$92,598
$97,754
$174,405
$82,958
$80,963
$90,705
$95,320
$163,945
$80,963
2020 Income
Levelb
$7,500,000
$470,000
$82,958
$92,598
$97,754
$174,405
$82,958
$80,963
$90,705
$95,320
$163,945
$80,963
2030 Income
Levelb
$7,700,000
$490,000
$82,958
$92,598
$97,754
$174,405
$82,958
$80,963
$90,705
$95,320
$163,945
$80,963
Derivation of Estimates
Point estimate is the mean of a normal distribution with a 95 percent
confidence interval between $1 and $10 million (in 2000$). Confidence
interval is based on two meta-analyses of the wage-risk VSL literature:
$ 1 million represents the lower end of the interquartile range from the
Mrozek and Taylor (2002)52 meta-analysis and $10 million represents
the upper end of the interquartile range from the Viscusi and Aldy
(2003)53 meta-analysis. Adjusted for 2005$, the mean equals
approximately $6.2 million. The VSL represents the value of a small
change in mortality risk aggregated over the affected population.
Point estimate 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]54)
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.
Age-specific cost-of-illness values reflect lost earnings and direct
medical costs over a 5-year period following a nonfatal MI. L0,,t
earnings estimates are based on Cropper and Krupnick (1990). Direct
medical costs are based on simple average of estimates from Russell et
al. (1998)56 and Wittels et al. (1990).57
Lost earnings:
Cropper and Krupnick (1990). Present discounted value of 5 years of
lost earnings:
age of onset: , -,<,, at 7%
25-44 $10,880 $9,740
45-54 $1M36 $14,357
55-65 $92,685 $82,958
Direct medical expenses: An average of:
1. Wittels et al. (1990) ($127,296— no discounting)
2. Russell et al. (1998), 5-year period ($27,690 at 3% discount rate;
$26,180 at 7% discount rate)
(continued)
8-31
-------
Final Regulatory Impact Analysis
Table 8.5-1. Unit Values Used for Economic Valuation of Health Endpoints (2005$)a (continued)
Health Endpoint
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
2020 Income
T 1b
Level
2030 Income
T 1b
Level
Derivation of Estimates
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
$15,345
$18,219
$8,226
$22,800
$355
$15,345
$18,219
$8,226
$22,800
$355
$15,345
$18,219
$8,226
$22,800
$355
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code-level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total COPD
category illnesses) reported in Agency for Healthcare Research and
CQ
Quality (2000) (www.ahrq.gov).
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code-level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total
pneumonia category illnesses) reported in Agency for Healthcare
Research and Quality (2000) (www.ahrq.gov).
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code-level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total asthma
category illnesses) reported in Agency for Healthcare Research and
Quality (2000) (www.ahrq.gov).
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code-level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total
cardiovascular category illnesses) reported in Agency for Healthcare
Research and Quality (2000) (www.ahrq.gov).
Simple average of two unit COI values:
(1) $386.32, from Smith et al. (1997)59 and
(2) $323.23, from Stanford et al. (1999).60
(continued)
8-32
-------
Cost-Benefit Analysis
Table 8.5-1. Unit Values Used for Economic Valuation of Health Endpoints (2005$)a (continued)
Health Endpoint
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
2020 Income
T 1b
Level
2030 Income
T 1b
Level
Derivation of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
$28
$30
$30
Combinations of the three symptoms for which WTP estimates are
available that closely match those listed by Pope et al. result in seven
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (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 seven different types of URS.
Lower Respiratory Symptoms
(LRS)
$18
$19
$19
Combinations of the four symptoms for which WTP estimates are
available that closely match those listed by Schwartz et al. result in
11 different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 11 different types of LRS.
Asthma Exacerbations
$47
$51
$51
Asthma exacerbations are valued at $47 per incidence (2005$), based
on the mean of average WTP estimates for the four severity
definitions of a "bad asthma day," described in Rowe and Chestnut
fO
(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.
Acute Bronchitis
$407
$434
$438
Assumes a 6-day episode, with daily value equal to the average of
low and high values for related respiratory symptoms recommended
in Neumann et al. (1994).
(continued)
8-33
-------
Final Regulatory Impact Analysis
Table 8.5-1. Unit Values Used for Economic Valuation of Health Endpoints (2005$)a (continued)
Health Endpoint
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
2020 Income
Level
2030 Income
Level
Derivation of Estimates
Restricted Activity and Work/School Loss Days
Work Loss Days (WLDs)
School Absence Days
Worker Productivity
Minor Restricted Activity Days
(MRADs)
Variable
(national
median = )
$85
$1.07 per
worker per
10% change in
ozone per day
$58
$85
$1.07 per
worker per
10% change
in ozone per
day
$61
$85
$1.07 per
worker per
10% change in
ozone per day
$62
County -specific median annual wages divided by 50 (assuming 2 weeks
of vacation) and then by 5 — to get median daily wage. U.S. Year 2000
Census, compiled by Geolytics, Inc.
Based on expected lost wages from parent staying home with child.
Estimated daily lost wage (if a mother must stay at home with a sick
child) is based on the median weekly wage among women age 25 and
older in 2000 (U.S. Census Bureau, Statistical Abstract of the United
States: 2001, Section 12: Labor Force, Employment, and Earnings,
Table No. 621). This median wage is $551. Dividing by 5 gives an
estimated median daily wage of $103..
The expected loss in wages due to a day of school absence in which the
mother would have to stay home with her child is estimated as the
probability that the mother is in the workforce times the daily wage she
would lose if she missed a day = 72.85% of $103, or $75 ($85 in 2005$)
Based on $68 ($77 in 2005$) - median daily earnings of workers in
farming, forestry and fishing - from Table 621, Statistical Abstract of
the United States ("Full-Time Wage and Salary Workers - Number and
Earnings: 1985 to 2000") (Source of data in table: U.S. Bureau of
Labor Statistics, Bulletin 2307 and Employment and Earnings, monthly).
Median WTP estimate to avoid one MRAD from Tolley et al. (1986).
All annual benefit estimates associated with the final standards have been inflated to reflect values in year 2005 dollars. We use the Consumer Price
Indexes to adjust both WTP- and COI-based benefits estimates to 2005 dollars from 2000 dollars. For WTP-based estimates, we use an inflation factor of
1.13 based on the CPI-U for "all items." For COI-based estimates, we use an inflation factor of 1.24 based on the CPI-U for medical care.
Our analysis accounts for expected growth in real income over time. Economic theory argues that WTP for most goods (such as environmental protection)
will increase if real incomes increase. Benefits are therefore adjusted by multiplying the unadjusted benefits by the appropriate adjustment factor to account
for income growth over time. For a complete discussion of how these adjustment factors were derived, we refer the reader to the PM NAAQS regulatory
impact analysis. Note that similar adjustments do not exist for cost-of-illness-based unit values. For these, we apply the same unit value regardless of the
future year of analysis.
8-34
-------
Cost-Benefit Analysis
8.6 Benefits Analysis Results for the Final Standards
Applying the impact and valuation functions described previously in this chapter to
the estimated changes in PM2.5 and ozone associated with the final standards results in
estimates of the changes in health damages (e.g., premature mortalities, cases, admissions)
and the associated monetary values for those changes. Estimates of physical health impacts
are presented in Table 8.6-1. Monetized values for those health endpoints are presented in
Table 8.6-2. Total aggregate monetized benefits are presented in Table 8.6-3 and Table 8.6-4
using either a 3 percent or 7 percent discount rate, respectively. All of the monetary benefits
are in constant-year 2005 dollars. For each endpoint presented in Tables 8.6-1 and 8.6-2, we
provide both the mean estimate and the 90% confidence interval.
In addition to omitted benefits categories such as air toxics and various welfare
effects, not all known PM2.5- and ozone-related health and welfare effects could be quantified
or monetized. The estimate of total monetized health benefits of the final standards is thus
equal to the subset of monetized PM2.5- and ozone-related health benefits we are able to
quantify plus the sum of the nonmonetized health and welfare benefits. We believe the total
benefits are therefore likely underestimated.
Total monetized benefits are dominated by benefits of mortality risk reductions. We
provide results for particulate matter based on PM2.5 concentration response functions from
the American Cancer Society Study (ACS), Six Cities, and Expert Elicitation to give an
indication of the sensitivity of the benefits estimates to alternative assumptions. Following the
recommendations of the NRC report (NRC, 2002), we identify those estimates which are
based on empirical data, and those which are based on expert judgments. EPA recently asked
its Science Advisory Board (SAB) to evaluate how EPA has incorporated expert elicitation
results into the benefits analysis, and the extent to which they find the presentation in this RIA
responsive to the NRC (2002) guidance to incorporate uncertainty into the main analysis and
further, whether the agency should move toward presenting a central estimate with
uncertainty bounds or continue to provide separate estimates for each of the 12 experts as well
as from the ACS and Six Cities studies. EPA has not yet had a chance to incorporate the
results of the SAB's July 11, 2008 report (EPA-COUNCIL-08-002).
Using the ACS and Six-Cities results, we estimate that the final standards would result
in between 150 and 340 cases of avoided PM2.5-related premature deaths annually in 2020 and
between 230 and 510 avoided premature deaths annually in 2030. When the range of expert
opinion is used, we estimate between 80 and 840 fewer premature mortalities in 2020 and
between 120 and 1,300 fewer premature mortalities in 2030. Note that in the case of the
premature mortality estimates derived from the expert elicitation, we report the 90% credible
interval, which encompasses a broader representation of uncertainty relative to the statistical
confidence intervals provided for the effect estimates derived from the epidemiology
literature.
The range of ozone benefits associated with the final standards is based on risk
reductions estimated using several sources of ozone-related mortality effect estimates. This
8-35
-------
Final Regulatory Impact Analysis
analysis presents four alternative estimates for the association based upon different functions
reported in the scientific literature, derived from both the National Morbidity, Mortality, and
Air Pollution Study (NMMAPS) (Bell et al., 2004) and from a series of recent meta-analyses
(Bell et al., 2005, Ito et al., 2005, and Levy et al., 2005). This approach is not inconsistent
with recommendations provided by the NRC in their recent report (NRC, 2008) on the
estimation of ozone-related mortality risk reductions, "The committee recommends that the
greatest emphasis be placed on estimates from new systematic multicity analyses that use
national databases of air pollution and mortality, such as in the NMMAPS, without excluding
consideration of meta-analyses of previously published studies."
Prior to the publication of the NRC ozone mortality report, EPA considered the
possibility that the observed associations between ozone and mortality may not be causal in
nature. The report, however, recommended that EPA give "little or no weight to the
assumption that there is no causal association between ozone exposure and premature
mortality." Because EPA has yet to develop a coordinated response to the NRC report's
findings and recommendations, we have retained the approach to estimating ozone-related
premature mortality used in RIA for the final Ozone NAAQS. EPA will specifically address
the report's findings and recommendations in future rulemakings.
For ozone-related premature mortality, we estimate a range of between 46 to 210
fewer premature mortalities as a result of the final rule in 2020 and between 77 to 350 in
2030, assuming that there is a causal relationship between ozone exposure and mortality. The
increase in annual benefits from 2020 to 2030 reflects additional emission reductions from the
final standards, as well as increases in total population and the average age (and thus baseline
mortality risk) of the population.
Our estimate of total monetized benefits in 2020 for the final standards, using the ACS
and Six-Cities PM mortality studies and the range of ozone mortality assumptions, is between
$1.2 billion and $4.0 billion, assuming a 3 percent discount rate, or between $1.1 billion and
$3.8 billion, assuming a 7 percent discount rate. In 2030, we estimate the monetized benefits
to be between $1.8 billion and $6.4 billion, assuming a 3 percent discount rate, or between
$1.6 billion and $6.1 billion, assuming a 7 percent discount rate. The monetized benefit
associated with reductions in the risk of both ozone- and PM2.s-related premature mortality
ranges between 90 to 98 percent of total monetized health benefits, in part because we are
unable to quantify a number benefits categories (see Table 8.4-1). These unqualified
benefits may be substantial, although their magnitude is highly uncertain.
The next largest benefit is for reductions in chronic illness (chronic bronchitis and
nonfatal heart attacks), although this value is more than an order of magnitude lower than for
premature mortality. Hospital admissions for respiratory and cardiovascular causes, minor
restricted activity days, and work loss days account for the majority of the remaining benefits.
The remaining categories each account for a small percentage of total benefit; however, they
represent a large number of avoided incidences affecting many individuals. A comparison of
the incidence table to the monetary benefits table reveals that there is not always a close
correspondence between the number of incidences avoided for a given endpoint and the
monetary value associated with that endpoint. For example, there are over 100 times more
8-36
-------
Cost-Benefit Analysis
work loss days than PM-related premature mortalities (based on the ACS study), yet work
loss days account for only a very small fraction of total monetized benefits. This reflects the
fact that many of the less severe health effects, while more common, are valued at a lower
level than the more severe health effects. Also, some effects, such as hospital admissions, are
valued using a proxy measure of willingness-to-pay (e.g., cost-of-illness). As such, the true
value of these effects may be higher than that reported in Table 8.6-2.
Following these tables, we also provide a more comprehensive presentation of the
distributions of incidence generated using the available information from empirical studies
and expert elicitation. Tables 8.6-5 and 8.6-6 present the distributions of the reduction in
PM2.s-related premature mortality based on the C-R distributions provided by each expert, as
well as that from the data-derived health impact functions, based on the statistical error
associated with the ACS study (Pope et al., 2002) and the Six-cities study (Laden et al., 2006).
The 90% confidence interval for each separate estimate of PM-related mortality is also
provided.
The effect estimates of five of the twelve experts included in the elicitation panel fall
within the empirically-derived range provided by the ACS and Six-Cities studies. One of the
experts fall below this range and six of the experts are above this range. Although the overall
range across experts is summarized in these tables, the full uncertainty in the estimates is
reflected by the results for the full set of 12 experts. The twelve experts' judgments as to the
likely mean effect estimate are not evenly distributed across the range illustrated by arraying
the highest and lowest expert means.
8-37
-------
Final Regulatory Impact Analysis
Table 8.6-1. Estimated Reduction in Incidence of Adverse Health Effects Related to the
Final Standards21
Health Effect
2020
2030
Mean Incidence Reduction
(5th_95th0/oile)
PM-Related Endpoints
Premature Mortality -
Derived from Epidemiology
Literature
Premature Mortality -
Derived from Expert
Elicitationb
Adult, age 30+ - ACS cohort
study (Pope et al., 2002)
Adult, age 25+ - Six-Cities
study (Laden et al., 2006)
Infant, age <1 year -
Woodruff etal. 1997
Adult, age 25+ - Lower
Bound (Expert K)
Adult, age 25+ - Upper Bound
(Expert E)
Chronic bronchitis (adult, age 26 and over)
Acute myocardial infarction (adults, age 18 and older)
Hospital admissions — respiratory (all ages)0
Hospital admissions — cardiovascular (adults, age >18)d
Emergency room visits for asthma (age 18 years and younger)
Acute bronchitis (children, age 8-12)
Lower respiratory symptoms (children, age 7-14)
Upper respiratory symptoms (asthmatic children, age 9-18)
Asthma exacerbation (asthmatic children, age 6-18)
Work loss days (adults, age 18-65)
Minor restricted-activity days (adults, age 18-65)
150
(60 - 240)
340
(190-500)
0
(0-1)
81
(0 - 380)
840
(420 - 1,300)
150
(28 - 270)
330
(180-480)
40
(20 - 59)
81
(50-110)
150
(85-210)
400
(-14-810)
2,700
(1,300-4,000)
1,900
(610-3,300)
2,400
(270 - 7,000)
17,000
(15,000 - 19,000)
100,000
(86,000 - 120,000)
230
(88 - 360)
510
(280 - 740)
1
(0-1)
120
(0 - 580)
1,300
(650 - 1,900)
220
(40 - 400)
530
(280 - 770)
61
(30 - 88)
130
(82 - 180)
210
(120-300)
580
(-20 - 1,200)
3,800
(1,800 - 5,800)
2,800
(880 - 4,700)
3,500
(380 - 10,000)
23,000
(20,000 - 26,000)
140,000
(120,000 - 160,000)
Ozone-Related Endpoints
Premature Mortality, All ages
- Derived from NMMAPS
Premature Mortality, All ages
- Derived from Meta-analyses
Bell et al., 2004
Bell et al., 2005
Ito etal., 2005
Levy etal., 2005
Premature Mortality - Assumption that association between
ozone and mortality is not causal6
Hospital admissions- respiratory causes (children, under 2;
46
(20-72)
150
(84-210)
200
(140-270)
210
(160-260)
0
540
77
(34 - 120)
250
(140-360)
340
(230 - 450)
350
(260 - 440)
0
1,000
8-38
-------
Cost-Benefit Analysis
adult, 65 and older)
Emergency room visit for asthma (all ages)
Minor restricted activity days (adults, age 18-65)
School absence days
(170 - 900)
200
(0-510)
310,000
(160,000-460,000)
110,000
(40,000 - 200,000)
(290 - 1,700)
320
(0-810)
450,000
(230,000 - 670,000)
180,000
(62,000 - 320,000)
Incidence is rounded to two significant digits. PM and ozone estimates represent impacts from the
final standards nationwide.
Based on effect estimates derived from the full-scale expert elicitation assessing the uncertainty in the
concentration-response function for PM-related premature mortality (lEc, 2006).66 The effect estimates of five
of the twelve experts included in the elicitation panel fall within the empirically-derived range provided by the
ACS and Six-Cities studies. One of the experts fall below this range and six of the experts are above this range.
Although the overall range across experts is summarized in this table, the full uncertainty in the estimates is
reflected by the results for the full set of 12 experts. The twelve experts' judgments as to the likely mean effect
estimate are not evenly distributed across the range illustrated by arraying the highest and lowest expert means.
Respiratory hospital admissions for PM include admissions for chronic obstructive pulmonary disease
(COPD), pneumonia, and asthma.
Cardiovascular hospital admissions for PM include total cardiovascular and subcategories for
ischemic heart disease, dysrhythmias, and heart failure.
e A recent report published by the National Research Council (NRC, 2008) recommended that EPA
"give little or no weight to the assumption that there is no causal association between estimated reductions in
premature mortality and reduced ozone exposure."
Respiratory hospital admissions for ozone include admissions for all respiratory causes and
subcategories for COPD and pneumonia.
8-39
-------
Final Regulatory Impact Analysis
Table 8.6-2. Estimated Monetary Value in Reductions in Incidence of Health and
Welfare Effects (in millions of 2005$)
a,b
PM2 5-Related Health Effect
Premature Mortality -
Derived from
Epidemiology
Studies0'4
Premature mortality -
Derived from Expert
Elicitationc'd'e
Adult, age 30+ - ACS study
(Pope et al., 2002)
3% discount rate
7% discount rate
Adult, age 25+ - Six-cities study
(Laden et al., 2006)
3% discount rate
7% discount rate
Infant Mortality, <1 year -
(Woodruff etal. 1997)
3% discount rate
7% discount rate
Adult, age 25+ - Lower bound
(Expert K)
3% discount rate
7% discount rate
Adult, age 25+ - Upper bound
(Expert E)
3% discount rate
7% discount rate
Chronic bronchitis (adults, 26 and over)
Non-fatal acute myocardial infarctions
3% discount rate
7% discount rate
Hospital admissions for respiratory causes
Hospital admissions for cardiovascular causes
Emergency room visits for asthma
Acute bronchitis (children, age 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthma, 9-11)
2020
2030
Estimated Mean Value of Reductions
(5th and 95th %ile)
$1,000
($240 -$2,100)
$910
($220 -$1,900)
$2,300
($630 - $4,400)
$2,100
($570 - $3,900)
$3.2
($0.8 - $6.2)
$2.9
($0.8 - $5.6)
$540
($0 - $2,600)
$490
($0 - $2,400)
$5,600
($1,500 -$11,000)
$5,100
($1,400 -$10,000)
$70
($5.7 - $230)
$34
($10 -$72)
$33
($10 -$70)
$0.8
($0.4 -$1.2)
$2.2
($1.3 -$2.9)
$0.05
($0.03 - $0.08)
$0.2
($0 - $0.4)
$0.05
($0.02 - $0.09)
$0.06
$1,600
($370 - $3,200)
$1,400
($330 - $2,800)
$3,500
($970 - $6,700)
$3,200
($870 - $6,000)
$3.9
($1.0 -$7.7)
$3.5
($0.9 - $6.9)
$850
($0 -$4,100)
$760
($0 - $3,700)
$8,800
($2,400 - $17,000)
$8,000
($2, 100 -$16,000)
$110
($8.6 - $350)
$52
($15 -$110)
$51
($14 -$110)
$1.3
($0.6 -$1.8)
$3.5
($2.2 - $4.7)
$0.07
($0.04 -$0.1)
$0.2
($0 - $0.6)
$0.07
($0.03 -$0.1)
$0.08
8-40
-------
Cost-Benefit Analysis
Asthma exacerbations
Work loss days
Minor restricted-activity days (MRADs)
Recreational Visibility, 86 Class I areas
($0.02 -$0.1)
$0.1
($0.01 - $0.4)
$2.5
($2.2 - $2.8)
$2.9
($0.3 -$5.7)
$17
(na)f
($0.02 - $0.2)
$0.2
($0.02 - $0.5)
$3.4
($3.0 -$3.8)
$4.0
($0.4 - $7.7)
$7
(na)
Ozone-related Health Effect
Premature Mortality, All
ages - Derived from
NMMAPS
Premature Mortality, All
ages - Derived from Meta-
analyses
Bell et al., 2004
Bell etal., 2005
Ito et al., 2005
Levy et al., 2005
Premature Mortality - Assumption that association between
ozone and mortality is not causalf
Hospital admissions- respiratory causes (children, under 2;
adult, 65 and older)
Emergency room visit for asthma (all ages)
Minor restricted activity days (adults, age 18-65)
School absence days
Worker Productivity
$340
($86 - $680)
$1,100
($3 10 -$2,100)
$1,500
($450 - $2,800)
$1,600
($470 - $2,700)
$0
$8.7
($2.1 -$15)
$0.07
($0 - $0.2)
$19
($8.5 -$31)
$9.7
($3.4 - $17)
$3.1
(na)g
$590
($150 -$1,200)
$1,900
($530 - $3,600)
$2,600
($760 - $4,700)
$2,600
($800 - $4,700)
$0
$17
($3. 8 -$31)
$0.1
($0 - $0.3)
$27
($13 - $46)
$15
($5.4 - $27)
$5.1
(na)g
Monetary benefits are rounded to two significant digits for ease of presentation and computation. PM
and ozone benefits are nationwide.
Monetary benefits adjusted to account for growth in real GDP per capita between 1990 and the
analysis year (2020 or 2030)
Valuation assumes discounting over the SAB recommended 20 year segmented lag structure. Results
reflect the use of 3 percent and 7 percent discount rates consistent with EPA and OMB guidelines for preparing
economic analyses (EPA, 2000; OMB, 2003).
The valuation of adult premature mortality, derived either from the epidemiology literature or the
expert elicitation, is not additive. Rather, the valuations represent a range of possible mortality benefits.
e Based on effect estimates derived from the full-scale expert elicitation assessing the uncertainty in the
concentration-response function for PM-related premature mortality (lEc, 2006). The effect estimates of five of
the twelve experts included in the elicitation panel fall within the empirically-derived range provided by the ACS
and Six-Cities studies. One of the experts fall below this range and six of the experts are above this range.
Although the overall range across experts is summarized in this table, the full uncertainty in the estimates is
reflected by the results for the full set of 12 experts. The twelve experts' judgments as to the likely mean effect
estimate are not evenly distributed across the range illustrated by arraying the highest and lowest expert means.
f A recent report published by the National Research Council (NRC, 2008) recommended that EPA
"give little or no weight to the assumption that there is no causal association between estimated reductions in
premature mortality and reduced ozone exposure."
8-41
-------
Final Regulatory Impact Analysis
8 We are unable at this time to characterize the uncertainty in the estimate of benefits of worker
productivity and improvements in visibility at Class I areas. As such, we treat these benefits as fixed and add
them to all percentiles of the health benefits distribution.
Table 8.6-3 Total Monetized Benefits of the Final Small SI and Marine SI Engine Rule -
3% Discount Rate
Total Ozone and PM Benefits (billions, 2005$) -
PM Mortality Derived from the ACS and Six Cities Studies
2020
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Belletal.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$1.5 -$2.8
$2.3 -$3.6
$2.7 - $4.0
$2.7 - $4.0
$1.2 -$2.5
2030
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Bell et al.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$2.4 - $4.3
$3.7 -$5.6
$4.4 - $6.4
$4.4 - $6.4
$1.8 -$3.8
Total Ozone and PM Benefits (billions, 2005$) -
PM Mortality Derived from Expert Elicitation (Lowest and Highest Estimate)
2020
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Belletal.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$1.1 -$6.1
$1.8 -$6.9
$2.2 - $7.3
$2.3 - $7.4
$0.7 -$5.8
2030
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Bell et al.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$1.7 -$9.7
$3.0 -$11
$3.7 -$12
$3.7 -$12
$1.1 -$9.1
A recent report published by the National Research Council (NRC, 2008) recommended that EPA "give little or
no weight to the assumption that there is no causal association between estimated reductions in premature
mortality and reduced ozone exposure."
8-42
-------
Cost-Benefit Analysis
Table 8.6-4 Total Monetized Benefits of the Final Small SI and Marine SI Engine Rule -
7% Discount Rate
Total Ozone and PM Benefits (billions, 2005$) -
PM Mortality Derived from the ACS and Six Cities Studies
2020
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Belletal.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal,
2005
association is
Mean Total
Benefits
$1.4 -$2.6
$2.2 -$3.4
$2.6 -$3.7
$2.6 -$3.8
$1.1 -$2.2
2030
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Bell et al.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$2.2 - $4.0
$3. 5 -$5.3
$4.2 - $6.0
$4.3 -$6.1
$1.6 -$3.4
Total Ozone and PM Benefits (billions, 2005$) -
PM Mortality Derived from Expert Elicitation (Lowest and Highest Estimate)
2020
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Belletal.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$1.0 -$5.6
$1.8 -$6.4
$2.2 - $6.8
$2.2 - $6.8
$0.7 - $5.2
2030
Ozone
Mortality
Function
NMMAPS
Meta-analysis
Assumption that
not causal3
Reference
Bell et al.,
2004
Belletal.,
2005
Ito et al., 2005
Levy etal.,
2005
association is
Mean Total
Benefits
$1.6 -$8.8
$2.9 -$10
$3.6 -$11
$3.7 -$11
$1.0 -$8.2
A recent report published by the National Research Council (NRC, 2008) recommended that EPA "give little or
no weight to the assumption that there is no causal association between estimated reductions in premature
mortality and reduced ozone exposure."
8-43
-------
Final Regulatory Impact Analysis
Table 8.6-5. Results of Application of Expert Elicitation: Annual Reductions in
Premature Mortality in 2020 Associated with the Final Standards
Source of Mortality
Estimate
Pope et al. (2002)
Laden et al. (2006)
Expert A
Expert B
Expert C
Expert D
Expert E
Expert F
Expert G
Expert H
Expert I
Expert J
Expert K
Expert L
2020 Primary Option
5th Percentile
59
190
120
64
92
74
420
320
0
1
80
120
0
45
Mean
150
340
670
510
510
350
840
460
300
380
500
410
81
350
95th Percentile
240
500
1,200
1,100
1,100
580
1,300
670
550
870
900
900
380
690
Table 8.6-6. Results of Application of Expert Elicitation: Annual Reductions in
Premature Mortality in 2030 Associated with the Final Standards
Source of Mortality
Estimate
Pope et al. (2002)
Laden et al. (2006)
Expert A
Expert B
Expert C
Expert D
Expert E
Expert F
Expert G
Expert H
Expert I
Expert J
Expert K
Expert L
2030 Primary Option
5th Percentile
88
280
190
97
140
110
650
490
0
2
120
190
0
67
Mean
230
510
1,000
780
780
540
1,300
700
450
580
770
620
120
530
95th Percentile
360
740
1,900
1,700
1,700
890
1,900
1,000
840
1,300
1,400
1,400
580
1,100
8-44
-------
Cost-Benefit Analysis
8.7 Comparison of Costs and Benefits
In estimating the net benefits of the final standards, the appropriate cost measure is
'social costs.' Social costs represent the welfare costs of a rule to society. These costs do not
consider transfer payments (such as taxes) that are simply redistributions of wealth. Table
8.7-1 contains the estimates of monetized benefits and estimated social welfare costs for the
final rule and each of the final control programs. The annual social welfare costs of all
provisions of this final rule are described more fully in Chapter 9 of this RIA.
The results in Table 8.7-1 suggest that the 2020 monetized benefits of the final
standards are greater than the expected social welfare costs. Specifically, the annual benefits
of the total program will range between $1.2 to $4.0 billion annually in 2020 using a three
percent discount rate, or between $1.1 to $3.8 billion assuming a 7 percent discount rate,
compared to estimated social costs of approximately $210 million in that same year. These
benefits are expected to increase to between $1.8 and $6.4 billion annually in 2030 using a
three percent discount rate, or between $1.6 and $6.1 billion assuming a 7 percent discount
rate, while the social costs are estimated to be approximately $190 million. Though there are
a number of health and environmental effects associated with the final standards that we are
unable to quantify or monetize (see Table 8.4-1), the benefits of the final standards far
outweigh the projected costs. When we examine the benefit-to-cost comparison for the rule
standards separately, we also find that the benefits of the specific engine standards outweigh
their projected costs.
Using a conservative benefits estimate, the 2020 benefits outweigh the costs by a
factor of 5. Using the upper end of the benefits range, the benefits could outweigh the costs
by a factor of 19. Likewise, in 2030 benefits outweigh the costs by at least a factor of 8 and
could be as much as a factor of 34. Thus, even taking the most conservative benefits
assumptions, benefits of the final standards clearly outweigh the costs.
8-45
-------
Final Regulatory Impact Analysis
Table 8.7-1. Summary of Annual Benefits and Costs of the Final Standards3
(Millions of 2005 dollars)
Description
Estimated Social Costs
Small SI
Marine SI
Total Social Costs
Estimated Health Benefits of the Final Standards0' 'e'
Small SI
3 percent discount rate
7 percent discount rate
Marine SI
3 percent discount rate
7 percent discount rate
Total Benefits
3 percent discount rate
7 percent discount rate
Annual Net Benefits (Total Benefits - Total Costs)
3 percent discount rate
7 percent discount rate
2020
$163
$44
$210
$860 to $2,600
$790 to $2,500
$340 to $1,400
$310 to $1,300
$1,200 to $4,000
$1,100 to $3,800
$990 to $3,800
$890 to $3,600
2030
$185
$0.8
$190
$820 to $2,900
$710 to $2,800
$980 to $3,500
$890 to $3,300
$1,800 to $6,400
$1,600 to $6, 100
$1,600 to $6,200
$1,400 to $5,900
a All estimates represent annualized benefits and costs anticipated for the years 2020 and 2030. Totals
may not sum due to rounding.
b The calculation of annual costs does not require amortization of costs over time. Therefore, the
estimates of annual cost do not include a discount rate or rate of return assumption (see Chapter 9 of the RIA).
In Chapter 9, however, we use both a 3 percent and 7 percent social discount rate to calculate the net present
value of total social costs consistent with EPA and OMB guidelines for preparing economic analyses (US EPA,
2000 and OMB, 2003).
0 Total includes ozone and PM2 5 benefits. Range was developed by adding the estimate from the ozone
premature mortality function, including an assumption that the association is not causal, to PM2 5-related
premature mortality derived from the ACS (Pope et al., 2002) and Six Cities (Laden et al., 2006) studies.
d Annual benefits analysis results reflect the use of a 3 percent and 7 percent discount rate in the
valuation of premature mortality and nonfatal myocardial infarctions, consistent with EPA and OMB guidelines
for preparing economic analyses (US EPA, 2000 and OMB, 2003).
e Valuation of premature mortality based on long-term PM exposure assumes discounting over the SAB
recommended 20-year segmented lag structure described in the Regulatory Impact Analysis for the Final Clean
Air Interstate Rule (March, 2005).
f Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit
categories that have not been quantified and monetized are listed in Table 8.4-1.
8-46
-------
Cost-Benefit Analysis
Chapter 8 References
1 U.S. Environmental Protection Agency. March 2008. Final Ozone NAAQS Regulatory Impact Analysis.
Prepared by: Office of Air and Radiation, Office of Air Quality Planning and Standards.
2 U.S. Environmental Protection Agency. October 2006. Final Regulatory Impact Analysis (RIA)for the
Proposed National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and
Radiation. Available atHTTP://www.epa.gov/ttn/ecas/ria.html.
3 U.S. Environmental Protection Agency (2006) Air quality criteria for ozone and related photochemical
oxidants (second external review draft) Research Triangle Park, NC: National Center for Environmental
Assessment; report no. EPA/600R-05/004aB-cB, Sv.Available:
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=137307[March 2006]
4 National Research Council (NRC), 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. The National Academies Press: Washington, D.C.
5 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. The National Academies Press: Washington, D.C.
6 GeoLytics Inc. 2002. Geolytics CensusCD® 2000 Short Form Blocks. CD-ROM Release 1.0. GeoLytics,
Inc. East Brunswick, NJ. Available: http://www.geolytics.com/ [accessed 29 September 2004].
7 Woods & Poole Economics Inc. 2001. Population by Single Year of Age CD. CD-ROM. Woods &Poole
Economics, Inc. Washington, D.C.
8 U.S. Environmental Protection Agency (2006) Air quality criteria for ozone and related photochemical
oxidants (second external review draft) Research Triangle Park, NC: National Center for Environmental
Assessment; report no. EPA/600R-05/004aB-cB, Sv.Available:
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=137307[March 2006]
9 U.S. Environmental Protection Agency, 2004. Air Quality Criteria for Particulate Matter Volume II of II.
National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC EPA/600/P-99/002bF
10 World Health Organization (WHO). 2003. Health Aspects of Air Pollution with Particulate Matter, Ozone
and Nitrogen Dioxide: Report on a WHO Working Group. World Health Organization. Bonn, Germany.
EUR/03/5042688.
11 Anderson HR, Atkinson RW, Peacock JL, Marston L, Konstantinou K. 2004. Meta-analysis of time-series
studies and panel studies of Particulate Matter (PM) and Ozone (O3): Report of a WHO task group. Copenhagen,
Denmark: World Health Organization.
12 U.S. Environmental Protection Agency (2006) Air quality criteria for ozone and related photochemical
oxidants (second external review draft) Research Triangle Park, NC: National Center for Environmental
Assessment; report no. EPA/600R-05/004aB-cB, Sv.Available:
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=137307[March 2006]
8-47
-------
Final Regulatory Impact Analysis
13 Bell, M.L., et al. 2004. Ozone and short-term mortality in 95 US urban communities, 1987-2000. Jama, 2004.
292(19): p. 2372-8.
14 Bell, M.L., F. Dominici, and J.M. Samet. A meta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology, 2005. 16(4): p. 436-45.
15
Ito, K., S.F. De Leon, and M. Lippmann. Associations between ozone and daily mortality: analysis and meta-
analysis. Epidemiology, 2005. 16(4): p. 446-57.
16
Levy, J.I., S.M. Chemerynski, and J.A. Sarnat. 2005. Ozone exposure and mortality: an empiric bayes
metaregression analysis. Epidemiology, 2005. 16(4): p. 458-68.
17 Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. 2002. "Lung
Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Paniculate Air Pollution." Journal of the
American Medical Association 287:1132-1141.
18 Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. Reduction in Fine Paniculate Air Pollution
and Mortality. American Journal of Respiratory and Critical Care Medicine. 173: 667-672.
19 Industrial Economics, Incorporated (IEc). 2006. Expanded Expert Judgment Assessment of the
Concentration-Response Relationship Between PM2.5 Exposure and Mortality. Peer Review Draft. Prepared
for: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle
Park, NC. August.
20
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected Causes of
Postneonatal Infant Mortality and Paniculate Air Pollution in the United States." Environmental Health
Perspectives 105(6):608-612.
21 Abbey, D.E., B.L. Hwang, R.J. Burchette, T. Vancuren, and P.K. Mills. 1995. "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.
22
Peters, A., D.W. Dockery, J.E. Muller, and M.A. Mittleman. 2001. "Increased Paniculate Air Pollution and
the Triggering of Myocardial Infarction." Circulation 103:2810-2815.
23 Schwartz J. 1995. Short term fluctuations in air pollution and hospital admissions of the elderly for
respiratory disease. Thorax 50(5):531-538.
24 Schwartz J. 1994a. PM(10) Ozone, and Hospital Admissions For the Elderly in Minneapolis St Paul,
Minnesota. Arch Environ Health 49(5):366-374.
25 Schwartz J. 1994b. Air Pollution and Hospital Admissions For the Elderly in Detroit, Michigan. Am J Respir
Crit Care Med 150(3):648-655.
26 Moolgavkar SH, Luebeck EG, Anderson EL. 1997. Air pollution and hospital admissions for respiratory
causes in Minneapolis St. Paul and Birmingham. Epidemiology 8(4):364-370.
27 Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, et al. 2001. Association between
ozone and hospitalization for acute respiratory diseases in children less than 2 years of age. Am J Epidemiol
153(5):444-452.
8-48
-------
Cost-Benefit Analysis
Moolgavkar, S.H. 2003. "Air Pollution and Daily Deaths and Hospital Admissions in Los Angeles and Cook
Counties." In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report. Boston,
MA: Health Effects Institute.
29
Ito, K. 2003. "Associations of Paniculate Matter Components with Daily Mortality and Morbidity in Detroit,
Michigan." In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report. Health
Effects Institute, Boston, MA.
30 Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the Circulatory System in
Three U.S. Metropolitan Areas." Journal of the Air and Waste Management Association 50:1199-1206.
31 Sheppard, L. 2003. "Ambient Air Pollution and Nonelderly Asthma Hospital Admissions in Seattle,
Washington, 1987-1994." In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special
Report. Boston, MA: Health Effects Institute.
32 Jaffe DH, Singer ME, Rimm AA. 2003. Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. Environ Res 91(l):21-28.
33 Peel, J. L., P. E. Tolbert, M. Klein, et al. 2005. Ambient air pollution and respiratory emergency department
visits. Epidemiology. Vol. 16 (2): 164-74.
34 Wilson, A. M., C. P. Wake, T. Kelly, et al. 2005. Air pollution, weather, and respiratory emergency room
visits in two northern New England cities: an ecological time-series study. Environ Res. Vol. 97 (3): 312-21.
35 Norris, G., S.N. YoungPong, J.Q. Koenig, T.V. Larson, L. Sheppard, and J.W. Stout. 1999. "An Association
between Fine Particles and Asthma Emergency Department Visits for Children in Seattle." Environmental
Health Perspectives 107(6):489-493.
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.
37 Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory Health and PM10
Pollution: A Daily Time Series Analysis." American Review of Respiratory Diseases 144:668-674.
38 Schwartz, J., and L.M. Neas. 2000. "Fine Particles are More Strongly Associated than Coarse Particles with
Acute Respiratory Health Effects in Schoolchildren." Epidemiology 11:6-10.
39
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, andM. White. 2001. "Air Pollution and Exacerbation of
Asthma in African-American Children in Los Angeles." Epidemiology 12(2):200-208.
40
Vedal, S., J. Petkau, R. White, and J. Blair. 1998. "Acute Effects of Ambient Inhalable Particles in Asthmatic
and Nonasthmatic Children." American Journal of Respiratory and Critical Care Medicine 157(4): 1034-1043.
41
Ostro, B.D. 1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal of Environmental
Economics Management 14:87-98.
8-49
-------
Final Regulatory Impact Analysis
42 Gilliland FD, Berhane K, Rappaport EB, Thomas DC, Avol E, Gauderman WJ, et al. 2001. The effects of
ambient air pollution on school absenteeism due to respiratory illnesses. Epidemiology 12(l):43-54.
43 Chen L, Jennison BL, Yang W, Omaye ST. 2000. Elementary school absenteeism and air pollution. Inhal
Toxicoll2(ll):997-1016.
44
Ostro, B.D. and S. Rothschild. 1989. "Air Pollution and Acute Respiratory Morbidity: An Observational
Study of Multiple Pollutants." Environmental Research 50:238-247.
45 U.S. Science Advisory Board. 2004. Advisory Plans for Health Effects Analysis in the Analytical Plan for
EPA's Second Prospective Analysis -Benefits and Costs of the Clean Air Act, 1990—2020. EPA-SAB-
COUNCIL-ADV-04-004.
46 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press.
47 U.S. Environmental Protection Agency. 1999. The Benefits and Costs of the Clean Air Act, 1990-2010.
Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of Policy Analysis and Review,
Washington, DC, November; EPA report no. EPA-410-R-99-001.
48 U.S. Environmental Protection Agency, 2006. Air Quality Criteria for Ozone and Related Photochemical
Oxidants. National Center for Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC EPA 600/R-05/004aF
49 U.S. Environmental Protection Agency (2006) Air quality criteria for ozone and related photochemical
oxidants (second external review draft) Research Triangle Park, NC: National Center for Environmental
Assessment; report no. EPA/600R-05/004aB-cB, Sv.Available:
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=137307[March 2006]
50 National Research Council (NRC), 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. The National Academies Press: Washington, D.C.
51 U.S. Environmental Protection Agency. March 2008. Final Ozone NAAQS Regulatory Impact Analysis.
Prepared by: Office of Air and Radiation, Office of Air Quality Planning and Standards.
52 Mrozek, J.R., and L.O. Taylor. 2002. "What Determines the Value of Life? A Meta-Analysis." Journal of
Policy Analysis and Management 21(2):253-270.
53 Viscusi, V.K., and J.E. Aldy. 2003. "The Value of a Statistical Life: A Critical Review of Market Estimates
Throughout the World." Journal of Risk and Uncertainty 27(l):5-76.
54 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.
55 Cropper, M.L., and A.J. Krupnick. 1990. "The Social Costs of Chronic Heart and Lung Disease." Resources
for the Future. Washington, DC. Discussion Paper QE 89-16-REV.
56 Russell, M.W., D.M. Huse, S. Drowns, B.C. Hamel, and S.C. Hartz. 1998. "Direct Medical Costs of
Coronary Artery Disease in the United States." American Journal of Cardiology 81(9): 1110-1115.
8-50
-------
Cost-Benefit Analysis
57 Wittels, E.H., J.W. Hay, and A.M. Gotto, Jr. 1990. "Medical Costs of Coronary Artery Disease in the United
States." American Journal of Cardiology 65(7):432-440.
58 Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and Utilization
Project. Rockville, MD. http://www.ahrq.gov/HCUPnet/.
59 Smith, D.H., D.C. Malone, K.A. Lawson, LJ. Okamoto, C. Battista, and W.B. Saunders. 1997. "A National
Estimate of the Economic Costs of Asthma." American Journal of Respiratory and Critical Care Medicine
156(3 Pt l):787-793.
60 Stanford, R., T. McLaughlin, and LJ. Okamoto. 1999. "The Cost of Asthma in the Emergency Department
and Hospital." American Journal of Respiratory and Critical Care Medicine 160(1):211-215.
61 Industrial Economics, Incorporated (IEc). March 31, 1994. Memorandum to Jim DeMocker, Office of Air
and Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection Agency.
62 Rowe, R.D., and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits Analysis—
Executive Summary." Prepared by Energy and Resource Consultants, Inc. Report to the U.S. Environmental
Protection Agency, Office of Policy Analysis. EPA-230-09-86-018. Washington, DC.
63 Neumann, J.E., M.T. Dickie, and R.E. Unsworth. March 31, 1994. "Linkage Between Health Effects
Estimation and Morbidity Valuation in the Section 812 Analysis—Draft Valuation Document." Industrial
Economics Incorporated (IEc) Memorandum to Jim DeMocker, U.S. Environmental Protection Agency, Office
of Air and Radiation, Office of Policy Analysis and Review.
64 Tolley, G.S. et al. January 1986. Valuation oj'Reductions in Human Health Symptoms and Risks. University
of Chicago. Final Report for the U.S. Environmental Protection Agency.
65 Council of Economic Advisors. 2005. The Annual Report of the Council of Economic Advisors. In:
Economic Report of the President. Table B-60. U.S. Government Printing Office: Washington, DC.
66 Industrial Economics, Incorporated (IEc). 2006. Expanded Expert Judgment Assessment of the
Concentration-Response Relationship Between PM2.5 Exposure and Mortality. Peer Review Draft. Prepared
for: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle
Park, NC. August.
8-51
-------
Final Regulatory Impact Analysis
Appendix 8A: Sensitivity Analyses of Key Parameters in the Benefits Analysis
The primary analysis presented in Chapter 8 is based on our current interpretation of the
scientific and economic literature. That interpretation requires judgments regarding the best
available data, models, and modeling methodologies and the assumptions that are most
appropriate to adopt in the face of important uncertainties and resource limitations. The
majority of the analytical assumptions used to develop the primary estimates of benefits have
been used to support similar rulemakings and approved by EPA's Science Advisory Board
(SAB). Both EPA and the SAB recognize that data and modeling limitations as well as
simplifying assumptions can introduce significant uncertainty into the benefit results and that
alternative choices exist for some inputs to the analysis, such as the mortality C-R functions.
This appendix supplements our primary estimates of benefits with a series of sensitivity
calculations that use other sources of health effect estimates and valuation data for key
benefits categories. The supplemental estimates examine sensitivity to both valuation issues
and for physical effects issues. These supplemental estimates are not meant to be
comprehensive. Rather, they reflect some of the key issues identified by EPA or commenters
as likely to have a significant impact on total benefits. The individual adjustments in the
tables should not simply be added together because: 1) there may be overlap among the
alternative assumptions; and 2) the joint probability among certain sets of alternative
assumptions may be low.
8.A.I Premature Mortality - Alternative Threshold Analysis
To consider the impact of a threshold in the response function for the chronic mortality
endpoint, we have constructed a sensitivity analysis by assigning different cutpoints below
which changes in PM2 5 are assumed to have no impact on premature mortality. In applying
the cutpoints, we have adjusted the mortality function slopes accordingly.A Five cutpoints
(including the base case assumption) were included in the sensitivity analysis: (a) 14 |ig/m3
(assumes no impacts below the alternative annual NAAQS), (b) 12 |ig/m3 (c) 10 |ig/m3
(reflects comments from CASAC, 2005) l, (d) 7.5 |ig/m3 (reflects recommendations from
SAB-HES to consider estimating mortality benefits down to the lowest exposure levels
considered in the Pope 2002 study used as the basis for modeling chronic mortality) 2 and (e)
background or 3 |ig/m3 (reflects NRC recommendation to consider effects all the way to
background).3 We repeat this sensitivity analysis for the RIA of the final standards, the
results of which can be found in Table 8A-1.
Note that this analysis only adjusted the mortality slopes for the 10 ug/m3, 12 ug/m3 and 14 ug/m3 cutpoints
since the 7.5 ug/m3 and background cutpoints were at or below the lowest measured exposure levels reported in
the Pope et al. (2002) study for the combined exposure dataset.
8-52
-------
Cost-Benefit Analysis
Table 8A-1. PM-Related Mortality Benefits of the Final Standards: Cutpoint Sensitivity
Analysis Using the ACS Study (Pope et al., 2002)a
Certainty that Benefits
are At Least Specified
Value
More C
Benefits
as
^
Less C
Benefits
as
Certain that
Are at Least
Large
V
ert£
Are
Lar
7
dn that
at Least
ge
Level of
Assumed
Threshold
14 ug/m3
12 ug/m3
10 ug/m3 c
7.5 ug/m3
3 ug/m3
PM Mortality Incidence
2020
6
29
150
220
250
2030
7
40
230
340
380
a Note that this table only presents the effects of a cutpoint on PM-related mortality incidence.
b Alternative annual PM NAAQS.
'Primary threshold assumption based onCASAC (2005).85
d SAB-HES (2004)86
e NAS (2002)87
8.A.2 Premature Mortality - Alternative Lag Structures
Over the last ten years, there has been a continuing discussion and evolving advice regarding
the timing of changes in health effects following changes in ambient air pollution. It has been
hypothesized that some reductions in premature mortality from exposure to ambient PM2.5
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.
The SAB-HES has recognized this lack of direct evidence. However, in early advice, 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 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," (EPA-SAB-COUNCIL-ADV-00-001, 1999, p. 9).4
In recent advice, the SAB-HES suggests that appropriate lag structures may be developed
based on the distribution of cause-specific deaths within the overall all-cause estimate (EPA-
SAB-COUNCIL-ADV-04-002, 2004). They suggest that diseases with longer progressions
should be characterized by longer-term lag structures, while air pollution impacts occurring in
populations with existing disease may be characterized by shorter-term lags.
A key question is the distribution of causes of death within the relatively broad categories
analyzed in the long-term cohort studies. Although it may be reasonable to assume the
8-53
-------
Final Regulatory Impact Analysis
cessation lag for lung cancer deaths mirrors the long latency of the disease, it is not at all clear
what the appropriate lag structure should be for cardiopulmonary deaths, which include both
respiratory and cardiovascular causes. Some respiratory diseases may have a long period of
progression, while others, such as pneumonia, have a very short duration. In the case of
cardiovascular disease, there is an important question of whether air pollution is causing the
disease, which would imply a relatively long cessation lag, or whether air pollution is causing
premature death in individuals with preexisting heart disease, which would imply very short
cessation lags.
The SAB-HES provides several recommendations for future research that could support the
development of defensible lag structures, including using disease-specific lag models and
constructing a segmented lag distribution to combine differential lags across causes of death
(EPA-SAB-COUNCIL-ADV-04-002, 2004). The SAB-HES indicated support for using "a
Weibull distribution or a simpler distributional form made up of several segments to cover the
response mechanisms outlined above, given our lack of knowledge on the specific form of the
distributions," (EPA-SAB-COUNCIL-ADV-04-002, 2004, p. 24). However, they noted that
"an important question to be resolved is what the relative magnitudes of these segments
should be, and how many of the acute effects are assumed to be included in the cohort effect
estimate," (EPA-SAB-COUNCIL-ADV-04-002, 2004, p. 24-25). Since the publication of
that report in March 2004, EPA has sought additional clarification from this committee. In its
follow-up advice provided in December 2004, the SAB suggested that until additional
research has been completed, EPA should assume a segmented lag structure characterized by
30 percent of mortality reductions occurring in the first year, 50 percent occurring evenly over
years 2 to 5 after the reduction in PM2.5, and 20 percent occurring evenly over the years 6 to
20 after the reduction in PM2.5 (EPA-COUNCIL-LTR-05-001, 2004).5 The distribution of
deaths over the latency period is intended to reflect the contribution of short-term exposures
in the first year, cardiopulmonary deaths in the 2- to 5-year period, and long-term lung disease
and lung cancer in the 6- to 20-year period. Furthermore, in their advisory letter, the SAB-
HES recommended that EPA include sensitivity analyses on other possible lag structures. In
this appendix, we investigate the sensitivity of premature mortality-reduction related benefits
to alternative cessation lag structures, noting that ongoing and future research may result in
changes to the lag structure used for the primary analysis.
In previous advice from the SAB-HES, they recommended an analysis of 0-, 8-, and 15-year
lags, as well as variations on the proportions of mortality allocated to each segment in the
segmented lag structure (EPA-SAB-COUNCIL-ADV-00-001, 1999,
(EPA-COUNCIL-LTR-05-001, 2004). 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.
(1995)6 and Dockery et al. (1993)7 studies, respectively.8 However, neither the Pope et al. nor
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. analyses do not support or refute 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
FF 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.
8-54
-------
Cost-Benefit Analysis
assume that all premature mortalities occur at the end of the study periods (i.e., at 8 and 15
years).
In addition to the simple 8- and 15-year lags, we have added two additional sensitivity
analyses examining the impact of assuming different allocations of mortality to the segmented
lag of the type suggested by the SAB-HES. The first sensitivity analysis assumes that more of
the mortality impact is associated with chronic lung diseases or lung cancer and less with
acute cardiopulmonary causes. This illustrative lag structure is characterized by 20 percent of
mortality reductions occurring in the first year, 50 percent occurring evenly over years 2 to 5
after the reduction in PM2.5, and 30 percent occurring evenly over the years 6 to 20 after the
reduction in PM2.5. The second sensitivity analysis assumes the 5-year distributed lag
structure used in previous analyses, which is equivalent to a three-segment lag structure with
50 percent in the first 2-year segment, 50 percent in the second 3-year segment, and 0 percent
in the 6- to 20-year segment.
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. ACS impact
function) are presented in Table 8A-2. These estimates are based on the value of statistical
lives saved approach (i.e., $5.5 million per incidence) and are presented using both a 3 percent
and 7 percent discount rate over the lag period.
8-55
-------
-------
Cost-Benefit Analysis
Table 8A-2. Sensitivity of Benefits of Premature Mortality Reductions to Alternative Lag Assumptions
(Relative to Primary Benefits Estimates of the Final Standards)
Avoided Incidences Value
(ACS; Pope et al., 2002)a (million 2006$)b
Description of Sensitivity Analysis 2020 2030 2020 2030
Alternative Lag Structures for PM-Related Premature Mortality
30 percent of incidences occur in 1st
year, 50 percent in years 2 to 5, and
Primary 20 percent in years 6 to 20
3% Discount Rate 150 230 $1,000 $1,600
7% Discount Rate 150 230 $900 $1,400
None Incidences all occur in the first year 150 230 $1,100 $1,700
8-year Incidences all occur in the 8th year
3% Discount Rate 150 230 $910 $1,400
7% Discount Rate 150 230 $690 $1,100
15-year Incidences all occur in the 15th year
3% Discount Rate 150 230 $740 $1,100
7% Discount Rate 150 230 $430 $660
20 percent of incidences occur in 1st
Alternative year, 50 percent in years 2 to 5, and
Segmented 30 percent in years 6 to 20
3% Discount Rate 150 230 $1,100 $1,500
7% Discount Rate 150 230 $1,000 $1,300
50 percent of incidences occur in
5-Year years 1 and 2 and 50 percent in years
Distributed 2 to 5
3% Discount Rate 150 230 $980 $1,600
7% Discount Rate 150 230 $850 $1,500
a Incidences rounded to two significant digits.
Dollar values rounded to two significant digits. The alternative lag structure analysis presents benefits calculated
using both a 3 percent and 7 percent discount rate.
The results of the scaled alternative lag sensitivity analysis demonstrate that choice of lag
structure can have a large impact on benefits. Because of discounting of delayed benefits, the lag
structure may have a large downward impact on monetized benefits if an extreme assumption
that no effects occur until after 15 years is applied. However, for most reasonable distributed lag
structures, differences in the specific shape of the lag function have relatively small impacts on
overall benefits.
8.A.3 Visibility Benefits in Additional Class I Areas
The Chestnut and Rowe (1990)vm study from which the primary visibility valuation estimates are
derived only examined WTP for visibility changes in Class I areas (national parks and wilderness
areas) in the southeast, southwest, and California. To obtain estimates of WTP for visibility
changes at national parks and wilderness areas in the northeast, northwest, and central regions of
the U.S., we have to transfer WTP values from the studied regions. 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
8-59
-------
Final Regulatory Impact Analysis
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.
Based on this benefits transfer methodology (implemented within the preference calibration
framework discussed in Chapter 5 and Appendix I of the final PM NAAQS RIA), estimated
additional visibility benefits in the northwest, central, and northeastern U.S. are provided in
Table8.A-3.
Table 8.A-3: Monetary Benefits Associated with Improvements in Visibility in Additional Federal Class I
Areas in 2020 and 2030 (in millions of 2006$)a
Year
2020
2030
Northwesf
$3.9
$15
Central0
$1.7
$17
Northeast
$9.2
$12
Total
$15
$44
a All estimates are rounded to 2 significant digits. All rounding occurs after final summing of unrounded
estimates. As such, totals will not sum across columns
b Northwest Class I areas include Crater Lake, Mount Rainier, North Cascades, and Olympic national parks, and
Alpine Lakes, Diamond Peak, Eagle Cap, Gearhart Mountain, Glacier Peak, Goat Rocks, Hells Canyon,
Kalmiopsis, Mount Adams, Mount Hood, Mount Jefferson, Mount Washington, Mountain Lakes, Pasayten,
Strawberry Mountain, and Three Sisters wilderness areas.
0 Central Class I areas include Craters of the Moon, Glacier, Grand Teton, Theodore Roosevelt, Badlands, Wind
Cave, and Yellowstone national parks, and Anaconda-Pintlar, Bob Marshall, Bridger, Cabinet Mountains,
Fitzpatrick, Gates of the Mountain, Lostwood, Medicine Lake, Mission Mountain, North Absaroka, Red Rock
Lakes, Sawtooth, Scapegoat, Selway-Bitterroot, Teton, U.L. Bend, and Washakie wilderness areas.
d Northeast Class I areas include Acadia, Big Bend, Guadalupe Mountains, Isle Royale,
Voyageurs, and Boundary Waters Canoe national parks, and Brigantine, Caney Creek, Great
Gulf, Hercules-Glades, Lye Brook, Mingo, Moosehorn, Presidential Range-Dry Roosevelt
Campobello, Seney, Upper Buffalo, and Wichita Mountains wilderness areas.
8-60
-------
Cost-Benefit Analysis
Appendix 8B: Health-Based Cost-Effectiveness of Reductions in Ambient Os and PM2.s
Associated with the Final Small SI and Recreational Marine Engine Rule
8B.1 Introduction
Health-based cost-effectiveness analysis (CEA) and cost-utility analysis (CUA) have been used
to analyze numerous health interventions but have not been widely adopted as tools to analyze
environmental policies. Analyses of environmental regulations have typically used benefit-cost
analysis to characterize impacts on social welfare. Benefit-cost analyses allow for aggregation of
the benefits of reducing mortality risks with other monetized benefits of reducing air pollution,
including reduced risk of acute and chronic morbidity, and non-health benefits. One of the great
advantages of the benefit-cost paradigm is that a wide range of quantifiable benefits can be
compared to costs to evaluate the economic efficiency of particular actions. However, alternative
paradigms such as CEA and CUA analyses may also provide useful insights. CEA involves
estimation of the costs per unit of benefit (e.g., lives or life years saved). CUA is a special type
of CEA using preference-based measures of effectiveness, such as quality-adjusted life years
(QALYs).
QALYs were developed to evaluate the effectiveness of individual medical treatments, and EPA
is still evaluating the appropriate methods for CEA for environmental regulations. Agency
concerns with the standard QALY methodology include the treatment of people with fewer years
to live (the elderly); fairness to people with preexisting conditions that may lead to reduced life
expectancy and reduced quality of life; and how the analysis should best account for non-health
benefits.
The Office of Management and Budget (OMB) recently issued Circular A-4 guidance on
regulatory analyses, requiring federal agencies to "prepare a CEA for all major rulemakings for
which the primary benefits are improved public health and safety to the extent that a valid
effectiveness measure can be developed to represent expected health and safety outcomes."
Environmental quality improvements may have multiple health and ecological benefits, however,
making application of CEA more difficult and less straightforward.
The Institute of Medicine (a member institution of the National Academies of Science)
established the Committee to Evaluate Measures of Health Benefits for Environmental, Health,
and Safety Regulation to assess the scientific validity, ethical implications, and practical utility
of a wide range of effectiveness measures used or proposed in CEA. This committee prepared a
report titled "Valuing Health for Regulatory Cost-Effectiveness Analysis" which concluded that
CEA is a useful tool for assessing regulatory interventions to promote human health and safety,
although not sufficient for informed regulatory decisions (Miller, Robinson, and Lawrence,
2006). They emphasized the need for additional data and methodological improvements for CEA
analyses, and urged greater consistency in the reporting of assumptions, data elements, and
analytic methods. They also provided a number of recommendations for the conduct of
regulatory CEA analyses. EPA is evaluating these recommendations and will determine a
response for upcoming analyses.
8-61
-------
Final Regulatory Impact Analysis
CEA and CUA are most useful for comparing programs that have similar goals, for example,
alternative medical interventions or treatments that can save a life or cure a disease. They are less
readily applicable to programs with multiple categories of benefits, such as those reducing
ambient air pollution, because the cost-effectiveness calculation is based on the quantity of a
single benefit category. In other words, we cannot readily convert non-health benefits, such as
visibility improvements associated with reductions in PM2.5 or increases in worker productivity
associated with reductions in 63, to a health metric such as life years saved. For these reasons,
environmental economists prefer to present results in terms of monetary benefits and net
benefits.
However, QALY-based CUA has been widely adopted within the health economics literature
(Neumann, 2003; Gold et al., 1996) and in the analysis of public health interventions (US FDA,
2004). QALY-based analyses have not been as accepted in the environmental economics
literature because of concerns about the theoretical consistency of QALYs with individual
preferences (Hammitt, 2002), treatment of nonhuman health benefits, and a number of other
factors (Freeman, Hammitt, and De Civita, 2002). For environmental regulations, benefit-cost
analysis has been the preferred method of choosing among regulatory alternatives in terms of
economic efficiency. Recently several academic analyses have proposed the use of life years-
based benefit-cost or CEAs of air pollution regulations (Cohen, Hammitt, and Levy, 2003; Coyle
et al., 2003; Rabl, 2003; Carrothers, Evans, and Graham, 2002). In addition, the World Health
Organization has adopted the use of disability-adjusted life years, a variant on QALYs, to assess
the global burden of disease due to different causes, including environmental pollution (Murray
et al., 2002; de Hollander et al., 1999).
One of the ongoing controversies in health impact assessment regards whether reductions in
mortality risk should be reported and valued in terms of statistical lives saved or in terms of
statistical life years saved. Life years saved measures differentiate among premature mortalities
based on the remaining life expectancy of affected individuals. In general, under the life years
approach, older individuals will gain fewer life years than younger individuals for the same
reduction in mortality risk during a given time period, making interventions that benefit older
individuals seem less beneficial relative to similar interventions benefiting younger individuals.
A further complication in the debate is whether to apply quality adjustments to life years lost.
Under this approach, individuals with preexisting health conditions would have fewer QALYs
lost relative to healthy individuals for the same loss in life expectancy, making interventions that
primarily benefit individuals with poor health seem less beneficial than similar interventions
affecting primarily healthy individuals.
In this CEA, based largely on a report prepared under contract with Abt Associates,3 we
calculated both life years saved and statistical lives saved. Following the methodology used in
the CEAs for the PM and Os NAAQS RIAs, we did not assign QALY weights to the life years
saved - i.e., we calculated life years saved, rather than QALYs gained from mortality avoided.
Put another way, we assumed weights of 1.0 for all life years saved. Life years saved in the
future, however, were discounted to reflect people's time preference (i.e., a benefit received now
3 The full report prepared by Abt Associates is included in the docket for the Final Small SI and Recreational Marine
Engine Rule (EPA-HQ-OAR-2004-0008).
8-62
-------
Cost-Benefit Analysis
is worth more than the same benefit received in the future). We used discount rates of 3 percent
and 7 percent.
Where possible, benefits that could not be quantified in the denominator of our cost-effectiveness
ratios were monetized and subtracted from the cost of the regulation in the numerator. For
example, developing QALYs for acute health effects is problematic (Bala and Zarkin, 2000).
Therefore, rather than try to derive QALYs for the acute morbidity endpoints, we instead applied
valuation estimates and subtracted the total monetized value of all avoided acute morbidity
effects from the cost of the regulation, in the numerator of the cost-effectiveness ratios. The
monetized benefits of non-health improvements, where they were estimated, were similarly
subtracted from the cost of the regulation. Finally, although QALY estimates were derived for
the (PM2.5-related) chronic morbidity endpoints, the medical and opportunity costs associated
with these chronic illnesses were also subtracted from the cost of the regulation.
PM2.5-related benefits derive not only from avoided cases of premature mortality and acute
morbidity, but from avoided cases of chronic morbidity (chronic bronchitis and non-fatal
myocardial infarction) as well. In the CEAs for the PM and O3 NAAQS RIAs, EPA derived
QALYs for these two chronic morbidity endpoints (see, for example, Appendix G of the PM
NAAQS RIA, http://www.epa.gov/ttn/ecas/regdata/RIAs/Appendix%20G-
Health%20Based%20Cost%20Effectiveness%20Analysis.pdf) and used an alternative aggregate
effectiveness metric, Morbidity Inclusive Life Years (MILYs), to address some of the concerns
about aggregation of life extension and quality-of-life impacts. MILYs represent the sum of life
years gained due to reductions in premature mortality and the QALYs gained due to reductions
in chronic morbidity. This measure may be preferred to existing QALY aggregation approaches
because it does not devalue life extensions in individuals with preexisting illnesses that reduce
quality of life. However, the MILY measure is still based on life years and thus still inherently
gives more weight to interventions that reduce mortality and morbidity impacts for younger
populations with higher remaining life expectancy.
For this analysis, we present several metrics: lives saved, life years saved, cost of the regulation
(net of the monetized benefits not included in the denominator) per life saved and per life year
saved, and MILYs gained and the cost of the regulation (net of the monetized benefits not
included in the denominator) per MILY gained.
Note that, like future life years saved, future QALYs gained from avoided cases of chronic
bronchitis and myocardial infarction are discounted. All costs and monetized benefits are in
2005 dollars.
Monte Carlo simulation methods as implemented in the Crystal Ball™ software program were
used to propagate uncertainty in several of the model parameters throughout the analysis. In
particular, we incorporated uncertainty surrounding the coefficients in the concentration-
response (C-R) functions, the unit values for the various morbidity endpoints included in the
analysis, and the quality of life weights for the two chronic morbidity endpoints for which we
developed QALYs.
8-63
-------
Final Regulatory Impact Analysis
We characterized overall uncertainly in the results with 95 percent credible or confidence
intervals based on the Monte Carlo simulations. In addition, we examined the impacts on the cost
effectiveness metrics of changing key parameters and/or assumptions, including
• the discount rate (for the cost of the regulation in the numerator and future lives or life
years saved and QALYs gained in the denominator);
• the C-R functions for (Vrelated and PM2.5-related mortality ; and
• the life expectancies (and therefore years of potential life lost) of individuals who die as a
result of exposure to 63 (as explained in Section 8B.4 below).
The methodology presented in this appendix is not intended to stand as precedent either for
future air pollution regulations or for other EPA regulations where it may be inappropriate. It is
intended solely to demonstrate one particular approach to estimating the cost-effectiveness of
reductions in ambient PM2 5 and O?, in achieving improvements in public health. Reductions in
ambient PM2.5 and Os are estimated to have other health and environmental benefits that will not
be reflected in this CEA. Other EPA regulations affecting other aspects of environmental quality
and public health may require additional data and models that may preclude the development of
similar health-based CEAs. A number of additional methodological issues must be considered
when conducting CEAs for environmental policies, including treatment of non-health effects,
aggregation of acute and long-term health impacts, and aggregation of life extensions and
quality-of-life improvements in different populations. The appropriateness of health-based CEA
should be evaluated on a case-by-case basis subject to the availability of appropriate data and
models, among other factors.
The remainder of this appendix provides an overview of the methods used to derive the cost
effectiveness metrics developed for this CEA and presents the resulting metrics. Section 8B.2
provides an overview of effectiveness measures. Section 8B.3 discusses general issues in
constructing cost-effectiveness ratios. Section 8B.4 presents methods and results. Finally,
Section 8B.5 presents concluding remarks.
8B.2 Effectiveness Measures
For the purposes of CEA, we focus the effectiveness measures on the quantifiable health impacts
of the reductions in PM2.5 and Os estimated to occur as a result of this rule. If the main impact of
interest is reductions in mortality risk from air pollution, the effectiveness measures are relatively
straightforward to develop. Mortality impacts can be characterized similar to the benefits
analysis, by counting the number of premature deaths avoided, or can be characterized in terms
of increases in life expectancy or life years.4 Estimates of premature mortality have the benefit
4 Life expectancy is an ex ante concept, indicating the impact on an entire population's
expectation of the number of life years they have remaining, before knowing which individuals
will be affected. Life expectancy thus incorporates both the probability of an effect and the
impact of the effect if realized. Life years is an ex post concept, indicating the impact on
individuals who actually die from exposure to air pollution. Changes in population life
expectancy will always be substantially smaller than changes in life years per premature
mortality avoided, although the total life years gained in the population will be the same. This is
8-64
-------
Cost-Benefit Analysis
of being relatively simple to calculate, are consistent with the benefit-cost analysis, and do not
impose additional assumptions on the degree of life shortening. However, some have argued that
counts of premature deaths avoided are problematic because a gain in life of only a few months
would be considered equivalent to a gain of many life years, and the true effectiveness of an
intervention is the gain in life expectancy or life years (Rabl, 2003; Miller and Hurley, 2003).
Calculations of changes in life years and life expectancy can be accomplished using standard life
table methods (Miller and Hurley, 2003). However, the calculations require assumptions about
the baseline mortality risks for each age cohort affected by air pollution. A general assumption
may be that air pollution mortality risks affect the general mortality risk of the population in a
proportional manner. However, some concerns have been raised that air pollution affects mainly
those individuals with preexisting cardiovascular and respiratory disease, who may have reduced
life expectancy relative to the general population. This issue is explored in more detail below.
Air pollution is also associated with a number of significant chronic and acute morbidity
endpoints. Failure to consider these morbidity effects may understate the cost-effectiveness of
air pollution regulations or give too little weight to reductions in particular pollutants that have
large morbidity impacts but no effect on life expectancy. The QALY approach explicitly
incorporates morbidity impacts into measures of life years gained and is often used in health
economics to assess the cost-effectiveness of medical spending programs (Gold et al., 1996).
Using a QALY rating system, health quality ranges from 0 to 1, where 1 may represent full
health, 0 death, and some number in between (e.g., 0.8) an impaired condition. QALYs thus
measure morbidity as a reduction in quality of life over a period of life. QALYs assume that
duration and quality of life are equivalent, so that 1 year spent in perfect health is equivalent to 2
years spent with quality of life half that of perfect health. QALYs can be used to evaluate
environmental rules under certain circumstances, although some very strong assumptions
(detailed below) are associated with QALYs. The U.S. Public Health Service Panel on Cost
Effectiveness in Health and Medicine recommended using QALYs when evaluating medical and
public health programs that primarily reduce both mortality and morbidity (Gold et al., 1996).
Although there are significant non-health benefits associated with air pollution regulations, over
90 percent of quantifiable monetized benefits are health-related. Thus, it can be argued that
QALYs are more applicable for these types of regulations than for other environmental policies.
However, the value of non-health benefits should not be ignored. As discussed below, we have
chosen to subtract the value of non-health benefits from the costs in the numerator of the cost-
effectiveness ratio.
The use of QALYs is predicated on the assumptions embedded in the QALY analytical
framework. As noted in the QALY literature, QALYs are consistent with the utility theory that
underlies most of economics only if one imposes several restrictive assumptions, including
independence between longevity and quality of life in the utility function, risk neutrality with
respect to years of life (which implies that the utility function is linear), and constant
proportionality in trade-offs between quality and quantity of life (Pliskin, Shepard, and
Weinstein, 1980; Bleichrodt, Wakker, and Johannesson, 1996). To the extent that these
assumptions do not represent actual preferences, the QALY approach will not provide results
because life expectancy gains average expected life years gained over the entire population,
while life years gained measures life years gained only for those experiencing the life extension.
8-65
-------
Final Regulatory Impact Analysis
that are consistent with a benefit-cost analysis based on the Kaldor-Hicks criterion.5 Even if the
assumptions are reasonably consistent with reality, because QALYs represent an average
valuation of health states rather than the sum of societal WTP, there are no guarantees that the
option with the highest QALY per dollar of cost will satisfy the Kaldor-Hicks criterion (i.e.,
generate a potential Pareto improvement [Garber and Phelps, 1997]).
Benefit-cost analysis based on WTP is not without potentially troubling underlying structures as
well, incorporating ability to pay (and thus the potential for equity concerns) and the notion of
consumer sovereignty (which emphasizes wealth effects). Table 8B-1 compares the two
approaches across a number of parameters. For the most part, WTP allows parameters to be
determined empirically, while the QALY approach imposes some conditions a priori.
Table 8B-1. Comparison of QALY and WTP Approaches
Parameter QALY WTP
Risk aversion Risk neutral Empirically determined
Relation of duration and quality Independent Empirically determined
Proportionality of duration/ quality trade-off Constant Variable
Treatment of time/age in utility function Utility linear in time Empirically determined
Preferences Community/Individual Individual
Source of preference data Stated Revealed and stated
Treatment of income and prices Not explicitly considered Constrains choices
8B.3 Construction of Cost-Effectiveness Ratios: General Issues
8B. 3.1 Dealing with Morbidity Health Effects and Non-health Effects
Health effects from exposure to PM2.5 and O3 air pollution encompass a wide array of chronic
and acute conditions in addition to premature mortality. EPA's Ozone and PM Criteria
Documents outline numerous health effects known or suspected to be linked to exposure to
ambient ozone and PM (US EPA, 2006; US EPA, 2005; Anderson et al., 2004). Although
chronic conditions and premature mortality generally account for the majority of monetized
benefits, acute symptoms can affect a broad population or sensitive populations (e.g., asthma-
related emergency room visits among asthmatics). In addition, reductions in air pollution may
result in a broad set of non-health environmental benefits, including improved worker
productivity, improved visibility in national parks, increased agricultural and forestry yields,
reduced acid damage to buildings, and a host of other impacts. Lives saved, life years saved, and
5 The Kaldor-Hicks efficiency criterion requires that the "winners" in a particular case be
potentially able to compensate the "losers" such that total societal welfare improves. In this
case, it is sufficient that total benefits exceed total costs of the regulation. This is also known as
a potential Pareto improvement, because gains could be allocated such that at least one person in
society would be better off while no one would be worse off.
8-66
-------
Cost-Benefit Analysis
QALYs gained address only health impacts, and the OMB guidance notes that "where regulation
may yield several different beneficial outcomes, a cost-effectiveness comparison becomes more
difficult to interpret because there is more than one measure of effectiveness to incorporate in the
analysis."
With regard to acute health impacts, Bala and Zarkin (2000) suggest that QALYs are not
appropriate for valuing acute symptoms, because of problems with both measuring utility for
acute health states and applying QALYs in a linear fashion to very short duration health states.
Johnson and Lievense (2000) suggest using conjoint analysis to get healthy-utility time
equivalences that can be compared across acute effects, but it is not clear how these can be
combined with QALYs for chronic effects and loss of life expectancy. There is also a class of
effects that EPA has traditionally treated as acute, such as hospital admissions, which may also
result in a loss of quality of life for a period of time following the effect. For example, life after
asthma hospitalization has been estimated with a utility weight of 0.93 (Bell et al., 2001;
Kerridge, Glasziou, andHillman, 1995).
How should these effects be combined with QALYs for chronic and mortality effects? One
method would be to convert the acute effects to QALYs; however, as noted above, there are
problems with the linearity assumption (i.e., if a year with asthma symptoms is equivalent to 0.7
year without asthma symptoms, then 1 day without asthma symptoms is equivalent to 0.0019
QALY gained). This is troubling from both a conceptual basis and a presentation basis. An
alternative approach is simply to treat acute health effects like non-health benefits and subtract
the dollar value (based on WTP or COI) from compliance costs in the CEA.
To address the issues of incorporating acute morbidity and non-health benefits, OMB suggests
that agencies "subtract the monetary estimate of the ancillary benefits from the gross cost
estimate to yield an estimated net cost." As with benefit-cost analysis, any unquantified benefits
and/or costs should be noted and an indication of how they might affect the cost-effectiveness
ratio should be described. We followed this recommended "net cost" approach, specifically in
netting out the benefits of health improvements other than reduced mortality and improved
quality of life from avoided chronic illness - in particular, the monetized benefits of acute
morbidity avoided, the medical and opportunity costs ("cost of illness") of avoided chronic
illness, and the benefits of non-health improvements, including increases in worker productivity
associated with reductions in Os and visibility improvements at national parks associated with
reductions in PM2.5 (see Chapter 8 for more details on these benefit categories).
8B.3.2 Should Life Years Gained Be Adjusted for Initial Health Status?
The methods outlined below in Section 8B.4 provide estimates of the total number of life years
gained in a population, regardless of the quality of those life years, or equivalently, assuming that
all life years gained are in perfect health. In some CEAs (Cohen, Hammitt, and Levy, 2003;
Coyle et al., 2003), analysts have adjusted the number of life years gained to reflect the fact that
1) the general public is not in perfect health and thus "healthy" life years are less than total life
years gained and 2) those affected by air pollution may be in a worse health state than the general
population and therefore will not gain as many "healthy" life years adjusted for quality, from an
air pollution reduction. This adjustment, which converts life years gained into QALYs, raises a
number of serious ethical issues. Proponents of QALYs have promoted the nondiscriminatory
8-67
-------
Final Regulatory Impact Analysis
nature of QALYs in evaluating improvements in quality of life (e.g., an improvement from a
score of 0.2 to 0.4 is equivalent to an improvement from 0.8 to 1.0), so the starting health status
does not affect the evaluation of interventions that improve quality of life. However, for life-
extending interventions, the gains in QALYs will be directly proportional to the baseline health
state (e.g., an individual with a 30-year life expectancy and a starting health status of 0.5 will
gain exactly half the QALYs of an individual with the same life expectancy and a starting health
status of 1.0 for a similar life-extending intervention). This is troubling because it imposes an
additional penalty for those already suffering from disabling conditions. Brock (2002) notes that
"the problem of disability discrimination represents a deep and unresolved problem for resource
priori tization."
OMB (2003) has recognized this issue in their Circular A-4 guidance, which includes the
following statement:
When CEA is performed in specific rulemaking contexts, you should be prepared to
make appropriate adjustments to ensure fair treatment of all segments of the
population. Fairness is important in the choice and execution of effectiveness
measures. For example, if QALYs are used to evaluate a life saving rule aimed at a
population that happens to experience a high rate of disability (i.e., where the rule is
not designed to affect the disability), the number of life years saved should not
necessarily be diminished simply because the rule saves the lives of people with life-
shortening disabilities. Both analytic simplicity and fairness suggest that the
estimated number of life years saved for the disabled population should be based on
average life expectancy information for the relevant age cohorts. More generally,
when numeric adjustments are made for life expectancy or quality of life, analysts
should prefer use of population averages rather than information derived from
subgroups dominated by a particular demographic or income group, (p. 13)
This suggests two adjustments to the standard QALY methodology: one adjusting the relevant
life expectancy of the affected population, and the other affecting the baseline quality of life for
the affected population.
In addition to the issue of fairness, potential measurement issues are specific to the air pollution
context that might argue for caution in applying quality-of-life adjustments to life years gained
due to air pollution reductions. A number of epidemiological and toxicological studies link
exposure to air pollution with chronic diseases, such as CB and atherosclerosis (Abbey et al.,
1995; Schwartz, 1993; Suwa et al., 2002). If these same individuals with chronic disease caused
by exposure to air pollution are then at increased risk of premature death from air pollution, there
is an important dimension of "double jeopardy" involved in determining the correct baseline for
assessing QALYs lost to air pollution (see Singer et al. [1995] for a broader discussion of the
double-jeopardy argument).
Analyses estimating mortality from acute exposures that ignore the effects of long-term exposure
on morbidity may understate the health impacts of reducing air pollution. Individuals exposed to
chronically elevated levels of air pollution may realize an increased risk of death and chronic
disease throughout life. If at some age they contract heart (or some other chronic) disease as a
result of the exposure to air pollution, they will from that point forward have both reduced life
8-68
-------
Cost-Benefit Analysis
expectancy and reduced quality of life. The benefit to that individual from reducing lifetime
exposure to air pollution would be the increase in life expectancy plus the increase in quality of
life over the full period of increased life expectancy. If the QALY loss is determined based on
the underlying chronic condition and life expectancy without regard to the fact that the person
would never have been in that state without long-term exposure to elevated air pollution, then the
person is placed in double jeopardy. In other words, air pollution has placed more people in the
susceptible pool, but then we penalize those people in evaluating policies by treating their
subsequent deaths as less valuable, adding insult to injury, and potentially downplaying the
importance of life expectancy losses due to air pollution. If the risk of chronic disease and risk
of death are considered together, then there is no conceptual problem with measuring QALYs,
but this has not been the case in recent applications of QALYs to air pollution (Carrothers,
Evans, and Graham, 2002; Coyle et al., 2003). The use of QALYs thus highlights the need for a
better understanding of the relationship between chronic disease and long-term exposure and
suggests that analyses need to consider morbidity and mortality jointly, rather than treating each
as a separate endpoint (this is an issue for current benefit-cost approaches as well).
Because of the fairness and measurement concerns discussed above, for the purposes of this
analysis, we do not reduce the number of life years gained to reflect any differences in
underlying health status that might reduce quality of life in remaining years. Thus, we maintain
the assumption that all direct gains in life years resulting from mortality risk reductions will be
assigned a weight of 1.0. The U.S. Public Health Service Panel on Cost Effectiveness in Health
and Medicine recommends that "since lives saved or extended by an intervention will not be in
perfect health, a saved life year will count as less than 1 full QALY" (Gold et al., 1996).
However, for the purposes of this analysis, we propose an alternative to the traditional aggregate
QALY metric that keeps separate quality adjustments to life expectancy and gains in life
expectancy. As such, we do not make any adjustments to life years gained to reflect the less than
perfect health of the general population. Gains in quality of life will be addressed as they accrue
because of reductions in the incidence of chronic diseases. This is an explicit equity choice in
the treatment of issues associated with quality-of-life adjustments for increases in life expectancy
that still capitalizes on the ability of QALYs to capture both morbidity and mortality impacts in a
single effectiveness measure.
8B. 3.3 Constructing Cost-Effectiveness Ratios
Construction of cost-effectiveness ratios requires estimates of effectiveness (in this case
measured by lives saved, life years gained, or MILYs gained) in the denominator and estimates
of costs in the numerator. The estimate of costs in the numerator should include both the direct
costs of the controls necessary to achieve the reduction in ambient concentrations of the air
pollutant and the avoided costs (cost savings) associated with the reductions in morbidity (Gold
et al., 1996). In general, because reductions in air pollution do not require direct actions by the
affected populations, there are no specific costs to affected individuals (aside from the overall
increases in prices that might be expected to occur as control costs are passed on by affected
industries). Likewise, because individuals do not engage in any specific actions to realize the
health benefit of the pollution reduction, there are no decreases in utility (as might occur from a
medical intervention) that need to be adjusted for in the denominator. Thus, the elements of the
numerator are direct costs of controls minus the avoided costs of illness (COI) associated with
chronic illnesses. In addition, as noted above, to account for the value of reductions in acute
8-69
-------
Final Regulatory Impact Analysis
health impacts and non-health benefits, we netted out the monetized value of these benefits from
the numerator to yield a "net cost" estimate.
The denominators of the cost-effectiveness ratios we calculated are either lives saved, life years
saved, or MTLYs gained. For the MILY aggregate effectiveness measure, the denominator is
simply the sum of life years gained from increased life expectancy and QALYs gained from the
reductions in incidence of chronic illnesses associated with PM2.5 - chronic bronchitis (CB) and
nonfatal acute myocardial infarction (AMI).
8B.4 Cost Effectiveness Metrics
In this section we describe the development of cost effectiveness metrics. To generate health
outcomes, we used the same framework as for the benefit-cost analysis described in Chapter 8.
For convenience, we summarize the basic methodologies here. For more details, see Chapter 8
and the Environmental Benefits Mapping and Analysis Program (BenMAP) user's manual
(http://www.epa.gov/ttn/ecas/benmodels.html).
BenMAP uses health impact functions to generate changes in the incidence of health effects.
Health impact functions are derived from the C-R functions reported in the epidemiology
literature. A standard health impact function has four components: an effect estimate from a
particular epidemiological study, a baseline incidence rate for the health effect (obtained from
either the epidemiology study or a source of public health statistics, such as CDC), the affected
population, and the estimated change in the relevant pollutant summary measure.
A typical health impact function might look like this:
where yo is the baseline incidence, equal to the baseline incidence rate times the potentially
affected population; P is the effect estimate; Ax is the estimated change in the pollutant (e.g.,
PM2.5 or 63) and Ay is the estimated change in incidence of the health effect (e.g., the number of
deaths avoided) associated with the change in the pollutant, Ax. There are other functional
forms, but the basic elements remain the same.
8B.4.1 Reductions in 0 ^-Related Premature Deaths
To calculate (Vrelated life years saved under the Final Small SI and Recreational Marine Engine
Rule (hereafter, Final SSI & RME Rule), we first calculated the numbers of Os-related statistical
lives saved within 5-year age groups, using BenMAP. (For more details on the calculation of
statistical lives saved using BenMAP, see Chapter 8 or the BenMAP user's manual
(http://www.epa.gov/ttn/ecas/benmodels.html). We used two studies used in the benefit analysis
for the Final SSI & RME Rule RIA - Bell et al. (2004) and Levy et al. (2005). Both studies
report estimated C-R functions of the association between premature mortality and short-term
exposures to ambient 63. Bell et al. (2004) is a multi-city study of 95 cities, and as such may
avoid the potential for publication bias that may be inherent in single-city studies or meta-
8-70
-------
Cost-Benefit Analysis
analyses of single-city studies. This study provides the lowest estimate of Os-related premature
deaths among the mortality studies included in the Final SSI & RME Rule RIA benefit analysis.
An upper bound estimate of O3-related premature deaths in the Final SSI & RME Rule RIA
benefit analysis was provided by Levy et al. (2005). More extensive discussions of these studies
are given in Chapter 8.
We checked to confirm that the total number of Os-related statistical lives saved, summed across
all age groups, equals the corresponding number calculated in the Final SSI & RME Rule RIA
benefit analysis. Age group-specific (Vrelated premature deaths avoided under the Final SSI &
RME Rule in 2020 and in 2030 are given in Table 8B-2.
8-71
-------
Final Regulatory Impact Analysis
Table 8B-2. Estimated Reduction in Incidence of O3-Related Premature Mortality Under the Final
SSI & RME Rule in 2020 and 2030
Age
Interval
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+
Total:
Reduction in O3-Related Premature Mortality
(95% Cl)*
2020
Bell et al. (2004)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-1)
0
(0-1)
1
(0-2)
1
(0-2)
3
(1-5)
3
(1-5)
6
(2-9)
4
(1-7)
7
(2-12)
5
(2-8)
15
(5 - 25)
46
(15-77)
Levy et al. (2005)
1
(1-1)
0
(0-1)
0
(0-1)
1
(0-1)
1
(1-2)
2
(1-2)
2
(1-2)
3
(2-3)
2
(2-3)
5
(3-6)
5
(4-7)
13
(9-18)
13
(9-17)
25
(17-33)
20
(14-26)
31
(22-41)
20
(14-26)
65
(45 - 85)
210
(140-270)
2030
Bell et al. (2004)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
0
(0-0)
1
(0-1)
1
(0-1)
1
(0-2)
1
(0-2)
3
(1-6)
4
(1-6)
9
(3-16)
8
(3-13)
15
(5 - 26)
9
(3-15)
24
(8 - 40)
77
(25- 130)
Levy et al. (2005)
1
(1-2)
1
(0-1)
1
(0-1)
1
(1-1)
2
(1-2)
2
(2-3)
2
(1-3)
4
(3-5)
3
(2-4)
7
(5-9)
5
(4-7)
16
(1 1 - 20)
16
(11 -21)
42
(29 - 54)
35
(24 - 46)
67
(46 - 88)
39
(27-51)
100
(72- 140)
350
(240 - 460)
*95 percent confidence or credible intervals (CIs) are based on the uncertainty about the
coefficient in the mortality C-R functions. All estimates rounded to two significant figures.
8-72
-------
Cost-Benefit Analysis
8B.4.2 Life Years Saved as a Result of Reductions in Os-Related Mortality Risk
The number of life years saved depends not only on the number of statistical lives saved, but also
on the life expectancies associated with those statistical lives. As was pointed out in the CEAs
for the PM and Os NAAQS RIAs, age-specific life expectancies for the general population are
calculated from mortality rates for the general population, and these reflect the prevalence of
chronic disease, which shortens life expectancies. The only reason one might use lower life
expectancies than those for the general population in the CEA for the Final SSI & RME Rule
RIA is if the population at risk from exposure to Os was limited solely or disproportionately to
individuals with preexisting chronic illness, whose life expectancies were, on average, shorter
than those of the general population (unless all of those individuals had preexisting chronic
illness because of long-term exposure to 63).
It is reasonable to assume that someone who dies from exposure to an air pollutant is already in a
compromised state. However, there are both acute and chronic compromised states. If an
individual has an acute illness (e.g., pneumonia) that puts him at risk of mortality when exposed
to a high concentration of an air pollutant, then in the absence of that high concentration he could
be expected to recover from the illness and go on to live the expected number of years for
someone his age - i.e., he would have the age-specific life expectancy of the general population.
If an individual has a chronic illness that makes him vulnerable to a high concentration of an air
pollutant, then an important question is whether or not he would have had that chronic illness if
he had not been exposed over the long term to high levels of the air pollutant.
We can categorize individuals who are at risk of dying because of exposure to an air pollutant
into three groups:
• those who are vulnerable because of a preexisting acute condition;
• those who are vulnerable because of a preexisting chronic condition that they would not
have had, had they not been exposed over the long term to high levels of the air pollutant;
and
• those who are vulnerable because of a preexisting chronic condition that they would have
had even in the absence of long term exposure to high levels of the air pollutant.
The age-specific life expectancies of the general population should apply to the first two groups,
and the age-specific life expectancies of the subpopulation with the relevant chronic condition(s)
should apply to the third group. If we knew the proportions of people who die from exposure to
63 who are in each group, and the life expectancies of people in the third group, we could
calculate the number of life years saved as follows:
Total life years saved = ^M. * (pu * LEt + p2i *LEt + p3i *LE*)
i
where
Mi denotes the number of O3-related deaths of individuals age /',
7 denotes the general population life expectancy for age /',
8-73
-------
Final Regulatory Impact Analysis
i denotes the life expectancy for age /' of the subpopulation with the relevant chronic
condition(s) - i.e., the third group;
pn denotes the proportion of the M CVrelated deaths that are in the first group;
P2i denotes the proportion of the M CVrelated deaths that are in the second group; and
PSJ denotes the proportion of the M CVrelated deaths that are in the third group.
Unlike for PM2.5 (discussed below), we currently lack information that would allow us to
estimate the relevant proportions necessary to estimate the set of life expectancies that would be
appropriate to apply to (Vrelated deaths. Although there is substantial evidence linking
premature mortality to short-term exposures to Os, there is currently not similar evidence for
long-term exposures. We therefore do not know if the second group above is relevant in the case
of (Vrelated mortality. Nor do we know what proportion of CVrelated deaths can be attributed
to preexisting acute conditions (the first group) versus preexisting chronic conditions that these
individuals would have had even in the absence of long term exposure to Os (the third group).
Because we currently lack the necessary information to determine the appropriate set of life
expectancies to use in calculating life years saved associated with CVrelated premature mortality
avoided, we calculated life years saved based on four different underlying assumptions:
• A lower bound assumption of zero life years saved, based on the hypothesis that the
observed statistical association between premature mortality and short-term exposures to
Os is not actually a causal relationship;
• An upper bound assumption that an O3-related premature death of an individual of a
given age will result in a loss of life years equal to the life expectancy in the general
population of that age;
• Two intermediate assumptions: That the proportions of CVrelated premature deaths in
the three groups delineated above (pn,p2t, and/?^) are such that, on average, the age-
specific life expectancies among people who die (Vrelated premature deaths are those
of
o people with severe preexisting chronic conditions, whose life expectancies are
substantially shorter than those of the general population; and
o people with preexisting chronic conditions of a range of severities, whose life
expectancies are somewhat shorter than those of the general population.
Life years saved based on the upper bound assumption were calculated from age-specific
mortality probabilities for the general population taken from the Centers for Disease Control
(CDC) National Vital Statistics Reports, Vol. 56, No. 9, December 28, 2007, Table 1. Life table
for the total population: United States, 2004.6 We used a simplified method of calculating life
expectancies from these age-specific mortality probabilities that yielded life expectancies that
were close to the life expectancies derived using the more complicated method employed by the
6 http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_09.pdf
8-74
-------
Cost-Benefit Analysis
CDC.7 In particular, starting with a cohort of size 1,000,000 at birth, we calculated the life-years
lived between ages x and (x+1), for x = 0, 1,2, ..., 99, using the age-specific mortality
probabilities taken from the CDC Vital Statistics Report (see above) and assuming that all deaths
that occurred between ages x and (x+1) occurred midway through the year (i.e., we assigned 0.5
life-year to each year of death). The life expectancy at age n was then calculated as the sum of
the life-years lived from age n through age 100 divided by the cohort size at age n. The life
expectancy at age n is the number of life years lost due to an O3-related premature mortality of
an individual age n.
To estimate life years saved under the two intermediate assumptions about the life years lost as a
result of O3-related premature mortality, we turned to the epidemiological evidence of a
statistically significant association between short-term exposures to 63 and respiratory hospital
admissions. This evidence suggests that these short-term exposures may exacerbate respiratory
conditions that were preexisting. It is reasonable to suppose that some of these hospitalizations
for respiratory illnesses on days of relatively high O3 concentrations might result in death. It
may also be the case that some individuals who did not go to the hospital might also die. We
therefore looked for information on life expectancies of people with chronic respiratory
conditions.
While there is information readily available in vital statistics sources on rates of death from
chronic respiratory diseases, there is not similarly available information on rates of death among
that subpopulation who suffer from those diseases. It is the latter rate - the rate of death among
that subpopulation who suffers from those diseases - that is of interest.
A recent study of people with and without chronic obstructive pulmonary disease (COPD)
provided data from which we were able to construct estimates of the mortality rates of interest.
Mannino et al. (2006) followed a cohort of 15,440 subjects ages 43 to 66 for up to 11 years. The
cohort subjects were selected from the larger cohort of the Atherosclerosis Risk in Communities
(ARIC) study, which selected its subjects from the population of four U.S. communities by
probability sampling.8 The subjects in the Mannino study were limited to the ARIC participants
who provided baseline information on respiratory symptoms and diagnoses, who underwent
pulmonary function testing, and for whom follow-up data were available.
Using a modification of the criteria developed by the Global Initiative on Obstructive Lung
Disease (GOLD), Mannino et al. (2006) classified the study subjects into COPD severity groups
(or stages), with GOLD stage 0 (presence of respiratory symptoms in the absence of any lung
function abnormality) being the least severe COPD group, and GOLD stages 3 and 4 being the
most severe. The unadjusted death rates of the study participants (taken from Table 1 of
Mannino et al., 2006), ratios of (unadjusted) death rates, and hazard ratios, based on Cox
7 We calculated life expectancies from the mortality probabilities rather than using the life
expectancies given in the CDC table because we were going to also calculate life expectancies
for the subpopulations with severe COPD and with "average" COPD by adjusting the age-
specific mortality probabilities and then calculating life expectancies using these adjusted
probabilities.
8 In one of the four communities probability sampling was used to select African-Americans
only.
8-75
-------
Final Regulatory Impact Analysis
proportional hazard regressions, which took into account several covariates (including, among
others, age, sex, race, smoking status, and education level) are shown in the table below. In
addition, the right-most column of the table below shows the proportion of COPD subjects in the
study in each GOLD category.
Table 8B-3. Death Rates and Hazard Ratios for Subjects with Varying Degrees of Severity of
COPD (from Mannino et al., 2006)
GOLD* Category
GOLD 3 or 4
GOLD 2
GOLD 1
GOLDO
Restricted
Normal
Total
N
271
1,484
1,679
2,244
1,101
8,661
1 5,440
Deaths
92
232
137
204
150
427
1,242
(%)
33.9%
15.6%
8.2%
9.1%
13.6%
4.9%
8.0%
Person-
Years
2,143
12,852
15,031
20,191
9,644
79,317
139,178
Death Rate
per 1,000
Person-Years
42.9
18.1
9.1
10.1
15.6
5.4
8.9
Ratio of Death Rate
to Death Rate for
Normal Population
7.97
3.35
1.69
1.88
2.89
1.00
Hazard
Ratio**
5.7
2.4
1.4
1.5
2.3
1.0
Proportion of COPD
Subjects in GOLD
Category
4.77%
26.14%
29.57%
39.52%
'Global Inititative on Obstructive Lung Disease (GOLD) guidelines for the staging of COPD severity.
"See Mannino et al. (2006), p. 117.
The ratios of unadjusted death rates are somewhat larger than the corresponding hazard ratios
because these ratios were not adjusted for age. COPD is a progressive disease, so it would be
expected that the proportion of older individuals would increase as the stages (and severity)
increased, and this was indeed the case in the Mannino study. The hazard ratios, being based on
regressions that took age into account, avoid this problem. We therefore used the hazard ratios
to derive age-specific mortality rates for individuals with (1) severe COPD and (2) COPD of
"average" severity. In particular, to derive age-specific mortality probabilities for the
subpopulation with severe COPD, we multiplied each age-specific mortality probability for the
general population by 5.7 (the hazard ratio for GOLD 3 or 4); to derive age-specific mortality
probabilities for the subpopulation with "average" COPD, we multiplied each age-specific
mortality probability for the general population by a weighted average of the GOLD category-
specific hazard ratios, where the weight for a GOLD category was the proportion of COPD
subjects in that GOLD category (given in the right-most column of Table 1 above). The
weighted average hazard ratio was 1.906. Age-specific life expectancies were then derived for
the severe COPD and "average" COPD subpopulations using these adjusted mortality
probabilities and the method for calculating life expectancies described above.
Once an appropriate set of life expectancies has been determined (e.g., life expectancies for the
general population or life expectancies for a subpopulation with severe COPD), these then
provide the number of life years lost for an individual who dies at a given age. This information
can then be combined with the estimated number of Os-related premature deaths at each age
calculated with BenMAP (see previous subsection). Because BenMAP calculates numbers of
premature deaths avoided within age intervals, we can either allocate the premature deaths
avoided within an age interval uniformly to the ages within the interval or, alternatively, we can
calculate average life expectancies for the age intervals. We illustrate the first approach in
8-76
-------
Cost-Benefit Analysis
calculating Os-related life years saved and the second approach in calculating PM2 5-related life
years saved (see Section 8B.4.4).
Total O3-related life years gained was calculated as the sum of life years gained at each age:
N
Total life years gained = = ^ LEt x Mi
where LEt is the remaining life expectancy for age /', Mt is the number of premature deaths
avoided among individuals age /', and TV is the oldest age considered.
For the purposes of determining cost effectiveness, it is also necessary to consider the time-
dependent nature of the gains in life years. Standard economic theory suggests that benefits
occurring in future years should be discounted relative to benefits occurring in the present. OMB
and EPA guidance suggest discount rates of three and seven percent. Selection of a 3 percent
discount rate is also consistent with recommendations from the U.S. Public Health Service Panel
on Cost Effectiveness in Health and Medicine (Gold et al., 1996).
Discounted total life years gained is calculated as follows:
fLE
Discounted LY '= e dt
Jo
where r is the discount rate, t indicates time, and LE is the life expectancy at the time when the
premature death would have occurred. Because (Vrelated premature mortality is associated
only with short-term exposures, all Os-related premature deaths are assumed to occur in the year
of exposure. We therefore did not discount Os-related premature deaths avoided.
Undiscounted age-specific life expectancies, and age-specific life expectancies using discount
rates of 3 percent and 7 percent are given for the general population, the subpopulation of
individuals with severe COPD, and the subpopulation of individuals with COPD of average
severity in Tables 8B-4, 8B-5, and 8B-6, respectively. The O3-related (discounted) life years
saved, based on each of the two (Vmortality studies and each of the assumptions about relevant
life expectancies, are given, using 3 percent and 7 percent discount rates, in Tables 8B-7 and 8B-
8, respectively. The Os-related (discounted) life years saved, under the first assumption - that
the observed statistical association between premature mortality and short-term exposures to O3
is not actually a causal relationship - is zero in all cases (i.e., regardless of the mortality study
used and the scenario considered), and is therefore not shown in these Tables.
8-77
-------
Final Regulatory Impact Analysis
Table 8B-4. Undiscounted and Discounted Age-Specific Life Expectancies for the General
Population
Age at
Beginning
of Year
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Mortality
Probability*
0.006799
0.000483
0.000297
0.000224
0.000188
0.000171
0.000161
0.000151
0.000136
0.000119
0.000106
0.000112
0.000149
0.000227
0.000337
0.000460
0.000579
0.000684
0.000763
0.000819
0.000873
0.000926
0.000960
0.000972
0.000969
0.000960
0.000954
0.000952
0.000958
0.000973
0.000994
0.001023
0.001063
0.001119
0.001192
0.001275
0.001373
0.001493
0.001634
0.001788
0.001945
0.002107
0.002287
0.002494
0.002727
0.002982
0.003246
Cohort Size
1 ,000,000
993,201
992,721
992,427
992,204
992,017
991 ,847
991 ,688
991 ,538
991 ,403
991 ,286
991,180
991 ,070
990,922
990,697
990,363
989,907
989,334
988,657
987,902
987,093
986,231
985,318
984,372
983,415
982,462
981,519
980,583
979,650
978,712
977,759
976,787
975,788
974,750
973,659
972,499
971 ,259
969,925
968,477
966,895
965,166
963,290
961 ,260
959,062
956,670
954,061
951,216
Deaths in
Year
6,799
480
295
222
187
170
159
149
135
118
105
111
148
225
333
456
573
677
755
809
862
913
946
957
953
943
936
933
939
952
972
999
1,038
1,091
1,160
1,240
1,334
1,448
1,582
1,729
1,877
2,029
2,198
2,392
2,609
2,845
3,088
Life-Years
in Year
996,600
992,961
992,574
992,315
992,111
991 ,932
991 ,768
991,613
991,471
991 ,345
991 ,233
991,125
990,996
990,809
990,530
990,135
989,621
988,996
988,280
987,498
986,662
985,775
984,845
983,893
982,939
981,991
981,051
980,117
979,181
978,235
977,273
976,287
975,269
974,205
973,079
971 ,879
970,592
969,201
967,686
966,031
964,228
962,275
960,161
957,866
955,366
952,639
949,672
Age-Specific
Life
Expectancy
77.8
77.3
76.4
75.4
74.4
73.4
72.4
71.4
70.4
69.5
68.5
67.5
66.5
65.5
64.5
63.5
62.5
61.6
60.6
59.7
58.7
57.8
56.8
55.9
54.9
54.0
53.0
52.1
51.1
50.2
49.2
48.3
47.3
46.4
45.4
44.5
43.5
42.6
41.7
40.7
39.8
38.9
38.0
37.0
36.1
35.2
34.3
3% Discounted
Remaining Life
Expectancy
30.9
30.8
30.7
30.6
30.5
30.4
30.3
30.2
30.1
29.9
29.8
29.7
29.5
29.4
29.2
29.1
28.9
28.8
28.6
28.4
28.3
28.1
27.9
27.8
27.6
27.4
27.2
27.0
26.8
26.5
26.3
26.1
25.9
25.6
25.4
25.1
24.9
24.6
24.3
24.0
23.7
23.5
23.2
22.8
22.5
22.2
21.9
7% Discounted
Remaining Life
Expectancy
15.2
15.2
15.2
15.2
15.2
15.2
15.2
15.2
15.2
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.0
15.0
15.0
15.0
15.0
15.0
14.9
14.9
14.9
14.9
14.8
14.8
14.8
14.7
14.7
14.7
14.6
14.6
14.5
14.5
14.4
14.4
14.3
14.3
14.2
14.1
14.0
14.0
13.9
13.8
8-78
-------
Cost-Benefit Analysis
Table 8B-4. Undiscounted and Discounted Age-Specific Life Expectancies for the General
Population (cont'd)
Age at
Beginning
of Year
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Mortality
Probability*
0.003520
0.003799
0.004088
0.004404
0.004750
0.005113
0.005488
0.005879
0.006295
0.006754
0.007280
0.007903
0.008633
0.009493
0.010449
0.011447
0.012428
0.013408
0.014473
0.015703
0.017081
0.018623
0.020322
0.022104
0.024023
0.026216
0.028745
0.031561
0.034427
0.037379
0.040756
0.044764
0.049395
0.054471
0.059772
0.065438
0.071598
0.078516
0.085898
0.093895
0.102542
0.111875
0.121928
0.132733
0.144318
0.156707
0.169922
0.183975
0.198875
0.214620
0.231201
0.248600
0.266786
1 .000000
Cohort Size
948,129
944,792
941 ,203
937,355
933,227
928,794
924,045
918,974
913,571
907,820
901 ,689
895,125
888,051
880,384
872,027
862,915
853,037
842,435
831,140
819,111
806,249
792,477
777,719
761,915
745,073
727,174
708,110
687,756
666,050
643,120
619,080
593,849
567,266
539,246
509,873
479,397
448,026
415,949
383,290
350,366
317,468
284,915
253,040
222,187
192,695
164,886
139,047
115,420
94,186
75,454
59,260
45,559
34,233
25,100
Deaths in
Year
3,337
3,589
3,848
4,128
4,433
4,749
5,071
5,403
5,751
6,131
6,564
7,074
7,667
8,357
9,112
9,878
10,601
1 1 ,295
12,029
12,863
13,771
14,758
15,805
16,841
17,899
19,064
20,355
21 ,706
22,930
24,039
25,231
26,583
28,020
29,373
30,476
31,371
32,078
32,659
32,924
32,897
32,554
31 ,875
30,853
29,492
27,809
25,839
23,627
21 ,234
18,731
16,194
13,701
1 1 ,326
9,133
25,100
Life-Years
in Year
946,460
942,997
939,279
935,291
931,010
926,419
921,510
916,273
910,696
904,755
898,407
891 ,588
884,217
876,205
867,471
857,976
847,736
836,788
825,126
812,680
799,363
785,098
769,817
753,494
736,124
717,642
697,933
676,903
654,585
631,100
606,465
580,558
553,256
524,560
494,635
463,712
431 ,987
399,619
366,828
333,917
301,192
268,977
237,613
207,441
178,791
151,967
127,234
104,803
84,820
67,357
52,410
39,896
29,667
12,550
Age-Specific
Life
Expectancy
33.5
32.6
31.7
30.8
30.0
29.1
28.2
27.4
26.6
25.7
24.9
24.1
23.3
22.5
21.7
20.9
20.1
19.4
18.6
17.9
17.2
16.5
15.8
15.1
14.4
13.7
13.1
12.5
11.9
11.3
10.7
10.1
9.6
9.0
8.5
8.1
7.6
7.1
6.7
6.3
5.9
5.5
5.1
4.8
4.4
4.1
3.7
3.4
3.0
2.7
2.3
1.8
1.2
0.5
3% Discounted
Remaining Life
Expectancy
21.6
21.2
20.9
20.5
20.2
19.8
19.4
19.1
18.7
18.3
17.9
17.5
17.1
16.7
16.2
15.8
15.4
15.0
14.5
14.1
13.7
13.2
12.8
12.3
11.9
11.5
11.0
10.6
10.2
9.7
9.3
8.9
8.5
8.1
7.7
7.3
6.9
6.5
6.2
5.8
5.5
5.1
4.8
4.5
4.2
3.9
3.6
3.3
3.0
2.6
2.2
1.8
1.2
0.5
7% Discounted
Remaining Life
Expectancy
13.7
13.6
13.5
13.4
13.3
13.2
13.0
12.9
12.7
12.6
12.4
12.3
12.1
11.9
11.8
11.6
11.4
11.2
11.0
10.7
10.5
10.3
10.0
9.8
9.5
9.3
9.0
8.7
8.4
8.2
7.9
7.6
7.3
7.0
6.7
6.4
6.1
5.8
5.6
5.3
5.0
4.7
4.5
4.2
3.9
3.7
3.4
3.1
2.8
2.5
2.2
1.8
1.2
0.5
'Mortality probabilities for the general population taken from Table 1
United States, 2004. CDC National Vital Statistics Reports, Vol. 56,
http://www.cdc.qov/nchs/data/nvsr/nvsr56/nvsr56 09.pdf
Life table for the total population:
No. 9, December 28, 2007
8-79
-------
Final Regulatory Impact Analysis
Table 8B-5. Undiscounted and Discounted Age-Specific Life Expectancies for the Subpopulation
with Severe COPD
Age at
Beginning
of Year
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Mortality
Probability*
0.038755
0.002752
0.001692
0.001277
0.001074
0.000978
0.000916
0.000859
0.000777
0.000677
0.000606
0.000636
0.000850
0.001295
0.001918
0.002625
0.003301
0.003901
0.004351
0.004671
0.004976
0.005278
0.005472
0.005542
0.005522
0.005470
0.005436
0.005425
0.005461
0.005547
0.005668
0.005830
0.006061
0.006380
0.006792
0.007269
0.007827
0.008510
0.009312
0.010191
0.011084
0.012008
0.013035
0.014215
0.015546
0.016996
0.018503
Cohort Size
1 ,000,000
961 ,245
958,599
956,977
955,755
954,729
953,796
952,923
952,104
951 ,365
950,721
950,145
949,540
948,733
947,505
945,687
943,205
940,092
936,424
932,350
927,995
923,377
918,504
913,478
908,415
903,399
898,458
893,573
888,726
883,873
878,970
873,988
868,893
863,626
858,117
852,289
846,094
839,472
832,328
824,577
816,174
807,128
797,436
787,041
775,854
763,792
750,811
Deaths in
Year
38,755
2,646
1,622
1,222
1,026
933
873
819
739
644
576
605
807
1,229
1,818
2,482
3,113
3,667
4,075
4,355
4,618
4,873
5,026
5,063
5,016
4,942
4,884
4,847
4,853
4,903
4,982
5,095
5,266
5,510
5,828
6,195
6,622
7,144
7,750
8,403
9,047
9,692
10,395
11,187
12,061
12,981
13,892
Life-Years
in Year
980,622
959,922
957,788
956,366
955,242
954,263
953,359
952,513
951 ,734
951 ,043
950,433
949,842
949,137
948,119
946,596
944,446
941 ,648
938,258
934,387
930,172
925,686
920,941
915,991
910,947
905,907
900,928
896,016
891,150
886,300
881 ,422
876,479
871 ,440
866,260
860,872
855,203
849,191
842,783
835,900
828,452
820,376
811,651
802,282
792,238
781 ,447
769,823
757,301
743,865
Age-Specific
Life
Expectancy
54.5
55.7
54.9
53.9
53.0
52.1
51.1
50.2
49.2
48.2
47.3
46.3
45.3
44.4
43.4
42.5
41.6
40.8
39.9
39.1
38.3
37.5
36.7
35.9
35.1
34.2
33.4
32.6
31.8
31.0
30.1
29.3
28.5
27.6
26.8
26.0
25.2
24.4
23.6
22.8
22.0
21.3
20.5
19.8
19.1
18.4
17.7
3% Discounted
Remaining Life
Expectancy
27.5
27.7
27.5
27.4
27.2
27.0
26.8
26.5
26.3
26.1
25.8
25.6
25.3
25.1
24.8
24.6
24.3
24.0
23.8
23.5
23.3
23.0
22.7
22.4
22.2
21.9
21.6
21.2
20.9
20.6
20.2
19.9
19.5
19.2
18.8
18.4
18.0
17.6
17.2
16.8
16.4
16.0
15.6
15.2
14.8
14.4
14.0
7% Discounted
Remaining Life
Expectancy
14.9
14.9
14.9
14.9
14.9
14.8
14.8
14.8
14.7
14.7
14.7
14.6
14.6
14.5
14.5
14.4
14.4
14.3
14.3
14.2
14.1
14.1
14.0
13.9
13.9
13.8
13.7
13.6
13.5
13.4
13.3
13.2
13.1
12.9
12.8
12.7
12.5
12.3
12.2
12.0
11.8
11.7
11.5
11.3
11.1
10.9
10.7
8-80
-------
Cost-Benefit Analysis
Table 8B-5. Undiscounted and Discounted Age-Specific Life Expectancies for the Subpopulation
with Severe COPD (cont'd)
Age at
Beginning
of Year
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
Mortality
Probability*
0.020061
0.021652
0.023303
0.025103
0.027075
0.029144
0.031280
0.033512
0.035880
0.038497
0.041497
0.045046
0.049211
0.054108
0.059560
0.065249
0.070839
0.076425
0.082495
0.089507
0.097361
0.106149
0.115833
0.125993
0.136933
0.149433
0.163847
0.179896
0.196231
0.213062
0.232309
0.255152
0.281552
0.310486
0.340699
0.372994
0.408108
0.447543
0.489619
0.535199
0.584489
0.637689
0.694992
0.756579
0.822612
0.893232
Cohort Size
736,919
722,135
706,500
690,036
672,714
654,500
635,425
615,549
594,921
573,575
551,494
528,609
504,797
479,956
453,986
426,947
399,089
370,818
342,478
314,225
286,100
258,245
230,833
204,094
178,380
153,954
130,948
109,493
89,795
72,175
56,797
43,603
32,477
23,333
16,089
10,607
6,651
3,937
2,175
1,110
516
214
78
24
6
1
Deaths in
Year
14,784
15,636
16,464
17,322
18,214
19,075
19,876
20,628
21 ,346
22,081
22,885
23,812
24,842
25,969
27,040
27,858
28,271
28,340
28,253
28,125
27,855
27,413
26,738
25,714
24,426
23,006
21 ,455
19,697
17,621
15,378
13,194
11,125
9,144
7,245
5,481
3,956
2,714
1,762
1,065
594
302
137
54
18
5
0
Life-Years
in Year
729,527
714,317
698,268
681 ,375
663,607
644,963
625,487
605,235
584,248
562,535
540,052
516,703
492,376
466,971
440,467
413,018
384,953
356,648
328,352
300,163
272,173
244,539
217,463
191,237
166,167
142,451
120,220
99,644
80,985
64,486
50,200
38,040
27,905
19,711
13,348
8,629
5,294
3,056
1,642
813
365
146
51
15
3
0
Age-Specific
Life
Expectancy
17.0
16.3
15.7
15.0
14.4
13.8
13.2
12.6
12.0
11.5
10.9
10.3
9.8
9.3
8.8
8.3
7.9
7.4
7.0
6.6
6.2
5.8
5.4
5.1
4.7
4.4
4.1
3.8
3.5
3.2
3.0
2.7
2.5
2.3
2.1
1.9
1.7
1.5
1.4
1.3
1.1
1.0
0.9
0.8
0.6
0.0
3% Discounted
Remaining Life
Expectancy
13.6
13.1
12.7
12.3
11.9
11.5
11.1
10.7
10.3
9.9
9.5
9.0
8.6
8.2
7.9
7.5
7.1
6.8
6.4
6.1
5.7
5.4
5.1
4.8
4.5
4.2
3.9
3.6
3.4
3.1
2.9
2.7
2.4
2.2
2.0
1.9
1.7
1.5
1.4
1.3
1.1
1.0
0.9
0.8
0.6
0.0
7% Discounted
Remaining Life
Expectancy
10.4
10.2
10.0
9.8
9.5
9.3
9.0
8.8
8.5
8.2
8.0
7.7
7.4
7.1
6.9
6.6
6.3
6.0
5.8
5.5
5.2
5.0
4.7
4.4
4.2
3.9
3.7
3.5
3.2
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.7
1.5
1.4
1.2
1.1
1.0
0.9
0.8
0.6
0.0
*Mortality probabilities derived from mortality probabilities for the general population by multiplying
by the hazard ratio (5.7) for GOLD 3 or 4, from Mannino et al. (2006).
8-81
-------
Final Regulatory Impact Analysis
Table 8B-6. Undiscounted and Discounted Age-Specific Life Expectancies for the Subpopulation
with COPD of Average Severity
Age at
Beginning
of Year
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Mortality
Probability*
0.01 2960
0.000920
0.000566
0.000427
0.000359
0.000327
0.000306
0.000287
0.000260
0.000226
0.000203
0.000213
0.000284
0.000433
0.000642
0.000878
0.001104
0.001304
0.001455
0.001562
0.001664
0.001765
0.001830
0.001853
0.001846
0.001829
0.001818
0.001814
0.001826
0.001855
0.001896
0.001949
0.002027
0.002133
0.002271
0.002431
0.002617
0.002846
0.003114
0.003408
0.003707
0.004016
0.004359
0.004753
0.005199
0.005683
0.006187
Cohort Size
1 ,000,000
987,040
986,132
985,574
985,153
984,799
984,477
984,176
983,893
983,638
983,415
983,216
983,006
982,727
982,302
981 ,671
980,810
979,727
978,449
977,025
975,499
973,876
972,157
970,378
968,580
966,792
965,023
963,269
961 ,521
959,766
957,985
956,169
954,305
952,371
950,339
948,181
945,876
943,400
940,716
937,786
934,591
931,127
927,388
923,345
918,956
914,179
908,983
Deaths in
Year
12,960
908
558
421
354
322
301
283
256
223
199
209
279
426
630
862
1,083
1,278
1,424
1,526
1,623
1,719
1,779
1,798
1,788
1,769
1,754
1,747
1,756
1,780
1,816
1,864
1,934
2,032
2,158
2,305
2,476
2,685
2,929
3,196
3,464
3,739
4,042
4,389
4,777
5,196
5,624
Life-Years
in Year
993,520
986,586
985,853
985,363
984,976
984,638
984,326
984,034
983,765
983,526
983,315
983,111
982,867
982,514
981 ,986
981,241
980,268
979,088
977,737
976,262
974,688
973,017
971 ,268
969,479
967,686
965,907
964,146
962,395
960,643
958,875
957,077
955,237
953,338
951 ,355
949,260
947,028
944,638
942,058
939,251
936,189
932,859
929,257
925,366
921,151
916,567
911,581
906,171
Age-Specific
Life
Expectancy
69.6
69.5
68.6
67.6
66.7
65.7
64.7
63.7
62.7
61.8
60.8
59.8
58.8
57.8
56.8
55.9
54.9
54.0
53.1
52.1
51.2
50.3
49.4
48.5
47.6
46.7
45.7
44.8
43.9
43.0
42.1
41.1
40.2
39.3
38.4
37.5
36.6
35.7
34.8
33.9
33.0
32.1
31.2
30.4
29.5
28.7
27.8
3% Discounted
Remaining Life
Expectancy
29.9
29.9
29.8
29.7
29.5
29.4
29.3
29.1
29.0
28.8
28.6
28.5
28.3
28.1
27.9
27.8
27.6
27.4
27.2
27.0
26.8
26.6
26.4
26.1
25.9
25.7
25.5
25.2
25.0
24.7
24.4
24.2
23.9
23.6
23.3
23.0
22.7
22.4
22.0
21.7
21.4
21.0
20.7
20.3
20.0
19.6
19.2
7% Discounted
Remaining Life
Expectancy
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.1
15.0
15.0
15.0
15.0
15.0
14.9
14.9
14.9
14.9
14.8
14.8
14.8
14.7
14.7
14.7
14.6
14.6
14.5
14.5
14.5
14.4
14.3
14.3
14.2
14.1
14.1
14.0
13.9
13.8
13.7
13.6
13.5
13.4
13.3
13.2
13.1
13.0
8-82
-------
Cost-Benefit Analysis
Table 8B-6. Undiscounted and Discounted Age-Specific Life Expectancies for the Subpopulation
with COPD of Average Severity (cont'd)
Age at
Beginning
of Year
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Mortality
Probability*
0.006709
0.007241
0.007793
0.008395
0.009054
0.009746
0.010460
0.011207
0.011999
0.012874
0.013877
0.015064
0.016456
0.018094
0.019917
0.021820
0.023689
0.025557
0.027587
0.029932
0.032558
0.035497
0.038735
0.0421 33
0.045791
0.049971
0.054791
0.060158
0.065621
0.071249
0.077685
0.085324
0.0941 52
0.103828
0.113932
0.124731
0.136473
0.149661
0.163731
0.178974
0.195456
0.213247
0.232409
0.253004
0.275086
0.298702
0.323890
0.350677
0.379078
0.409089
0.440695
0.473858
0.508523
1 .000000
Cohort Size
903,359
897,298
890,801
883,860
876,440
868,505
860,040
851 ,044
841 ,507
831,410
820,707
809,318
797,127
784,009
769,823
754,490
738,028
720,545
702,130
682,760
662,324
640,760
618,015
594,076
569,046
542,989
515,855
487,591
458,258
428,187
397,679
366,785
335,489
303,902
272,349
241,319
211,219
182,394
155,096
129,702
106,489
85,675
67,405
51 ,740
38,649
28,017
1 9,649
13,285
8,626
5,356
3,165
1,770
931
458
Deaths in
Year
6,060
6,497
6,942
7,420
7,935
8,464
8,996
9,537
10,097
10,703
1 1 ,389
12,191
13,118
14,186
15,333
1 6,463
1 7,483
18,415
19,370
20,436
21,564
22,745
23,939
25,030
26,057
27,134
28,264
29,333
30,071
30,508
30,894
31,296
31,587
31,554
31,029
30,100
28,826
27,297
25,394
23,213
20,814
18,270
15,666
13,090
10,632
8,369
6,364
4,659
3,270
2,191
1,395
839
474
458
Life-Years
in Year
900,329
894,050
887,331
880,150
872,472
864,273
855,542
846,276
836,458
826,058
815,012
803,222
790,568
776,916
762,157
746,259
729,286
71 1 ,337
692,445
672,542
651 ,542
629,388
606,046
581,561
556,017
529,422
501 ,723
472,924
443,223
412,933
382,232
351,137
319,696
288,125
256,834
226,269
196,806
168,745
142,399
118,096
96,082
76,540
59,572
45,194
33,333
23,833
1 6,467
10,955
6,991
4,261
2,468
1,351
695
229
Age-Specific
Life
Expectancy
27.0
26.2
25.3
24.5
23.7
23.0
22.2
21.4
20.6
19.9
19.1
18.4
17.7
17.0
16.3
15.6
14.9
14.3
13.6
13.0
12.4
11.8
11.2
10.6
10.1
9.6
9.0
8.5
8.0
7.6
7.1
6.7
6.2
5.8
5.5
5.1
4.8
4.4
4.1
3.8
3.5
3.3
3.0
2.8
2.6
2.4
2.2
2.0
1.9
1.7
1.5
1.3
1.0
0.5
3% Discounted
Remaining Life
Expectancy
18.9
18.5
18.1
17.7
17.3
16.9
16.5
16.1
15.7
15.3
14.8
14.4
14.0
13.5
13.1
12.7
12.3
11.8
11.4
11.0
10.5
10.1
9.7
9.3
8.9
8.4
8.0
7.6
7.3
6.9
6.5
6.1
5.8
5.4
5.1
4.8
4.5
4.2
3.9
3.7
3.4
3.2
3.0
2.7
2.5
2.4
2.2
2.0
1.8
1.7
1.5
1.3
1.0
0.5
7% Discounted
Remaining Life
Expectancy
12.8
12.7
12.5
12.4
12.2
12.1
11.9
11.7
11.5
11.3
11.1
10.9
10.7
10.4
10.2
10.0
9.7
9.5
9.2
8.9
8.7
8.4
8.1
7.8
7.6
7.3
7.0
6.7
6.4
6.1
5.8
5.6
5.3
5.0
4.7
4.5
4.2
4.0
3.7
3.5
3.3
3.1
2.8
2.7
2.5
2.3
2.1
2.0
1.8
1.6
1.5
1.3
1.0
0.5
'Mortality probabilities derived from mortality probabilities for the general population (see Table 2) by multiplying
by the weighted average of hazard ratios for the GOLD severity categories (1.906) from Mannino et al. (2006).
8-83
-------
Final Regulatory Impact Analysis
Table 8B-7. Estimated O3-Related Life Years Saved in 2020 and in 2030 Under the Final SSI &
RME Rule, Using a 3 Percent Discount Rate
Estimated O3-Related Life Years Saved
(95% Cl)*
Mortality Study:
Assuming Life Expectancies of the
General Population
Assuming Life Expectancies of the Sub-
Population with COPD of Average
Severity
Assuming Life Expectancies of the Sub-
Population with Severe COPD
2020
Bell et al (2004)
500
(150-800)
360
(120-600)
190
(60 - 320)
Levy et al. (2005)
2,200
(1,500-2,900)
1,700
(1,200-2,200)
1,000
(700-1,300)
2030
Bell et al (2004)
700
(250-1,200)
560
(180-900)
290
(100-490)
Levy et al. (2005)
3,500
(2,400 - 4,600)
2,700
(1,800-3,500)
1,500
(1,000-1,900)
*95 percent confidence or credible intervals are based on the uncertainty about the coefficient in the mortality C-R
functions. All estimates rounded to two significant figures.
Table 8B-8. Estimated O3-Related Life Years Saved in 2020 and in 2030 Under the Final SSI & RME
Rule, Using a 7 Percent Discount Rate
Estimated O3-Related Life Years Saved
(95% Cl)*
Mortality Study:
Assuming Life Expectancies of the
General Population
Assuming Life Expectancies of the Sub-
Population with COPD of Average
Severity
Assuming Life Expectancies of the Sub-
Population with Severe COPD
2020
Bell et al (2004)
360
(120-600)
290
(90 - 500)
170
(50 - 280)
Levy et al. (2005)
1,700
(1,200-2,200)
1,400
(900-1,800)
800
(600-1,100)
2030
Bell et al (2004)
590
(190-1,000)
460
(150-800)
250
(80 - 430)
Levy et al. (2005)
2,700
(1,900-3,500)
2,100
(1,500-2,800)
1,200
(800-1,600)
*95 percent confidence or credible intervals are based on the uncertainty about the coefficient in the mortality C-R
functions. All estimates rounded to two significant figures.
8B.4.3 Reductions in PM2.s-Related Premature Deaths
To generate PM2.s-related health outcomes, we used the same framework as for the benefit-cost
analysis described in Chapter 8 and briefly summarized above in the introductory portion of
Section 8B.4.
As in several recent air pollution health impact assessments (e.g., Kunzli et al., 2000; EPA,
2004), we focused on the prospective cohort long-term exposure studies in deriving the health
impact function for the estimate of premature mortality. Cohort analyses are better able to
capture the full public health impact of exposure to air pollution over time (Kunzli et al., 2001;
NRC, 2002). We selected an effect estimate from the extended analysis of the ACS cohort (Pope
et al., 2002) as well as from the Harvard Six City Study (Laden et al., 2006). Given the focus in
8-84
-------
Cost-Benefit Analysis
this analysis on developing a broader expression of uncertainties in the benefits estimates, and
the weight that was placed on both the ACS and Harvard Six-city studies by experts participating
in the PM2.5 mortality expert elicitation, we elected to provide estimates derived from both Pope
et al. (2002) and Laden et al. (2006).
This latest re-analysis of the ACS cohort data (Pope et al, 2002) provides additional refinements
to the analysis of PM-related mortality by (a) extending the follow-up period for the ACS study
subjects to 16 years, which triples the size of the mortality data set; (b) substantially increasing
exposure data, including consideration for cohort exposure to PM2.5 following implementation of
PM2 5 standard in 1999; (c) controlling for a variety of personal risk factors including
occupational exposure and diet; and (d) using advanced statistical methods to evaluate specific
issues that can adversely affect risk estimates, including the possibility of spatial autocorrelation
of survival times in communities located near each other. The effect estimate from Pope et al.
(2002) quantifies the relationship between annual mean PM2 5 levels and all-cause mortality in
adults 30 and older. We selected the effect estimate estimated using the measure of PM
representing average exposure over the follow-up period, calculated as the average of 1979-1984
and 1999-2000 PM2 5 levels. The effect estimate from this study is 0.0058, which is equivalent
to a relative risk of 1.06 for a 10 |_ig change in PM2.s.
A recent follow up to the Harvard 6-city study (Laden et al., 2006) both confirmed the effect size
from the first study and provided additional confirmation that reductions in PM2 5 directly result
in reductions in the risk of premature death. This additional evidence stems from the observed
reductions in PM2.5 in each city during the extended follow-up period. Laden et al. (2006) found
that mortality rates consistently went down at a rate proportionate to the observed reductions in
PM2.s. The effect estimate obtained from Laden et al. (2006) is 0.0148, which is equivalent to a
relative risk of 1.16 for a 10 ug/m3 change in PM2 5.
Age, cause, and county-specific mortality rates were obtained from CDC for the years 1996
through 1998. CDC maintains an online data repository of health statistics, CDC Wonder,
accessible at http://wonder.cdc.gov/. The mortality rates provided are derived from U.S. death
records and U.S. Census Bureau postcensal population estimates. Mortality rates were averaged
across 3 years (1996 through 1998) to provide more stable estimates. When estimating rates for
age groups that differed from the CDC Wonder groupings, we assumed that rates were uniform
across all ages in the reported age group. For example, to estimate mortality rates for individuals
ages 30 and up, we scaled the 25- to 34-year old death count and population by one-half and then
generated a population-weighted mortality rate using data for the older age groups.
The reductions in incidence of PM25-related premature mortality within each age group
associated with the Final SSI & RME Rule in 2020 and 2030 are summarized in Table 8B-9.
8-85
-------
Final Regulatory Impact Analysis
Table 8B-9: Estimated Reduction in Incidence of PM2.5-Related All-Cause Premature Mortality
Under the Final SSI & RME Rule in 2020 and 2030
Age
Interval
25-29
30-34
35-44
45-54
55-64
65-74
75-84
85+
Total:
Reduction in PM2 5- Related Premature Mortality
(90% Cl)*
2020
Pope et al. (2002)
—
1
(0-2)
3
(1-5)
5
(2-9)
13
(5-21)
19
(7 - 30)
31
(12-49)
47
(18-75)
120
(47- 190)
Laden et al. (2006)
3
(2-5)
3
(2-4)
7
(4-10)
12
(7-18)
29
(16-43)
42
(23-61)
69
(38-100)
110
(57-150)
270
(150-390)
2030
Pope et al. (2002)
—
2
(1-3)
5
(2-8)
9
(3-14)
22
(9 - 35)
51
(20-81)
69
(27-110)
68
(27- 110)
230
(88 - 360)
Laden et al. (2006)
4
(2-5)
4
(2-5)
11
(6-17)
20
(1 1 - 29)
50
(27 - 72)
110
(62- 170)
160
(85 - 230)
150
(84 - 220)
510
(280 - 750)
*90 percent confidence or credible intervals (CIs) are based on the uncertainty about the
coefficient in the mortality C-R functions. All estimates rounded to two significant figures.
8B. 4.4 Life Years Saved as a Result of Reductions in PM2.5-Related Mortality Risk
To calculate life years saved associated with a given change in air pollution, we used a life table
approach coupled with age-specific estimates of reductions in premature mortality. We began
with the complete unabridged life table for the United States in 2000, obtained from CDC (CDC,
2002). For each 1-year age interval (e.g., zero to one, one to two) the life table provides
estimates of the baseline probability of dying during the interval, person years lived in the
interval, and remaining life expectancy. From this unabridged life table, we constructed an
abridged life table to match the age intervals for which we have predictions of changes in
incidence of premature mortality. We used the abridgement method described in CDC (2002).
Table 8B-10 presents the abridged life table for 10-year age intervals for adults over 30 (to match
the Pope et al. [2002] study population). Note that the abridgement actually includes one 5-year
interval, covering adults 30 to 34, with the remaining age intervals covering 10 years each. This
is to provide conformity with the age intervals available for mortality rates.
From the abridged life table (Table 8B-10), we obtained the remaining life expectancy for each
age cohort, conditional on surviving to that age. This is then the number of life years lost for an
individual in the general population dying during that age interval. This information can then be
combined with the estimated number of premature deaths in each age interval calculated with
8-86
-------
Cost-Benefit Analysis
BenMAP (see previous subsection). Total life years gained will then be the sum of life years
gained in each age interval:
TotalLife Years = LEt x M i .
where LEt is the remaining life expectancy for age interval /', Mt is the change in incidence of
mortality in age interval /', and TV is the number of age intervals.
As noted above, for the purposes of determining cost-effectiveness, it is also necessary to
consider the time-dependent nature of the gains in life years. Standard economic theory suggests
that benefits occurring in future years should be discounted relative to benefits occurring in the
present. OMB and EPA guidance suggest discount rates of three and seven percent. Selection of
a 3 percent discount rate is also consistent with recommendations from the U.S. Public Health
Service Panel on Cost Effectiveness in Health and Medicine (Gold et al., 1996).
Table 8B-10. Abridged Life Table for the Total Population, United States, 2000
Age Interval
Probability
of Dying
Between
Ages x to
x+1
Number
Surviving to
Age x
Number
Dying
Between
Ages x to
x+1
Person
Years
Lived
Between
Ages x to
x+1
Total
Number of
Person
Years
Lived
Above Age
x
Expectation
of Life at
Age x
Start
Age
30
35
45
55
65
75
85
95
100+
End
Age
35
45
55
65
75
85
95
100
qx
0.00577
0.01979
0.04303
0.09858
0.21779
0.45584
0.79256
0.75441
1.00000
Ix
97,696
97,132
95,210
91,113
82,131
64,244
34,959
7,252
1,781
dx
564
1,922
4,097
8,982
17,887
29,285
27,707
5,471
1,781
Lx
487,130
962,882
934,026
872,003
740,927
505,278
196,269
20,388
4,636
Tx
4,723,539
4,236,409
3,273,527
2,339,501
1,467,498
726,571
221 ,293
25,024
4,636
ex
48.3
43.6
34.4
25.7
17.9
11.3
6.3
3.5
2.6
Unlike Os-related premature deaths, PM2.s-related premature deaths are associated with long-
term exposures. We therefore did not assume that these deaths all occur in 2020 or 2030. The
PM2.5-related premature deaths avoided and associated life years saved are thus further
discounted to account for the lag between the reduction in ambient PM2 5 and the corresponding
reduction in mortality risk. We used the same 20-year segmented lag structure that is used in the
benefit-cost analysis (see Chapter 8).
8-87
-------
Final Regulatory Impact Analysis
The most complete estimate of the impacts of PM2.5 on life years is calculated using the Pope et
al. (2002) C-R function relating all-cause mortality in adults 30 and over with ambient PM2.5
concentrations averaged over the periods 1979-1983 and 1999-2000. Use of all-cause mortality
is appropriate if there are no differences in the life expectancy of individuals dying from air
pollution-related causes and those dying from other causes. The argument that long-term
exposure to PM2.5 may affect mainly individuals with serious preexisting illnesses is not
supported by current empirical studies. For example, the Krewski et al. (2000) ACS reanalysis
suggests that the mortality risk is no greater for those with preexisting illness at time of
enrollment in the study. Life expectancy for the general population in fact includes individuals
with serious chronic illness. Mortality rates for the general population then reflect prevalence of
chronic disease, and as populations age the prevalence of chronic disease increases.
The only reason one might use a lower life expectancy is if the population at risk from air
pollution was limited solely to those with preexisting disease. Also, note that the OMB Circular
A-4 notes that "if QALYs are used to evaluate a lifesaving rule aimed at a population that
happens to experience a high rate of disability (i.e., where the rule is not designed to affect the
disability), the number of life years saved should not necessarily be diminished simply because
the rule saves lives of people with life-shortening disabilities. Both analytic simplicity and
fairness suggest that the estimate number of life years saved for the disabled population should
be based on average life expectancy information for the relevant age cohorts." As such, use of a
general population life expectancy is preferred over disability-specific life expectancies. Our
primary life years calculations are thus consistent with the concept of not penalizing individuals
with disabling chronic health conditions by assessing them reduced benefits of mortality risk
reductions. PM2.5-Related life years saved under the Final SSI & RME Rule in 2020 and 2030
are given in Table 8B-11.
Table 8B-11. Estimated PM2.5-Related Life Years Saved Under the Final SSI & RME Rule in 2020
and 2030
Estimated PM2.5-Related Life Years Saved
(95% Cl)*
Discounted back to 2020 or 2030,
using a 3 percent discount rate:
Discounted back to 2020 or 2030,
using a 7 percent discount rate:
2020
Pope et al (2002)
1,100
(400-1,800)
800
(300-1,200)
Laden et al. (2006)
2,600
(1,400-4,000)
1,800
(1,000-2,500)
2030
Pope et al (2002)
2,200
(900-3,500)
1,500
(600-2,400)
Laden etal. (2006)
5,000
(2,700-7,000)
3,500
(1,900-5,100)
*95 percent confidence or credible intervals (CIs) are based on the uncertainty about the coefficient in the mortality C-R
functions. All estimates rounded to two significant figures.
For this analysis, direct impacts on life expectancy are measured only through the estimated
change in mortality risk based on the Pope et al. (2002) C-R function. The SAB-HES has
advised against including additional gains in life expectancy due to reductions in incidence of
chronic disease or nonfatal heart attacks (EPA-SAB-COUNCIL-ADV-04-002). Although
reductions in these endpoints are likely to result in increased life expectancy, the HES has
suggested that the cohort design and relatively long follow-up period in the Pope et al. study
-------
Cost-Benefit Analysis
should capture any life-prolonging impacts associated with those endpoints. Impacts of CB and
nonfatal heart attacks on quality of life will be captured separately in the QALY calculation as
years lived with improved quality of life. The methods for calculating this benefit are discussed
below.
8B.4.5 Calculating Changes in the Quality of Life Years (PM2.s-Related Chronic Morbidity)
In addition to directly measuring the quantity of life gained, measured by life years, it may also
be informative to measure gains in the quality of life. The indirect reductions in levels of PM2 5
also lead to reductions in serious illnesses that affect quality of life. These include chronic
bronchitis (CB) and cardiovascular disease, for which we are able to quantify changes in the
incidence of nonfatal heart attacks. To capture these important benefits in the measure of
effectiveness, they must first be converted into a life-year equivalent so that they can be
combined with the direct gains in life expectancy.
For the cost effectiveness analyses for the PM and Os NAAQS RIAs, we developed estimates of
the QALYs gained from reductions in the incidence of CB and nonfatal heart attacks associated
with reductions in ambient PM2.5. In general, QALY calculations require four elements:
1. the estimated change in incidence of the health condition,
2. the duration of the health condition,
3. the quality-of-life weight with the health condition, and
4. the quality-of-life weight without the health condition (i.e., the baseline health state).
The first element is derived using the health impact function approach. The second element is
based on the medical literature for each health condition. The third and fourth elements are
derived from the medical cost-effectiveness and cost-utility literature. In the following two
subsections, we discuss the choices of elements for CB and nonfatal heart attacks.
The preferred source of quality-of-life weights are those based on community preferences, rather
than patient or clinician ratings (Gold et al., 1996). Several methods are used to estimate quality-
of-life weights. These include rating scale, standard gamble, time trade-off, and person trade-off
approaches (Gold, Stevenson, and Fryback, 2002). Only the standard gamble approach is
completely consistent with utility theory. However, the time trade-off method has also been
widely applied in eliciting community preferences (Gold, Stevenson, and Fryback, 2002).
Quality-of-life weights can be directly elicited for individual specific health states or for a more
general set of activity restrictions and health states that can then be used to construct QALY
weights for specific conditions (Horsman et al., 2003; Kind, 1996). For this analysis, we used
weights based on community-based preferences, using time trade-off or standard gamble when
available. In some cases, we used patient or clinician ratings when no community preference-
based weights were available. Sources for weights are discussed in more detail below. Table 8B-
12 summarizes the key inputs for calculating QALYs associated with chronic health endpoints.
8-89
-------
Final Regulatory Impact Analysis
Table 8B-12. Summary of Key Parameters Used in QALY Calculations for Chronic Disease
Endpoints
Parameter
Value(s)
Source(s)
Discount rate
Quality of life preference
score for chronic
bronchitis
Duration of acute phase
of acute myocardial
infarction (AMI)
Probability of CHF post
AMI
Probability of angina post
AMI
Quality-of-life preference
score for post-AMI with
CHF (no angina)
Quality-of-life preference
score for post-AMI with
CHF and angina
Quality-of-life preference
score for post-AMI with
angina (no CHF)
Quality-of-life preference
score for post-AMI (no
angina, no CHF)
0.03 (0.07
sensitivity
analysis)
0.5-0.7
5.5 days - 22
days
0.2
0.51
0.80-0.89
0.76-0.85
0.7-0.89
0.93
Gold et al. (1996), U.S. EPA (2000), U.S. OMB (2003)
Triangular distribution centered at 0.7 with upper bound at
0.9 (Vos, 1999a) (slightly better than a mild/moderate case)
and a lower bound at 0.5 (average weight for a severe case
based on Vos [1999a] and Smith and Peske [1994])
Uniform distribution with lower bound based on average
length of stay for an AMI (AHRQ, 2000) and upper bound
based on Vos (1999b).
Vos, 1999a (WHO Burden of Disease Study, based on
Cowie etal., 1997)
American Heart Association, 2003
(Calculated as the population with angina divided by the
total population with heart disease)
Uniform distribution with lower bound at 0.80 (Stinnett et
al., 1996) and upper bound at 0.89 (Kuntzetal., 1996).
Both studies used the time trade-off elicitation method.
Uniform distribution with lower bound at 0.76 (Stinnett et
al., 1996, adjusted for severity) and upper bound at 0.85
(Kuntz et al., 1996). Both studies used the time trade-off
elicitation method.
Uniform distribution with lower bound at 0.7, based on the
standard gamble elicitation method (Pliskin, Stason, and
Weinstein, 1981) and upper bound at 0.89, based on the
time trade-off method (Kuntz et al., 1996).
Only one value available from the literature. Thus, no
distribution is specified. Source of value is Kuntz et al.
(1996).
8B. 4. 5. 1 Calculating QALYs Associated with Reductions in the Incidence of Chronic Bronchitis
CB is characterized by mucus in the lungs and a persistent wet cough for at least 3 months a year
for several years in a row. CB affects an estimated 5 percent of the U.S. population (American
Lung Association, 1999). For gains in quality of life resulting from reduced incidences of PM-
induced CB, discounted QALYs are calculated as
DISCOUNTED ALYGAINED =
D*
(wt
8-90
-------
Cost-Benefit Analysis
where ACB; is the number of incidences of CB avoided in age interval i, wt is the average QALY
weight for the rth age interval, wfB is the QALY weight associated with CB in the rth age
interval, and/)* is the discounted duration of life with CB for individuals with onset of disease in
A
the rth age interval, equal to \e~rtdt , where A is the duration of life with CB for individuals
with onset of disease the rth age interval.
A limited number of studies have estimated the impact of air pollution on new incidences of CB.
Schwartz (1993) and Abbey et al. (1995) provide evidence that long-term PM exposure gives
rise to the development of CB in the United States. Only the Abbey et al. (1995) study was used,
because it is the only study focusing on the relationship between PM2.5 and new incidences of
CB. The number of cases of CB in each age interval was derived by applying the impact
function from Abbey et al. (1995) to the population in each age interval with the appropriate
baseline incidence rate.9 The effect estimate from the Abbey et al. (1995) study is 0.0137,
which, based on the logistic specification of the model, is equivalent to a relative risk of 1.15 for
a 10 |j,g change in PM2.5. Table 8B-13 presents the estimated reduction in new incidences of CB
associated with the Final SSI & RME Rule in 2020 and 2030.
CB is assumed to persist for the remainder of an affected individual's lifespan. Duration of CB
will thus equal life expectancy conditioned on having CB. CDC has estimated that COPD (of
which CB is one element) results in an average loss of life years equal to 4.26 per COPD death,
relative to a reference life expectancy of 75 years (CDC, 2003). Thus, we subtracted 4.26 from
the remaining life expectancy for each age group, up to age 75. For age groups over 75, we
applied the ratio of 4.26 to the life expectancy for the 65 to 74 year group (0.237) to the life
expectancy for the 75 to 84 and 85 and up age groups to estimate potential life years lost and
then subtracted that value from the base life expectancy.
9 Prevalence rates for CB were obtained from the 1999 National Health Interview Survey
(American Lung Association, 2002). Prevalence rates were available for three age groups: 18-
44, 45-64, and 65 and older. Prevalence rates per person for these groups were 0.0367 for 18-
44, 0.0505 for 45-64, and 0.0587 for 65 and older. The incidence rate for new cases of CB
(0.00378 per person) was taken directly from Abbey et al. (1995).
8-91
-------
Final Regulatory Impact Analysis
Table 8B-13. Estimated Reduction in Incidence of Chronic Bronchitis Under the Final SSI & RME
Rule in 2020 and 2030
Age Interval
27-34
35-44
45-54
55-64
65-74
75-84
85+
Total:
Reduction in PM2 5-Related Chronic Bronchitis
(90% Cl)*
2020
18
(3 - 33)
15
(3 - 28)
14
(3 - 26)
16
(3 - 28)
11
(2-19)
6
(1-11)
3
(1-5)
84
(16-150)
2030
22
(4 - 40)
26
(5 - 48)
22
(4 - 40)
22
(4 - 40)
21
(4 - 38)
12
(2 - 22)
4
(1-8)
130
(24 - 240)
*90 percent confidence or credible intervals (CIs) are based on the uncertainty
about the coefficient in the mortality C-R functions. All estimates rounded to
two significant figures.
Quality of life with chronic lung diseases has been examined in several studies. In an analysis of
the impacts of environmental exposures to contaminants, de Hollander et al. (1999) assigned a
weight of 0.69 to years lived with CB. This weight was based on physicians' evaluations of
health states similar to CB. Salomon and Murray (2003) estimated a pooled weight of 0.77
based on visual analogue scale, time trade-off, standard gamble, and person trade-off techniques
applied to a convenience sample of health professionals. The Harvard Center for Risk Analysis
catalog of preference scores reports a weight of 0.40 for severe COPD, with a range from 0.2 to
0.8, based on the judgments of the study's authors (Bell et al., 2001). The Victoria Burden of
Disease (BoD) study used a weight of 0.47 for severe COPD and 0.83 for mild to moderate
COPD, based on an analysis by Stouthard et al. (1997) of chronic diseases in Dutch populations
(Vos, 1999a). Based on the recommendations of Gold et al. (1996), quality-of-life weights based
on community preferences are preferred for CEA of interventions affecting broad populations.
Use of weights based on health professionals is not recommended. It is not clear from the
Victoria BoD study whether the weights used for COPD are based on community preferences or
judgments of health professionals. The Harvard catalog score is clearly identified as based on
author judgment. Given the lack of a clear preferred weight, we selected a triangular distribution
centered at 0.7 with an upper bound at 0.9 (slightly better than a mild/moderate case defined by
the Victoria BoD study) and a lower bound at 0.5 based on the Victoria BoD study. We will
need additional empirical data on quality of life with chronic respiratory diseases based on
community preferences to improve our estimates.
8-92
-------
Cost-Benefit Analysis
Selection of a reference weight for the general population without CB is somewhat uncertain. It
is clear that the general population is not in perfect health; however, there is some uncertainty as
to whether individuals' ratings of health states are in reference to a perfect health state or to a
generally achievable "normal" health state given age and general health status. The U.S. Public
Health Service Panel on Cost Effectiveness in Health and Medicine recommends that "since
lives saved or extended by an intervention will not be in perfect health, a saved life year will
count as less than 1 full QALY" (Gold et al., 1996). Following Carrothers, Evans, and Graham
(2002), we assumed that the reference weight for the general population without CB is 0.95. To
allow for uncertainty in this parameter, we assigned a triangular distribution around this weight,
bounded by 0.9 and 1.0. Note that the reference weight for the general population is used solely
to determine the incremental quality-of-life improvement applied to the duration of life that
would have been lived with the chronic disease. For example, if CB has a quality-of-life weight
of 0.7 relative to a reference quality-of-life weight of 0.9, then the incremental quality-of-life
improvement in 0.2. If the reference quality-of-life weight is 0.95, then the incremental quality-
of-life improvement is 0.25. As noted above, the population is assumed to have a reference
weight of 1.0 for all life years gained due to mortality risk reductions.
We present discounted QALYs over the duration of the lifespan with CB using a 3 percent
discount rate. Based on the assumptions defined above, we used Monte Carlo simulation
methods as implemented in the Crystal Ball™ software program to develop the distribution of
QALYs gained per incidence of CB for each age interval.10 Based on the assumptions defined
above, the mean 3 percent discounted QALY gained per incidence of CB for each age interval
along with the 95 percent confidence interval resulting from the Monte Carlo simulation is
presented in Table 8B-14. Table 8B-14 presents both the undiscounted and discounted QALYs
gained per incidence, using a 3 percent discount rate.
10 Monte Carlo simulation uses random sampling from distributions of parameters to characterize
the effects of uncertainty on output variables. For more details, see Gentile (1998).
8-93
-------
Final Regulatory Impact Analysis
Table 8B-14. QALYs Gained per Avoided Incidence of CB
Age Interval
Start Age End Age
25 34
35 44
45 54
55 64
65 74
75 84
85+
QALYs Gained per Incidence3
Undiscounted
12.15
(4.40-19.95)
9.91
(3.54-16.10)
7.49
(2.71-12.34)
5.36
(1.95-8.80)
3.40
(1 .22-5.64)
2.15
(0.77-3.49)
0.79
(0.27-1.29)
Discounted (3%)
6.52
(2.36-10.71)
5.94
(2.12-9.66)
5.03
(1 .82-8.29)
4.03
(1.47-6.61)
2.84
(1.02-4.71)
1.92
(0.69-3.13)
0.77
(0.26-1.25)
a Mean of Monte Carlo generated distribution; 95% confidence interval presented in parentheses.
8B.4.5.2 Calculating QALYs Associated with Reductions in the Incidence ofNonfatal
Myocardial Infarctions
Nonfatal heart attacks, or acute myocardial infarctions, require more complicated calculations to
derive estimates of QALY impacts. The actual heart attack, which results when an area of the
heart muscle dies or is permanently damaged because of oxygen deprivation, and subsequent
emergency care are of relatively short duration. Many heart attacks result in sudden death.
However, for survivors, the long-term impacts of advanced coronary heart disease (CHD) are
potentially of long duration and can result in significant losses in quality of life and life
expectancy.
In this phase of the analysis, we did not independently estimate the gains in life expectancy
associated with reductions in nonfatal heart attacks. Based on recommendations from the SAB-
HES, we assumed that all gains in life expectancy are captured in the estimates of reduced
mortality risk provided by the Pope et al. (2002) analysis. We estimated only the change in
quality of life over the period of life affected by the occurrence of a heart attack. This may
understate the QALY impacts of nonfatal heart attacks but ensures that the overall QALY impact
estimates across endpoints do not double-count potential life-year gains.
Our approach adapts a CHD model developed for the Victoria Burden of Disease study (Vos,
1999b). This model accounts for the lost quality of life during the heart attack and the possible
health states following the heart attack. Figure 8B-1 shows the heart attack QALY model in
diagrammatic form.
The total gain in QALYs is calculated as:
8-94
-------
Cost-Benefit Analysis
DISCOUNTED AMI QALY GAINED =
where AAMI; is the number of nonfatal acute myocardial infarctions avoided in age interval /',
WAMI js ^e QJ^LY weight associated with the acute phase of the AMI, PJ is the probability of
being in the/th post-AMI status, w'os is the QALY weight associated with post-AMI health
,DAM
status /, w; is the average QALY weight for age interval i, A = } , e "^ , the discounted
DpaOAM
value of Df11, the duration of the acute phase of the AMI, and A- P°s = ]t=1 e " ^, is the
discounted value of DjostAMI, the duration of post-AMI health status/
8-95
-------
Final Regulatory Impact Analysis
Acute Treatment Stage
Chronic Post-AMI Follow up Stage
Post AMI QALY with Angina and CHF
<£>
Nonfatal AMI
Post AMI QALY with CHF without Angina
Post AMI QALY with Angina without CHF
Post AMI QALY without Angina or CHF
Figure 8B-1. Decision Tree Used in Modeling Gains in QALYs from Reduced Incidence of
Nonfatal Acute Myocardial Infarctions
Nonfatal heart attacks have been linked with short-term exposures to PM2.5 in the United States
(Peters et al., 2001) and other countries (Poloniecki et al., 1997). We used a recent study by
Peters et al. (2001) as the basis for the impact function estimating the relationship between PM2.5
and nonfatal heart attacks. Peters et al. is the only available U.S. study to provide a specific
estimate for heart attacks. Other studies, such as Samet et al. (2000) and Moolgavkar (2000),
show a consistent relationship between all cardiovascular hospital admissions, including for
nonfatal heart attacks, and PM. Given the lasting impact of a heart attack on longer-term health
costs and earnings, we chose to provide a separate estimate for nonfatal heart attacks based on
the single available U.S. effect estimate. The finding of a specific impact on heart attacks is
consistent with hospital admission and other studies showing relationships between fine particles
and cardiovascular effects both within and outside the United States. These studies provide a
weight of evidence for this type of effect. Several epidemiologic studies (Liao et al., 1999; Gold
et al., 2000; Magari et al., 2001) have shown that heart rate variability (an indicator of how much
8-96
-------
Cost-Benefit Analysis
the heart is able to speed up or slow down in response to momentary stresses) is negatively
related to PM levels. Heart rate variability is a risk factor for heart attacks and other CHDs
(Carthenon et al., 2002; Dekker et al., 2000; Liao et al., 1997, Tsuji et al., 1996). As such,
significant impacts of PM on heart rate variability are consistent with an increased risk of heart
attacks.
The number of avoided nonfatal AMI in each age interval was derived by applying the impact
function from Peters et al. (2001) to the population in each age interval with the appropriate
baseline incidence rate.11 The effect estimate from the Peters et al. (2001) study is 0.0241,
which, based on the logistic specification of the model, is equivalent to a relative risk of 1.27 for
a 10 |j,g change in PM2.5. Table 8B-15 presents the estimated reduction in nonfatal AMI
associated with the Final SSI & RME Rule in 2020 and 2030.
Table 8B-15. Estimated Reduction in Nonfatal Acute Myocardial Infarctions Under the Final SSI &
RME Rule in 2020 and 2030
Age Interval
18-24
25-29
35-44
45-54
55-64
65-74
75-84
85+
Total:
Reduction in PM2 5-Related Acute Myocardial Infarction
(90% Cl)*
2020
0
(0-0)
1
(1-2)
10
(5-14)
29
(16-42)
68
(37 - 98)
94
(51 - 140)
48
(26 - 69)
42
(23 - 62)
290
(160-420)
2030
0
(0-0)
2
(1-3)
16
(9 - 23)
43
(23 - 63)
99
(53- 140)
160
(84 - 230)
140
(76-210)
67
(36 - 98)
530
(280 - 770)
*90 percent confidence or credible intervals (CIs) are based on the uncertainty
about the coefficient in the mortality C-R functions. All estimates rounded to
two significant figures.
11 Daily nonfatal myocardial infarction incidence rates per person were obtained from the 1999
National Hospital Discharge Survey (assuming all diagnosed nonfatal AMI visit the hospital).
Age-specific rates for four regions are used in the analysis. Regional averages for populations 18
and older are 0.0000159 for the Northeast, 0.0000135 for the Midwest, 0.0000111 for the South,
and 0.0000100 for the West.
8-97
-------
Final Regulatory Impact Analysis
Acute myocardial infarction results in significant loss of quality of life for a relatively short
duration. The WHO Global Burden of Disease study, as reported in Vos (1999b), assumes that
the acute phase of an acute myocardial infarction lasts for 0.06 years, or around 22 days. An
alternative assumption is the acute phase is characterized by the average length of hospital stay
for an AMI in the United States, which is 5.5 days, based on data from the Agency for
Healthcare Research and Quality's Healthcare Cost and Utilization Project (HCUP).12 We
assumed a distribution of acute phase duration characterized by a uniform distribution between
5.5 and 22 days, noting that due to earlier discharges and in-home therapy available in the United
States, duration of reduced quality of life may continue after discharge from the hospital. In the
period during and directly following an AMI (the acute phase), we assigned a quality of life
weight equal to 0.605, consistent with the weight for the period in treatment during and
immediately after an attack (Vos, 1999b).
During the post-AMI period, a number of different health states can determine the loss in quality
of life. We chose to classify post-AMI health status into four states defined by the presence or
absence of angina and congestive heart failure (CHF). This makes a very explicit assumption
that without the occurrence of an AMI, individuals would not experience either angina or CHF.
If in fact individuals already have CHF or angina, then the quality of life gained will be
overstated. We do not have information about the percentage of the population have been
diagnosed with angina or CHF with no occurrence of an AMI. Nor do we have information on
what proportion of the heart attacks occurring due to PM exposure are first heart attacks versus
repeat attacks. Probabilities for the four post-AMI health states sum to one.
Given the occurrence of a nonfatal AMI, the probability of congestive heart failure is set at 0.2,
following the heart disease model developed by Vos (1999b). The probability is based on a
study by Cowie et al. (1997), which estimated that 20 percent of those surviving AMI develop
heart failure, based on an analysis of the results of the Framingham Heart Study.
The probability of angina is based on the prevalence rate of angina in the U.S. population. Using
data from the American Heart Association, we calculated the prevalence rate for angina by
dividing the estimated number of people with angina (6.6 million) by the estimated number of
people with CHD of all types (12.9 million). We then assumed that the prevalence of angina in
the population surviving an AMI is similar to the prevalence of angina in the total population
with CHD. The estimated prevalence rate is 51 percent, so the probability of angina is 0.51.
Combining these factors leads to the probabilities for each of the four health states as follows:
I. Post AMI with CHF and angina = 0.102
II. Post AMI with CHF without angina = 0.098
III. Post AMI with angina without CHF = 0.408
IV. Post AMI without angina or CHF = 0.392
12 Average length of stay estimated from the HCUP data includes all discharges, including those
due to death. As such, the 5.5-day average length of stay is likely an underestimate of the
average length of stay for AMI admissions where the patient is discharged alive.
8-98
-------
Cost-Benefit Analysis
Duration of post-AMI health states varies, based in part on assumptions regarding life
expectancy with post-AMI complicating health conditions. Based on the model used for
established market economies (EME) in the WHO Global Burden of Disease study, as reported
in Vos (1999b), we assumed that individuals with CHF have a relatively short remaining life
expectancy and thus a relatively short period with reduced quality of life (recall that gains in life
expectancy are assumed to be captured by the cohort estimates of reduced mortality risk).
Table 8B-16 provides the duration (both discounted and undiscounted) of CHF assumed for post-
AMI cases by age interval.
Table 8B-16. Assumed Duration of Congestive Heart Failure
Age Interval
Duration of Heart Failure (years)
Start Age
18
25
35
45
55
65
75
85+
End Age
24
34
44
54
64
74
84
Undiscounted
7.11
6.98
6.49
5.31
1.96
1.71
1.52
1.52
Discounted (3%)
6.51
6.40
6.00
4.99
1.93
1.69
1.50
1.50
Duration of health states without CHF is assumed to be equal to the life expectancy of
individuals conditional on surviving an AMI. Ganz et al. (2000) note that "Because patients with
a history of myocardial infarction have a higher chance of dying of CHD that is unrelated to
recurrent myocardial infarction (for example, arrhythmia), this cohort has a higher risk for death
from causes other than myocardial infarction or stroke than does an unselected population."
They go on to specify a mortality risk ratio of 1.52 for mortality from other causes for the cohort
of individuals with a previous (nonfatal) AMI. The risk ratio is relative to all-cause mortality for
an age-matched unselected population (i.e., general population). We adopted the same ratios and
applied them to each age-specific all-cause mortality rate to derive life expectancies (both
discounted and undiscounted) for each age group after an AMI, presented in Table 8B-17. These
life expectancies were then used to represent the duration of non-CHF post-AMI health states (III
and IV).
8-99
-------
Final Regulatory Impact Analysis
Table 8B-17. Assumed Duration of Non-CHF Post-AMI Health States
Age Interval
Post-AMI Years of Life Expectancy (non-CHF)
Start Age
18
25
35
45
55
65
75
85+
End Age
24
34
44
54
64
74
84
Undiscounted
55.5
46.1
36.8
27.9
19.8
12.8
7.4
3.6
Discounted (3%)
27.68
25.54
22.76
19.28
15.21
10.82
6.75
3.47
For the four post-AMI health states, we used QALY weights based on preferences for the
combined conditions characterizing each health state. A number of estimates of QALY weights
are available for post-AMI health conditions.
The first two health states are characterized by the presence of CHF, with or without angina.
The Harvard Center for Risk Analysis catalog of preference scores provides several specific
weights for CHF with and without mild or severe angina and one set specific to post-AMI CHF.
Following the Victoria Burden of Disease model, we assumed that most cases of angina will be
treated and thus kept at a mild to moderate state. We thus focused our selection on QALY
weights for mild to moderate angina. The Harvard database includes two sets of community
preference-based scores for CHF (Stinnett et al., 1996; Kuntz et al., 1996). The scores for CHF
with angina range from 0.736 to 0.85. The lower of the two scores is based on angina in general
with no delineation by severity. Based on the range of the scores for mild to severe cases of
angina in the second study, one can infer that an average case of angina has a score around 0.96
of the score for a mild case. Applying this adjustment raises the lower end of the range of
preference scores for a mild case of angina to 0.76. We selected a uniform distribution over the
range 0.76 to 0.85 for CHF with mild angina, with a midpoint of 0.81. The same two studies in
the Harvard catalog also provide weights for CHF without angina. These scores range from
0.801 to 0.89. We selected a uniform distribution over this range, with a midpoint of 0.85.
The third health state is characterized by angina, without the presence of CHF. The Harvard
catalog includes five sets of community preference-based scores for angina, one that specifies
scores for both mild and severe angina (Kuntz et al., 1996), one that specifies mild angina only
(Pliskin, Stason, and Weinstein, 1981), one that specifies severe angina only (Cohen, Breall, and
Ho, 1994), and two that specify angina with no severity classification (Salkeld, Phongsavan, and
Oldenburg, 1997; Stinnett et al., 1996). With the exception of the Pliskin, Stason, and Weinstein
score, all of the angina scores are based on the time trade-off method of elicitation. The Pliskin,
Stason, and Weinstein score is based on the standard gamble elicitation method. The scores for
the nonspecific severity angina fall within the range of the two scores for mild angina
specifically. Thus, we used the range of mild angina scores as the endpoints of a uniform
distribution. The range of mild angina scores is from 0.7 to 0.89, with a midpoint of 0.80.
8-100
-------
Cost-Benefit Analysis
For the fourth health state, characterized by the absence of CHF and/or angina, there is only one
relevant community preference score available from the Harvard catalog. This score is 0.93,
derived from a time trade-off elicitation (Kuntz et al., 1996). Insufficient information is
available to provide a distribution for this weight; therefore, it is treated as a fixed value.
Similar to CB, we assumed that the reference weight for the general population without AMI is
0.95. To allow for uncertainty in this parameter, we assigned a triangular distribution around this
weight, bounded by 0.9 and 1.0.
Based on the assumptions defined above, we used Monte Carlo simulation methods as
implemented in the Crystal Ball™ software program to develop the distribution of QALYs
gained per incidence of nonfatal AMI for each age interval. For the Monte Carlo simulation, all
distributions were assumed to be independent. The mean QALYs gained per incidence of
nonfatal AMI for each age interval is presented in Table 8B-18, along with the 95 percent
confidence interval resulting from the Monte Carlo simulation. Table 8B-18 presents both the
undiscounted and discounted QALYs gained per incidence.
Table 8B-18. QALYs Gained per Avoided Nonfatal Myocardial Infarction
Age Interval
QALYs Gained per Incidence3
Start Age
18
25
35
45
55
65
75
85+
End Age
24
34
44
54
64
74
84
Undiscounted
4.18
(1 .24-7.09)
3.48
(1 .09-5.87)
2.81
(0.88-4.74)
2.14
(0.67-3.61)
1.49
(0.42-2.52)
0.97
(0.30-1.64)
0.59
(0.20-0.97)
0.32
(0.13-0.50)
Discounted (3%)
2.17
(0.70-3.62)
2.00
(0.68-3.33)
1.79
(0.60-2.99)
1.52
(0.51-2.53)
1.16
(0.34-1.95)
0.83
(0.26-1.39)
0.54
(0.19-0.89)
0.31
(0.13-0.49)
1 Mean of Monte Carlo generated distribution; 95% confidence interval presented in parentheses.
8B.4.6 Aggregating Life Expectancy and Quality-of-Life Gains
Given the estimates of changes in life expectancy and quality of life, the next step is to aggregate
life expectancy and quality-of-life gains to form an effectiveness measure that can be compared
to costs to develop cost-effectiveness ratios. This section discusses the proper characterization of
the combined effectiveness measure for the denominator of the cost-effectiveness ratio.
8-101
-------
Final Regulatory Impact Analysis
To develop an integrated measure of changes in health, we simply sum together the gains in life
years from reduced mortality risk in each age interval with the gains in QALYs from reductions
in incidence of chronic morbidity endpoints (CB and acute myocardial infarctions). The
resulting measure of effectiveness then forms the denominator in the cost-effectiveness ratio.
This combined measure of effectiveness is not a QALY measure in a strict sense, because we
have not adjusted life-expectancy gains for preexisting health status (quality of life). It is
however, an effectiveness measure that adds a scaled morbidity equivalent to the standard life
years calculation. Thus, we term the aggregate measure morbidity inclusive life years, or
MILYs. Alternatively, the combined measure could be considered as QALYs with an
assumption that the community preference weight for all life-expectancy gains is 1.0. If one
considers that this weight might be considered to be a "fair" treatment of those with preexisting
disabilities, the effectiveness measure might be termed "fair QALY" gained. However, this
implies that all aspects of fairness have been addressed, and there are clearly other issues with
the fairness of QALYs (or other effectiveness measures) that are not addressed in this simple
adjustment. The MILY measure violates some of the properties used in deriving QALY weights,
such as linear substitution between quality of life and quantity of life. However, in aggregating
life expectancy and quality-of-life gains, it merely represents an alternative social weighting that
is consistent with the spirit of the recent OMB guidance on CEA. The guidance notes that
"fairness is important in the choice and execution of effectiveness measures" (OMB, 2003). The
resulting aggregate measure of effectiveness will not be consistent with a strict utility
interpretation of QALYs; however, it may still be a useful index of effectiveness.
Applying the life expectancies and distributions of QALYs per incidence for CB and AMI to
estimated distributions of incidences yields distributions of life expectancy and QALYs gained
under the Final SSI & RME Rule. These distributions reflect both the quantified uncertainty in
estimates of avoided incidence and the quantified uncertainty in QALYs gained per incidence
avoided.
Tables 8B-19 and 8B-20 present the discounted life years, QALYs, and MILYs gained, based on
each combination of Os-mortality study, PM2.5-mortality study, and life expectancy assumption
for Os-related life years saved used for the analysis, using a 3 percent discount rate, for 2020 and
2030, respectively. Tables 8B-21 and 8B-22 present the corresponding results using a 7 percent
discount rate.
8-102
-------
Cost-Benefit Analysis
Table 8B-19. Estimated Gains in Discounted MILYs Under the Final SSI & RME Rule in 2020, Using a 3 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levyetal. (2005)
Levyetal. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levyetal. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality
Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden etal. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-
Related Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
OS-Related Life Years
Gained from Mortality
Risk
Reductions
(95% Cl)
500
(200 - 800)
400
(100-600)
200
(100-300)
2,200
(1 ,500 - 2,900)
1,700
(1 ,200 - 2,200)
1,000
(700 - 1 ,300)
500
(200 - 800)
400
(100-600)
200
(100-300)
2,200
(1 ,500 - 2,900)
1,700
(1 ,200 - 2,200)
1,000
(700 - 1 ,300)
PM2.5-Related Life
Years Gained from
Mortality Risk
Reductions
(95% Cl)
1,100
(400 - 1 ,800)
2,600
(1 ,400 - 4,000)
QALYs Gained from
Reductions in PM2.5-
Related Chronic
Bronchitis
(95% Cl)
390
(50 - 900)
QALYs Gained from
Reductions in PM2.5-
Related Non-Fatal
Myocardial Infarction
(95% Cl)
250
(70-510)
Total MILYs
Gained
(95% Cl)
5,500
(2,600 - 8,000)
5,400
(2,500 - 8,000)
5,200
(2,400 - 8,000)
7,000
(4,300-10,000)
7,000
(3,800-10,000)
6,000
(3,100-9,000)
1 1 ,000
(6,300-16,000)
1 1 ,000
(6,100-16,000)
1 1 ,000
(6,000-15,000)
13,000
(7,900-17,000)
12,000
(7,000-17,000)
1 1 ,000
(6,800-16,000)
"Life years, QALYs, and MILYs are discounted back to 2020. 95% confidence or credible intervals (CIs) around the point estimates are based on the uncertainty surrounding the effect
estimates (coefficients) in the C-R functions and, for QALYs and MILYs, the uncertainty surrounding the quality of life weights. All estimates rounded to two significant figures.
8-103
-------
Final Regulatory Impact Analysis
Table 8B-20. Estimated Gains in Discounted MILYs Under the Final SSI & RME Rule in 2030, Using a 3 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levyetal. (2005)
Levyetal. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levyetal. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality
Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden etal. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-
Related Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
OS-Related Life Years
Gained from Mortality
Risk
Reductions
(95% Cl)
700
(200 - 1 ,200)
600
(200 - 900)
300
(100-500)
3,500
(2,400 - 4,600)
2,700
(1 ,800 - 3,500)
1,500
(1 ,000 - 1 ,900)
700
(200 - 1 ,200)
600
(200 - 900)
300
(100-500)
3,500
(2,400 - 4,600)
2,700
(1 ,800 - 3,500)
1,500
(1 ,000 - 1 ,900)
PM2.5-Related Life
Years Gained from
Mortality Risk
Reductions
(95% Cl)
2,200
(900 - 3,500)
5,000
(2,700 - 7,000)
QALYs Gained from
Reductions in PM2.5-
Related Chronic
Bronchitis
(95% Cl)
590
(80 - 1 ,400)
QALYs Gained from
Reductions in PM2.5-
Related Non-Fatal
Myocardial Infarction
(95% Cl)
430
(110-880)
Total MILYs
Gained
(95% Cl)
6,100
(3,100-9,000)
6,000
(3,000 - 9,000)
5,700
(2,700 - 9,000)
9,000
(5,800-12,000)
8,000
(5,000 - 1 1 ,000)
6,800
(3,900-10,000)
1 1 ,600
(6,800-16,000)
1 1 ,000
(6,600-16,000)
1 1 ,000
(6,400-16,000)
1 4,000
(9,400-19,000)
1 4,000
(9,000-18,000)
12,000
(7,500-17,000)
"Life years, QALYs, and MILYs are discounted back to 2030. 95% confidence or credible intervals (CIs) around the point estimates are based on the uncertainty surrounding the effect
estimates (coefficients) in the C-R functions and, for QALYs and MILYs, the uncertainty surrounding the quality of life weights. All estimates rounded to two significant figures.
8-104
-------
Cost-Benefit Analysis
Table 8B-21. Estimated Gains in Discounted MILYs Under the Final SSI & RME Rule in 2020, Using a 7 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality
Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden etal. (2006)
Laden etal. (2006)
Laden etal. (2006)
Laden etal. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-
Related Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
OS-Related Life Years
Gained from Mortality
Risk
Reductions
(95% Cl)
360
(120-600)
290
(90 - 500)
170
(50 - 280)
1,700
(1 ,200 - 2,200)
1,400
(900 - 1 ,800)
800
(600-1,100)
360
(120-600)
290
(90 - 500)
170
(50 - 280)
1,700
(1 ,200 - 2,200)
1,400
(900 - 1 ,800)
800
(600- 1,100)
PM2.5-Related Life
Years Gained from
Mortality Risk
Reductions
(95% Cl)
800
(300 - 1 ,200)
1,800
(1 ,000 - 2,500)
QALYs Gained from
Reductions in PM2.5-
Related Chronic
Bronchitis
(95% Cl)
300
(30 - 600)
QALYs Gained from
Reductions in PM2.5-
Related Non-Fatal
Myocardial Infarction
(95% Cl)
200
(50 - 400)
Total MILYs
Gained
(95% Cl)
3,800
(1 ,800 - 5,700)
3,700
(1 ,800 - 5,700)
3,600
(1 ,600 - 5,500)
5,100
(3,100-7,000)
4,800
(2,700 - 7,000)
4,200
(2,300 - 6,200)
7,500
(4,300 - 1 1 ,000)
7,500
(4,200 - 1 1 ,000)
7,300
(4,100-11,000)
9,000
(5,600-12,000)
9,000
(5,300-12,000)
8,000
(4,800 - 1 1 ,000)
"Life years, QALYs, and MILYs are discounted back to 2020. 95% confidence or credible intervals (CIs) around the point estimates are based on the uncertainty surrounding the effect
estimates (coefficients) in the C-R functions and, for QALYs and MILYs, the uncertainty surrounding the quality of life weights. All estimates rounded to two significant figures.
8-105
-------
Final Regulatory Impact Analysis
Table 8B-22. Estimated Gains in Discounted MILYs Under the Final SSI & RME Rule in 2030, Using a 7 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality
Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden etal. (2006)
Laden etal. (2006)
Laden etal. (2006)
Laden etal. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-
Related Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
OS-Related Life Years
Gained from Mortality
Risk
Reductions
(95% Cl)
590
(190-1,000)
460
(150-800)
250
(80 - 430)
2,700
(1 ,900 - 3,500)
2,100
(1 ,500 - 2,800)
1,200
(800 - 1 ,600)
590
(190-1,000)
460
(150-800)
250
(80 - 430)
2,700
(1 ,900 - 3,500)
2,100
(1 ,500 - 2,800)
1,200
(800 - 1 ,600)
PM2.5-Related Life
Years Gained from
Mortality Risk
Reductions
(95% Cl)
800
(300 - 1 ,200)
1,800
(1 ,000 - 2,500)
QALYs Gained from
Reductions in PM2.5-
Related Chronic
Bronchitis
(95% Cl)
400
(50 - 900)
QALYs Gained from
Reductions in PM2.5-
Related Non-Fatal
Myocardial Infarction
(95% Cl)
340
(90 - 700)
Total MILYs
Gained
(95% Cl)
4,300
(2,200 - 6,300)
4,100
(2,100-6,200)
3,900
(1 ,900 - 5,900)
6,400
(4,200 - 9,000)
5,800
(3,700 - 8,000)
4,900
(2,900 - 7,000)
8,100
(4,800 - 1 1 ,000)
8,000
(4,600 - 1 1 ,000)
7,800
(4,500 - 1 1 ,000)
10,000
(6,800- 14,000)
10,000
(6,300-13,000)
9,000
(5,400-12,000)
"Life years, QALYs, and MILYs are discounted back to 2030. 95% confidence or credible intervals (CIs) around the point estimates are based on the uncertainty surrounding the effect
estimates (coefficients) in the C-R functions and, for QALYs and MILYs, the uncertainty surrounding the quality of life weights. All estimates rounded to two significant figures.
8-106
-------
Cost-Benefit Analysis
8B. 4.7 Estimating the Avoided Costs of Chronic Illness
Construction of cost-effectiveness ratios requires estimates of effectiveness (in this case
measured by lives saved, life years gained, or MILYs gained) in the denominator and estimates
of costs in the numerator. As noted above (see Section 8B.3.1), our estimate of costs in the
numerator is net of the avoided costs (cost savings) associated with the reductions in morbidity
(Gold et al., 1996). Among the morbidity costs subtracted from the direct costs of controls in the
numerator are the avoided costs of illness (COI) associated with PM2.s-related CB and nonfatal
AMI.
Avoided costs for CB and nonfatal AMI are based on estimates of lost earnings and medical
costs.13 Using age-specific annual lost earnings and medical costs estimated by Cropper and
Krupnick (1990) and a 3 percent discount rate, we estimated a lifetime present discounted value
(in 2005$) due to CB of $179,305 for someone between the ages of 27 and 44; $116,892 for
someone between the ages of 45 and 64; and $13,741 for someone over 65. The corresponding
age-specific estimates of lifetime present discounted value (in 2005$) using a 7 percent discount
rate are $102,300, $86,359, and $11,190, respectively. These estimates assumed that 1) lost
earnings continue only until age 65, 2) medical expenditures are incurred until death, and 3) life
expectancy is unchanged by CB.
Because the costs associated with a myocardial infarction extend beyond the initial event itself,
we consider costs incurred over several years. Using age-specific annual lost earnings estimated
by Cropper and Krupnick (1990) and a 3 percent discount rate, we estimated a present
discounted value in lost earnings (in 2005$) over 5 years due to a myocardial infarction of
$10,389 for someone between the ages of 25 and 44, $15,313 for someone between the ages of
45 and 54, and $88,508 for someone between the ages of 55 and 65. The corresponding age-
specific estimates of lost earnings (in 2005$) using a 7 percent discount rate are $9,301, $13,709,
and $79,241, respectively. Cropper and Krupnick (1990) do not provide lost earnings estimates
for populations under 25 or over 65. Thus, we do not include lost earnings in the cost estimates
for these age groups.
Two estimates of the direct medical costs of myocardial infarction are used. The first estimate is
from Wittels, Hay, and Gotto (1990), which estimated expected total medical costs of MI over 5
years to be $51,211 (in 1986$) for people who were admitted to the hospital and survived
hospitalization (there does not appear to be any discounting used). Using the CPI-U for medical
care, the Wittels estimate is $135,667 in year 2005$. This estimated cost is based on a medical
cost model, which incorporated therapeutic options, projected outcomes, and prices (using
"knowledgeable cardiologists" as consultants). The model used medical data and medical
13 Gold et al. (1996) recommend not including lost earnings in the cost-of-illness estimates,
suggesting that in some cases, they may be already be counted in the effectiveness measures.
However, this requires that individuals fully incorporate the value of lost earnings and reduced
labor force participation opportunities into their responses to time-tradeoff or standard-gamble
questions. For the purposes of this analysis and for consistency with the way costs-of-illness are
calculated for the benefit-cost analysis, we have assumed that individuals do not incorporate lost
earnings in responses to these questions. This assumption can be relaxed in future analyses with
improved understanding of how lost earnings are treated in preference elicitations.
8-107
-------
Final Regulatory Impact Analysis
decision algorithms to estimate the probabilities of certain events and/or medical procedures
being used. The second estimate is from Russell et al. (1998), which estimated first-year direct
medical costs of treating nonfatal myocardial infarction of $15,540 (in 1995$), and $1,051
annually thereafter. Converting to year 2005$, that would be $27,674 for a 5-year period (using
a 3 percent discount rate).
The two estimates from these studies are substantially different, and we have not adequately
resolved the sources of differences in the estimates. Because the wage-related opportunity cost
estimates from Cropper and Krupnick (1990) cover a 5-year period, we used estimates for
medical costs that similarly cover a 5-year period. We used a simple average of the two 5-year
estimates, or $81,671, and add it to the 5-year opportunity cost estimate. The resulting estimates
are given in Table 8B-23.
Table 8B-23. Estimated Costs Over a 5-Year Period (in 2005$) of a Nonfatal Myocardial Infarction
Age of Onset
0-24
25-44
45-54
55-65
>65
Opportunity Cost1
$0
$10,389
$15,313
$88,508
$0
Medical Cost2
$81,671
$81,671
$81,671
$81,671
$81,671
Total Cost*
$81,671
$92,060
$96,984
$170,179
$81,671
1 Positive opportunity costs are based on Cropper and Krupnick (1990), using a 3 percent
discount rate.
2 An average of the 5-year costs estimated by Wittels, Hay, and Gotto (1990) and Russell et al.
(1998).
The total avoided COI by age group associated with the reductions in CB and nonfatal acute
myocardial infarctions (using a 3 percent discount rate) is provided in Table 8B-24. The total
avoided COI associated with the Final SSI & RME Rule (using a 3 percent discount rate) is
about $42 million in 2020 and about $71 million in 2030. Note that these estimates do not
include any direct avoided medical costs associated with premature mortality. Nor do they
include any medical costs that occur more than 5 years from the onset of a nonfatal AMI.
Therefore, they are likely underestimates of the true avoided COI associated with the Final SSI
& RME Rule in 2020 and 2030.
8-108
-------
Cost-Benefit Analysis
Table 8B-24. Avoided Costs of Illness Associated with Reductions in Chronic Bronchitis and
Nonfatal Acute Myocardial Infarctions Under the Final SSI & RME Rule in 2020 and
2030
Age Interval
18-24
25-29
35-44
45-54
55-64
65-74
75-84
85+
Total:
Avoided Cost of Illness (in millions of 2005$)*
2020
Chronic Bronchitis
—
$4.1
$3.4
$2.1
$2.2
$0.2
$0.1
$0.1
$12.1
Nonfatal Acute
Myocardial Infarction
$0.0
$0.1
$0.9
$2.8
$11.5
$7.7
$3.9
$3.5
$30.4
2030
Chronic Bronchitis
—
$4.9
$6.0
$3.2
$3.3
$0.4
$0.2
$0.1
$18.1
Nonfatal Acute
Myocardial Infarction
$0.0
$0.2
$1.5
$4.2
$16.8
$12.8
$11.6
$5.5
$52.5
'Discounted using a 3 percent discount rate.
8B.4.8 Cost-Effectiveness Ratios
Construction of cost-effectiveness ratios requires estimates of effectiveness (in this case
measured by lives saved, life years gained, or MILYs gained) in the denominator and estimates
of costs in the numerator. As noted above (see Section 8B.3.1), the estimate of costs in the
numerator should include both the direct costs of the controls necessary to achieve the reduction
in ambient PM2.5 and 63 and the avoided costs (cost savings) associated with the reductions in
morbidity (Gold et al., 1996). In general, because reductions in air pollution do not require direct
actions by the affected populations, there are no specific costs to affected individuals (aside from
the overall increases in prices that might be expected to occur as control costs are passed on by
affected industries). Likewise, because individuals do not engage in any specific actions to
realize the health benefit of the pollution reduction, there are no decreases in utility (as might
occur from a medical intervention) that need to be adjusted for in the denominator. Thus, the
elements of the numerator are direct costs of controls minus the avoided COI associated with CB
and nonfatal AMI. In addition, to account for the value of reductions in (V and PM2.5-related
acute health impacts and non-health benefits, we netted out the monetized value of these benefits
from the numerator to yield a "net cost" estimate. For the MILY aggregate effectiveness
measure, the denominator is simply the sum of (Os- and PM2.s-related) life years gained from
increased life expectancy and QALYs gained from the (PM2.5-related) reductions in CB and
nonfatal AMI. The separate (V and PM2.5-related inputs to the denominators of the cost-
effectiveness ratios are summarized above in Tables 8B-19 through 8B-22. The cost-
effectiveness ratios and 95 percent confidence (credible) intervals resulting from all of the
sources of uncertainty considered, using Monte Carlo procedures as implemented in the Crystal
Ball™ software program and incorporating both the (V and PM2.5-related benefits are shown in
the tables below. Tables 8B-25 and 8B-26 show cost per life saved, using a 3 percent and 7
percent discount rate, respectively. Tables 8B-27 and 8B-28 show cost per life year saved at the
two discount rates; and Tables 8B-29 and 8B-30 show cost per MILY gained.
8-109
-------
Final Regulatory Impact Analysis
Table 8B-25. Estimated Net Cost (2005$) per O3- and PM2.5-Related Life Saved Under the Final SSI
& RME Rule in 2020 and 2030, Using a 3 Percent Discount Rate
O3 Mortality Study
Bell etal. (2004)
Bell etal. (2004)
Levy etal. (2005)
Levy etal. (2005)
PM2 5 Mortality Study
Pope et al. (2002)
Laden etal. (2006)
Pope et al. (2002)
Laden etal. (2006)
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
Life Saved*
(95% Cl)**
2020
$260
($110 -$580)
$110
($54 - $220)
$180
($85 - $320)
$96
($48 -$170)
2030
$74
($-99 - $280)
$34
($-44 -$120)
$44
($-58 -$140)
$26
($-34 - $83)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. PM25-related avoided
deaths are discounted back to 2020 or 2030. O3-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals incorporate uncertainty surrounding the O3 and PM2 5 coefficients in the
mortality and morbidity C-R functions as well as the uncertainty surrounding unit values of morbidity endpoints. All
estimates rounded to two significant figures.
Table 8B-26. Estimated Net Cost (2005$) per O3- and PM2.5-Related Life Saved Under the Final SSI
& RME Rule in 2020 and 2030, Using a 7 Percent Discount Rate
O3 Mortality Study
Bell etal. (2004)
Bell etal. (2004)
Levy etal. (2005)
Levy etal. (2005)
PM2 5 Mortality Study
Pope et al. (2002)
Laden etal. (2006)
Pope et al. (2002)
Laden etal. (2006)
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
Life Saved*
(95% Cl)**
2020
$300
($130 -$660)
$130
($67 - $250)
$200
($100 -$350)
$110
($58 -$190)
2030
$99
($-87 - $330)
$47
($-39 -$140)
$57
($-49 -$160)
$35
($-30 - $95)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. PM25-related avoided
deaths are discounted back to 2020 or 2030. O3-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals incorporate uncertainty surrounding the O3 and PM2.5 coefficients in the
mortality and morbidity C-R functions as well as the uncertainty surrounding unit values of morbidity endpoints. All
estimates rounded to two significant figures.
8-110
-------
Cost-Benefit Analysis
Table 8B-27. Estimated Net Cost (2005$) per O3- and PM2.5-Related Life Year Saved Under the Final SSI & RME Rule in 2020 and 2030,
Using a 3 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-Related
Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
Life Year Saved*
(95% Cl)**
2020
$23
($9.9 -$54)
$24
($10 -$56)
$25
($10 -$61)
$16
($7.8 -$30)
$18
($8.3 -$34)
$21
($9.2 - $44)
$10
($5 - $20)
$11
($5 -$21)
$11
($5.1 -$21)
$8.8
($4.4 -$16)
$9.2
($4.5 -$17)
$9.9
($4.8 -$19)
2030
$6.8
($-9 - $26)
$7.1
($-9.5 -$27)
$7.6
($-10 -$30)
$4.1
($-5.5 -$13)
$4.7
($-6.2 -$15)
$5.8
($-7.6 - $20)
$3.1
($-4.2 -$11)
$3.2
($-4.2 -$11)
$3.3
($-4.4 -$11)
$2.4
($-3.2 -$7.7)
$2.6
($-3.4 -$8.3)
$2.9
($-3.9 -$9.5)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. All life years are discounted back to the year of death. PM2 5-related
avoided deaths are discounted back to 2020 or 2030. Q s-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals (CIs) incorporate uncertainty surrounding the O3 and PM2 5 coefficients in the mortality and morbidity C-R functions as well as
the uncertainty surrounding unit values of morbidity endpoints. All estimates rounded to two significant figures.
8-111
-------
Final Regulatory Impact Analysis
Table 8B-28. Estimated Net Cost (2005$) per O3- and PM2.5-Related Life Year Saved Under the Final SSI & RME Rule in 2020 and 2030,
Using a 7 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-Related
Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
Life Year Saved*
(95% Cl)**
2020
$36
($16 -$81)
$37
($16 -$85)
$39
($17 -$93)
$24
($12 -$44)
$26
($13 -$49)
$31
($15 -$62)
$16
($7.8 -$30)
$16
($7.9 -$31)
$17
($8 - $32)
$13
($6.8 -$23)
$14
($7.1 -$24)
$15
($7.5 -$28)
2030
$12
($-1 1 - $42)
$13
($-1 1 - $44)
$14
($-12 -$50)
$7
($-6 - $20)
$8.0
($-6.7 - $23)
$10
($-8.3 -$31)
$5.6
($-4.7 -$17)
$5.7
($-4.8 -$17)
$5.9
($-5 -$18)
$4.3
($-3.5 -$12)
$4.6
($-3.9 -$13)
$5.1
($.4.4 .$15)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. All life years are discounted back to the year of death. PM25-related
avoided deaths are discounted back to 2020 or 2030. g s-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals (CIs) incorporate uncertainty surrounding the O3 and PM2.5 coefficients in the mortality and morbidity C-R functions as well as
the uncertainty surrounding unit values of morbidity endpoints. All estimates rounded to two significant figures.
8-112
-------
Cost-Benefit Analysis
Table 8B-29. Estimated Net Cost (2005$) per O3- and PM2.5-Related MILY Gained Under the Final SSI & RME Rule in 2020 and 2030,
Using a 3 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-Related
Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
MILY Gained*
(95% Cl)**
2020
$20
($9 - $42)
$21
($9.3 -$44)
$21
($9.4 - $46)
$15
($7.2 - $26)
$16
($7.6 -$29)
$18
($8.4 -$36)
$9.8
($4.7 -$18)
$9.9
($4.8 -$19)
$10
($4.8 -$19)
$8.3
($4.2 -$14)
$8.7
($4.3 -$16)
$9.3
($4.5 -$17)
2030
$5.5
($-7.4 -$19)
$5.6
($-7.6 - $20)
$6.0
($-8.1 -$21)
$3.6
($-4.9 -$11)
$4.0
($-5.4 -$13)
$4.8
($-6.4 -$16)
$2.8
($-3.8 -$9.4)
$2.9
($-3.8 -$9.6)
$3.0
($-3.9 -$10)
$2.2
($-2.9- $7)
$2.4
($-3.2 -$7.5)
$2.7
($-3.5 -$8.6)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. PM25-related avoided deaths are discounted back to 2020 or
2030. O3-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals (CIs) incorporate uncertainty surrounding the O3 and PM2 5 coefficients in the mortality and morbidity C-R
functions as well as the uncertainty surrounding unit values of morbidity endpoints. All estimates rounded to two significant figures.
8-113
-------
Final Regulatory Impact Analysis
Table 8B-30. Estimated Net Cost (2005$) per O3- and PM2.5-Related MILY Gained Under the Final SSI & RME Rule in 2020 and 2030,
Using a 7 Percent Discount Rate
O3 Mortality Study
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
Bell et al. (2004)
Bell et al. (2004)
Bell et al. (2004)
Levy et al. (2005)
Levy et al. (2005)
Levy et al. (2005)
PM2.5 Mortality Study
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Pope et al. (2002)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Laden et al. (2006)
Life Expectancy Assumption for O3-Related
Mortality
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
General Population
Subpopulation with Average COPD
Subpopulation with Severe COPD
Cost Effectiveness Ratio: Net Cost (in Thousand $) per
MILY Gained*
(95% Cl)**
2020
$110
($58 -$190)
$33
($15 -$70)
$31
($15 -$64)
$31
($15 -$66)
$27
($13 -$51)
$22
($12 -$38)
$23
($12 -$42)
$15
($7.6 -$29)
$15
($7.4 -$27)
$15
($7.5 -$28)
$14
($7.1 -$25)
$13
($6.5 -$21)
2030
$31
($15 -$64)
$31
($15 -$66)
$33
($15 -$70)
$22
($12 -$38)
$24
($12 -$42)
$27
($13 -$51)
$15
($7.4 -$27)
$15
($7.5 -$28)
$15
($7.6 -$29)
$13
($6.5 -$21)
$13
($6.8 -$22)
$14
($7.1 -$25)
The cost of the regulation is estimated to be $207.4 million in 2020 and $185.5 million in 2030. PM25-related avoided deaths are discounted back to 2020 or
2030. Os-related deaths are assumed to occur in 2020 or 2030.
**95 percent confidence or credible intervals (CIs) incorporate uncertainty surrounding the O3 and PM2 5 coefficients in the mortality and morbidity C-R
functions as well as the uncertainty surrounding unit values of morbidity endpoints. All estimates rounded to two significant figures.
8-114
-------
-------
Final Regulatory Impact Analysis
8B.5 Conclusions
We estimated the cost effectiveness of attaining the Final Small SI and Recreational Marine
Engine Rule in 2020 and in 2030, based on reductions in premature deaths and incidence of
chronic disease. We measured effectiveness using several different metrics, including lives
saved, life years saved, and QALYs gained (for improvements in quality of life due to reductions
in incidence of chronic disease). We suggested a new metric for aggregating life years saved and
improvements in quality of life, morbidity inclusive life years (MILY) which assumes that
society assigns a weight of one to years of life extended regardless of preexisting disabilities or
chronic health conditions.
CEA of environmental regulations that have substantial public health impacts may be
informative in identifying programs that have achieved cost-effective reductions in health
impacts and can suggest areas where additional controls may be justified. However, the overall
efficiency of a regulatory action can only be judged through a complete benefit-cost analysis that
takes into account all benefits and costs, including both health and non-health effects. The
benefit-cost analysis for the Final Small SI and Recreational Marine Engine Rule, provided in
Chapter 8, shows that the rule has potentially large net benefits, indicating that implementation
of the Final Small SI and Recreational Marine Engine Rule will likely result in improvements in
overall public welfare.
8-60
-------
Cost-Benefit Analysis
8B.6 References
Abbey, D.E., B.L. Hwang, RJ. Burchette, T. Vancuren, and P.K. Mills. 1995. "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.
Agency for Healthcare Research and Quality. 2000. HCUPnet, Healthcare Cost and Utilization
Project.
American Lung Association. 2002. Trends in Chronic Bronchitis and Emphysema: Morbidity
and Mortality. American Lung Association, Best Practices and Program Services, Epidemiology
and Statistics Unit.
American Heart Association. 2003. Heart Disease and Stroke Statistics—2003 Update. Dallas,
Texas: American Heart Association.
Anderson HR, Atkinson RW, Peacock JL, Marston L, Konstantinou K. 2004. Meta-analysis of time-series
studies and panel studies of Particulate Matter (PM) and Ozone (O3): Report of a WHO task group.
Copenhagen, Denmark: World Health Organization.
Bala, M.V., and G.A. Zarkin. 2000. "Are QALYs an Appropriate Measure for Valuing
Morbidity in Acute Diseases?" Health Economics 9:177-180.
Bell, C.M., R.H. Chapman, P.W. Stone, E.A. Sandberg, and P.J. Neumann. 2001. "An Off-the-
Shelf Help List: A Comprehensive Catalog of Preference Scores from Published Cost-Utility
Analyses." Medical Decision Making 21:288-94.
Bleichrodt, H., P.P. Wakker, and M. Johannesson. 1996. "Characterizing QALYs by Risk
Neutrality." Journal of Risk and Uncertainty 15:107-114.
Brock, D.W. 2002. "Chapter 14.3: Fairness and Health." In Summary Measures of Population
Health: Concepts, Ethics, Measurement and Applications, C.J.L. Murray, J.A. Salomon, C.D.
Mathers, and A.D. Lopez (eds.). Geneva: World Health Organization.
Carrothers, T. J., J.S. Evans, and J.D. Graham. 2002. "The Lifesaving Benefits of Enhanced Air
Quality." Harvard Center for Risk Analysis Working Paper.
Carnethon, M.R., D. Liao, G.W. Evans, W.E. Cascio, L.E. Chambless, W.D. Rosamond, and G.
Heiss. 2002. "Does the Cardiac Autonomic Response to Postural Change Predict Incident
Coronary Heart Disease and Mortality? The Atherosclerosis Risk in Communities Study."
American Journal of Epidemiology 155(l):48-56.
Centers for Disease Control and Prevention (CDC) Health, United States. 2003. Table 30.
Years of Potential Life Lost before Age 75 for Selected Causes of Death, According to Sex,
Race, and Hispanic Origin: United States, Selected Years 1980-2000.
Centers for Disease Control and Prevention (CDC). 2002. National Vital Statistics Reports
51(3).
8-61
-------
Final Regulatory Impact Analysis
Cohen, D.J., J.A. Breall, and K.K. Ho. 1994. "Evaluating the Potential Cost-Effectiveness of
Stenting as a Treatment for Symptomatic Single-Vessel Coronary Disease: Use of a Decision-
Analytic Model." Circulation 89(4): 1859-1874.
Cohen, IT., J.K. Hammitt, and J.I. Levy. 2003. "Fuels for Urban Transit Buses: A Cost-
effectiveness Analysis." Environmental Science and Technology 37:1477-1484.
Cowie, M.R., A. Mosterd, D.A. Wood, et al. 1997. "The Epidemiology of Heart Failure."
European Heart Journal 18:208-25.
Coyle, D., D. Stieb, R.T. Burnett, P. DeCivita, D. Krewski, Y. Chen, and MJ. Thun. 2003.
"Impact of Particulate Air Pollution on Quality Adjusted Life Expectancy in Canada." Journal
of Toxicology and Environmental Health, Part A 66:1847-1863.
Cropper, M.L., and AJ. Krupnick. 1990. "The Social Costs of Chronic Heart and Lung
Disease." Discussion Paper QE 89-16-REV. Washington, DC: Resources for the Future.
de Hollander, A.E.M., J.M. Melse, E. Lebret, and P.G.N. Kramers. 1999. "An Aggregate Public
Health Indicator to Represent the Impact of Multiple Environmental Exposures." Epidemiology
10:606-617.
Dekker, J.M., R.S. Crow, A.R. Folsom, PJ. Hannan, D. Liao, C.A. Swenne, and E.G. Schouten.
2000. "Low Heart Rate Variability in a 2-Minute Rhythm Strip Predicts Risk of Coronary Heart
Disease and Mortality From Several Causes: The ARIC Study." Circulation 2000 102:1239-
1244.
Freeman, A.M., J.K. Hammitt, and P. De Civita. 2002. "On Quality Adjusted Life Years
(QALYs) and Environmental/Consumer Safety Valuation." AERENewsletter 22(1):7-12.
Ganz, D.A., K.M. Kuntz, G.A. Jacobson, and J. Avorn. 2000. "Cost-Effectiveness of 3-
Hydroxy-3-Methylglutaryl Coenzyme A Reductase Inhibitor Therapy in Older Patients with
Myocardial Infarction." Annals of Internal Medicine 132:780-787.
Garber, A.M., and C.E. Phelps. 1997. "Economic Foundations of Cost-Effectiveness Analysis."
Journal of Health Economics 16:1-31.
Gentile, J.E. 1998. Random Number Generation and Monte Carlo Methods. New York, NY:
Springer Verlag.
Gold, M.R., J.E. Siegel, L.B. Russell, and M.C. Weinstein. 1996. Cost-effectiveness in Health
andMedicine. New York: Oxford University Press.
Gold, D.R., A. Litonjua, J. Schwartz, E. Lovett, A. Larson, B. Nearing, G. Allen, M. Verrier, R.
Cherry, and R. Verrier. 2000. "Ambient Pollution and Heart Rate Variability." Circulation
101(11):1267-1273.
8-62
-------
Cost-Benefit Analysis
Gold, M.R., D. Stevenson, and D.G. Fryback. 2002. "HALYS and QALYS and DALYS, Oh
My: Similarities and Differences in Summary Measures of Population Health." Annual Review
of Public Health 23:115-34.
Hammitt, J.K. 2002. "QALYs versus WTP." Risk Analysis 22(5):985-1001.
Harvard Center for Risk Analysis, Catalog of Preference Scores. Available at:
http://www.hcra.harvard.edu/pdf/preferencescores.pdf.
Horsman, I, W. Furlong, D. Feeny, and G. Torrance. 2003. "The Health Utilities Index (HUI):
Concepts, Measurement Properties, and Applications." Health and Quality of Life Outcomes
1:54.
Johnson, F.R., and K. Lievense. 2000. Stated-Preference Indirect Utility and Quality-Adjusted
Life Years. Report prepared for Health Canada, July.
Kerridge, R.K., P.P. Glasziou, and K.M. Hillman. 1995. "The Use of "Quality-Adjusted Life
Years" (QALYs) to Evaluate Treatment in Intensive Care." Anaesthia and Intensive Care
23:322-31.
Kind, P. 1996. The EuroQoL Instrument: An Index of Health-Related Quality of Life. Quality
of Life and Pharmacoeconomics in Clinical Trials, Second Edition, B. Spilker (ed.), pp. 191-201.
Philadelphia, PA: Lippincott-Raven Publishers.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M.
Abrahamowicz, and W.H. White. 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.
Kuntz, K.M, J. Tsevant, L. Goldman, and M.C. Weinstein. 1996. "Cost-Effectiveness of
Routine Coronary Angiography after Acute Myocardial Infarction." Circulation 94(5):957-65.
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.
Kunzli, N., S. Medina, R. Kaiser, P. Quenel, F. Horak Jr, and M. Studnicka. 2001. "Assessment
of Deaths Attributable to Air Pollution: Should We Use Risk Estimates Based on Time Series or
on Cohort Studies?" American Journal of Epidemiology 153(11): 1050-55.
Liao, D., J. Cai, W.D. Rosamond, R.W. Barnes, R.G. Hutchinson, E.A. Whitsel, P. Rautaharju,
and G. Heiss. 1997. "Cardiac Autonomic Function and Incident Coronary Heart Disease: A
Population-Based Case-Cohort Study. The ARIC Study. Atherosclerosis Risk in Communities
Study." American Journal of Epidemiology 145(8):696-706.
8-63
-------
Final Regulatory Impact Analysis
Liao, D., J. Creason, C. Shy, R. Williams, R. Watts, and R. Zweidinger. 1999. "Daily Variation
of Parti culate Air Pollution and Poor Cardiac Autonomic Control in the Elderly." Environmental
Health Perspectives 107:521-5.
Magari, S.R., R. Hauser, J. Schwartz, P.L. Williams, T.J. Smith, and D.C. Christiani. 2001.
"Association of Heart Rate Variability with Occupational and Environmental Exposure to
Particulate Air Pollution." Circulation 104(9):986-91.
Mannino, D.M., D. E. Dohertya, A. S. Buist. 2006. Global Initiative on Obstructive Lung
Disease (GOLD) Classification of Lung Disease and Mortality: Findings from the
Atherosclerosis Risk in Communities (ARIC) Study. Respiratory Medicine 100: 115-122.
Miller, E.G., and J.F. Hurley. 2003. "Life Table Methods for Quantitative Impact Assessments
in Chronic Mortality." Journal of Epidemiology and Community Health 57:200-206.
Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the Circulatory
System in Three U.S. Metropolitan Areas." Journal of the Air Waste Management Association
50:1199-206.
Murray, C.J.L., J.A. Salomon, C.D. Mathers, and A.D. Lopez. 2002. "Summary Measures of
Population Health: Concepts, Ethics, Measurement and Applications." Geneva: World Health
Organization.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air
Pollution Regulations. Washington, DC: The National Academies Press.
Neumann, P. 2003. "A Web-based Registry of Cost-utility Analyses." Risk in Perspective
11(3).
Peters, A., D.W. Dockery, I.E. Muller, and M.A. Mittleman. 2001. "Increased Particulate Air
Pollution and the Triggering of Myocardial Infarction." Circulation 103:2810-2815.
Pliskin, J.S., D.S. Shepard, and M.C. Weinstein. 1980. "Utility Functions for Life Years and
Health Status." Operations Research 28:206-224.
Pliskin, J.S., W.B. Stason, and M.C. Weinstein. 1981. "Coronary Artery Bypass Graft Surgery:
Clinical Decision Making and Cost-Effective Analysis." Medical Decision Making 1(1): 10-28.
Poloniecki, J.D., R.W. Atkinson, A.P. de Leon, and H.R. Anderson. 1997. "Daily Time Series
for Cardiovascular Hospital Admissions and Previous Day's Air Pollution in London, UK."
Occupational and Environmental Medicine 54(8):535-40.
Pope, D.A., R.T. Burnett, MJ. Thun, E.E. Cale, D. Krewski, K. Ito, and G.D. Thurston. 2002.
"Lung Cancer, Cardiopulmonary Mortality, and Long-Term Exposure to Fine Particulate Air
Pollution." Journal of the American Medical Association 287:1132-1141.
8-64
-------
Cost-Benefit Analysis
Rabl, A. 2003. "Interpretation of Air Pollution Mortality: Number of Deaths or Years of Life
Lost?" Journal of the Air and Waste Management Association 53:41-50.
Russell, M.W., D.M. Huse, S. Drowns, B.C. Hamel, and S.C. Hartz. 1998. "Direct Medical
Costs of Coronary Artery Disease in the United States." American Journal of Cardiology
Salkeld, G., P. Phongsavan, and B. Oldenburg. 1997. "The Cost-Effectiveness of a
Cardiovascular Risk Reduction Program in General Practice." Health Policy 41 : 105-19.
Salomon, J.A., and C.J.L. Murray. 2003. "A Multi-Method Approach to Measuring Health-
State Valuations." Health Economics., online.
Samet, J.M., S.L. Zeger, F. Dominici, F. Curriero, I. Coursac, D.W. Dockery, J. Schwartz, and
A. Zanobetti. June 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.
Cambridge, MA: Health Effects Institute.
Schwartz, J. 1993. "Parti culate Air Pollution and Chronic Respiratory Disease." Environmental
Research 62:7 '-13.
Singer, P., J. McKie, H. Kuhse, J. Richardson, and J. Harris. 1995. "Double Jeopardy and the
Use of QALYs in Health Care Allocation." Journal of Medical Ethics 21 : 144-157.
Smith, K.J., and R.R. Pesce. 1994. "Pulmonary Artery Catheterization in Exacerbations of
COPD Requiring Mechanical Ventilation: A Cost-Effectiveness Analysis." Respiratory Care
39:961-7.
Stinnett, A.A., M.A. Mittleman, M.C. Weinstein, K.M. Kuntz, D.J. Cohen, L.W. Williams, P.A.
Goldman, D.O. Staiger, M.G.M. Hunink, J. Tsevat, A.N.A. Tosteson, and L. Goldman. 1996.
"The Cost-effectiveness of Dietary and Pharmacologic Therapy for Cholesterol Reduction in
Adults." In Cost-effectiveness in Health and Medicine, Gold, M.E., J. E. Siegel, L.B. Russell,
and M.C. Weinstein (eds.). New York: Oxford University Press.
Stouthard, M.E.A., J.L. Essink-Bot, G.J. Bonsel, J.J. Barendregt, P.G.N. Kramers, H.P.A. Van de
Water, and L.J. Gunning-Schepers. 1997. "Disability Weights for Diseases in the Netherlands."
Rotterdam: Department of Public Health, Erasmus University Rotterdam.
Suwa, T. J.C. Hogg, K.B. Quinlan, A. Ohgami, R. Vincent, and S.F. van Eeden. 2002.
"Parti culate Air Pollution Induces Progression of Atherosclerosis." Journal of the American
College of Cardiology 39:935-942.
Tsuji, H., M.G. Larson, F.J. Venditti, Jr., E.S. Manders, J.C. Evans, C.L. Feldman, and D. Levy.
1996. "Impact of Reduced Heart Rate Variability on Risk for Cardiac Events. The Framingham
Heart Study." Circulation 94(ll):2850-2855.
U.S. Environmental Protection Agency (EPA). 2000. Guidelines for Preparing Economic
Analyses. EPA240-R-00-003.
8-65
-------
Final Regulatory Impact Analysis
U.S. Environmental Protection Agency (EPA). 2004. Final Regulatory Analysis: Control of
Emissions from Nonroad Diesel Engines.
U.S. Environmental Protection Agency, 2006. Final Regulatory Impact Analysis: PM2.5 NAAQS.
Prepared by Office of Air and Radiation. Available: http://www.epa.gov/ttn/ecas/ria.html
[accessed 10 March 2008].
U.S. Environmental Protection Agency (EPA) Science Advisory Board (SAB). EPA-SAB-
COUNCIL-ADV-04-002. March 2004. Advisory on Plans for Health Effects Analysis in the
Analytical Plan for EPA's Second Prospective Analysis—Benefits and Costs of the Clean Air
Act, 1990-2020: Advisory by the Health Effects Subcommittee of the Advisory Council on
Clean Air Compliance Analysis.
U.S. Environmental Protection Agency (EPA) Science Advisory Board (SAB). 2005. EPA's
Review of the National Ambient Air Quality Standards for Particulate Matter (Second Draft PM
Staff Paper, January 2005). EPA-SAB-CASAC-05-007. June.
U.S. Food and Drug Administration (FDA). 2004. Bar Code Label Requirements for Human
Drug Products and Biological Products; Final Rule. Federal Register: February 26, 2004
Volume 69 (38): 9119-9171.
U.S. Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis,
September 17, 2003. Available at: http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf.
Vos, T. 1999a. Final worksheet VM1 COPD.xls, prepared for Victoria Burden of Disease
Study. Available at: http://www.dhs.vic.gov.au/phd/bod/daly.htm.
Vos, T. 1999b. Final worksheet VL2 IHD.xls, prepared for Victoria Burden of Disease Study.
Available at: http://www.dhs.vic.gov.au/phd/bod/daly.htm.
Wittels, E.H., J.W. Hay, and A.M. Gotto, Jr. 1990. "Medical Costs of Coronary Artery Disease
in the United States." American Journal of Cardiology 65(7):432-40.
1 Clean Air Science Advisory Committee. June 2005. EPA's Review of the National Ambient Air Quality Standards
for P articulate Matter (Second Dr aft PM Staff Paper, January 2005). A Review by the PM Review Panel of the EPA
Clean Air Science Advisory Committee. EPASAB-CASAC-05-007.
2 EPA-SAB-COUNCIL_ADV_04-002. March 2004. Advisory on Plans for Health Effects Analysis in the
Analytical Plan for EPA 's Second Prospective Analysis - Benefits and Costs of the Clean Air Act, 1990-2020:
Advisory by the Health Effects Subcommittee of the Advisory Council on Clean Air Compliance Analysis.
8-66
-------
Cost-Benefit Analysis
3 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press.
4 EPA-SAB-COUNCIL-ADV-00-001. October 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.
5EPA-COUNCIL-LTR-05-001. December 2004. Advisory Council on Clean Air Compliance Analysis Response to
Agency Request on Cessation Lag.
6 Pope, C.A., III, M.J. Thun, M.M. Namboodiri, D.W. Dockery, IS. Evans, F.E. Speizer, and C.W. Heath, Jr. 1995.
"Paniculate 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.
7 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.
V111 Chestnut, L.G., and R.D. Rowe. 1990. Preservation Values for Visibility Protection at the National Parks: Draft
Final Report. Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC and Air Quality Management Division, National Park Service, Denver, CO.
8-67
-------
Economic Impact Analysis
CHAPTER 9: Economic Impact Analysis
We prepared an Economic Impact Analysis (EIA) to estimate the economic impacts of
the final emission control program on the Small SI and Marine SI engine and equipment
markets. In this chapter we describe the Economic Impact Model (EIM) developed to estimate
the market-level changes in price and outputs for affected markets and the social costs of the
program as well as the expected distribution of those costs across affected economic sectors. We
also present the results of our analysis.
We estimate the net social costs of the final program to be about $186 million in 2030.l
This estimate reflects the estimated compliance costs associated with the Small SI and Marine SI
engine standards and the expected fuel savings from improved evaporative controls. When the
fuel savings are not taken into account, the results of the economic impact modeling suggest that
the social costs of these programs are expected to be about $459 million in 2030. Consumers of
Small SI and Marine products are expected to bear about 86 percent of these costs. Engine and
equipment manufacturers are expected to bear 3.3 percent and 10.3 percent, respectively. We
estimate fuel savings of about $273 million in 2030, which will accrue to consumers.
With regard to market-level impacts in 2030, the average price increase for Small SI
engines is expected to be about 7.4 percent ($12 per unit). The average price increase for Marine
SI engines is expected to be about 1.9 percent ($213 per unit). The largest average price increase
for Small SI equipment is expected to be about 5.6 percent ($15 per unit) for Class I equipment.
The largest average price increase for Marine SI vessels is expected to be about 2.4 percent
($204 per unit) for Personal Watercraft.
9.1 Overview and Results
9.1.1 What is an Economic Impact Analysis?
An Economic Impact Analysis (EIA) is prepared to inform decision makers about the
potential economic consequences of a regulatory action. The analysis consists of estimating the
social costs of a regulatory program and the distribution of these costs across stakeholders.
These estimated social costs can then be compared with estimated social benefits (as presented in
Chapter 8). As defined in EPA's Guidelines for Preparing Economic Analyses (EPA 2000, p
113), social costs are the value of the goods and services lost by society resulting from a) the use
of resources to comply with and implement a regulation and b) reductions in output. In this
analysis, social costs are explored in two steps. In the market analysis, we estimate how prices
and quantities of goods affected by the final emission control program can be expected to
'All estimates presented in this section are in 2005$. The fuel savings in this net social cost is calculated by
2005 gasoline price. 2005 Petroleum Marketing Annual (Table 31). U.S. Department of Energy, Energy Information
Administration (DoE 2005).
9-1
-------
Final Regulatory Impact Analysis
change once the program goes into effect. In the economic welfare analysis, we look at the total
social costs associated with the program and their distribution across stakeholders.
9.1.2 What Methodology Did EPA Use in this Economic Impact Assessment?
The Economic Impact Model (EIM) is a behavioral model developed for this proposal to
estimate price and quantity changes and total social costs associated with the emission controls
under consideration. The model relies on basic microeconomic theory to simulate how
producers and consumers of affected products can be expected to respond to an increase in
production costs as a result of the final emission control program. The economic theory that
underlies the model is described in detail in Section 9.2.
The EIM is designed to estimate the economic impacts of the final program by simulating
economic behavior. This is done by creating a model of the initial, pre-control market for a
product, shocking it by the estimated compliance costs, and observing the impacts on the market.
At the initial, pre-control market equilibrium, a market is characterized by a price and quantity
combination at which consumers are willing to purchase the same amount of a product that
producers are willing to produce at that price (demand is equal to supply). The control program
under consideration would increase the production costs of affected goods by the amount of the
compliance costs. This generates a "shock" to the initial equilibrium market conditions.
Producers of affected products will try to pass some or all of the increased costs on to the
consumers of these goods through price increases. In response to the price increases, consumers
will decrease their demand for the affected goods. Producers will react to the decrease in
quantity demanded by decreasing the quantity they produce; the market will react by setting a
higher price for those fewer units. These interactions continue until a new market equilibrium
price and quantity combination is achieved. The amount of the compliance costs that can be
passed on to consumers is ultimately limited by the price sensitivity of purchasers and producers
in the relevant market (price elasticity of demand and supply). The EIM explicitly models these
behavioral responses and estimates new equilibrium prices and output and the resulting
distribution of social costs across these stakeholders (producers and consumers).
The EIM is a behavioral model. The estimated social costs of this emission control
program are a function of the ways in which producers and consumers of the engines and
equipment affected by the standards change their behavior in response to the costs incurred in
complying with the standards. These behavioral responses are incorporated in the EIM through
the price elasticity of supply and demand (reflected in the slope of the supply and demand
curves), which measure the price sensitivity of consumers and producers. An "inelastic" price
elasticity (less than one) means that supply or demand is not very responsive to price changes (a
one percent change in price leads to less than one percent change in supply or demand ). An
"elastic" price elasticity (more than one) means that supply or demand is sensitive to price
changes (a one percent change in price leads to more than one percent change in supply or
demand). A price elasticity of one is unit elastic, meaning there is a one-to-one correspondence
between a change in price and change in demand. The price elasticities used in this analysis are
described in Section 9.3 and were estimated using well-established econometric methods. It
9-2
-------
Economic Impact Analysis
should be noted that demand in the engine markets is internally derived from the Small SI
equipment and Marine SI vessel markets as part of the process of running the model. This is an
important feature of the EIM, which allows it to link the engine and equipment components of
each model and simulate how compliance costs can be expected to ripple through the affected
market.
9.1.3 What Economic Sectors are Included in the Economic Impact Model?
There are two broad economic sectors affected by the emission control program
described in this proposal: (1) Small SI engines and equipment, and (2) Marine SI engines and
equipment. For Small SI engines and equipment we model one integrated handheld engine and
equipment category. On the nonhandheld side, the model distinguishes between 9 engine
categories, depending on engine class and useful life (Class I: UL125, UL250, and UL500;
Class I -snowblower: UL 125, UL250, and UL 500; Class II: UL250, UL500, UL1000), and 8
nonhandheld equipment categories (agriculture/construction/ general industrial; utility and
recreational vehicles; lawn mowers; tractors; other lawn and garden; gensets/welders;
pumps/compressors/pressure washers; and snowblowers). For Marine SI engines and
equipment, the model distinguishes between sterndrives and inboards (SD/I), outboards (OB),
and personal watercraft (PWC); SD/I and OB are further classified by whether they are luxury or
not. These markets are described in Section 9.3 and in more detail in the industry
characterizations prepared for this proposal.
This analysis assumes that the all of these products are purchased and used by residential
households. This means that to model the behavior change associated with final standards we
model all uses as residential lawn and garden care, power generation (Small SI) or personal
recreation (Marine SI). We do not explicitly model commercial uses (how the costs of
complying with the final programs may affect the production of goods and services that use
Small SI or Marine SI engines or equipment as production inputs); we treat all commercial uses
as if they were residential uses. We believe this approach is reasonable because the commercial
share of the end use markets for both Small SI and Marine SI equipment is very small (see
Section 9.3.1.1). In addition, for any commercial uses of these products the share of the cost of
these products to total production costs is also small (e.g., the cost of a Small SI generator is only
a very small part of the total production costs for a construction firm). Therefore, a price
increase of the magnitude anticipated for this control program is not expected to have a
noticeable impact on prices or quantities of goods or services produced using Small SI or Marine
SI equipment as inputs (e.g., commercial turf care, construction, or fishing).
In the ELM the Small SI and Marine SI markets are not linked (there is no feedback
mechanism between the Small SI and Marine SI market segments). This is appropriate because
the affected equipment is not interchangeable and because there is very little overlap between the
engine producers in each market. These two sectors represent different aspects of economic
activity (lawn and garden care and power generation as opposed to recreational marine) and
production and consumption of one product is not affected by the other. In other words, an
increase in the price of lawnmowers is not expected to have an impact on the production and
supply of personal watercraft, and vice versa. Production and consumption of each of these
9-3
-------
Final Regulatory Impact Analysis
products are the results of other factors that have little crossover impacts (the need for residential
garden upkeep or power generation; the desire for personal recreation).
Consistent with the final emission controls, this Economic Impact Analysis covers
engines sold in 49 states. California engines are not included because California has its own
state-level controls for Small SI and Marine SI engines. The sole exceptions are Small SI
engines used in agriculture and construction applications in California: these engines are
included in the control program of this analysis because the Clean Air Act preempts California
from setting standards for those engines.
Table 9.1-1 summarizes the markets included in this Economic Impact Analysis. More
detailed information on the markets and model data inputs is provided in Section 9.3.3, and in
the industry profiles prepared for this proposal (See Chapter 1, & RTI, 2006 ).
9-4
-------
Economic Impact Analysis
Table 9.1-1: Summary of Markets in Economic Impact Model
Model Dimension
Small SI
Marine SI
Description of Markets
HANDHELD
No distinction between engine and
equipment types for this analysis
NONHANDHELD
Engine types
Class I (125, 250, 500 hours)
Class II (250, 500, 1000 hours)
Equipment types
Lawn mowers
Lawn and garden tractors
Pumps/compressors/pressure washers
Agriculture/construction/industrial
Other lawn and garden
Gensets/welders
Snowblowers
Utility and recreational vehicles
Engine and equipment types
SDfl recreational (runabouts,
airboats, jetboats)
SD/I luxury (yachts, cruisers, offshore)
OB recreational (runabouts, pontoons,
fishing)
OB luxury (yacht, cruiser, express
fish)
Personal watercraft (PWC)
Engine sizes
Less than 25 hp
26 to 50 hp
51 to 100 hp
101 to 175 hp
176 to 300 hp
Greater than 3 00 hp
Geographic scope
49 state, plus agriculture and
construction for California
49 state
(no California engines or equipment)
Market structure
Competitive
Competitive
Baseline population
EPA certification database
PSR OE Link sales database
EPA and CARB certification database
NMMA published statistical data
Growth projections
EPA's 2005 Nonroad model
EPA's 2005 Nonroad model
Supply elasticity
Econometric estimate (elastic)
Econometric estimate (elastic)
Demand elasticity
Econometric estimate
Gensets, all handheld: elastic
Lawn mowers & other LG: inelastic
All others: unit elastic
Econometric estimate (elastic)
Regulatory shock
Handheld (integrated market): direct
compliance costs (fixed + variable)
cause shift in supply function
Nonhandheld:
Engine: direct compliance costs
cause shift in supply function
Equipment (Class I): no direct
compliance costs but higher engine
prices cause shift in supply function
Equipment (Class II): direct
compliance costs plus higher engine
prices cause shift in supply function
PWC (integrated): direct compliance
costs (fixed + variable) cause shift in
supply function
SD/I and Outboard luxury:
Engine: direct compliance costs
cause shift in supply function
Vessel: direct compliance costs plus
higher engine prices cause shift in
supply function
Outboard recreational:
Engine: direct compliance costs
cause shift in supply function
Vessel: direct compliance costs
cause shift in supply function
-------
Final Regulatory Impact Analysis
9.1.4 Summary of Results
The EIA consists of two parts: a market analysis and welfare analysis. The market
analysis looks at expected changes in prices and quantities for affected products. The welfare
analysis looks at economic impacts in terms of annual and present value changes in social costs.
We performed a market analysis for all years and all engines and equipment markets. In
this section we present summarized results for selected markets and years. More detail can be
found in the appendices to this chapter and in the docket for this rule (Li, 2007). Also, included
in Appendix 9H are sensitivity analyses for several key inputs.
In this analysis, initial market equilibrium conditions are shocked by either the fixed cost
or the variable cost. For the market analysis, this leads to a small increase in estimated price
impacts for the years 2008 through 2014, the period during which the costs change over time
reflecting the phase-in of either the different costs (variable and fixed costs) or the different
standards. The increase is small because, for many elements of the program, annual per unit
compliance costs are relatively smaller than engine or equipment per unit price . For the
welfare analysis, applying both fixed and variable costs means that the burden of the social costs
attributable to producers and consumers remains fixed throughout the period of analysis. This is
because producers pass the fixed costs to consumers at the same rate as the variable costs instead
of having to absorb them internally.
9.1.4.1 Market Analysis Results
In the market analysis, we estimate how prices and quantities of goods affected by the
final emission control program can be expected to change once the program goes into effect.
The analysis relies on the initial market equilibrium prices and quantities for each type of
equipment and the price elasticity of supply and demand. It predicts market reactions to the
increase in production costs due to the new compliance costs (variable and fixed). It should be
noted that this analysis does not allow any other factors of production to vary. In other words, it
does not consider that manufacturers may adjust their production processes or marketing
strategies in response to the control program. Also, as explained above, while the markets are
shocked by both fixed and variable costs, the market shock is not offset by fuel savings.
A summary of the estimated market impacts is presented in Table 9.1-2 for 2014, 2018,
and 2030. These years were chosen because 2014 is the year of highest compliance cost; the
market impacts reflect the compliance costs for all the programs as well as growth in equipment
population; 2018 is the year in which the learning curve is expected to be applied to the variable
cost; and 2030 illustrates the long-term impacts of the program.
Market level impacts are reported for the engine and equipment markets separately. This
is because the EIM is a two-level model that treats these markets separately. However, changes
in equipment prices and quantities are due to impacts of both direct equipment compliance costs
9-6
-------
Economic Impact Analysis
and indirect engine compliance costs that are passed through to the equipment market from the
engine market through higher engine prices.
The average market-level impacts presented in this section are designed to provide a
broad overview of the expected market impacts that is useful when considering the impacts of
the rule on the economy as a whole. The average price impacts are product-weighted averages
of the results for the individual engine and equipment categories included in that sub-sector (e.g.,
the estimated Marine SI engine price and quantity changes are weighted averages of the
estimated results for all of the Marine SI engine markets). The average quantity impacts are the
sum of the decrease in units produced units across sub-markets. Price increases and quantity
decreases for specific types of engines and equipment are likely to be different.
Although each of the affected equipment in this analysis generally require one engine
(the exception being Marine SI sterndrive/inboards), the estimated decrease in the number of
engines produced in Table 9.1-2 is less than the estimated decrease in the number of equipment
produced. At first glance, this result seems counterintuitive because it does not reflect the
approximate one-to-one correspondence between engines and equipment. This discrepancy
occurs because the engine market-level analysis examines only output changes for engines that
are produced by independent engine manufacturers and subsequently sold to independent
equipment manufacturers. Engines produced and consumed by vertically integrated
equipment/engine manufactures are not explicitly modeled. Therefore, the market-level analysis
only reflects engines sold on the "open market," and estimates of output changes for engines
consumed internally are not reflected in this number.2 Despite the fact that changes in
consumption of internally consumed engines in not directly reported in the market-level analysis
results, the costs associated with these engines are included in the market-level analysis (as
supply shift for the equipment markets). In addition, the cost and welfare analyses include the
compliance costs associated with internally consumed engines.
9.1.4.1.1 Marine SI Market Analysis
The average price increase for Marine SI engines in 2014, the high cost year, is estimated
to be about 2.4 percent, or $266. By 2018, this average price increase is expected to decline to
about 1.9 percent, or $213, and remain at that level for later years. The market impact analysis
predicts that with these increases in engine prices the expected average decrease in total sales in
2014 is about 2.7 percent, or 10,883 engines. This decreases to about 2.2 percent in 2018, or
about 9,055 engines.
On the vessel side, the average price change reflects the direct equipment compliance
costs plus the portion of the engine costs that are passed on to the equipment purchaser (via
higher engine prices). The average price increase in 2014 is expected to be about 1.6 percent, or
2For example, PWC and handheld equipment producers generally integrate equipment and engine
manufacturing processes and are included in the EIM as one-level equipment markets. Since there is no engine
market for these engines, the EIM does not include PWC and handheld engine consumption changes in engine
market-level results.
9-7
-------
Final Regulatory Impact Analysis
$285. By 2018, this average price increase is expected to decline to about 1.3 percent, or $231.
These price increases are expected to vary across vessel categories. The category with the
largest price increase in 2014 is expected to be personal watercraft, with an estimated price
increase of about 3.0 percent in 2014; this is expected to decrease to 2.4 percent in 2018. The
smallest expected change in 2014 is expected to be for sterndrive/inboards vessels, which are
expected to see price increases of about 0.9 percent. The market impact analysis predicts that
with these increases in vessel prices the expected average decrease in quantity produced in 2014
is about 3.2 percent, or 12,230 vessels. This is expected to decrease to about 2.6 percent in
2018, or about 10,145 vessels. The personal watercraft category is expected to experience the
largest decline in 2014, about 6.0 percent (4,800 vessels). The smallest percentage decrease in
production is expected for sterndrive/inboards at 1.7 percent (1,580 vessels); the smallest
absolute decrease in quantity is expected for outboard recreational vessels, at 144 vessels (2.0
percent).
9.1.4.1.2 Small SI Market Analysis
The average price increase for Small SI engines in 2014, the high cost year, is estimated to
be about 8.3 percent, or $14. By 2018, this average price increase is expected to decline to about
7.4 percent, or $12, and remain at that level for later years. The market impact analysis predicts
that with these increases in engine prices the expected average decrease in total sales in 2014 is
expected to be about 1.9 percent, or 304,000 engines. This is expected to decrease to about 1.7
percent in 2018, or about 285,000 engines.
On the equipment side, the average price change reflects the direct equipment compliance
costs plus the portion of the engine costs that are passed on to the equipment purchaser (via
higher engine prices). The average price increase for all Small SI equipment in 2014 is expected
to be about 2.6 percent, or $10. By 2018, this average price increase is expected to decline to
about 2.3 percent, or $8. The average price increase and quantity decrease differs by category of
equipment. As shown in Table 9.1-2, the price increase for Class I equipment is estimated to be
about 6.2 percent ($17) in 2014, decreasing to 5.6 percent ($15) in 2018. The market impact
analysis predicts that with these increases in equipment prices the expected average decrease in
the quantity of Class I equipment produced in 2014 is about 2.1 percent, or 209,000 units. This
is expected to decrease to about 1.9 percent in 2018, or about 200,000 units. For Class II
equipment, a higher price increase is expected, about 2.6 percent ($24) in 2014, decreasing to 2.2
percent ($20) in 2018. The expected average decrease in the quantity of Class II equipment
produced in 2014 is about 2.8 percent, or 101,000 units, decreasing to 2.4 percent, or about
92,000 units, in 2018.
For the handheld equipment market, prices are expected to increase about 0.2 percent for
all years, and quantities are expected to decrease about 0.3 percent.
9-8
-------
Economic Impact Analysis
Table 9.1-2: Summary of Estimated Market Impacts for 2014, 2018, 2030 (2005$)
Market
Change
Absolute
in Price
Percent
Change
Absolute
in Quantity
Percent
2014
Marine
Engines
Equipment
SD/I
OB Recreational
OB Luxury
PWC
Small SI
Engines
Equipment
Class I
Class II
HH
$266
$285
$299
$870
$271
$253
$14
$10
$17
$24
$0.3
2.4%
1.6%
0.9%
1.0%
1.4%
3.0%
8.3%
2.6%
6.2%
2.6%
0.2%
-10,883
-12,229
-1,578
-144
-5,666
-4,841
-303,992
-360,310
-209,284
-101,104
-49,922
-2.7%
-3.2%
-1.7%
-2.0%
-2.8%
-6.0%
-1.9%
-1.4%
-2.1%
-2.8%
-0.3%
2018
Marine
Engines
Equipment
SD/I
OB Recreational
OB Luxury
PWC
Small SI
Engines
Equipment
Class I
Class II
HH
$213
$231
$244
$702
$218
$204
$12
$8
$15
$20
$0.3
1.9%
1.3%
0.7%
0.8%
1.1%
2.4%
7.4%
2.3%
5.6%
2.2%
0.2%
-9,055
-10,145
-1,318
-119
-4,697
-4,010
-284,995
-347,189
-200,155
-91,871
-55,164
-2.2%
-2.6%
-1.4%
-1.6%
-2.3%
-4.8%
-1.7%
-1.2%
-1.9%
-2.4%
-0.3%
2030
Marine
Engines
Equipment
SD/I
OB Recreational
OB Luxury
PWC
Small SI
Engines
Equipment
Class I
Class II
HH
$213
$231
$244
$702
$218
$204
$12
$8
$15
$20
$0.3
1.9%
1.3%
0.7%
0.8%
1.1%
2.4%
7.4%
2.3%
5.6%
2.2%
0.2%
-9,802
-10,981
-1,426
-129
-5,085
-4,341
-338,346
-412,103
-237,485
-109,120
-65,498
-2.2%
-2.6%
-1.4%
-1.6%
-2.3%
-4.8%
-1.7%
-1.2%
-1.9%
-2.4%
-0.3%
9-9
-------
Final Regulatory Impact Analysis
9.1.4.2 Economic Welfare Results
In the economic welfare analysis we look at the costs to society of the final program in
terms of losses to consumer and producer surplus. These surplus losses are combined with
estimated fuel savings to estimate the net economic welfare impacts of the program. Estimated
annual net social costs for selected years are presented in Table 9.1-3. This table shows that total
social costs for each year are slightly less than the total engineering costs. This is because the
total engineering costs do not reflect the decreased sales of engines and equipment that are
incorporated in the total social costs.
9-10
-------
Economic Impact Analysis
Table
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
NPV at 3%a
NPV at 7%a
9.1-3: Estimated Annual Engineering and Social Costs Through 2037
(2005$, Smillion)
Total
Engineering
Costs
$53.8
$126.8
$271.0
$328.7
$441.6
$445.4
$450.0
$411.6
$408.8
$398.3
$403.3
$408.3
$413.4
$418.4
$423.4
$428.4
$433.4
$438.4
$443.5
$448.5
$453.6
$458.6
$463.7
$468.7
$473.8
$478.8
$483.9
$488.9
$494.0
$499.0
$7,705.3
$4,559.3
Total Social
Costs
$53.8
$126.2
$267.4
$324.1
$435.5
$439.3
$443.8
$406.8
$404.2
$393.8
$398.8
$403.8
$408.8
$413.7
$418.6
$423.6
$428.6
$433.6
$438.6
$443.6
$448.6
$453.5
$458.6
$463.6
$468.6
$473.6
$478.5
$483.6
$488.6
$493.6
$7,616.6
$4,506.2
Fuel Savings
$3.2
$8.1
$19.6
$43.9
$70.8
$95.7
$115.9
$134.3
$150.9
$165.3
$178.3
$190.4
$201.4
$211.1
$220.5
$229.3
$237.1
$244.2
$250.8
$256.9
$262.7
$268.1
$273.0
$277.6
$281.9
$285.8
$289.6
$293.3
$296.8
$300.1
$3,374.6
$1,774.7
Net Engineering
Costs
(including fuel
savings)
$50.7
$118.7
$251.5
$284.8
$370.8
$349.7
$334.1
$277.3
$258.0
$233.0
$225.0
$217.9
$212.0
$207.2
$202.9
$199.0
$196.3
$194.2
$192.7
$191.6
$190.8
$190.5
$190.6
$191.1
$191.9
$193.0
$194.2
$195.6
$197.2
$198.9
$4,330.7
$2,784.6
Net Social
Costs
(including fuel
savings)
$50.6
$118.1
$247.8
$280.2
$364.7
$343.6
$327.9
$272.5
$253.3
$228.6
$220.5
$213.4
$207.4
$202.6
$198.2
$194.3
$191.5
$189.3
$187.8
$186.6
$185.8
$185.4
$185.5
$185.9
$186.7
$187.7
$188.9
$190.3
$191.8
$193.5
$4,242.0
$2,731.4
aEPA presents the present value of cost and benefits estimates using both a three percent and a seven percent
social discount rate. According to OMB Circular A-4, "the 3 percent discount rate represents the 'social rate
of time preference' . . . [which] means the rate at which 'society' discounts future consumption flows to their
present value " ; "the seven percent rate is an estimate of the average before-tax rate of return to private capital
in the U.S. economy . . . [that] approximates the opportunity cost of capital."
9-11
-------
Final Regulatory Impact Analysis
Figure 9.1-1: Estimated Engineering, Total Social, Net Social Costs and Fuel Savings
600
2008 2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036
• Total Engineering Costs
- Fuel Savings
Total Social Costs
•Net Social Costs (includes fuel savings)
Table 9.1-4 shows how total social costs are expected to be shared across stakeholders, for
selected years. According to these results, consumers in the Marine SI market are expected to
bear approximately 76 percent of the cost of the Marine SI program. This is expected to be offset
by the fuel savings. Vessel manufacturers are expected to bear about 17 percent of that program,
and engine manufacturers the remaining 6 percent. In the Small SI market, consumers are
expected to bear 91 percent of the cost of the Small SI program. This will also be offset by the
fuel savings. Equipment manufacturers are expected to bear about 7 percent of that program, and
engine manufacturers the remaining 2 percent. The estimated percentage changes in surplus are
the same for all years because the initial equilibrium conditions are shocked by both fixed and
variable costs; producers would pass the fixed costs to consumers at the same rate as the variable
costs.
9-12
-------
Economic Impact Analysis
Table 9.1-4: Summary of Estimated Social Costs for 2014, 2018, 2030 (2005$, Smillion)
Market
Absolute Change
in Surplus
Percent Change in
Surplus
Fuel
Savings
total Change in
Surplus
2014
Marine SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
Small SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
TOTAL
-$10.5
-$29.7
-$130.0
-$170.2
-$5.4
-$18.1
-$250.2
-$273.6
-$443.8
6%
17%
76%
2%
7%
91%
$45.4
$70.4
$115.9
-$10.5
-$29.7
-$84.6
-$124.8
-$5.4
-$18.1
-$179.7
-$203.2
-$327.9
2018
Marine SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
Small SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
TOTAL
-$8.7
-$25.0
-$108.2
-$142.0
-$5.0
-$16.9
-$235.0
-$256.8
-$398.8
6%
18%
76%
2%
7%
91%
-
$82.7
$95.6
$178.3
-$8.7
-$25.0
-$25.6
-$59.3
-$5.0
-$16.9
-$139.4
-$161.2
-$220.5
2030
Marine SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
Small SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
TOTAL
-$9.4
-$27.1
-$117.2
-$153.7
-$5.9
-$20.0
-$278.9
-$304.9
-$458.6
6%
18%
76%
2%
7%
91%
$152.9
$120.1
$273.0
-$9.4
-$27.1
$35.8
-$0.8
-$5.9
-$20.0
-$158.8
-$184.8
-$185.5
Table 9.1-5 contains more detailed information on the sources of the social costs for 2014.
This table shows that engines and equipment manufacturers are expected to bear more of the
burden of the program than end users. The loss of producer surplus for the small SI equipment
and vessel manufacturers has two sources. First, they would bear part of the burden of the
equipment costs. Second, they would also bear part of the engine costs, which are passed on to
vessel manufacturers in the form of higher engine prices. In comparing with small SI equipment
manufactures, marine SI vessel manufacturers would be able to pass along a relatively smaller
share of compliance costs to end consumers due to the elastic price elasticity of demand for
consumers of these vessels. As indicated in Table 9.3-22, the price elasticity of small SI
equipment demand is inelastic while the price elasticity of vessel demand is very elastic.
9-13
-------
Final Regulatory Impact Analysis
Table 9.1-5: Estimated Surplus Changes by Market and Stakeholder for 2014
(2005$, Smillion)
Scenario
Engine Manufacturers
Equipment
Manufacturers
Engine Price
Changes
Equipment Cost
Changes
End User
(Households)
Engine Price
Changes
Equipment Price
Changes
Subtotal
Small SI
Engine
Manufacturers
Equipment
Manufacturers
Engine Price
Changes
Equipment Cost
Changes
End User
(Households)
Engine Price
Changes
Equipment Cost
Changes
Subtotal
TOTAL
Engineering
Compliance Producer Consumer Total Fuel
Costs Surplus Surplus Surplus Savings
Marine SI
$118.7 -$10.5 -$10.5
$55.1 -$29.7 -$29.7
-$13.2
-$16.5
-$130.0 -$130.0 $45.4
-$93.6
-$36.4
$173.8 -$40.2 -$130.0 -$170.2 $45.4
$227.2 -$5.4 -$5.4
$49.0 -$18.1 -$18.1
-$13.0
-$5.1
-$250.1 -$250.1 $70.4
-$206.6
-$43.5
$276.2 -$23.6 -$250.1 -$273.6 $70.4
$450.0 -$63.7 -$380.1 -$443.8 $115.9
Net
Surplus
-$10.5
-$29. 7
-$84.6
-$124.8
-$5.4
-$18.1
-$179.6
-$203.2
-$327.9
The present value of net social costs of the final standards through 2037 at a 3 percent
discount rate, shown in Table 9.1-6, is estimated to be $4.2 billion, taking the fuel savings into
account. We also performed an analysis using a 7 percent social discount rate. Using that
discount rate, the present value of the net social costs through 2037 is estimated to be $2.7 billion,
including the fuel savings.
9-14
-------
Economic Impact Analysis
Table 9.1-6. Estimated Net Social Costs Through 2037 by Stakeholder (2005$, Smillion)
Market
Marine SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
Small SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
TOTAL
Marine SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
Small SI
Engine Manufacturers
Equipment Manufacturers
End User (Households)
Subtotal
TOTAL
Total Change in Percentage Change
Surplus in Total Surplus
Net
-$167.0
-$474.5
-$2,079.0
-$2,720.5
-$94.1
-$329.9
-$4,472.1
-$4,896.1
-$7,616.6
Net
-$100.8
-$285.2
-$1,257.1
-$1,643.2
-$54.8
-$195.4
-$2,612.8
-$2,863.0
-$4,506.2
Present Value 3%
6%
17%
76%
2%
7%
91%
Present Value 7%
6%
17%
77%
2%
7%
91%
Fuel Savings
$1,730.8
$1,730.8
$1,643.8
$1,643.8
$3,374.6
$881.0
$881.0
$893.8
$893.8
$1,774.7
Net Change in
Surplus
-$167.0
-$474.5
-$348.1
-$989.6
-$94.1
-$329.9
-$2,828.3
-$3,252.3
-$4,242.0
-$100.8
-$285.2
-$376.1
-$762.2
-$54.8
-$195.4
-$1,719.1
-$1,969.2
-$2,731.4
9-15
-------
Final Regulatory Impact Analysis
9.2 Economic Methodology
Economic impact analysis uses a combination of theory and econometric modeling to
evaluate potential behavior changes associated with a new regulatory program. As noted above,
the goal is to estimate the impact of the regulatory program on producers and consumers. This is
done by creating a mathematical model based on economic theory and populating the model using
publically available price and quantity data. A key factor in this type of analysis is the
responsiveness of the quantity of engines and equipment demanded by consumers or supplied by
producers to a change in the price of that product. This relationship is called the elasticity of
demand or supply.
The EIM's methodology is rooted in applied microeconomic theory and was developed
following the OAQPS Economic Analysis Resource Document (EPA 1999). This section
discusses the economic theory underlying the modeling for this EIA and several key issues that
affect the way the model was developed.
9.2.1 Behavioral Economic Models
Models incorporating different levels of economic decision making can generally be
categorized as w/Y/z-behavior responses or without-behavior responses. The EIM is a behavioral
model.
Engineering cost analysis is an example of the latter and provides detailed estimates of the
cost of a regulation based on the projected number of affected units and engineering estimates of
the annualized costs. The result is an estimate of the total compliance costs for a program.
However, these models do not attempt to estimate how a regulatory program will change the prices
or output of an affected industry. Therefore, the results may over-estimate the total costs of a
program because they do not take decreases in quantity produced into account.
The w/Y/z-behavior response approach builds on the engineering cost analysis and
incorporates economic theory related to producer and consumer behavior to estimate changes in
market conditions. As Bingham and Fox (1999) note, this framework provides "a richer story" of
the expected distribution of economic welfare changes across producers and consumers. In
behavioral models, manufacturers of goods affected by a regulation are economic agents that can
make adjustments, such as changing production rates or altering input mixes, that will generally
affect the market environment in which they operate. As producers change their production levels
in response to a new regulation, consumers of the affected goods are typically faced with changes
in prices that cause them to alter the quantity that they are willing to purchase. These changes in
price and output resulting from the market adjustments are used to estimate the distribution of
social costs between consumers and producers.
If markets are competitive and per-unit regulatory costs are small, the behavioral approach
will yield approximately the same total cost impact as the engineering cost approach. However,
the advantage of the w/Y/z-behavior response approach is that it illustrates how the costs flow
9-16
-------
Economic Impact Analysis
through the economic system and it identifies which stakeholders, producers, and consumers are
most likely to be affected.
9.2.2 What Is the Economic Theory Underlying the EIM?
The EIM is a multi-market partial-equilibrium numerical simulation model that estimates
price and quantity changes in the intermediate run under competitive market conditions. Each of
these model features is described in this section.
9.2.2.1 Partial Equilibrium Multi-Market Model
In the broadest sense, all markets are directly or indirectly linked in the economy, and a
new regulatory program will theoretically affect all commodities and markets to some extent.
However, not all regulatory programs have noticeable impacts on all markets. For example, a
regulation that imposes significant per unit compliance costs on an important manufacturing input,
such as steel, will have a larger impact on the national economy. A regulation that imposes a
small direct compliance cost on an important input, or any direct compliance costs on an input that
is only a small share of production costs, would be expected to have less of an impact on all
markets in the economy.
The appropriate level of market interactions to be included in an economic impact analysis
is determined by the number of industries directly affected by the requirements and the ability of
affected firms to pass along the regulatory costs in the form of higher prices. There are at least
three alternative approaches for modeling interactions between economic sectors, that reflect three
different levels of analysis.
In ^partial equilibrium model, individual markets are modeled in isolation. The only
factor affecting the market is the cost of the regulation on facilities in the industry being modeled;
there are no interaction effects with other markets. Conditions in other markets are assumed either
to be unaffected by a policy or unimportant for cost estimation.
In a multi-market model, a subset of related markets is modeled together, with sector
linkages, and hence selected interaction effects, explicitly specified. This approach represents an
intermediate step between a simple, single-market partial equilibrium approach and a full general
equilibrium approach. This technique has most recently been referred to in the literature as
"partial equilibrium analysis of multiple markets" (Berck and Hoffmann, 2002).
In a general equilibrium model, all sectors of the economy are modeled together,
incorporating interaction effects between all sectors included in the model. General equilibrium
models operationalize neoclassical microeconomic theory by modeling not only the direct effects
of control costs but also potential input substitution effects, changes in production levels
associated with changes in market prices across all sectors, and the associated changes in welfare
economy-wide. A disadvantage of general equilibrium modeling is that substantial time and
resources are required to develop a new model or tailor an existing model for analyzing regulatory
alternatives.
9-17
-------
Final Regulatory Impact Analysis
This analysis uses a partial equilibrium approach in that it models only those markets that
are directly affected by the final emission control program: the Small SI and Marine SI markets.
In addition, these markets are modeled separately. This approach is appropriate because the Small
SI and Marine SI sector represent different activities (residential garden care and personal
recreation), and production and consumption of one is not affected by the other. In other words,
an increase in the price of lawnmowers is not expected to have an impact on the production and
supply of recreational marine vessels, and vice versa. Production and consumption of these
products are the result of other factors that have little cross-over impacts.
The EIM uses a single-market approach for some sectors (Small SI handheld, Class I
nonhandheld, personal watercraft, outboards recreational) and a two-market approach for the
others (Small SI Class II nonhandheld; sterndrive/inboards; and outboards luxury) reflecting
whether the markets are integrated and whether the controls affect only engines or both engines
and equipment. The advantage of a two-market approach is that it allows us to describe the
expected distribution of the program's effects across equipment and engine markets as well as the
effects on purchasers of these engines and equipment. To simulate these relationships, the EIM
consists of a series of standard partial equilibrium models that are linked through interactions
between the equipment and engine markets. As a result, the model estimates changes in prices and
quantities across all markets simultaneously for each of the linked engine and equipment markets.
The EIM does not specifically estimate potential price and quantity impacts on final goods
and services that may be produced by equipment that would be subject to the final controls in the
agricultural and construction sectors. This is appropriate because the vast majority of engines and
equipment that would be subject to the final standards are purchased for residential use
(recreational marine; home lawn and garden and residential utility uses; see Section 9.3 and the
industry characterization prepared for this rule). Not only is the share of commercial users of this
equipment small, but such equipment represents only a small portion of the total production costs
for application markets such as agriculture, construction or manufacturing. The final standards
would affect only a very small part of total inputs for those markets and would not be expected to
result in an adverse impact on output and prices of goods produced in these commercial
application sectors.
It should also be noted that the economic impact model employed for this analysis
estimates the market-level economic impacts of the rule. It is not a firm-level analysis and
therefore the impact for any particular manufacturer may be greater or less than the average impact
for the market as a whole. This difference can be important, particularly where the rule affects
different firms' costs over different volumes of production. However, to the extent there are
differential effects, EPA believes that the wide array of flexibilities provided in this rule are
adequate to address any cost inequities that are likely to arise.
9.2.2.2 Competitive Market Structure Model
In a market oriented economic analysis, the analyst must determine the market structure
according to most appropriate characteristics of the market under study. This market structure will
form the basis of the economic impact model and determine the economic relationship to be
9-18
-------
Economic Impact Analysis
reflected in the model. There are several types of market structures in economics: perfect
competition, oligopoly, monopolistic competition, and monopoly. The typical economic impact
analysis assumes a competitive market structure, although circumstance may require relaxing this
assumption.3
The assumption of a competitive market is not about the number of firms in a market. It is
about wether producers in the market are price takers or wether they have sufficient market power
to influence the market price. In a competitive market, producers are price takers. Indicators of a
competitive market include absence of barriers to entry, absence of strategic behavior among firms
in the market, and product differentiation.4 In addition, according to contestable market theory,
oligopolies and even monopolies will behave very much like firms in a competitive market if it is
possible to enter particular markets costlessly (i.e., there are no sunk costs associated with market
entry or exit). This would be the case, for example, when products are substantially similar.
In imperfectly competitive markets, producers have some ability to influence the market
price of output they produce. One of the classic reasons firms may be able to do this is their ability
to produce commodities with unique attributes that differentiate them from competitors' products.
This allows them to limit supply, which in turn increases the market price, given the traditional
downward-sloping demand curve. Decreasing the quantity produced increases the monopolist's
profits but decreases total social surplus because a less than optimal amount of the product is being
consumed. In the monopolistic equilibrium, the value society (consumers) places on the marginal
product exceeds the marginal cost to society (producers) of producing the last unit. Thus, social
welfare would be increased by inducing the monopolist to increase production. Social cost
estimates associated with a final regulation are larger with monopolistic market structures and
other forms of imperfect competition because the regulation exacerbates the existing social
inefficiency of too little output from a social perspective. The Office of Management and Budget
(OMB) explicitly mentions the need to consider these market power-related welfare costs in
evaluating regulations under Executive Order 12866 (OMB, 1996).
This EIA is based on a competitive market structure. This is appropriate because the
markets under analysis do not exhibit evidence of noncompetitive behavior: there are no
indications of barriers to entry, the firms in these markets are not price setters, and there is no
evidence of high levels of strategic behavior in the price and quantity decisions of the firms.
As described in the industry profiles for this final regulation (RTI, 2004), several of the
recreational marine and Small SI sectors are highly concentrated and thus have the potential for the
emergence of imperfect competition and price-setting behavior. Nonetheless, our analysis
suggests that mitigating factors will limit this potential for raising price above marginal cost and
thus that the assumption of a competitive market structure is justified. Among the mitigating
factors are the presence of substantial import competition, relative ease of entry, existing excess
3U.S. EPA. 2000. Guidelines for Preparing Economics Analyses, EPA-240-R-00-003, page 126; 1999.
OAQPS Economic Analysis Resource Document, page 5-8.
4The number of firms in a market is not a necessary condition for a perfectly competitive market. See
Robert H. Frank, Microeconomics and Behavior, 1991, McGraw-Hill, Incl., p. 33.
9-19
-------
Final Regulatory Impact Analysis
production capacity, and a historical tendency of market participants to compete on price. These
markets are also mature markets, as evidenced by unit sales growing at the rate of population
increases. Pricing power in such markets is typically limited, and empirical data indicates that
price pressure has existed in these markets for years and firms in these markets are price takers.5
In addition, the products produced within each market are somewhat homogeneous in that engines
and equipment from one firm can be purchased instead of engines and equipment from another
firm, enhancing competition.
According to contestable market theory, oligopolies and even monopolies will behave very
much like firms in a competitive market if it is possible to enter particular markets costlessly (i.e.,
there are no sunk costs associated with market entry or exit). This is the case with these markets
as there is significant excess production capacity in both the Small SI and Marine SI industries, in
part due to improved productivity and efficiency in current plants. Data on domestic plant
capacity utilization rates are published by the U.S. Census (U.S. Census, 2005). The full
production capability is defined as "the maximum level of production that an establishment could
reasonably expect to attain under normal and realistic operating conditions fully utilizing the
machinery and equipment in place." Recent domestic data for 2000 to 2004 indicate the internal
combustion engine industry (NAICS 333618 Other Equipment Manufacturing) operated at 53 to
73 percent of full production capability. Similar data for vessels (NAICS 336612 Boat Building)
indicate this industry operated between 59 and 62 percent of full production capability. The small
SI equipment industry (NAICS 333112, lawn & garden tractor and home & lawn garden
equipment manufacturing) operated at 50 to 65 percent of full production capability. Idle
production capacity also limits the ability of firms to raise prices, since competitors can easily
capture market share by increasing their production at the expense of a producer that increases its
prices.
Finally, domestic producers face substantial competition from foreign manufacturers (RTI,
2006). These overseas firms may have strong incentives to compete vigorously on price with the
well-established U.S. firms. For all of these reasons it is appropriate to use a competitive market
structure model to estimate the economic impacts of this proposal.
9.2.2.3 Intermediate-Run Model
In developing the multi-market partial equilibrium model, the choices available to
producers must be considered. For example, are producers able to increase their factors of
production (e.g., increase production capacity) or alter their production mix (e.g., substitution
between materials, labor, and capital)? These modeling issues are largely dependent on the time
horizon for which the analysis is performed. Three benchmark time horizons are discussed below:
the very short run, the long run, and the intermediate run. This discussion relies in large part on
the material contained in the OAQPSEconomic Analysis Resource Guide (U.S. EPA, 1999).
5 RTI (2006). Historical Market Data and Trends, Industry Profile for Small SI Engines and Equipment,
Section 2.5. Draft Report
9-20
-------
Economic Impact Analysis
The EIM models market impacts in the intermediate run. The use of the intermediate run
means that some factors of production are fixed and some are variable. This modeling period
allows analysis of the economic effects of the rule's compliance costs on current producers. As
described below, a short-run analysis imposes all compliance costs on producers, while a long-run
analysis imposes all costs on consumers. The use of the intermediate time frame is consistent with
economic practices for this type of analysis.
In the very short run, all factors of production are assumed to be fixed, leaving the directly
affected entity with no means to respond to increased costs associated with the regulation (e.g.,
they cannot adjust labor or capital inputs). Within a very short time horizon, regulated producers
are constrained in their ability to adjust inputs or outputs due to contractual, institutional, or other
factors and can be represented by a vertical supply curve, as shown in Figure 9.2-1. In essence,
this is equivalent to the nonbehavioral model described earlier. Neither the price nor quantity
changes and the manufacturer's compliance costs become fixed or sunk costs. Under this time
horizon, the impacts of the regulation fall entirely on the regulated entity. Producers incur the
entire regulatory burden as a one-to-one reduction in their profit. This is referred to as the
"full-cost absorption" scenario and is equivalent to the engineering cost estimates. Although there
is no hard and fast rule for determining what length of time constitutes the very short run, it is
inappropriate to use this time horizon for this analysis because it assumes economic entities have
no flexibility to adjust factors of production.
Figure 9.2-1: Short Run: All Costs Born by Producers
Price
Output
In the long run, all factors of production are variable, and producers can be expected to
adjust production plans in response to cost changes imposed by a regulation (e.g., using a different
labor/capital mix). Figure 9.2-2 illustrates a typical, if somewhat simplified, long-run industry
supply function. The function is horizontal, indicating that the marginal and average costs of
9-21
-------
Final Regulatory Impact Analysis
production are constant with respect to output.6 This horizontal slope reflects the fact that, under
long-run constant returns to scale, technology and input prices ultimately determine the market
price, not the level of output in the market.
Market demand is represented by the standard downward-sloping curve. The market is
assumed here to be competitive; equilibrium is determined by the intersection of the supply and
demand curves. In this case, the upward shift in the market supply curve represents the
regulation's effect on production costs. The shift causes the market price to increase by the full
amount of the per-unit control cost (i.e., from P to P'). With the quantity demanded sensitive to
price, the increase in market price leads to a reduction in output in the new with-regulation
equilibrium (i.e., Q to Q'). As a result, consumers incur the entire regulatory burden as
represented by the loss in consumer surplus (i.e., the area P ac P'). In the nomenclature of EIAs,
this long-run scenario is typically referred to as "full-cost pass-through" and is illustrated in Figure
9.2-2.
Figure 9.2-2: Long Run: Full-Cost Pass-Through
$ d
- With Regulation
Price / 1
Increase^
P
^^
\ Unit Cost Increase
a ^
\SQ Witho
D
Q.
Output
Taken together, impacts modeled under the long-run/full-cost-pass-through scenario reveal
an important point: under fairly general economic conditions, a regulation's impact on producers is
transitory. Ultimately, the costs are passed on to consumers in the form of higher prices.
6 The constancy of marginal costs reflects an underlying assumption of constant returns to scale of
production, which may or may not apply in all cases.
9-22
-------
Economic Impact Analysis
However, this does not mean that the impacts of a regulation will have no impact on producers of
goods and services affected by a regulation. For example, the long run may cover the time taken
to retire all of today's capital vintage, which could take decades. Therefore, transitory impacts
could be protracted and could dominate long-run impacts in terms of present value. In addition, to
evaluate impacts on current producers, the long-run approach is not appropriate. Consequently a
time horizon that falls between the very short-run/full-cost-absorption case and the
long-run/full-cost-pass-through case is most appropriate for this EIA.
The intermediate run time frame allows examination of impacts of a regulatory program
during the transition between the short run and the long run. In the intermediate run, some factors
are fixed; some are variable. In other words, producers can adjust some, but not all, factors of
production, meaning they will bear some portion of the costs of the regulatory program. The
existence of fixed production factors generally leads to diminishing returns to those fixed factors.
This typically manifests itself in the form of a marginal cost (supply) function that rises with the
output rate, as shown in Figure 9.2-3.
Figure 9.2-3: Intermediate Run: Partial-Cost Pass-Through
Price
Increase
With Regulation
Cost Increase
Without Regulation
Output
Again, the regulation causes an upward shift in the supply function. The lack of resource
mobility may cause producers to suffer profit (producer surplus) losses in the face of regulation;
however, producers are able to pass through some of the associated costs to consumers, to the
extent the market will allow. As shown, in this case, the market-clearing process generates an
increase in price (from P to P') that is less than the per-unit increase in costs, so that the regulatory
9-23
-------
Final Regulatory Impact Analysis
burden is shared by producers (net reduction in profits) and consumers (rise in price). In other
words, there is a loss of both producer and consumer surplus.
Consistent with other economic impact analyses performed by EPA, this EIM uses an
intermediate run approach. This approach allows us to examine the market and social welfare
impacts of the program as producers adjust their output and consumers adjust their consumption of
affected products in response to the increased production costs. During this period, the
distribution of the welfare losses between producer and consumer depends in large part on the
relative supply and demand elasticity parameters used in the model. For example, if demand for
Small SI equipment is relatively inelastic (i.e., demand does not decrease much as price increases),
then most of the direct compliance cost on refiners will be passed along to Small SI equipment
consumers in the form of higher prices.
9.2.3 How is the EIM Used to Estimate Economic Impacts?
9.2.3.1 Estimation of Market Impacts (Single Market)
A graphical representation of a general economic competitive model of price formation, as
shown in Figure 9.2-4(a), posits that market prices and quantities are determined by the
intersection of the market supply and market demand curves. Under the baseline scenario, a
market price and quantity (p,Q) are determined by the intersection of the downward-sloping
market demand curve (DM) and the upward-sloping market supply curve (SM). The market supply
curve reflects the sum of the domestic (Sd) and import (S;) supply curves.
9-24
-------
Economic Impact Analysis
Figure 9.2-4: Market Equilibrium without and with Regulation
= P
Q
Domestic Supply
Foreign Supply
a) Baseline Equilibrium
Market
P'
P
P'
P
Q' Q
Domestic Supply
Foreign Supply
b) With-Regulation Equilibrium
Market
With the regulation, the costs of production increase for suppliers. The imposition of these
regulatory control costs is represented as an upward shift in the supply curve for domestic and
import supply by the estimated compliance costs. As a result of the upward shift in the supply
curve, the market supply curve will also shift upward as shown in Figure 9.2-3(b) to reflect the
increased costs of production.
At baseline without the final rule, the industry produces total output, Q, at price, p, with
domestic producers supplying the amount qd and imports accounting for Q minus qd, or qf. With
the regulation, the market price increases from p to p', and market output (as determined from the
market demand curve) decreases from Q to Q'. This reduction in market output is the net result of
reductions in domestic and import supply.
9-25
-------
Final Regulatory Impact Analysis
As indicated in Figure 9.2-4, when the final standards are applied the supply curve will
shift upward by the amount of the estimated compliance costs. The demand curve, however, does
not shift in this analysis. This is explained by the dynamics underlying the demand curve. The
demand curve represents the relationship between prices and quantity demanded. Changes in
prices lead to changes in the quantity demanded and are illustrated by movements along a fixed
demand curve. In contrast, changes in any of the other variables would lead to change in demand
and are illustrated as shifts in the position of the demand curve.7 For example, an increase in the
number of consumers in a market would cause the demand curve to shift outward because there are
more individuals willing to buy the good at every price. Similarly, an exogenous increase in
nominal income would also lead the demand curve to shift outward as people choose to buy more
of a good at a given price. Changes in the prices of related good and tastes or preferences can also
lead to demand curve shifts.
The final standards are expected to increase the costs of production in the Small SI engine
and equipment and Marine SI engine vessel markets and ultimately lead to higher equilibrium
prices in the affected markets. As these prices increase, the quantity demanded falls (i.e., the price
change leads to a movement along the demand curve).8 However, the final program is not
expected to lead to shifts in the demand curve for several reasons. First, the assume the program
will not directly influence prices of related goods (i.e., prices of any potential substitutes remain
constant in the analysis). In addition, the program will not change nominal incomes through
public finance mechanisms (e.g., lump sum subsidies/taxes) or change labor supply decisions.
Finally, we assume tastes and preference will not change during the period of analysis. For all of
these reasons, it would be inappropriate to shift the demand curve for this analysis.
9.2.3.2 Incorporating Multi-Market Interactions
The above description is typical of the expected market effects for a single product markets
(e.g., Small SI handheld and Class I nonhandheld; personal watercraft) considered in isolation.
However, several of the markets considered in this EIA are more complicated because the engine
and equipment manufacturers are not integrated.
When both engine and equipment markets are considered separately, the regulatory
program will affect equipment producers in two ways. First, equipment producers are affected by
higher input costs (increases in the price of gasoline engines) associated with the rule. Second, the
standards will also impose additional production costs on equipment producers associated with
7 An accessible detailed discussion of these concepts can be found in Chapter 5-7 of Nicholson's (1998)
intermediate microeconomics textbook.
8 Nicholson (1998) provides an example of the effects of a price increase on the quantity consumed (p: 134-
135). Throughout this discussion, we use uncompensated Marshallian demand functions. As a result, a price
increase will also change an individual's "real" income and reinforce substitution quantity responses to a good's
price change through an "income" effect. Both substitution and (real) income effects are therefore built in the
Marshallian demand function used for this analysis. It is important to note, however, that this type of "income"
effect is conceptually different from an exogenous change in nominal income that leads to a shift in a demand
function.
9-26
-------
Economic Impact Analysis
equipment changes necessary to accommodate changes in engine design. In the sections that
follow, we describe the demand relationships between these markets and how they are
incorporated in the economic model.
In markets such as Class II nonhandheld or SD/I marine, the demand for engines is directly
linked to the production of equipment or vessels that uses those engines.9 This means that it is
reasonable to assume that the input-output relationship between the gasoline engines and the
equipment is strictly fixed and that the demand for engines varies directly with the demand for
equipment.10 A demand curve specified in terms of its downstream consumption is referred to as a
derived demand curve. Figure 9.2-5 illustrates how a derived demand curve is identified.
9 In marine applications, one or two engines are used per boat, depending on its intrinsic design, and this
configuration is insensitive to small changes in engine used. In the case of Small SI equipment, the one-to-one
correspondence is exact. Furthermore, there is no potential for technical substitution, i.e., to make gasoline
equipment one needs a gasoline engine.
10 This one-to-one relationship holds for engines sold on the market and for engines consumed internally by
integrated engine/equipment manufacturers.
9-27
-------
Final Regulatory Impact Analysis
Figure 9.2-5: Derived Demand for Engines
Price
Equipment
($/Q)
APC
AQr
Q - Equipment
Price
Engines
($/Q)
t
AP
eng
eng
Unit Cost Increase
Derived
Demand
AQ
Q - Engines
eng
= AQ
eng
Consider an event in the marine equipment market that causes the price of equipment to
increase by AP (such as an increase in the price of engines). This increase in the price of
equipment will cause the supply curve in the equipment market to shift up, leading to a decreased
quantity (AQE). The change in equipment production leads to a decrease in the demand for
engines (AQEng). The new point (QE - AQE, P - AP) traces out the derived demand curve. Note
that the supply and demand curves in the marine equipment markets are needed to identify the
derived demand in the engine market. All of the market supply and demand curves and the
elasticity parameters used in the EEVI are described in Appendix 9E
9-28
-------
Economic Impact Analysis
9.2.3.3 Estimation of Social Costs
The economic welfare implications of the market price and output changes with the
regulation can be examined by calculating consumer and producer net "surplus" changes
associated with these adjustments. This is a measure of the negative impact of an environmental
policy change and is commonly referred to as the "social cost" of a regulation. It is important to
emphasize that this measure does not include the benefits that occur outside of the market, that is,
the value of the reduced levels of air pollution with the regulation. Including this benefit will
reduce the net cost of the regulation and even make it positive.
The demand and supply curves that are used to project market price and quantity impacts
can be used to estimate the change in consumer, producer, and total surplus or social cost of the
regulation (see Figure 9.2-6).
The difference between the maximum price consumers are willing to pay for a good and
the price they actually pay is referred to as "consumer surplus." Consumer surplus is measured as
the area under the demand curve and above the price of the product. Similarly, the difference
between the minimum price producers are willing to accept for a good and the price they actually
receive is referred to as "producer surplus." Producer surplus is measured as the area above the
supply curve below the price of the product. These areas can be thought of as consumers' net
benefits of consumption and producers' net benefits of production, respectively.
In Figure 9.2-6, baseline equilibrium occurs at the intersection of the demand curve, D, and
supply curve, S. Price is P{ with quantity Q,. The increased cost of production with the regulation
will cause the market supply curve to shift upward to S'. The new equilibrium price of the product
is P2. With a higher price for the product there is less consumer welfare, all else being unchanged.
In Figure 9.2-6(a), area A represents the dollar value of the annual net loss in consumers' welfare
associated with the increased price. The rectangular portion represents the loss in consumer
surplus on the quantity still consumed due to the price increase, Q2, while the triangular area
represents the foregone surplus resulting from the reduced quantity consumed, Q[ - Q2.
In addition to the changes in consumers' welfare, there are also changes in producers'
welfare with the regulatory action. With the increase in market price, producers receive higher
revenues on the quantity still purchased, Q2. In Figure 9.2-6(b), area B represents the increase in
revenues due to this increase in price. The difference in the area under the supply curve up to the
original market price, area C, measures the loss in producer surplus, which includes the loss
associated with the quantity no longer produced. The net change in producers' welfare is
represented by area B - C.
9-29
-------
Final Regulatory Impact Analysis
Figure 9.2-6: Market Surplus Changes with Regulations
Consumer and Producer Surplus
$/Q
_l_
Q2 Q,
(a) Change in Consumer Surplus with
Regulation
Q/t
$/Q
Q2 Q1
(b) Change in Producer Surplus with
Regulation
Q/t
$/Q
Q2 Q1
(c) Net Change in Economic Welfare with
Regulation
Q/t
9-30
-------
Economic Impact Analysis
The change in economic welfare attributable to the compliance costs of the regulations is
the sum of consumer and producer surplus changes, that is, -(A) + (B-C). Figure 9.2-6(c) shows
the net (negative) change in economic welfare associated with the regulation as area D.
9.2.4 How Are Special Market Characteristics Addressed?
In addition to the general model features described in Section 9.2.2, there are several
specific characteristics of the Small SI and Marine SI markets that need to be addressed in the
EEVI. These are the treatment of fixed and variable costs, fuel savings, programmatic flexibilities,
and substitution, and distribution systems effects.
9.2.4.1 Fixed and Variable Costs in a Competitive Market
The estimated engineering compliance costs, consisting of fixed costs (R&D,
capital/tooling, certification costs), variable costs, and operating costs provide an initial measure of
total annual compliance costs without accounting for behavioral responses. The starting point for
assessing the market impacts of a regulatory action is to incorporate the regulatory compliance
costs into the production decision of the firm.
In general, shifting the supply curve by the total cost per unit implies that both capital and
operating costs vary with output levels. At least in the case of capital, this raises some questions.
In the long run, all inputs (and their costs) can be expected to vary with output. But a short(er)-run
analysis typically holds some capital factors fixed. For instance, to the extent that a market supply
function is tied to existing facilities, there is an element of fixed capital (or one-time R&D). As
indicated above, the current market supply function might reflect these fixed factors with an
upward slope. As shown in Figure 9.2-7, the marginal cost (MC) curve will only be affected, or
shift upwards, by the per-unit variable compliance costs (cj= TVCC/q), while the average total
cost (ATAC) curve will shift up by the per-unit total compliance costs (c2 = TCC/q). Thus, the
variable costs will directly affect the production decision (optimal output rate), and the fixed costs
will affect the closure decision by establishing a new higher reservation price for the firm (i.e.,
Pm). In other words, the fixed costs are important in determining whether the firm will stay in this
line of business (i.e., produce anything at all), and the variable costs determine the level (quantity)
of production.
9-31
-------
Final Regulatory Impact Analysis
$/q
Figure 9.2-7: Modeling Fixed Costs
' MC'
MC
Dm'
TVCC
TCC
q/t
(a) Upward-sloping supply function
Depending on the industry type, fixed costs associated with complying with a new
regulation can generally be treated differently in an analysis of market impacts. In a competitive
market, the industry supply curve is generally based on the market's marginal cost curve; fixed
costs do not influence production decisions at the margin. Therefore, the market anlaysis for a
competitive market is based on variable costs only.
The nature of the Small SI and Marine SI markets suggests the market supply curve shifts
in the model should include fixed compliance cost and variable compliance cost. This is because
Small SI and Marine SI engine and equipment manufacturers produce a product that changes very
little over time. These manufacturers may not engage in research and development to improve
their products on a continuous basis (as opposed to highway vehicles or nonroad engines and
equipment). In this case, the product changes that would be required to comply with the final
standards would require these manufacturers to devote new funds and resources to product
redesign and facilities changes. In this situation, Small SI and Marine SI engine and equipment
manufacturers would be expected to increase their prices in attempting to recover both fixed and
variable costs. This is in contrast to the nonroad diesel engine and equipment markets:
manufacturers in those markets generally allocate redesign resources each year to accommodate a
changing market. As stated in the section 9.3.3, this analysis applied fixed costs in the year in
which they occur prior to the rule taking effect, and variable costs during the years that the rule is
implemented. To reflect these conditions, the supply shift in this EIM is based on either fixed costs
prior to the rule or variable costs after the rule starts, even though the model assumes a competitive
market structure.
9-32
-------
Economic Impact Analysis
9.2.4.2 Fuel Savings and Fuel Taxes
If all the costs of the regulation are not reflected in the supply shift, then the producer and
consumer surplus changes reflected in Figure 9.2-6(c) will not capture the total social costs of the
regulation. This will be the case, for example, if there are cost savings attributable to a program
that are not readily apparent to consumers.
In this case, the final evaporative and exhaust controls are expected to result in fuel savings
for users. Small SI engine and equipment manufacturers are expected to use fuel injection
techniques to comply with the final standards for some of their two-cylinder Class II engines.
These fuel injected engines are expected to have better fuel efficiency than carbureted engines.
Marine SI manufacturers are expected to use 4-stroke and direct on-injection 2-stroke technology
for outboards and PWC. In addition, all sterndrive and inboard engines are expected to use fuel
injection. These technologies are expected to result in reductions in fuel consumption.
These fuel savings are not included in the market analysis for this economic impact
analysis. This is because all available evidence suggests that fuel savings do not affect consumer
decisions with respect to the purchase of this equipment. Unlike motor vehicles or other consumer
goods, neither Small SI nor Marine SI equipment is labeled with expected fuel consumption or
expected annual operating costs. Therefore, there is no information available for the consumer to
use or make this decision. Instead consumers base their purchase decision on other attributes of
the product for which the manufacturer provides information. For lawn mowers this may be the
horsepower of the engine, whether the machine has a bag or has a mulching feature, its blade size,
etc. For PWC it may be how many people it can carry, its maximum speed, its horsepower, etc. In
many cases, especially for Small SI equipment, the consumer may not even be aware of the fuel
savings when operating the equipment, especially if he or she uses the same portable fuel storage
container to fuel several different pieces of equipment.
These fuel savings are included in the social cost analysis. This is because they are savings
that accrue to society. These savings are attributed to consumers of the relevant equipment. As
explained in more detail in 9.3.5, the social cost analysis is based on the equivalent of the pre-tax
price of gasoline in that analysis. Although the consumer will realize a savings equal to the pump
price of gasoline (post-tax), part of that savings is offset by a tax loss to governmental agencies
and is thus a loss to consumers of the services supported by those taxes. This tax revenue loss,
considered a transfer payment in this analysis, does not affect the benefit-cost analysis results.
9.2.4.3 Flexibility Provisions
Consistent with the engineering cost estimates, the EIM does not include cost savings
associated with compliance flexibility provisions or averaging, banking, and trading provisions.
As a result, the results of this EIA can be viewed as somewhat conservative.
9.2.4.4 Substitution
Gasoline-powered SI engines convert the potential energy contained in the fuel into
mechanical energy, which can then be used to do useful work, to provide locomotion, and/or to
9-33
-------
Final Regulatory Impact Analysis
generate electricity. These machines are technologically similar compression-ignition engines
powered by diesel fuel, and often compete in the same equipment and applications markets.
Similarly, electric motors are capable of performing many of the same tasks as gasoline engines in
small and inexpensive equipment.
The relationships modeled in the EIM do not include substitution away from Small SI and
Marine SI engines and equipment to diesel or electric alternatives. This is appropriate because
consumers are not likely to make these substitutions. Diesel engines' superior efficiency in energy
conversion makes them more attractive for large engines, and for those with long required service
lives, whether measured in operating hours or years of service. Gasoline-powered engines, on the
other hand, have lower initial cost, and utilization in garden or recreational activities is not high
enough for diesel fuel efficiency to overcome this gasoline advantage. On the SI marine side, the
current population of recreational boats is overwhelmingly powered by gasoline engines, even in
the large horsepower classes where diesel's superior efficiency would seem to provide significant
cost advantages, and gasoline engines are the prevalent choice for garden equipment and
residential generators. On the Small SI side, substitution to diesel is not a viable option for most
residential consumers, either because diesel equipment does not exist (e.g., diesel string trimmers)
or because there would be a large price premium that would discourage the use of diesel
equipment (e.g., diesel lawnmowers and diesel recreational marine vessels). In addition, most
households are not equipped to handle the additional fuel type and misfueling would carry a high
cost. Finally, the lack of a large infrastructure system already in place like the one supporting the
use of gasoline equipment for residential and recreational purposes, including refueling and
maintenance, represents a large barrier to substitution from gasoline to diesel equipment. With
regard to electric alternatives, the impact of substitution to electric for Small SI equipment (there
are no comparable options for Marine SI) is also expected to be negligible. Gasoline is the power
source of choice for small and inexpensive equipment due to its low initial cost. Gasoline
equipment is also inherently portable, which make them more attractive to competing electric
equipment that must be connected with a power grid or use batteries that require frequent
recharging. Data that would allow investigation of the details of this clear consumer preference
are not available, but it is reasonable to assume that increases in the cost of gasoline engines of the
magnitude associated with this program would not cause widespread substitution to diesel or
electric alternatives.
9.2.4.5 Distribution System Effects
The market interactions modeled in the EIM are those between producers and consumers of
the specified engines and equipment that use those engines. The EIM does not consider sales
distribution networks or how the regulated goods are sold to final consumers through wholesalers
and/or retailers. This is appropriate because the final regulatory program does not impose
additional costs on the distribution networks and those relationships are not expected to change as
a result of the standards.
In the case of Small SI equipment, however, concerns have been raised about the potential
for dominant retailers (big box stores such as Wal-Mart, Sears and K-Mart) to affect market
equilibria and the ability of manufacturers to pass along cost increases associated with new
9-34
-------
Economic Impact Analysis
emission control requirements. Specifically, some Small SI equipment manufacturers assert that
Big Box stores impose a price structure that would force them to absorb the compliance costs
associated with the final standards. They contend that this is a relatively new phenomenon for
their market and that EPA should consider these effects in the economic impact analysis for this
proposal.
Dominant retailers are a fairly well-understood sector of the consumer good distribution
network, especially with regard to clothing and household goods. These stores reduce product
prices by exerting important influences on relevant producers. Specifically, they discipline
markets by encouraging manufacturers to compete on price, and force inefficient firms to cut costs
or leave the market.
Dominant retailers may also prevent efficient producers from passing on extra increases in
fixed costs to consumers, including R&D costs associated with engine or equipment redesign. So,
for example, it may be the case that if a particular firm redesigns a lawnmower to produce more
power a dominant retailer may not choose to change its pricing structure to account for that
redesign. Nevertheless, the firm may still choose to incorporate the design change in the hope of
capturing a greater share of the market and/or improve its name recognition.
It is unlikely, however, that a dominant retailer could prevent firms from passing on
market-wide increases in average costs or marginal costs in response to a regulatory program.
Profit maximizing manufacturers will continue to follow a marginal cost equals price pricing rule
regardless of the distribution arrangements. A dominant retailer could not force the manufacturer
to produce units where the marginal cost exceeds the price. If large retail distributors attempted to
prevent efficient manufacturers from raising prices in response to the standards, manufacturers
would likely respond to a retailer's price pressure by reducing output. This would result in large
excess demand in the equipment market which would ultimately have to be satisfied through some
sort of arbitrage mechanism to a new higher equilibrium price.
An individual manufacturing company has little, if any, ability to pass on a cost increase if
it is the only entity affected by that cost increase. In such a case, retailers would clearly have an
incentive to purchase comparable engines or equipment that were not affected by the cost increase,
placing the affected firm at a competitive disadvantage and reducing its market share. However, in
this case all engine manufacturers will face increased average costs or marginal costs of
production associated with the regulatory program. Therefore, the program does not necessarily
put one engine manufacturer at a competitive disadvantage, although manufacturers that can more
easily accommodate the new requirements will likely see lower costs than those who cannot.
9.3 EIM Data Inputs and Model Solution
9-35
-------
Final Regulatory Impact Analysis
The EIM is a computer model comprised of a series of spreadsheet modules that simulate
the supply and demand characteristics of the markets under consideration. The model equations,
presented in Appendix D to this chapter, are based on the economic relationships described in
Section 9.2. The EIM analysis consists of four basic steps:
• Define the initial equilibrium conditions of the markets under consideration
(equilibrium prices and quantities and behavioral parameters; these yield
equilibrium supply and demand curves).
Introduce a policy "shock" into the model based on estimated compliance costs that
shift the supply functions.
• Use a solution algorithm to estimate a new, with-regulation equilibrium price and
quantity for all markets.
• Estimate the change in producer and consumer surplus in all markets included in
the model.
Supply responses and market adjustments can be conceptualized as an interactive process.
Producers facing increased production costs due to compliance are willing to supply smaller
quantities at the baseline price. This reduction in market supply leads to an increase in the market
price that all producers and consumers face, which leads to further responses by producers and
consumers and thus new market prices, and so on. The new with-regulation equilibrium reflects
the new market prices where total market supply equals market demand.
The remainder of this section describes the data used to construct the EIM: initial
equilibrium market conditions (equilibrium prices and quantities), compliance cost inputs, and
model elasticity parameters. Also included is a brief discussion of the analytical expression used
to estimate with-regulation market conditions.
9.3.1 Description of Product Markets
This EIM estimates the behavioral responses of the Small SI and Marine SI markets to the
cost of complying with the final emission control program. Each of these markets is very briefly
described below. More information can be found in the industry characterizations prepared for
this proposal (Chapter 1 and RTI 2006).
9.3.1.1 Small SI Market
The Small SI market is the market for a variety of nonroad equipment powered by two-
stroke or four-stroke spark-ignition engines rated up to 19 kW (25 hp). This economic impact
assessment distinguishes between two Small SI market sectors: handheld and nonhandheld. The
handheld (HH) sector consists generally of equipment that is carried by the operator and is
operated multipositionally, although some equipment in this category may have two wheels. HH
equipment includes string trimmers, edgers, leaf blowers, and chain saws. The nonhandheld
(NHH) sector consists mostly of wheeled equipment such as lawn mowers, garden tractors, and
wheeled trimmers, blowers, and edgers. Also included in the Small SI market are generators,
9-36
-------
Economic Impact Analysis
compressors, and construction, agricultural, and small industrial equipment, as well as some
recreational and utility vehicles and snowblowers.
The HH market can be characterized as an integrated market in which producers
manufacture both the engine and the associated equipment. In the NHH market, in contrast, the
engine and equipment manufacturers are typically separate entities. Engines produced by a
manufacturer for use in its own equipment are called "captive" engines. Engines produced by
manufacturers for sale on the open market to anyone who wants to buy them are called "merchant"
engines. This distinction is important because compliance costs affect captive and merchant
engines differently. Engine-related compliance costs for captive engines are absorbed into the
equipment costs of integrated suppliers in their entirety. In contrast, nonintegrated suppliers who
buy merchant engines absorb only part of the engine compliance costs into their equipment costs;
the rest is borne by the engine manufacturer. Depending on the price sensitivity of demand in the
engine market, the pass-through of engine compliance costs to the equipment manufacturer may be
larger (more inelastic demand) or smaller (more elastic demand).
This analysis makes the simplifying assumption that virtually all Small SI equipment is
sold to residential end-users for their personal use and a negligible number are sold to commercial
entities for use as an input to the production of goods or services. This simplifying assumption
allows us to disregard the impact of the compliance costs on the production of goods and services
that would have Small SI equipment as an input. Any such impacts would be expected to be
negligible given the relative share of Small SI equipment to any such production processes. This
assumption is supported by data from the Outdoor Power Equipment and Engine Service
Association (OPEESA), contained in Table 9.3-1, which indicates that only about 3 percent of the
NHH products sold in 2003 and 2004 were sold to commercial users. The rest, 97 percent, were
sold to residential users. While this data reflects only NHH equipment, a similar situation likely
exists for HH equipment given the nature of that equipment (light-duty lawn and garden equipment
or gensets). Recent EPA certification data also supports this simplifying assumption. According
to model year 2005 data, about 5 percent of Class I and 7 percent of Class II engines were high
hour useful life (commercial) categories, or a total of about 9 percent of Classes I and II combined.
About 19 percent of HH engines were high useful life categories.
Table 9.3-1: Share of Residential and Commercial Small SI Shipments (Various years)
Total Commercial Turf Products
Total Consumer NHH Products
Commercial Unit Volume NHH Share
HH products (assumed consumer)
Commercial share - all Small SI
2003
297,085
8,598,901
3.3%
12,600,440
1.4%
2004
234,475
8.188,614
2.8%
11,949,557
1.2%
Source: Outdoor Power Equipment & Engine Service Association, 2004.
The analysis also assumes that there is a one-to-one correspondence between engines and
equipment (there is only one engine per equipment unit) and that there is no market for loose
9-37
-------
Final Regulatory Impact Analysis
engines. These assumptions are reasonable given the nature of this equipment and because owners
generally do not repower this equipment when the engine fails; instead, they repair the engine or
replace the equipment. This assumption makes it possible to estimate the number of engines
produced directly from the number of equipment.
9.3.1.1.1 Handheld Market
The HH engine market consists of Class III (< 20 cc), IV (20-50 cc) and V (>50 cc)
engines. These engines are used in similar types of equipment, all of which are small and
relatively lightweight. According to the industry profile prepared for this rule, the HH market is
an integrated market in that about 90 percent of HH engines are "captive" engines, with the engine
and equipment manufacturer being the same company (RTI, 2006). An integrated market means
the EIM can use a one-market approach.
For the purpose of this analysis, all HH engines and equipment are grouped into one
engine/equipment market. This is reasonable both because it is an integrated market and because
the estimated compliance costs for the HH standards are expected to be similar for all types of HH
engines and equipment regardless of size or application. The final standards for HH consist only
of evaporative emission controls and the cost to comply with the standards are primarily related to
fuel tank volume and fuel hose length, which do not vary significantly for most equipment.
9.3.1.1.2 NonhandheldMarket
The NHH engine market consists of Class I (<225 cc) and Class II (>225 cc) engines.
There are three useful life categories for each and the costs for complying with the exhaust
standards will vary by useful life category for each engine class. According to the industry profile
prepared for this rule, the NHH market is not integrated in that about 95 percent of Class I and
Class II NHH engines are merchant engines (RTI, 2006). The model thus explores the impacts on
engine producers and equipment producers separately. This means it is necessary to use a two-
market approach, with the engine and equipment markets sharing some of the compliance costs
and consumers bearing the rest.
Snowblowers engines are treated differently under EPA's final program. The final
program would impose only evaporative controls on these engines. Because Class I manufacturers
of snowblower engines make the whole engine as a set (i.e., including fuel tank and fuel lines), it
was decided to place all of the compliance costs on the engine manufacturer. These manufacturers
are expected to produce a separate snowblower engine to be used in this equipment. Class II
engines are commonly sold without fuel tanks, and so the evaporative controls for Class II
snowblowers are attributed to the equipment manufacturer.
The nine Small SI nonhandheld engine markets are summarized in Table 9.3-2.
Table 9.3-2: Small SI Nonhandheld Engine Categories
9-38
-------
Economic Impact Analysis
Class
Class I
Class I - Snowblower
Class II
Useful Life
125 hours
250 hours
500 hours
125 hours
250 hours
500 hours
250 hours
500 hours
1000 hours
The EIM includes eight types of NHH equipment, as described in Table 9.3-3. However,
because not all engine/equipment combinations are applicable, there are a total of 40
engine/equipment markets. Specifically, there are no Class II lawnmowers, there are no Class I
tractors, and all equipment in the "other lawn and garden" category using Class I engines are in the
UL125 grouping.
Table 9.3-3: Nonhandheld Equipment Categories
Equipment
Agriculture/construction/general industrial
Utility and recreational vehicles
Lawn mowers
Tractors
Lawn and garden, other
Gensets/welders
Pumps/compressors/pressure washers
Snowblowers
Class I
Yes
Yes
Yes
No
UL125 only
Yes
Yes
Yes
Class II
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
9.3.1.2 Marine SI market
The Marine SI market is the market for a variety of marine vessels powered by gasoline
engines. These final Marine SI standards discussed here are for propulsion engines only.
Auxiliary Marine SI engines <37 kW are included as Small SI engines for this rule. Larger
auxiliary Marine SI engines were covered in the new standards for Large SI engines. Many of the
auxiliary Marine SI engines are being designed with catalysts independent of the final standards,
so the final standards will codify what is already happening in the industry and force new entrants
9-39
-------
Final Regulatory Impact Analysis
in the market to employ the same types of emission controls. Given that the industry is already
using catalysts, the estimated costs of complying are with the final standards are negligible. These
engines typically use the same fuel tank as the propulsion engines so evaporative emission controls
for these engines impose a nominal cost that is already covered in the vessel costs since the vessel
costs include costs for hoses and tanks. The impact of treating marine Auxiliary Marine SI
engines in this way are expected to be minimal because the number of vessels with installed
auxiliary units is small and limited to sterndrive/inboard and outboard luxury vessels: about 21,700
out of a total of 356,300 vessels.
9.3.1.2.1 Marine SI Engine Markets
Unlike Small SI engines that can be used in a variety of different types of equipment,
Marine SI engines are designed and manufactured for specific applications. Engines used in
sterndrive or inboard vessels are different from those used in outboard applications, and are made
by different manufacturers. Outboards and SD/I engines produced for luxury vessels are different
from those produced for the general market. Personal watercraft, on the other hand, are generally
an integrated system. Taking this into consideration, there are 13 engine markets included in this
EIA, based on design and horsepower. These are described in Table 9.3-4.
Table 9.3-4: Marine SI Engine Markets
Engine Design
SD/I Recreation
SD/I Luxury
OB Recreational
OB Luxury
OB Loose
<25hp
XXX
XXX
25-50 hp
XXX
5 1-100 hp
XXX
101-175 hp
XXX
XXX
XXX
176-300 hp
XXX
XXX
XXX
XXX
>301 hp
XXX
XXX
Similar to the Small SI market, most marine SI engines are used for recreational purposes.
According to a 2000 study of the boat building industry, about 79 percent of Marine SI vessels are
used for recreational purposes and only 7 percent for commercial purposes, with the remaining 14
percent for other purposes (CCA, 2000).n The propulsion system of choice for commercial
marine vessels is diesel due to its greater reliability and lower fuel costs. The combustion
characteristics of diesel engines also make them a better choice for vessels that are likely to spend
large amounts of time at sea. While gasoline marine engines are used in applications such as
lifeboats, patrol boats and small fishing vessels, their numbers are not large enough to warrant
separate consideration in this Economic Impact Analysis.
"This study looked at NAICS 336612-establishments primarily engaged in building boats, defined as
watercraft not built in shipyards and typically of the type suitable or intended for personal use; it is not clear what is
meant by "other" in this study.
9-40
-------
Economic Impact Analysis
For the purposes of this analysis, all personal watercraft manufacturers are considered to be
integrated manufacturers, and thus the engines are "captive." This is reasonable because personal
watercraft are similar to land-based recreational vehicles in that the engines are produced by the
equipment manufacturer specifically for certain models.
The other two primary types of SI marine engines are outboards and sterndrives/inboards
(SD/I). For these engines, we model a merchant relationship between the engine manufacturers
and boat builders. This is reasonable because these engines are typically sold on the open market
(outboards) or sold internally but through a market-type relationship between the engine and the
equipment businesses (SD/I).
Outboard engines are typically produced by the engine manufacturer with little or no
knowledge of what vessels the engines will be used on. Outboards are a self-contained assembly,
with a power unit and drive unit, that can be fit to a wide range of boats. They may be used either
with a portable fuel tank or connected to a fuel system installed on a vessel. In most cases, the
engine manufacturer and boat builder are separate companies. However, it is becoming more
common for engine manufacturing companies to purchase boat builders. Based on conversations
with engine manufacturers and boat builders, we have received indications that this trend has not
significantly changed the relationship between the engine business units and the boat building
business units. The boat builders typically pay market price for the engines and there is little
integration of design beyond a typical manufacturer/supplier relationship. It seems that engine
manufacturers generally buy outboard vessel building companies to gain access to target markets
rather than to develop an integrated design. Generally, the vessel is sold without the engine and
the consumer chooses the engine at the point of sale. This means that the vessel builder may not
be involved in the transaction and that the distribution of the compliance costs is between the
engine builder and the end consumer rather than between the engine builder and the vessel builder.
The relationship between engine manufacturers and boat builders is similar for SD/I
engines as for outboard engines. One difference is that there are only two large businesses and
many small businesses producing SD/I engines. These small businesses typically do not produce
boats or own companies that do. SD/I engines are often sold to buyer groups created by boat
builders to gain volume discounts on engines. Because of this, SD/I engine manufacturers often
do not know what boats their engines are being used in. In the case where a large SD/I
manufacturer has purchased boat building companies, the relationship is similar to that for
outboards. Nevertheless, the distribution of compliance costs would be between the engine
manufacturer and the vessel builder, since the engine is integrated in the final vessel design.
9.3.1.2.2 Marine SI Equipment Markets
There are five types of marine vessel markets:
SD/I recreational (runabouts, airboats, jetboats)
• SD/I luxury (yachts, cruisers offshore)
• OB recreational (runabouts, pontoons, fishing)
• OB luxury (yacht, cruiser, express fish)
9-41
-------
Final Regulatory Impact Analysis
Personal watercraft
Of the 30 possible engine/vessel combinations, there are 15 combinations that are not
applicable. For example, SD/I vessels use engines above 100 hp only. Personal watercraft use
engines above 50 hp but do not use engines above 300 hp. This yields a total of 15 engine/vessel
markets.
Table 9.3-5: Marine SI Vessel Types
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
<25hp
XXX
25-50 hp
XXX
5 1-100 hp
XXX
XXX
101-175 hp
XXX
XXX
XXX
XXX
176-300 hp
XXX
XXX
XXX
XXX
XXX
>301 hp
XXX
XXX
Unlike Small SI equipment, there is not a one-to-one relationship between engines and
equipment. Some vessels may have more than one propulsion engine. Table 9.3-6 shows the
average number of engines per vessel assumed for the purposes of this analysis. In this table, OB
engines per boat sale represents the average number of engines per outboard vessel in general.
This average consists of three components: 1) some outboard vessels have more than one engine;
2) engines that are made as replacement engines; and 3) loose engines that are not sold with the
boat, such as "kicker" engines which are used for low speed trolling.
Table 9.3-6: Average Number of Marine SI Engines per Vessel (2005)
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
OB Engine/boat sale
<25hp
1.25
25-50
hp
1.25
51-100
hp
1.00
1.29
101-175
hp
1.00
1.00
1.29
2.50
176-300
hp
1.00
1.02
1.25
1.29
2.50
>301 hp
1.01
1.52
Averag<
1.00
1.01
1.39
1.28
2.50
1.47
9.3.1.3 Market Linkages
In the ELM, the Small SI and Marine SI markets are not linked (there is no feedback
mechanism between the Small SI and Marine SI market segments). This is appropriate because
the affected equipment is not interchangeable and because there is very little overlap between the
engine producers in each market. These two sectors represent different aspects of economic
activity (lawn and garden care and power generation as opposed to recreational marine) and
9-42
-------
Economic Impact Analysis
production and consumption of one product is not affected by the other. In other words, an
increase in the price of lawnmowers is not expected to have an impact on the production and
supply of personal watercraft, and vice versa. Production and consumption of each of these
productions are the results of other factors that have little cross-over impacts (the need for
residential garden upkeep or power generation; the desire for personal recreation).
9.3.2 Market Equilibrium Conditions
The starting point for the economic impact analysis is initial market equilibrium conditions
(prices and quantities) that exist prior to the implementation of new standards. At pre-control
market equilibrium conditions, consumers are willing to purchase the same amount of a product
that producers are willing to produce at the market price.
9.3.2.1 Small SI Initial Equilibrium Quantities and Prices
9.3.2.1.1 Small SI Engine and Equipment Initial Equilibrium Quantities
The EIM uses the same engine sales quantities that are used in the Small SI cost analysis
presented in Chapter 6. The sales numbers for 2005 are reproduced in Tables 9.3-7 and 9.3-8.
They are based on engine and equipment sales are for 49 states (all states except California) for
2005. However, the sales numbers include construction and agriculture equipment sold in
California, since that equipment is not covered by California's small engine program.
These engine sales numbers are taken from EPA's NONROAD 2005 emission inventory
model. To breakout the sales data by equipment, industry information from Power Systems
Research database-OELink was used to characterize the distribution of equipment by the eight
different equipment categories noted earlier. In addition, the sales within each equipment category
were apportioned to the different useful life categories based on the fraction of engines certified in
each class determined from EPA certification data for model year 2005.
Because of the one-to-one correspondence between Small SI engines and equipment, the
number of equipment is equal to the number of engines sold in a given year.
Table 9.3-7: Small SI Handheld Engine and Equipment Sales (2005)
Sales - All Handheld Engines, Equipment
8,153,106
9-43
-------
Final Regulatory Impact Analysis
Table 9.3-8: Small SI Nonhandheld Engine and Equipment Sales (2005)
Application
Agricultural/Construction/
General Industrial/
Material Handling Equip
Utility and Rec Vehicles
Lawn Mowers
Tractors
Lawn and Garden Other
Gensets/ Welders
Pumps/ Compressors/
Pressure Washers
Snowblowers
Total
Class I
UL125
71,682
81,703
5,895,706
NA
647,256
271,391
579,775
551,509
8,099,022
UL250
7,675
8,748
631,254
NA
NA
29,058
62,077
59,050
797,861
UL500
5,287
6,026
434,846
NA
NA
20,017
42,762
40,677
549,615
Class II
UL250
71,380
173,846
NA
1,701,351
127,915
605,169
253,971
475,353
3,408,985
UL500
15,503
37,757
NA
369,516
27,782
131,437
55,160
103,242
740,396
UL1000
17,585
42,828
NA
419,141
31,513
149,088
62,568
117,107
839,829
Total
189,112
350,908
6,961,805
2,490,008
834,465
1,206,160
1,056,313
1,346,938
14,435,709
9.3.2.1.2 Small SI Engine and Equipment Initial Equilibrium Prices
The initial equilibrium prices for Small SI engines and equipment are contained in Tables
9.3-9 and 9.3-10. The engine prices were prices estimated by EPA using prices compiled from
various websites and obtained from manufacturers. The engine prices were averaged for each
useful life category for each class. The equipment prices were gathered through a survey of
retailers, government dealers, and equipment websites (Caffrey, 2006).
For the handheld market, although all costs are placed on the engine manufacturer, the
engine and equipment manufacturers are integrated so only the equipment price is necessary for
the analysis.
Table 9.3-9: Small SI Handheld Engine and
Equipment Prices (2005$)
Equipment Price
$210
Table 9.3-10a: Small SI Nonhandheld Engine Prices (2005$)
Class I
UL 125
$125
UL250
$217
UL500
$218
Class II
UL250
$208
UL500
$409
UL 1000
$757
9-44
-------
Economic Impact Analysis
Table 9.3-10b: Small SI Nonhandheld Equipment Prices (2005$)
Application
Agricultural/Construction/ General
Industrial/ Material Handling Equip
Utility and Rec Vehicles
Lawn Mowers
Tractors
Lawn and Garden Other
Gensets/ Welders
Pumps/ Compressors/ Pressure Washers
Snowblowers
Class I
UL125
$1,108
$570
$218
$245
$999
$96
$324
UL250
$1,621
$750
$420
$1,428
$661
$480
UL500
$2,133
$931
$2,786
$1,856
$1,225
$637
Class II
UL250
$1,825
$2,894
$1,937
$312
$666
$349
$665
UL500
$3,538
$3,981
$5,241
$969
$1,414
$1,485
$890
UL 1000
$5,251
$5,068
$6,841
$1,626
$2,162
$2,834
$1,115
9.3.2.2 Marine SI Initial Equilibrium Quantities and Prices
9.3.2.2.1 Marine SI Engine and Equipment InitialEquilibrium Quantities
The EIM uses the same engine sales quantities that are used in the Marine SI cost analysis
presented in Chapter 6. The sales numbers for 2005 are reproduced in Tables 9.3-11 and 9.3-12.
The engine sales data are derived for 2003 from certification databases for EPA and the California
Air Resources Board and nationwide statistical data published by the National Marine
Manufacturers Association (Samulski, 2004). These 2003 sales were adjusted to 2005 and future
years using the growth rate described in 9.3.4.
Table 9.3-11: Marine SI Engine Sales (2005)
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
OB loose engines
Total
<25hp
35,756
30,317
66,073
25-50 hp
49,055
49,055
5 1-100 hp
19,327
73,393
92,720
101-175 hp
53,137
13,985
42,903
8,393
118,417
176-300 hp
3,496
33,101
8,877
39,609
8,393
93,476
>301 hp
24,106
12,027
36,133
Total
75,960
71,192
20,904
240,716
16,785
30,317
455,875
9-45
-------
Final Regulatory Impact Analysis
Table 9.3-12: Marine SI Vessel Sales (2005)
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
Total
<25hp
28,605
28,605
25-50 hp
39,244
39,244
5 1-100 hp
19,327
56,780
76,107
101-175 hp
53,137
13,985
33,191
3,357
103,670
176-300 hp
3,496
"51 1 O'}
Jz,JoJ
7,081
30,644
3,357
76,961
>301 hp
23,799
7,927
31,727
Total
75,960
70,168
15,009
188,464
6,714
356,314
9.3.2.2.2 Marine SI Engine and Vessel Initial Equilibrium Prices
The Marine SI engine and vessel initial equilibrium prices are contained in Tables 9.3-13
and 9.3-14. They are based on advertised prices in trade literatures and on the web and on
statistical data collected by the National Marine Manufacturers Association (Samulski, 2004). For
the estimated vessel prices, replacement engines are included but are discounted at 7 percent for
outboard recreational and luxury outboard and sterndrive vessels. The discount is used to account
for the assumption that replacement engines are purchased several years after the boat is
purchased. For this analysis, the discount is based on the average useful engine life estimates in
the NONROAD2005 model. The original price data was 2003 data; these were adjusted by
applying the Product Price Index Series published by the U.S. Bureau of Labor Statistics.12
Table 9.3-13: Marine SI Engine Prices (2005$)
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
OB loose engines
<25hp
$2,606
$2,491
25-50 hp
$5,693
5 1-100 hp
N/A
$9,114
101-175 hp
N/A
$7,577
$13,481
$26,001
176-300 hp
N/A
$12,604
$16,508
$20,786
$40,074
>301 hp
$18,715
$31,959
12For Marine SI engines, the PPI for Gasoline Engines (except aircraft, automobile, highway truck, bus, and
tank; PCU3336183336181) was used; the ratio for this index is 110.1/105.7= 1.042. For marine vessel, the PPI for
Boat Building (PCU 336612336612) was used; the ratio for this index is 206.7/194.2 = 1.064.
9-46
-------
Economic Impact Analysis
Table 9.3-14: Marine SI Vessel Prices* (2005$)
Vessel
PWC
SD/I Recreational
SD/I Luxury
OB Recreational
OB Luxury
<25hp
$3,658
25-50 hp
$10,884
5 1-100 hp
$7,566
$21,561
101-175
hp
$9,982
$16,549
$32,467
$65,097
176-300
hp
$11,960
$32,356
$58,024
$49,420
$104,562
>301 hp
$46,432
$205,658
"Includes replacement engines discounted at 7% for outboard recreational and luxury outboard
in sterndrive/inboard vessels.
9.3.3 Compliance Costs
The social costs of the final standards are estimated by shocking the initial market
equilibrium conditions by the amount of the compliance costs. The compliance costs used in this
analysis are the engineering compliance costs described in Chapters 6 of this RIA and are
summarized in this section.
This analysis applies fixed costs in the year in which they occur prior to the rule taking
effect. The small SI exhaust standards begin in 2011 for Class II and 2012 for Class I. Fixed costs
are applied during 3 years (2008 to 2010) for Class II or 4 years (2008-2011) for Class I. The
fixed costs include research and development, tooling, certification, and 1065 compliance. The
marine exhaust standards generally begin in 2010; however, there are some exceptions for SD/I
engines, where additional lead time was given in specific instances to provide regulatory
flexibility. All fixed costs associated with marine exhaust standards are applied in the year of
2008 and 2009. The implementation dates for the small SI evaporative emission standards are
staggered beginning in 2008, with regulatory flexibility providing some small delays until as late
as 2013. The implementation dates for the marine evaporative emission standards are staggered
beginning in 2009, with regulatory flexibility providing some small delays until as late as 2015.
Fixed costs due to evaporative emission standard are applied in the two years before the primary
effective date for each of the evaporative emission standards (tank permeation, hose permeation,
and diurnal emissions). For simplicity, all of the fixed costs associated with certification are
applied in 2008 and 2009. Variable costs for either exhaust or evaporative emission standards on
small SI and marine SI begin to be incurred only when the programs go into effect.
9-47
-------
Final Regulatory Impact Analysis
9.3.3.1 Small SI Market Compliance Costs
The Small SI engine and equipment compliance costs are summarized in Tables 9.3-15 and 9.3-
16. There is one set of compliance costs for HH engines, since there is only one market. There are
seven sets of engine compliance costs for NHH engines, one for each engine market. These costs
begin in 2008 for HH and NHH; the costs changes over time reflecting the phase-in of the
different standards.
There are no equipment compliance cost estimates for HH or for Class I NHH equipment.
Since the HH market is integrated, all costs are applied to engines. For NHH Class I equipment,
the engine manufacturers typically produce a complete engine and fuel system package. Therefore,
the final program is not expected to impose any additional costs on the equipment manufacturers.
Costs are provided for NHH Class II equipment, reflecting the need for evaporative and emission
controls. An average cost for Class II equipment was applied in this analysis to each of the
equipment categories.
Table 9.3-15: Compliance Costs per Engine - Small SI (2005$)
Class
Useful
life
Cost
type
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017+
Handheld
All engines
Variable
"ixed
Total
$0.00
$0.30
$0.30
$0.65
$0.30
$0.95
$0.65
$0.27
$0.92
$0.65
$0.26
$0.91
$0.84
$0.00
$0.84
$0.84
$0.00
$0.84
$0.71
$0.00
$0.71
$0.71
$0.00
$0.71
$0.71
$0.00
$0.71
$0.71
$O.OC
$0.71
lonhandheld
1
1
1
1
2
2
2
125
250
500
125/250/500
snow-
blower
250
500
1,000
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
?ixed
Total
Variable
?ixed
Total
Variable
?ixed
Total
Variable
?ixed
Total
$0.34
$0.51
$0.85
$0.34
$3.24
$3.58
$0.34
$4.44
$4.77
$0.34
$0.13
$0.47
$0.00
$1.40
$1.40
$0.00
$5.82
$5.82
$0.00
$11.41
$11.41
$0.34
$0.50
$0.84
$0.34
$3.19
$3.52
$0.34
$4.36
$4.70
$0.34
$0.13
$0.47
$0.00
$3.55
$3.55
$0.00
$13.16
$13.16
$0.00
$31.34
$31.34
$0.53
$1.90
$2.44
$0.53
$8.36
$8.89
$0.53
$12.09
$12.63
$0.53
$0.18
$0.71
$0.00
$3.49
$3.49
$0.00
$12.93
$12.93
$0.00
$30.80
$30.80
$0.53
$1.87
$2.40
$0.53
$8.21
$8.75
$0.53
$11.88
$12.41
$0.53
$0.18
$0.71
$16.80
$0.12
$16.92
$12.05
$0.83
$12.88
$30.56
$0.59
$31.15
$12.71
$0.00
$12.71
$15.11
$0.00
$15.11
$14.65
$0.00
$14.65
$2.94
$0.00
$2.94
$16.80
$0.00
$16.80
$12.05
$0.00
$12.05
$30.56
$0.00
$30.56
$12.58
$0.00
$12.58
$14.98
$0.00
$14.98
$14.52
$0.00
$14.52
$2.81
$0.00
$2.81
$16.80
$0.00
$16.80
$12.05
$0.00
$12.05
$30.56
$0.00
$30.56
$12.58
$0.00
$12.58
$14.98
$0.00
$14.98
$14.52
$0.00
$14.52
$2.81
$0.00
$2.81
$16.80
$0.00
$16.80
$12.05
$0.00
$12.05
$30.56
$0.00
$30.56
$12.58
$0.00
$12.58
$14.98
$0.00
$14.98
$14.52
$0.00
$14.52
$2.81
$0.00
$2.81
$14.24
$0.00
$14.24
$9.98
$0.00
$9.98
$25.71
$0.00
$25.71
$12.58
$0.00
$12.58
$14.98
$0.00
$14.98
$14.52
$0.00
$14.52
$2.81
$0.00
$2.81
$14.24
$0.00
$14.24
$9.98
$0.00
$9.98
$25.71
$0.00
$25.71
$11.3'
$O.OC
$11.3'
$13.6f
$O.OC
$13.6f
$13.21
$O.OC
$13.21
$2.33
$o.oc
$2.33
$14.2'
$O.OC
$14.2'
$9.9*
$O.OC
$9.9*
$25.71
$O.OC
$25.71
9-48
-------
Economic Impact Analysis
Table 9.3-16: Compliance Costs per Equipment - Small SI (2005$)
Class
Useful
life
Cost
type
2008 | 2009
2010 | 2011
2012 | 2013
2014 | 2015
2016 | 2017+
Handheld
All engines
Variable
"ixed
Total
No equipment costs for HH, all costs are allocated to engine manufacturers
lonhandheld
1
2
2
2
2
2
2
2
2
2
2
2
2
2
125-500
250
ag/const
250
tractor res
250
L&G other
250
pumps
250
utility
250
weld/press
500
ag/const
500
tractor com
500
L&G other
500
pumps
500
utility
500
weld/press
1,000
ag/const
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
No equipment costs for NHH Class I, all costs are allocated to engine manufacturers
$1.29
$0.26
$1.55
$1.21
$0.26
$1.47
$0.60
$0.26
$0.86
$0.96
$0.26
$1.22
$1.00
$0.26
$1.26
$1.34
$0.26
$1.60
$1.29
$0.26
$1.55
$1.21
$0.26
$1.47
$0.60
$0.26
$0.86
$0.96
$0.26
$1.22
$1.00
$0.26
$1.26
$1.34
$0.26
$1.60
$1.29
$0.26
$1.55
$1.29
$4.47
$5.76
$1.21
$4.84
$6.05
$0.60
$3.68
$4.28
$0.96
$4.32
$5.28
$1.00
$4.14
$5.14
$1.34
$4.58
$5.93
$1.29
$22.36
$23.65
$1.21
$22.73
$23.94
$0.60
$21.58
$22.18
$0.96
$22.21
$23.17
$1.00
$22.04
$23.04
$1.34
$22.48
$23.82
$1.29
$16.22
$17.51
$1.29
$4.13
$5.42
$1.21
$4.50
$5.71
$0.60
$3.36
$3.96
$0.96
$3.99
$4.95
$1.00
$3.81
$4.81
$1.34
$4.25
$5.59
$1.29
$21.72
$23.01
$1.21
$22.08
$23.29
$0.60
$20.95
$21.55
$0.96
$21.57
$22.53
$1.00
$21.39
$22.40
$1.34
$21.83
$23.17
$1.29
$15.68
$16.97
$9.60
$0.00
$9.60
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$9.60
$0.00
$9.60
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$9.60
$0.00
$9.60
$9.60
$0.00
$9.60
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$9.60
$0.00
$9.60
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$9.60
$0.00
$9.60
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$9.47
$0.00
$9.47
$8.16
$0.00
$8.16
$6.08
$0.00
$6.08
$4.63
$0.00
$4.63
$7.23
$0.00
$7.23
$6.76
$0.00
$6.76
$8.58
$0.00
$8.58
$8.16
$0.00
$8.16
$6.08
$0.00
$6.08
$4.63
$0.00
$4.63
$7.23
$0.00
$7.23
$6.76
$0.00
$6.76
$8.58
$0.00
$8.58
$8.16
$0.00
$8.16
$8.U
$O.OC
$8.U
$6.0*
$O.OC
$6.0*
$4.62
$O.OC
$4.62
$7.22
$o.oc
$7.22
$6.7f
$O.OC
$6.7f
$8.5*
$O.OC
$8.5*
$8.U
$O.OC
$8.U
$6.0*
$O.OC
$6.0*
$4.62
$O.OC
$4.62
$7.22
$o.oc
$7.22
$6.7f
$O.OC
$6.7f
$8.5*
$O.OC
$8.5*
$8.U
$O.OC
$8.U
9-49
-------
Final Regulatory Impact Analysis
Class
2
2
2
2
2
2
Useful
life
1,000
tractor com
1,000
L&G other
1,000
pumps
1,000
utility
1,000
weld/press
250/500/1000
snow-
blower
Cost
type
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
Variable
"ixed
Total
2008
$1.21
$0.26
$1.47
$0.60
$0.26
$0.86
$0.96
$0.26
$1.22
$1.00
$0.26
$1.26
$1.34
$0.26
$1.60
$0.51
$0.26
$0.77
2009
$1.21
$16.59
$17.80
$0.60
$15.44
$16.04
$0.96
$16.07
$17.03
$1.00
$15.89
$16.89
$1.34
$16.33
$17.68
$0.51
$0.47
$0.97
2010
$1.21
$16.04
$17.25
$0.60
$14.91
$15.51
$0.96
$15.53
$16.49
$1.00
$15.36
$16.36
$1.34
$15.79
$17.14
$0.51
$0.20
$0.71
2011
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$3.98
$0.00
$3.98
2012
$7.07
$0.00
$7.07
$5.44
$0.00
$5.44
$8.53
$0.00
$8.53
$7.96
$0.00
$7.96
$10.09
$0.00
$10.09
$3.98
$0.00
$3.98
2013
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$3.85
$0.00
$3.85
2014
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$3.85
$0.00
$3.85
2015
$6.94
$0.00
$6.94
$5.32
$0.00
$5.32
$8.41
$0.00
$8.41
$7.83
$0.00
$7.83
$9.96
$0.00
$9.96
$3.85
$0.00
$3.85
2016
$6.08
$0.00
$6.08
$4.63
$0.00
$4.63
$7.23
$0.00
$7.23
$6.76
$0.00
$6.76
$8.58
$0.00
$8.58
$3.27
$0.00
$3.27
2017+
$6.0*
$O.OC
$6.0*
$4.62
$O.OC
$4.62
$7.22
$o.oc
$7.22
$6.7f
$O.OC
$6.7f
$8.5*
$O.OC
$8.5*
$3.2"
$o.oc
$3.2"
9.3.3.2 Marine SI Market Compliance Costs
The Marine SI engine and equipment compliance costs are summarized in Tables 9.3-17
and 9.3-18. Cost estimates are given for each of the 15 engine/equipment combinations, plus cost
estimates for loose OB engines. The engine costs begin in 2008 and increase in 2010 when the
variable costs for exhaust emission standards begin to be incurred. In addition, we apply a one
time learning curve correction to the variable cost in the sixth year. The engine compliance costs
remain the same for 2015 and later years. The equipment costs are more complicated due to the
phase in of the different standards. They begin in 2009, increase in 2011 or 2012 , and then
decrease in 2015. Equipment compliance costs remain the same for 2015 and later years.
9-50
-------
Table 9.3-17: Compliance Costs per Engine - Marine SI (2005$)
Application
Category
PWC
PWC
PWC
SD/I
recreational
SD/I
recreational
SD/I
recreational
SD/I
luxury
SD/I
luxury
HP
Category
50-100
100-175
175-300
100-175
175-300
300+
175-300
300+
Cost
Type
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$0
$73
$73
$0
$34
$34
$0
$113
$113
$0
$45
$45
$0
$50
$50
$0
$57
$57
$0
$50
$50
$0
$57
$57
2009
$0
$73
$73
$0
$34
$34
$0
$113
$113
$0
$45
$45
$0
$50
$50
$0
$57
$57
$0
$50
$50
$0
$57
$57
2010
$870
$0
$870
$85
$0
$85
$1,290
$0
$1,290
$465
$0
$465
$320
$0
$320
$297
$0
$297
$320
$0
$320
$297
$0
$297
2011
$870
$0
$870
$85
$0
$85
$1,290
$0
$1,290
$465
$0
$465
$320
$0
$320
$297
$0
$297
$320
$0
$320
$297
$0
$297
2012
$870
$0
$870
$85
$0
$85
$1,290
$0
$1,290
$465
$0
$465
$320
$0
$320
$297
$0
$297
$320
$0
$320
$297
$0
$297
2013
$870
$0
$870
$85
$0
$85
$1,290
$0
$1,290
$465
$0
$465
$320
$0
$320
$297
$0
$297
$320
$0
$320
$297
$0
$297
2014
$870
$0
$870
$85
$0
$85
$1,290
$0
$1,290
$465
$0
$465
$320
$0
$320
$297
$0
$297
$320
$0
$320
$297
$0
$297
2015
$696
$0
$696
$68
$0
$68
$1,032
$0
$1,032
$372
$0
$372
$256
$0
$256
$238
$0
$238
$256
$0
$256
$238
$0
$238
2016
$696
$0
$696
$68
$0
$68
$1,032
$0
$1,032
$372
$0
$372
$256
$0
$256
$238
$0
$238
$256
$0
$256
$238
$0
$238
2017
$696
$0
$696
$68
$0
$68
$1,032
$0
$1,032
$372
$0
$372
$256
$0
$256
$238
$0
$238
$256
$0
$256
$238
$0
$238
2018-23
$696
$0
$696
$68
$0
$68
$1,032
$0
$1,032
$372
$0
$372
$256
$0
$256
$238
$0
$238
$256
$0
$256
$238
$0
$238
2024+
$696
$0
$696
$68
$0
$68
$1,032
$0
$1,032
$372
$0
$372
$256
$0
$256
$238
$0
$238
$256
$0
$256
$238
$0
$238
-------
Application
Category
OB
recreational
OB
recreational
OB
recreational
OB
recreational
OB
recreational
OB
luxury
OB
luxury
OB
loose engines
HP
Category
<25
25-50
50-100
100-175
175-300
100-175
175-300
<25
Cost
Type
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$0
$12
$12
$0
$14
$14
$0
$20
$20
$0
$37
$37
$0
$67
$67
$0
$37
$37
$0
$67
$67
$0
$12
$12
2009
$0
$12
$12
$0
$14
$14
$0
$20
$20
$0
$37
$37
$0
$67
$67
$0
$37
$37
$0
$67
$67
$0
$12
$12
2010
$69
$0
$69
$216
$0
$216
$203
$0
$203
$338
$0
$338
$690
$0
$690
$338
$0
$338
$690
$0
$690
$69
$0
$69
2011
$69
$0
$69
$216
$0
$216
$203
$0
$203
$338
$0
$338
$690
$0
$690
$338
$0
$338
$690
$0
$690
$69
$0
$69
2012
$69
$0
$69
$216
$0
$216
$203
$0
$203
$338
$0
$338
$690
$0
$690
$338
$0
$338
$690
$0
$690
$69
$0
$69
2013
$69
$0
$69
$216
$0
$216
$203
$0
$203
$338
$0
$338
$690
$0
$690
$338
$0
$338
$690
$0
$690
$69
$0
$69
2014
$69
$0
$69
$216
$0
$216
$203
$0
$203
$338
$0
$338
$690
$0
$690
$338
$0
$338
$690
$0
$690
$69
$0
$69
2015
$55
$0
$55
$173
$0
$173
$162
$0
$162
$270
$0
$270
$552
$0
$552
$270
$0
$270
$552
$0
$552
$55
$0
$55
2016
$55
$0
$55
$173
$0
$173
$162
$0
$162
$270
$0
$270
$552
$0
$552
$270
$0
$270
$552
$0
$552
$55
$0
$55
2017
$55
$0
$55
$173
$0
$173
$162
$0
$162
$270
$0
$270
$552
$0
$552
$270
$0
$270
$552
$0
$552
$55
$0
$55
2018-23
$55
$0
$55
$173
$0
$173
$162
$0
$162
$270
$0
$270
$552
$0
$552
$270
$0
$270
$552
$0
$552
$55
$0
$55
2024+
$55
$0
$55
$173
$0
$173
$162
$0
$162
$270
$0
$270
$552
$0
$552
$270
$0
$270
$552
$0
$552
$55
$0
$55
-------
Table 9.3-18: Compliance Costs per Equipment- Marine SI (2005$)
Application
Category
PWC
PWC
PWC
SD/I
recreational
SD/I
recreational
SD/I
recreational
SD/I
luxury
HP
Category
50-100
100-175
175-300
100-175
175-300
300+
175-300
Cost
Type
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$0.0
$0.9
$0.9
$0.0
$0.9
$0.9
$0.0
$0.9
$0.9
$0.0
$1.5
$1.5
$0.0
$1.5
$1.5
$0.0
$1.5
$1.5
$0.0
$1.9
$1.9
2009
$1.6
$15.5
$17.1
$1.9
$17.0
$18.9
$1.9
$17.0
$18.9
$3.8
$1.5
$5.3
$4.6
$1.5
$6.2
$5.2
$1.5
$6.8
$5.7
$1.9
$7.5
2010
$1.6
$14.7
$16.2
$1.9
$16.1
$18.1
$1.9
$16.1
$18.1
$3.8
$0.3
$4.1
$4.6
$0.3
$4.9
$5.2
$0.3
$5.5
$5.7
$0.4
$6.0
2011
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$31.4
$0.3
$31.7
$43.7
$0.3
$44.0
$71.6
$0.3
$71.9
$53.6
$0.4
$54.0
2012
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$67.2
$0.0
$67.2
$94.4
$0.0
$94.4
$157.6
$0.0
$157.6
$115.7
$0.0
$115.7
2013
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$67.2
$0.0
$67.2
$94.4
$0.0
$94.4
$157.6
$0.0
$157.6
$115.7
$0.0
$115.7
2014
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$67.2
$0.0
$67.2
$94.4
$0.0
$94.4
$157.6
$0.0
$157.6
$115.7
$0.0
$115.7
2015
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$67.2
$0.0
$67.2
$94.4
$0.0
$94.4
$157.6
$0.0
$157.6
$115.7
$0.0
$115.7
2016
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$61.7
$0.0
$61.7
$86.6
$0.0
$86.6
$144.3
$0.0
$144.3
$106.1
$0.0
$106.1
2017
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$56.3
$0.0
$56.3
$80.7
$0.0
$80.7
$137.4
$0.0
$137.4
$98.9
$0.0
$98.9
2018-23
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$56.3
$0.0
$56.3
$80.7
$0.0
$80.7
$137.4
$0.0
$137.4
$98.9
$0.0
$98.9
2024+
$9.7
$0.0
$9.7
$11.2
$0.0
$11.2
$11.2
$0.0
$11.2
$56.3
$0.0
$56.3
$80.7
$0.0
$80.7
$137.4
$0.0
$137.4
$98.9
$0.0
$98.9
-------
Application
Category
SD/I
luxury
OB
recreational
OB
recreational
OB
recreational
OB
recreational
OB
recreational
OB
luxury
OB
luxury
OB
loose
engine
HP
Category
300+
<25
25-50
50-100
100-175
175-300
100-175
175-300
<25
Cost
Type
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$0.0
$2.3
$2.3
$0.0
$0.7
$0.7
$0.0
$1.6
$1.6
$0.0
$1.7
$1.7
$0.0
$1.7
$1.7
$0.0
$1.7
$1.7
$0.0
$3.2
$3.2
$0.0
$3.2
$3.2
$0.0
$0.6
$0.6
2009
$7.8
$2.3
$10.1
$3.9
$10.8
$14.6
$5.5
$1.6
$7.2
$8.3
$1.7
$10.0
$10.0
$1.7
$11.7
$11.7
$1.7
$13.3
$19.4
$3.2
$22.6
$22.6
$3.2
$25.8
$3.1
$8.6
$11.7
2010
$7.8
$0.4
$8.3
$5.5
$10.1
$15.5
$5.5
$0.4
$5.9
$8.3
$0.4
$8.7
$10.0
$0.4
$10.4
$11.7
$0.4
$12.0
$19.4
$0.7
$20.1
$22.6
$0.7
$23.3
$4.4
$8.1
$12.4
2011
$107.2
$0.4
$107.7
$10.8
$0.0
$10.8
$21.6
$0.4
$22.0
$34.6
$0.4
$34.9
$52.4
$0.4
$52.8
$74.9
$0.4
$75.2
$101.4
$0.7
$102.1
$144.8
$0.7
$145.5
$8.6
$0.0
$8.6
2012
$236.1
$0.0
$236.1
$10.8
$0.0
$10.8
$38.6
$0.0
$38.6
$61.6
$0.0
$61.6
$95.4
$0.0
$95.4
$138.3
$0.0
$138.3
$184.5
$0.0
$184.5
$267.4
$0.0
$267.4
$8.6
$0.0
$8.6
2013
$236.1
$0.0
$236.1
$10.8
$0.0
$10.8
$38.6
$0.0
$38.6
$61.6
$0.0
$61.6
$95.4
$0.0
$95.4
$138.3
$0.0
$138.3
$184.5
$0.0
$184.5
$267.4
$0.0
$267.4
$8.6
$0.0
$8.6
2014
$236.1
$0.0
$236.1
$10.8
$0.0
$10.8
$38.6
$0.0
$38.6
$61.6
$0.0
$61.6
$95.4
$0.0
$95.4
$138.3
$0.0
$138.3
$184.5
$0.0
$184.5
$267.4
$0.0
$267.4
$8.6
$0.0
$8.6
2015
$236.1
$0.0
$236.1
$10.4
$0.0
$10.4
$38.6
$0.0
$38.6
$61.6
$0.0
$61.6
$95.4
$0.0
$95.4
$138.3
$0.0
$138.3
$184.5
$0.0
$184.5
$267.4
$0.0
$267.4
$8.3
$0.0
$8.3
2016
$216.2
$0.0
$216.2
$10.4
$0.0
$10.4
$35.4
$0.0
$35.4
$56.4
$0.0
$56.4
$86.9
$0.0
$86.9
$125.6
$0.0
$125.6
$168.1
$0.0
$168.1
$243.0
$0.0
$243.0
$8.3
$0.0
$8.3
2017
$205.8
$0.0
$205.8
$10.4
$0.0
$10.4
$29.5
$0.0
$29.5
$49.9
$0.0
$49.9
$79.8
$0.0
$79.8
$117.6
$0.0
$117.6
$154.3
$0.0
$154.3
$227.5
$0.0
$227.5
$8.3
$0.0
$8.3
2018-23
$205.8
$0.0
$205.8
$10.4
$0.0
$10.4
$29.5
$0.0
$29.5
$49.9
$0.0
$49.9
$79.8
$0.0
$79.8
$117.6
$0.0
$117.6
$154.3
$0.0
$154.3
$227.5
$0.0
$227.5
$8.3
$0.0
$8.3
2024+
$205.8
$0.0
$205.8
$10.4
$0.0
$10.4
$29.5
$0.0
$29.5
$49.9
$0.0
$49.9
$79.8
$0.0
$79.8
$117.6
$0.0
$117.6
$154.3
$0.0
$154.3
$227.5
$0.0
$227.5
$8.3
$0.0
$8.3
-------
Economic Impact Analysis
9.3.4 Growth Rates
The growth rates used in this analysis for future Small SI and Marine SI engines and
equipment sales are from EPA's Nonroad 2005 model and are the same the same as those use for
the cost analysis (EPA 2004b). Because the growth rates are linear, the annual growth rate
decreases over time. For Small SI, the growth rate is approximately 2 percent per year beginning
in 2008 to decreases to approximately 1.5 percent for 2020 and later years. The growth rate for
Marine SI is about 0.8 percent per year in the early years and 0.6 percent in later years.
9.3.5 Fuel Savings
As noted in Section 9.2.4.2, there are fuel savings attributable to the final emission control
program, reflecting the reduction in evaporative emissions and the use of more fuel-efficient
engine technology to meet the final engine exhaust standards. As explained in that section, these
savings are included in the economic welfare analysis as a separate line item. Consumers of Small
SI and Marine SI engines and equipment will realize an increase in their welfare equivalent to the
amount of gallons of gasoline saved multiplied by the retail price of the gasoline (post-tax price).
In the engineering cost analysis the fuel savings are estimated in this manner. However, in the
context of the social welfare analysis, some of this increase in consumer welfare is offset by lost
tax revenues to local, state, and federal governments. These welfare losses must be accounted for
as well. Therefore, the net change in social welfare is the difference between the increase in
consumer welfare and the lost tax revenues. This is equivalent to using the pre-tax price of
gasoline to estimate the fuel savings for the social welfare analysis.
The amount of gallons of gasoline fuel saved is composed of two parts. First, upgrades in
engine technology is expected to reduce fuel consumption rates. These fuel consumption
reductions were calculated using the NONROAD2005 model. In addition, fuel savings due to
evaporative emission control is estimated based on the VOC reductions attributable to these
controls. Tons of annual VOC reductions are translated to gallons of gasoline saved using a fuel
density of 6 Ibs per gallon (for lighter hydrocarbons which evaporate first).
Because the gallons of gasoline saved are based on estimated national reductions and were
not estimated by PADD, we estimated a national average retail gasoline price (RTI, Memorandum
on Calculation Motor Gasoline Prices in Small SI rule EIA, 2006). This estimate is the sum of the
weighted average of pre-tax gasoline prices by PADD and the weighted average gasoline tax by
PADD, using data from the 2005 Petroleum Marketing Annual (DoE 2005, Table 31). The results
of this analysis are shown in Tables 13.3-19 and 13.3-20.
9-55
-------
Table 9.3-19: Estimated National Average Fuel Prices (2005$)
PADD
PADD 1
PADD 2
PADD 3
PADD 4
PADD 5
(excluding CA)
Total
Weight
0.40
0.31
0.18
0.04
0.07
Pre-tax
Price/Gallon
$1.819
$1.792
$1.787
$1.848
$1.938
$1.814
Average State
Taxes
$0.207
$0.209
$0.194
$0.225
$0.198
Federal Tax
$0.184
$0.184
$0.184
$0.184
$0.184
Post-Tax
Price/Gallon
$2.210
$2.185
$2.165
$2.257
$2.320
$2.204
Source: 2005 Petroleum Marketing Annual (Table 31). U.S. Department of Energy, Energy Information
Administration (DoE 2005). Memorandum on Calculation Motor Gasoline Prices in Small SI Rule EM, RTI, 2006.
From 2008 until 2020 the estimated consumer savings associated with reduced gasoline
consumption from the small SI and marine SI programs gas can controls increases sharply, from
$3.2 million to $201 million. After 2020 the savings continue to accrue, but at a reduced rate as
the engines and equipment population turns over and fuel savings are due to the continuing
benefits of using compliant engines and equipment. Similarly, the tax revenue losses are expected
to be increased from $0.7 million in 2008 to $43 million in 2020.
-------
Economic Impact Analysis
Table 9.3-20: Estimated Fuel Savings and Tax Revenue Impacts (2005$)
Year
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
2037
Small SI
Gallons
1,748,394
4,060,953
6,304,081
14,982,484
24,473,425
32,954,248
38,823,924
43,779,918
47,746,869
50,580,035
52,702,739
54,429,770
55,917,371
57,171,954
58,309,088
59,361,292
60,376,870
61,369,887
62,353,284
63,326,668
64,292,904
65,253,036
66,207,356
67,157,496
68,105,401
69,050,694
69,993,828
70,935,570
71,874,840
72,811,766
Marine SI
Gallons
0
398,339
4,488,257
9,200,101
14,543,630
19,806,738
25,044,187
30,265,803
35,426,120
40,529,842
45,571,856
50,527,419
55,085,140
59,222,320
63,220,853
67,056,116
70,326,971
73,270,886
75,906,088
78,313,723
80,527,973
82,543,743
84,310,121
85,894,203
87,275,207
88,524,271
89,672,904
90,731,710
91,713,009
92,624,481
Tax
Consumer Revenue Net Fuel
Fuel Savings Impacts Savings
Total Gallons (Millions) (MillionS) (MillionS)
1,748,394
4,459,291
10,792,339
24,182,584
39,017,055
52,760,987
63,868,112
74,045,720
83,172,989
91,109,877
98,274,595
104,957,189
111,002,511
116,394,274
121,529,941
126,417,408
130,703,841
134,640,773
138,259,372
141,640,391
144,820,877
147,796,779
150,517,478
153,051,700
155,380,608
157,574,965
159,666,732
161,667,280
163,587,849
165,436,247
$3.9
$9.8
$23.8
$53.3
$86.0
$116.3
$140.8
$163.2
$183.3
$200.8
$216.6
$231.3
$244.6
$256.5
$267.9
$278.6
$288.1
$296.7
$304.7
$312.2
$319.2
$325.7
$331.7
$337.3
$342.5
$347.3
$351.9
$356.3
$360.5
$364.6
$0.7
$1.7
$4.2
$9.4
$15.2
$20.6
$24.9
$28.9
$32.4
$35.5
$OOO
Jo.J
$40.9
$43.3
$45.4
$47.4
$49.3
$51.0
$52.5
$53.9
$55.2
$56.5
$57.6
$58.7
$59.7
$60.6
$61.5
$62.3
$63.1
$63.8
$64.5
$3.2
$8.1
$19.6
$43.9
$70.8
$95.7
$115.9
$134.3
$150.9
$165.3
$178.3
$190.4
$201.4
$211.1
$220.5
$229.3
$237.1
$244.2
$250.8
$256.9
$262.7
$268.1
$273.0
$277.6
$281.9
$285.8
$289.6
$293.3
$296.7
$300.1
9-57
-------
Final Regulatory Impact Analysis
9.3.6 Supply and Demand Elasticity Estimates
The estimated market impacts and economic welfare costs of this emission control program
are a function of the ways in which producers and consumers of the Small SI and Marine SI
engines and equipment affected by the standards change their behavior in response to the costs
incurred in complying with the standards. These behavioral responses are incorporated in the EEVI
through the price elasticity of supply and demand (reflected in the slope of the supply and demand
curves), which measure the price sensitivity of consumers and producers.
Because we were unable to find published supply and demand elasticities for the Small SI
and Marine SI markets, we estimated these parameters using the procedures described in Appendix
9E. These methods are well-documented and are consistent with generally accepted econometric
practice. It should be noted that these elasticities reflect intermediate-run behavioral changes. In
the long run, supply and demand elasticities are expected to be more elastic. It should also be
noted that the aggregate data (6 digits NAICS code or 4 digit SIC code industry data) we used to
estimate elasticities include data on other markets as well as the Small SI or Marine SI markets. If
we had been able to obtain market-specific data for Small SI or Marine SI only, the estimated price
elasticities may have been different.
The estimated supply and demand elasticities were based on best data we could find. For
supply elasticities, we used the establishment-level or plant-level data from Census of
Manufactures, conducted by the U. S. Census Bureau to estimate the production function for
affected industries by this rule. The estimated coefficients of the production function were then
used to calculate the supply elasticity for the industry. Establishments-level data were selected
from 6 digit NAICS code industry for five Census years between 1972 and 1997. The supply
elasticities estimated from plant-level data are more elastic than industry-level data, as we
previously used for the EIA chapter in the NPRM
For demand elasitcities, we used the industry-level data published by the National Bureau
of Economic Research (NBER)-Center for Economic Studies (Bartlesman, Becker, and Gray,
2000). In addition to NBER data, we also used the Current Industrial Reports (CIR) series from
the U.S. Census Bureau to produce an annual summary of the production of motors and generators
and a summary of production of several types of lawn and garden equipment; both of these reports
include the number of units manufactured and the value of production (U.S. Census Bureau, 1998;
2000). For walk-behind lawnmowers, we used several data series reported in a study by Air
Improvement Resource, Inc., and National Economic Research Associates (AIR/NERA, 2003).
The U.S. Census Bureau publishes historical data on household income and housing starts (U.S.
Census Bureau, 2002; 2004), and we collected price, wage, and material cost indexes from the
Bureau of Labor Statistics (BLS, 2004a,b,c,d,e). In cases where a price index was not available,
we used the most recent implicit gross domestic product (GDP) price deflator reported by the U.S.
Bureau of Economic Analysis (BEA, 2004).13
13In estimating demand elasticity, all values are expressed in 1987$.
9-58
-------
Economic Impact Analysis
Tables 9.3-21 and 9.3-22 provide a summary of the demand and supply elasticities used to
estimate the economic impact of the final rule.
The estimated supply elasticities for all of the equipment and engine markets are elastic,
ranging from 3.8 for all recreational marine except PWC, to 8.8 for generators, 5.2 for PWCs and,
10 for all Small SI applications except generators, and 9.5 for engines. This means that quantities
supplied are expected to be fairly sensitive to price changes (e.g., a 1 percent change in price
yields a 8.8 percent change in quantity of generator producers are willing to supply ).
On the demand side, the Marine SI equipment market estimated demand elasticity is
elastic, at -2.0. This is consistent with the discretionary nature of purchases of recreational marine
vessels (consumers can easily decide to spend their recreational budget on other alternatives).
The estimated demand elasticity for handheld equipment is elastic, at -1.9. This suggests
that consumers are more sensitive to price changes for handheld equipment than for other Small SI
equipment. In other words, they are more likely to change their purchase decision for a small
change in the price of a string trimmer, perhaps opting for trimmer shears or deciding to forego
trimming altogether.
The estimated demand elasticity for lawnmowers is very inelastic at -0.2. This suggests
that consumers of this equipment are not very sensitive to price changes. Most of this equipment
is sold to individual homeowners, who are often required by local authorities to keep their lawns
trimmed. Household ownership of a gasoline lawnmower is often their least expensive option.
Lawncare services are more expensive since the price for these services includes labor and other
factors of production. Purchasing other equipment may also not be attractive, since electric and
diesel mowers are generally more expensive and often less convenient. Finally, the option of
using landscape alternatives (e.g., prairie, wildflower, or rock gardens) may not be attractive for
home homeowners who may also use their yards for recreational purposes. For all these reasons,
the price sensitivity of homeowners to lawnmower prices would be expected to be inelastic.
All the other demand elasticities, for gensets, welders, compressors, and agriculture/
construction equipment, are about unit elastic, at -1.0 meaning a 1 percent change in price is
expected to result in a 1 percent change in demand.
The demand elasticities for the engine markets are internally derived as part of the process
of running the model. This is an important feature of the ELM, which allows it to link the engine
and equipment components of each model and simulate how compliance costs can be expected to
ripple through the affected market. In actual markets, for example, the quantity of lawnmowers
produced in a particular period depends on the price of engines (the Small SI engine market) and
the demand for equipment by residential consumers. Similarly, the number of engines produced
depends on the demand for engines (the lawnmower market), which depends on consumer demand
for equipment. Changes in conditions in one of these markets will affect the others. By designing
the model to derive the engine demand elasticities, the EIM simulates these connections between
supply and demand among the product markets and replicates the economic interactions between
producers and consumers.
9-59
-------
Final Regulatory Impact Analysis
As discussed in 9.2.3.2, the EIM model uses a derived demand approach for the engine
market to incorporate the interaction between the equipment and engine markets. The demand
curve for the engine market is solely derived from the equipment market. The derived demand is
not affected by the product attributes that could shift the demand curve. In other words, as
explained in 9.2.3.1, the demand curves for either the equipment or engine markets do not shift in
response to any change in consumer preferences that may occur due to the compliance strategies of
the producers in the analysis. We explore the impacts of relaxing this assumption in a sensitivity
analysis (see 9.H.4). The engine and equipment changes needed to meet emission standards may
affect the demand because of the potential changes to fuel consumption and engine performance.
Section 9H. 1 contains a sensitivity analysis that evaluates the effect of increased or decreased
demand elasticities on the estimates of social cost of the rule. How this demand change affects the
total social welfare of the rule is outside the scope of the analysis presented in Chapter 9 because
the corresponding market failure (e.g., the health effects of air pollution) is not explicitly modeled;
for example the economic impact of this regulation through reducing selected air pollutants is
discussed separately in Chapter 8.
Because the elasticity estimates are a key input to the model, a sensitivity analysis for
supply and demand elasticity parameters was performed as part of this analysis in considering the
uncertainty involved in the estimated elasticities. The results are presented in Appendix 9H.
9-60
-------
Economic Impact Analysis
Table 9.3-21: Summary of Market Supply Elasticities Used in EIM
Market
Estimate
Source
Method
Input Data Source
Engine Markets
Small SI and Marine SI
9.5
EPA econometric
estimate
Cobb-Douglas
production function
Census of
Manufacture, US
Census Bureau; five
years between 1972
and 1997; NAICS
333618
Marine Equipment Markets
PWC 5.2
EPA econometric
estimate
Cobb-Douglas
production function
All other vessel types
EPA econometric
estimate
Cobb-Douglas
production function
Census of
Manufacture, US
Census Bureau; five
years between 1972
and 1997; NAICS
336999
Census of
Manufacture, US
Census Bureau; five
years between 1972
and 1997; NAICS
336612
Small SI Equipment Markets
Gensets/welders
EPA econometric
estimate
Cobb-Douglas
production function
All other Small SI
equipment (handheld
and nonhandheld)
10.0
EPA econometric
estimate
Cobb-Douglas
production function
Census of
Manufacture, US
Census Bureau; five
years between 1972
and 1997; NAICS
335312
Census of
Manufacture, US
Census Bureau; five
years between 1972
and 1997; NAICS
333618
9-61
-------
Final Regulatory Impact Analysis
Table 9.3-22: Summary of Market Demand Elasticities Used in EIM
Market
Estimate Source
Method
Input Data Source
Engine Markets
Small SI and Marine SI
Derived Demand
Marine Equipment Markets
All vessel types -2.0
EPA econometric
estimate
Simultaneous
equation (3 SLS)
Bartlesman et al (2000);
Manufacturing Industry
Data from US Census
Bureau: 1958-1996; SIC
3732
Small SI Equipment Markets
HANDHELD: All -1.9
NONHANDHELD
Lawn mowers
-0.2
Other lawn and garden -0.9
Gensets/welders - Class I -1.4
Gensets/welders - Class II -1.1
All other nonhandheld -1.0
EPA econometric Simultaneous
estimate equation (2SLS)
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
Simultaneous
equation (3 SLS)
Simultaneous
equation (2SLS)
Simultaneous
equation (2SLS)
Simultaneous
equation (2SLS)
Simultaneous
equation (2SLS)
U.S. Census Bureau,
Current Industrial
Reports, MA333A 2000
and selected previous
years; 1980-1997
AIR/NERA (2003);
1973-2002
Bureau, Current
Industrial Reports,
MA333A 2000 and
selected previous years;
1980-1997
Bureau, Current
Industrial Reports,
MA333A 2000 and
selected previous years;
1980-1997
Bureau, Current
Industrial Reports,
MA333A 2000 and
selected previous years;
1980-1997
U.S. Census Bureau,
Current Industrial
Reports, MA333A 2000
and selected previous
years; 1980-1997
9-62
-------
Economic Impact Analysis
9.3.7 Economic Impact Model Structure
9.3.7.1 Computing Baseline and With-Regulation Equilibrium Conditions
The economic impact analysis is conducted using the data and the supply and demand
framework described above. The price and quantity data, along with the supply and demand
elasticities, are used to identify the market supply and demand curves. The regulatory costs are
then used to shift the supply curve, and the resulting new equilibrium determines the market
impacts and distribution of social impacts.
Figure 9.3-1 illustrates the economic impact modeling structure. Point A represents the
initial baseline equilibrium price and quantity (corresponding to the prices and quantities presented
in section 9.3.2). The slope of the supply and demand curves passing through the baseline point A
are determined by applying the appropriate supply and demand elasticities presented in section
9.3.6. These slopes reflect the responsiveness of producers and consumers when prices change
and determine how much of the compliance costs producers are able to pass along to consumers in
the with-regulation equilibrium.
The compliance costs associated with the regulation (presented in Section 9.3.3) enter the
model expressed as per-unit costs and result in an upward shift in the supply curve from S0 to Sj in
Figure 9.3-1. Note that the demand curve does not shift because consumer preferences and income
are not affected by the regulation.
With the addition of the compliance costs, if prices were not allowed to adjust demanders
would still want to consume the quantity at point A, but suppliers would only be willing to supply
the quantity at point B (i.e., demand exceeds supply at the baseline price, P). The model then
solves for the new equilibrium price (P*) where the quantity demanded equals the quantity
supplied. The movement from the baseline equilibrium point A to with-regulation equilibrium
point C determines the market impacts (changes in price and quantity) as well as the distribution of
social costs. Appendix 9D describes the set of supply and demand equations included in the
model. Given the number of equations included in the model, the solution algorithm described
below is used to identify the new with-regulation set of equilibrium prices and quantities (Point C).
The analysis illustrated in Figure 9.3-1 is repeated for each year included in the period of
analysis. For future years, a projected time series of prices and quantities are developed and used
as the baseline (point A) from which market changes are evaluated. The engineering cost analysis
provides quantities for future years using historical annual growth rates. In contrast, there is much
more uncertainty surrounding future prices for these markets. As a result, we use a constant 2005
observed prices for the relevant markets during the period of analysis.
9.3.7.2 Solution Algorithm
Supply responses and market adjustments can be conceptualized as an interactive process.
Producers facing increased production costs due to compliance are willing to supply smaller
quantities at the baseline price. This reduction in market supply leads to an increase in the market
9-63
-------
Final Regulatory Impact Analysis
price that all producers and consumers face, which leads to further responses by producers and
consumers and thus new market prices, and so on. The new with-regulation equilibrium is the
result of a series of iterations in which price is adjusted and producers and consumers respond,
until a set of stable market prices arises where total market supply equals market demand. Market
price adjustment takes place based on a price-revision rule, described below, that adjusts price
upward (downward) by a given percentage in response to excess demand (excess supply).
The EIM model uses a similar type of algorithm for determining with-regulation equilibria
and the process can be summarized by six recursive steps:
1. Impose the control costs on affected supply segments, thereby affecting their supply
decisions.
2. Recalculate the market supply in each market. Excess demand currently exists.
3. Determine the new prices via a price revision rule. We use a rule similar to the
factor price revision rule described by Kimbell and Harrison (1986). P; is the market
price at iteration I, qd is the quantity demanded, and qs is the quantity supplied. The
parameter z influences the magnitude of the price revision and speed of
convergence. The revision rule increases the price when excess demand exists,
lowers the price when excess supply exists, and leaves the price unchanged when
market demand equals market supply. The price adjustment is expressed as follows:
(10.1)
4. Recalculate market supply with new prices,
5. Compute market demand in each market.
6. Compare supply and demand in each market. If equilibrium conditions are not
satisfied, go to Step 3, resulting in a new set of market prices. Repeat until
equilibrium conditions are satisfied (i.e., the ratio of supply and demand is arbitrarily
close to one). When the ratio is appropriately close to one, the market-clearing
condition of supply equals demand is satisfied.
9.3.7.3 Estimating Impacts
Using the static partial equilibrium analysis, the EIM model loops through each year
calculating new market equilibriums based on the projected baseline economic conditions and
compliance cost estimates that shift the supply curves in the model. The model calculates price
and quantity changes and uses these measures to estimate the social costs of the rule and partition
the impact between producers and consumers.
9.4 Methods for Describing Uncertainty
9-64
-------
Economic Impact Analysis
Every economic impact analysis examining the market and social welfare impacts of a
regulatory program is limited to some extent by limitations in model capabilities, deficiencies in
the economic literatures with respect to estimated values of key variables necessary to configure
the model, and data gaps. In this EIA, there are three main potential sources of uncertainty: (1)
uncertainty resulting from the way the EIM is designed, particularly from the use of a partial
equilibrium model; (2) uncertainty resulting from the values for key model parameters, particularly
the price elasticity of supply and demand; and (3) uncertainty resulting from the values for key
model inputs, particularly baseline equilibrium price and quantities. Sources of uncertainty that
have a bearing on the results of the EIA for the final program are listed and described in more
detail in Table 9.4-1.
The values used for the price elasticities of supply and demand are critical parameters in
the EIM. The values of these parameters have an impact on both the estimated change in price and
quantity produced expected as a result of compliance with the final standards and on how the
burden of the social costs will be shared among producer and consumer groups. In selecting the
values to use in the EIM it is important that they reflect the behavioral responses of the industries
under analysis.
The first source of values for elasticities of supply and demand is the published economic
literature. These estimates are peer reviewed and generally constitute reasonable estimates for the
industries in question. In this analysis, because we were unable to find published supply and
demand elasticities for the Small SI and Marine SI markets, we estimated these parameters
econometrically using the procedures described in Appendix 9E.
The previous estimates of supply elasticities reflect a production function approach using
data at the aggregate industry level. This method was chosen because of limitations with the
available data; we were not able to obtain firm-level or plant-level production data for companies
that operate in the affected sectors. However, the use of aggregate industry level data may not be
appropriate or an accurate way to estimate the price elasticity of supply compared to firm-level or
plant-level data. This is because, at the aggregate industry level, the size of the data sample is
limited to the time series of the available years and because aggregate industry data may not reveal
each individual firm or plant production function (heterogeneity). There may be significant
differences among the firms that may be hidden in the aggregate data but that may affect the
estimated elasticity. In addition, the use of time series aggregate industry data may introduce time
trend effects that are difficult to isolate and control.
To address these concerns, EPA has investigated estimates for the price elasticity of supply
for the affected industries for which published estimates are not available, using alternative
methods and data inputs. This research program used the cross-sectional data model at the
firm-level or plant level from the U.S. Census Bureau to estimate these elasticities. We used the
results of this research for the analysis.
Table 9 .4-1 Primary Sources of Uncertainty in the Economic Impact Analysis
9-65
-------
Final Regulatory Impact Analysis
Source of Uncertainty Description
Potential Impact
UNCERTAINTIES ASSOCIATED WITH ECONOMIC IMPACT MODEL STRUCTURE
Partial equilibrium
model
The EIM domain is limited to the economic sectors
directly affected by the emission control program;
impacts on secondary markets are not accounted for.
However, such impacts are not expected to be large
since directly affected products and services (small SI
equipment and marine SI vessels) are mostly used by
households and only a very small portion of these
engines and equipment are used as production inputs
to other industry (e.g., agriculture, manufacturing,
construction). In addition, Small SI engines and
equipment would not be a large share of total
production costs for final goods and services in those
commercial markets.
Results understate social costs;
magnitude of impact is
uncertain.
National level model
The EIM considers only national-level impacts;
regional impacts are not modeled. This is appropriate
because Small SI engine and equipment or Marine SI
engine and vessel markets are national markets. While
there may be some regional differences these are likely
to be small due to the competitive nature of the
manufacture industry.
Impacts uncertain
Supply side
assumptions
On the supply side, industries are assumed to be
mature and behave linearly within the range of
analysis; no substitution between production inputs.
This is appropriate because per unit compliance costs
are not large enough to prompt a major change in
product design or assembly.
Impacts uncertain
Demand side
assumption
On the demand side, end consumer's preferences or
consumption patterns are assumed to be constant and
behave linearly within the range of analysis. This is
appropriate because all other factors in the demand
function will not be changed by the final rule.
Impacts uncertain
Constant price
assumption
Prices are assumed to be constant across the period of
analysis. This is a reasonable assumption since it is
not possible to predict changes in these prices over
time (see Appendix G).
Impacts uncertain
Period of analysis
Each period of analysis is assumed to be independent
of previous period and producers are assumed to not
engage in long-term planning to smooth the
compliance costs over a longer period of time.
Because the new exhaust standards will not go into
effect for several years after the program is finalized,
producers may in fact take the full program into
account in production plans to minimize their costs.
Estimated price changes may
be too high for early periods,
too low for later periods;
magnitude of impact is
uncertain
9-66
-------
Economic Impact Analysis
Market shock
In the EIM, the market is shocked by either fixed or
variable compliance costs. This is appropriate because
producers in these industries may not engage in R&D
on a continuous basis and thus the product changes
that would be required to comply with the final
standards would require manufacturers to devote new
funds and resources to product redesign.
Results may overstate
distribution of social costs to
some producers, understate
market impacts; magnitude of
impact is uncertain
UNCERTAINTIES ASSOCIATED WITH PRICE ELASTICITY ESTIMATION
Uncertainty resulting from the functional form used in
the estimation, the data used (aggregate or firm-level),
the time period involved, sample size.
Impacts on distribution of
social costs among
stakeholders (e.g., higher
supply elasticity would result
in less social costs for
manufacturers and more social
costs for consumers)
Impacts on market analysis
(change in price, change in
quantity produced)
Magnitude of impact is
uncertain
UNCERTAINTIES ASSOCIATED WITH DATA INPUTS
Submarket groupings
Baseline equilibrium
prices
Baseline equilibrium
quantities
Submarket data is assumed to be representative and
capture the range of affected equipment. However,
the product groupings in NAICS or SIC 4-digit
categories may include other engines or equipment
that may not have the same production or consumption
characteristics; these groupings not behave the same
way as the directly-affected industries.
Estimated baseline equilibrium prices are assumed to
be representative and capture the range of affected
equipment, and reflect actual transaction prices.
However, the actual prices paid by consumers may be
different. Also, the mix of products included in price
analysis may not be representative of the population.
Estimated baseline equilibrium quantities and future
quantities assumed to be representative; these are the
same as the cost analysis.
Impacts on social welfare and
market analyses uncertain
Impacts on market analysis
uncertain
Impacts on market analysis
uncertain
To explore the effects of key sources of uncertainty, we performed a sensitivity analysis in
which we examine the results of using alternative values for the price elasticity of supply and
demand (using the upper and lower bound of at 95 percent confidence interval around the point
estimate for each elasticity estimate), and alternative baseline equilibrium prices for lawnmowers
9-67
-------
Final Regulatory Impact Analysis
and tractors. The results of these analyses are contained in Appendix 9H. A summary of the
results are presented in Table 9.4-2.
Table 9.4-2. Results of Sensitivity Analysis
Parameter
Year
Change in Value
Impact
Price Elasticity
of Supply
2014
More elastic
(upper bound of
95 percent
confidence
interval for each
elasticity estimate)
Negligible impact on expected price increase and quantity
decrease (less than 0.2 additional increase in price increase
compared to primary analysis; less than 0.2 additional
increase in quantity decrease compared to primary analysis)
More elasticity price elasticity of supply associated with
increase in social cost burden for users of Small SI and
Marine SI engines and equipment (shift of about 3.6 percent
of burden of compliance costs from producers to consumers in
Marine SI market; shift of about 2.5 percent of burden of
compliance costs from producers to consumers in Small SI
market)
2014
Less Elastic
(lower bound of
95 percent
confidence
interval for each
elasticity estimate)
Negligible impact on expected price increase and quantity
decrease (less than 0.2 additional increase in price increase
compared to primary analysis; less than 0.2 percent additional
increase in quantity decrease compared to primary analysis)
Higher value associated with increase in social cost burden
for producers of Small SI and Marine SI engines and
equipment (shift of about 6 percent of burden of compliance
costs from consumers to producers in Marine SI market; shift
of about 6 percent of burden of compliance costs from
consumers to producers in Small SI market)
Price Elasticity
of Demand
2014
More Elastic
(upper bound of
95 percent
confidence
interval for each
elasticity estimate)
Negligible impact on expected price increase and quantity
decrease (less than 0.5 percent additional increase in price
increase compared to primary analysis; less than 1.5 percent
additional increase in quantity decrease, compared to primary
analysis)
More elastic price elasticity of demand associated with
increase in social cost burden for producers of Small SI and
Marine SI engines and equipment (shift of about 11 percent of
burden of compliance costs from consumers to producers in
Marine SI market; shift of about 5 percent of burden of
compliance costs from consumers to producers in Small SI
market)
9-68
-------
Economic Impact Analysis
Alternative
Baseline
Equilibrium
Price -
Lawnmowers
and Tractors
2014
2014
Less Elastic
(lower bound of
95 percent
confidence
interval for each
elasticity estimate)
Lower baseline
equilibrium price
Negligible impact on expected price increase and quantity
decrease (less than 1 percent additional increase in price
increase compared to primary analysis; less than 3 additional
increase in quantity decrease, compared to primary analysis)
Less elastic price elasticity of demand associated with
increase in social cost burden for users of Small SI and
Marine SI engines and equipment (shift of about 22 percent of
burden of compliance costs from producers to consumers in
Marine SI market; shift of about 6 percent of burden of
compliance costs from producers to consumers in Small SI
market)
Larger percent increase in price and percent decrease in
quantity, although absolute changes are smaller (less than 2
percent additional price change for both sectors compared to
primary analysis; about 0.3 percent additional quantity
decrease for lawn mowers and about 1 percent additional
quantity decrease for tractors compared to primary analysis)
Social welfare impacts unchanged.
9-69
-------
Final Regulatory Impact Analysis
Chapter 9 References
Air Improvement Resource, Inc. and National Economic Research Associates, Inc. (AIR/NERA).
2003. "Cost-Effectiveness Analyses of Alternative California Air Resources Board Tier 3 Non-
Handheld Exhaust Emission Proposals." Prepared for Engine Manufacturers Association and
Outdoor Power Equipment Institute. Docket Identification EPA-HQ-OAR-2004-0008-0458
Allen, R.G.D. 1938. Mathematical Analysis for Economists. New York: St. Martin's Press.
Baumol, William. "Contestable Markets: An Uprising in the theory of Industry Structure,"
American Economic Review, 72, March 1982:1-15.
Baumol, William, John Panzer, and Robert Willig. 1982. Contestable Markets and the Theory of
Industry Structure, San Diego, CA: Harcourt, Brace, Jovanovich.
Berck, P., and S. Hoffmann. 2002. "Assessing the Employment Impacts." Environmental and
Resource Economics 22:133-156.
Bingham, T.H., and TJ. Fox. 1999. "Model Complexity and Scope for Policy Analysis."
Public Administration Quarterly 23(3).
Caffrey, C, and Cle Jackson, Zuimdie Guerra. September 2006. Memorandum to Docket EPA-
HQ-OAR-2004-0008 Re: Small SI Engine Sales and Price Estimates. A copy of this document is
available in Docket EPA-HQ-OAR-2004-0008. Docket Identification EPA-HQ-OAR-2004-0008-
0491.
Fullerton, D., and G. Metcalf. 2002. "Tax Incidence." In A. Auerbach and M. Feldstein, eds.,
Handbook of Public Economics, Vol.4, Amsterdam: Elsevier.
Harberger, Arnold C. 1974. Taxation and Welfare. Chicago: University of Chicago Press.
Hicks, J.R., 1961. Marshall's Third Rule: A Further Comment. Oxford Economic Papers
13:262-65. Docket Identification EPA-HQ-OAR-2004-0008-0459
Hicks, J.R., 1966. The Theory of Wages. 2nd Ed. New York: St. Martins Press, pp. 233-247.
Docket Identification EPA-HQ-OAR-2004-0008-0460
Kimbell, L.J., and G.W. Harrison. 1986. "On the Solution of General Equilibrium Models."
Economic Modeling 3:197-212.
Li, Chi. June 6, 2008. Memorandum to Docket EPA-HQ-OAR-2004-0008 Detailed Results
From Economic Impact Model for the Final Rule.
Li, Chi. May 19, 2008. Memorandum to Docket EPA-HQ-OAR-2004-0008 Supply Elasticity
Estimation Report.
9-70
-------
Economic Impact Analysis
NBER-CES. National Bureau of Economic Research and U.S. Census Bureau, Center for
Economic Research. 2002. NBER-CES Manufacturing Industry Database, 1958 - 1996.
http://www.nber.org/nberces/nbprod96.htm
Nicholson, W. 1998. Microeconomic Theory: Basic Principles and Extensions. Fort Worth:
Dryden Press.
Office Management and Budget (OMB). 1996. Executive Analysis of Federal Regulations
Under Executive Order 12866. Executive Office of the President, Office Management and
Budget. January 11, 1996. A copy of this document is available at
http ://www. whitehouse. gov/omb/inforeg/print/riaguide.html.
Outdoor Power Equipment & Engine Service Association. OPE-IN-TFIE-KNOW, Volume
LXXIV, December 30, 2004. "The Business of Outdoor Power Equipment". Docket Identification
EPA-HQ-OAR-2004-0008-0478.
Raboy, David. 1987. "Results of an Economic Analysis of Proposed Excise Taxes on Boats."
Washington, DC. Docket Identification EPA-HQ-OAR-2004-0008-0462.
RTI International (RTI). 2005. "Economic Impact Analysis (EIA) for Additional Tier of Emissions
Standards for Nonroad Spark Ignition Engines and Equipment: Revised Draft Analysis Plan."
EPA Contract No. 68-D-99-024. Docket Identification EPA-HQ-OAR-2004-0008-0463.
RTI International (RTI). 2006a. "Draft Industry Profile for Recreational Marine Industry."
Research Triangle Park, NC: RTI. EPA Contract No. 68-D-99-024. Docket Identification EPA-
HQ-OAR-2004-0008-0492
RTI International (RTI). 2006b. "Draft Industry Profile for Small Nonroad Spark-Ignition Engines
and Equipment." Research Triangle Park, NC: RTI. EPA Contract No. 68-D-99-024. Docket
Identification EPA-HQ-OAR-2004-0008-0469.
RTI International (RTI). 2006c. "Memorandum on Calculation Motor Gasoline Prices in Small SI
rule EIA". EPA Contract No. 68-D-99-024. A copy of this document is available in EPA-HQ-
OAR-2004-0008. Docket Identification EPA-HQ-OAR-2004-0008-0449.
Samulski, M. August 2004. Memorandum to Docket EPA-HQ-OAR-2004-0008 Re: Draft Marine
SI Sales and Prices Estimates. Docket Identification EPA-HQ-OAR-2004-0008-0448.
U.S. Bureau of Economic Analysis. 2004. Table 1.1.9, Implicit Price Deflators for Gross Domestic
Product. Last revised April 29, 2004. Docket Identification EPA-HQ-OAR-2004-0008-0473.
U.S. Bureau of Labor Statistics. 2004a. Producer Price Index Industry Data: (1) Plastic Material
and Resins Manufacturing. Series ID: PCU325211325211 (May 14, 2004), (2) Consumer
Nonriding Lawn, Garden, and Snow Equipment. Series ID: PCU3331123331121 (May 20, 2004),
9-71
-------
Final Regulatory Impact Analysis
(3) Other Engine Equipment Manufacturing. Series ID: PCU3336183336181 (May 20, 2004), (4)
Rotary, Push Type Gasoline Engine Powered. Series ID: WPU12660201 ( August 10, 2004). A
copy of this document is available in the Docket. Docket Identification EPA-HQ-OAR-2004-0008-
0474.
U.S. Bureau of Labor Statistics. 2004b. Average Hourly Earnings of Production Workers (Boat
Building: Series ID CEU3133661206; Farm Machinery and Equipment: Series ID EEU31352306).
Obtained May 18, 2004. Docket Identification EPA-HQ-OAR-2004-0008-0474.
U.S. Census Bureau. 1998 and selected other previous years. Current Industrial Reports. Motors
and Generators. MA335H (MA36H prior to 1998). Washington, DC: U.S. Government Printing
Office. Docket Identification EPA-HQ-OAR-2004-0008-0475.
U.S. Census Bureau. 2000 and selected other previous years. Current Industrial Reports. Farm
Machinery and Lawn and Garden Equipment. MA333A (MA35A prior to 1998). Washington, DC:
U.S. Government Printing Office. Docket Identification EPA-HQ-OAR-2004-0008-0476.
U.S. Census Bureau. 2005. Survey of Plant Capacity: 2004. Current Industrial Reports MQ-
Cl(04). Table 1. A copy of this document can be found at
. Table la.
U.S. Department of Energy, Energy Information Administration. 2006. "Petroleum Marketing
Annual 2005." A copy of this document can be found at
http://www.eia.doe.gov/oil_gas/petroleum/data_publications/petroleum_marketing_annual/pma.ht
ml. Table 31.
U.S. Environmental Protection Agency. 1999. OAQPSEconomic Analysis Resource Document.
Research Triangle Park, NC: EPA. A copy of this document can be found at
http://www.epa.gov/ttn/ecas/econdata/6807-305.pdf.
U.S. Environmental Protection Agency. 2000. Guidelines for Preparing Economic Analyses.
EPA-240-R-00-003, September 2000.
U.S. Environmental Protection Agency. 2004. Final Regulatory Analysis: Control of Emissions
from NonroadDiesel Engines. EPA-420-R-04-007. Research Triangle Park, NC: EPA.
9-72
-------
Economic Impact Analysis
Appendix 9A: Impacts on Small SI Markets
This appendix provides the time series of impacts from 2008 through 2037 for the
following Small SI engines and equipment markets; a complete set of results for all markets can be
found in the docket for this rule (Li, 2008). Results are presented for equipment in the Class I
UL125 and Class IIUL250 categories because those are the categories with the highest sales.
• Class I engines
• Class II engines
• Handheld equipment
• Agriculture/construcion/general industrial, UL125 andUL250
• Utility and recreational vehicles, UL125 and UL250
• Lawn mowers, UL125
• Tractors, UL250
• Lawn and garden other, UL125 and UL250
• Gensets/welders, UL125 and 250
• Pumps/compressors, pressure washers, UL125 and UL250
• Snowblowers, UL125 andUL250
Table 9A-1 through Table 9 A-17 provide the time series of impacts for each engine class
market and each selected equipment market, respectively, includes the following:
• average engine or equipment price
• average engineering costs (variable and fixed) per engine or equipment
• absolute change in the market price ($)
• relative change in market price (%)
• relative change in market quantity (%)
• total engineering costs associated with each engine or equipment market
• changes in producer surplus associated with each engine or equipment market
All prices and costs are presented in 2005 dollars and real engine or equipment prices are
assumed to be constant during the period of analysis. Net present values were estimated using
social discount rates of 3 percent and 7 percent over the period of analysis.
9-73
-------
Final Regulatory Impact Analysis
Table 9A-1: Impact on Small SI Engine Market
Class I (Average Price per Engine = $140)a>b
Small SI Engine (Class I)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$1
$1
$3
$3
$12
$12
$12
$12
$12
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
Absolute
Change in
Price
$1
$1
$3
$3
$12
$12
$12
$12
$12
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
Change in
Change in Quantity
Price (%) (%)
0.8% -0.1%
0.8% -0.1%
2.3% -0.4%
2.2% -0.4%
8.9%
8.8%
8.8%
8.8%
8.8%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
.9%
.9%
.9%
.9%
.9%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
.7%
Total
Engineering
Costs
(million $)
$11.4
$11.5
$32.3
$32.4
$131.3
$132.2
$134.4
$136.6
$138.8
$127.
$129.
$131.
$133.
$135.
$137.
$139.
$141.
$143.
$145.
$147.
$149.
$151.2
$153.2
$155.2
$157.2
$159.2
$161.3
$163.3
$165.3
$167.3
$2,340.8
$1,331.1
Change in Engine
Manufacturers
Surplus
(million $)
-$0.1
-$0.2
-$0.6
-$0.5
-$2.8
-$2.7
-$2.8
-$2.8
-$2.9
-$2.6
-$2.6
-$2.7
-$2.7
-$2.8
-$2.8
-$2.8
-$2.9
-$2.9
-$3.0
-$3.0
-$3.0
-$3.1
-$3.1
$o *•>
3.2
$o ^
3.2
-$3.2
-$3.3
-$3.3
-$3.4
-$3.4
-$47.5
-$27.0
a Figures are in 2005 dollars.
b Average price per engine is a weighted average price of engines by UL.
9-74
-------
Economic Impact Analysis
Table 9A-2. Impact on Small SI Engine Market
Class II (Average Price per Engine = $3 10)a'b
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$3
$9
$9
$19
$18
$18
$18
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
$16
Small
Absolute
Change in
Price
$3
$8
$8
$18
$18
$18
$18
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
$15
SI Engine (Class II)
Change in Change in
Price Quantity
(%) (%)
1.1% -0.3%
2.7% -0.9%
2.7% -0.9%
5.8%
5.7%
5.7%
5.7%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
4.8%
.9%
.9%
.9%
.9%
.7%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
Total
Engineering
Costs
(million $)
$15.4
$40.6
$40.6
$89.4
$89.7
$91.2
$92.8
$79.7
$80.9
$82.2
$83.5
$84.9
$86.2
$87.5
$88.8
$90.1
$91.4
$92.7
$94.0
$95.3
$96.6
$97.9
$99.2
$100.6
$101.9
$103.2
$104.5
$105.8
$107.1
$108.4
$1,633.5
$967.4
Change in Engine
Manufacturers
Surplus
(million $)
-$0.4
-$1.5
-$1.5
-$2.5
-$2.5
-$2.6
-$2.6
-$2.3
-$2.3
-$2.3
-$2.4
-$2.4
-$2.4
-$2.5
-$2.5
-$2.5
-$2.6
-$2.6
-$2.7
-$2.7
-$2.7
-$2.8
-$2.8
-$2.8
-$2.9
-$2.9
-$2.9
-$3.0
-$3.0
-$3.1
-$46.6
-$27.8
a Figures are in 2005 dollars.
b Average price per engine is a weighted average price of engines by UL.
9-75
-------
Final Regulatory Impact Analysis
Table 9A-3 : Impact on Small SI Equipment Market
Handheld (Average Price per Equipment = $210)a
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$0
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
Small SI
Absolute
Change in
Price
$0
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
$1
Equipment
Change
in Price
(%)
0.1%
0.4%
0.4%
0.4%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
(Handheld)
Change in
Quantity
(%)
-0.2%
-0.7%
-0.7%
-0.7%
-0.6%
-0.6%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$2.6
$8.3
$8.2
$8.3
$7.8
$7.9
$6.8
$6.9
$7.0
$7.1
$7.3
$7.4
$7.5
$7.6
$7.7
$7.8
$7.9
$8.0
$8.2
$8.3
$8.4
$8.5
$8.6
(TO H
4>0. /
$8.8
$8.9
$9.1
$9.2
$9.3
$9.4
$151.3
$92.8
-$0.4
-$1.3
-$1.3
-$1.3
-$1.2
-$1.3
-$1.1
-$1.1
-$1.1
-$1.1
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.5
-$1.5
-$1.5
-$1.5
-$24.2
-$14.8
Figures are in 2005 dollars.
9-76
-------
Economic Impact Analysis
Table 9A-4: Impact on Small SI Equipment Market: Class I Ag/Constr./Gen. Ind/ Material
Handling Equipment UL 125 (Average Price per Equipment = $l,108)a
Class 1 Agricultural/Construction/General Industrial/ Material
Handling Equipment UL 125
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Absolute
Change in
Price
$1
$1
$2
$2
$11
$11
$11
$11
$11
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
Change
in Price
0.1%
0.1%
0.2%
0.2%
.0%
.0%
.0%
.0%
.0%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
Change in
Quantity
-0.1%
-0.1%
-0.2%
-0.2%
-1.0%
-1.0%
-1.0%
-1.0%
-1.0%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$1.7
-$1.0
Figures are in 2005 dollars.
9-77
-------
Final Regulatory Impact Analysis
Table 9A-5:
Impact on Small SI Equipment Market: Class I Utility and Recreational Vehicles UL
125 (Average Price per Equipment = $570)a
Small SI Equipment (Class I Utility and Recreational Vehicles UL
125)
Year
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
2037
NPV (3%)
NPV (7%)
Engineering
Cost/Unit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Absolute Change Change in
Change in in Price Quantity
Price (%) (%)
$1 0.1% -0.1%
$1 0.1% -0.1%
$2 0.4% -0.4%
$2 0.4% -0.4%
$11 2.0% -2.0%
$11 2.0% -2.0%
$11 2.0% -2.0%
$11 2.0% -2.0%
$11 2.0% -2.0%
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
.8%
Total
Engineering
Costs
(million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Change in Equipment
Manufacturers Surplus
(million $)
$0.0
$0.0
$0.0
$0.0
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$1.9
-$1.1
Figures are in 2005 dollars.
9-78
-------
Economic Impact Analysis
Table
Year
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
2037
NPV (3%)
NPV (7%)
9A-6: Impact
Small S
Engineering
Cost/Unit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
on Small S
(Average
I Equipment
Absolute
Change in
Price
$1
$1
$2
$2
$12
$12
$12
$12
$12
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
$11
[ Equipment Market:
Price per Equipment
(Class I Lawn Mowers 1
Change Change in
in Price Quantity
(%) (%)
0.4% -0.1%
0.4% -0.1%
1.1% -0.2%
1.1% -0.2%
5.6% - .1%
5.5% - .1%
5.5% - .1%
5.5% - .1%
5.5% - .1%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
5.0% -0.9%
Class I Lawn
= $218)a
LJL 125)
Total
Engineering
Costs
(million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Mowers UL 125
Change in Equipment
Manufacturers Surplus
(million $)
-$0.1
-$0.1
-$0.3
-$0.3
-$1.6
-$1.6
-$1.7
-$1.7
-$1.7
-$1.6
-$1.6
-$1.6
-$1.6
-$1.7
-$1.7
-$1.7
-$1.7
-$1.8
-$1.8
-$1.8
-$1.8
-$1.9
-$1.9
-$1.9
-$1.9
-$2.0
-$2.0
-$2.0
-$2.0
-$2.1
-$28.6
-$16.2
Figures are in 2005 dollars.
9-79
-------
Final Regulatory Impact Analysis
Table 9A-7: Impact on Small SI Equipment Market: Class I Other Lawn and Garden Equipment
UL 125 (Average Price per Equipment = $245)a
Small SI Equipment (Class I Other Lawn and Garden Equipment
UL 125)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Absolute
Change in
Price
$1
$1
$2
$2
$11
$11
$11
$11
$11
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
Change
in Price
(%)
0.3%
0.3%
0.9%
0.9%
4.7%
4.6%
4.6%
4.6%
4.6%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
Change in
Quantity
(%)
-0.3%
-0.3%
-0.8%
-0.8%
-4.2%
-4.1%
-4.1%
-4.1%
-4.1%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
-3.7%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
-$0.1
-$0.1
-$0.1
-$0.1
-$0.7
-$0.8
-$0.8
-$0.8
-$0.8
-$0.7
-$0.7
-$0.7
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$1.0
-$13.1
-$7.4
Figures are in 2005 dollars.
9-80
-------
Economic Impact Analysis
Table 9A-8: Impact on Small SI Equipment Market: Class I Gensets/Welders UL 125
(Average Price per Equipment = $999)a
Small SI Equipment (Class I Gensets/Welders UL 125)
Year
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
2037
NPV (3%)
NPV (7%)
Engineering
Cost/Unit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Absolute Change Change in
Change in in Price Quantity
Price (%) (%)
$1 0.1% -0.1%
$1 0.1% -0.1%
$2 0.2% -0.3%
$2 0.2% -0.3%
$11
$11
$11
$11
$11
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
.1%
.1%
.1%
.1%
.1%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.5%
.5%
.5%
.5%
.5%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
.4%
Total
Engineering
Costs
(million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Change in Equipment
Manufacturers Surplus
(million $)
$0.0
$0.0
-$0.1
-$0.1
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$8.7
-$4.9
Figures are in 2005 dollars.
9-81
-------
Final Regulatory Impact Analysis
Table 9A-9: Impact on Small SI Equipment Market: Class I Pumps/Compressors/Pressure
Washers UL 125 (Average Price per Equipment = $96)a
Small SI Equipment (Class I Pumps/Compressors/Pressure
Washers UL 125)
Year
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
2037
NPV (3%)
NPV (7%)
Engineering
Cost/Unit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Absolute
Change in
Price
$1
$1
$2
$2
$11
$11
$11
$11
$11
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
$10
Change
in Price
(%)
0.8%
0.8%
2.3%
2.2%
11.8%
11.7%
11.7%
11.6%
11.6%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
10.5%
Change in
Quantity
(%)
-0.7%
-0.7%
-2.3%
-2.2%
-11.8%
-11.6%
-11.6%
-11.6%
-11.6%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
-10.4%
Total
Engineering
Costs
(million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Change in Equipment
Manufacturers Surplus
(million $)
$0.0
-$0.1
-$0.1
-$0.1
-$0.7
-$0.7
-$0.8
U.o
-$0.8
-$0.7
-$0.7
-$0.7
-$0.8
U.o
U.o
U.o
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$13.0
-$7.4
Figures are in 2005 dollars.
9-82
-------
Economic Impact Analysis
Table 9A-10: Impact on Small SI Equipment Market:
(Average Price per Equipment =
Small
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Class I Snowblowers UL 125
= $324)a
SI Equipment (Class I Snowblowers UL 125)
Absolute
Change in
Price
$0
$0
$1
$1
$3
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
$2
Change
in Price
(%)
0.1%
0.1%
0.2%
0.2%
0.8%
0.8%
0.8%
0.8%
0.8%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
Change in
Quantity
(%)
-0.1%
-0.1%
-0.2%
-0.2%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$2.8
-$1.6
Figures are in 2005 dollars.
9-83
-------
Final Regulatory Impact Analysis
Table 9A-1 1 : Impact on Small SI Equipment Market: Class II Agri/Constr./G. Ind/ Material
Handling Equipment UL 250 (Average Price per Equipment = $l,825)a
Small SI Equipment (Class II Agricultural/Construction/ General
Industrial/ Material Handling Equipment UL 250)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$2
$6
$5
$10
$10
$9
$9
$9
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
$8
Absolute
Change in
Price
$3
$8
$8
$24
$24
$23
$23
$21
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
Change
in Price
0.1%
0.5%
0.4%
1.3%
1.3%
1.3%
1.3%
1.2%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
Change in
Quantity
-0.1%
-0.5%
-0.4%
-1.3%
-1.3%
-1.3%
-1.3%
-1.2%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.5
$0.5
$0.8
$0.8
$0.8
$0.9
$0.9
$0.8
$0.8
$0.8
$0.8
$0.8
$0.8
$0.8
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
$15.7
$9.3
$0.0
-$0.1
-$0.1
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.3
-$0.3
-$0.3
-$3.7
-$2.2
Figures are in 2005 dollars.
9-84
-------
Economic Impact Analysis
Table 9A-12: Impact on Small SI Equipment Market: Class II Utility and Recreational Vehicle UL
250 (Average Price per Equipment = $2,894)a
Small SI Equipment (Class II Utility and Recreational Vehicle UL
250)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$1
$5
$5
(TO
ij>0
(TO
ij>0
(TO
ij>0
(TO
ij>0
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
Absolute
Change in
Price
$2
$8
$7
$22
$22
$22
$22
$20
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
Change
in Price
(%)
0.1%
0.3%
0.3%
0.8%
0.8%
0.8%
0.8%
0.7%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
Change in
Quantity
(%)
-0.1%
-0.3%
-0.3%
-0.8%
-0.8%
-0.8%
-0.8%
-0.7%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
-0.6%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.3
$1.0
$1.0
$1.7
$1.7
$1.7
$1.7
$1.8
$1.6
$1.6
$1.6
$1.6
$1.7
$1.7
$1.7
$1.7
$1.8
$1.8
$1.8
$1.8
$1.9
$1.9
$1.9
$1.9
$2.0
$2.0
$2.0
$2.0
$2.1
$2.1
$31.8
$19.0
-$0.1
-$0.2
-$0.2
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.4
-$0.4
-$0.4
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.6
-$0.6
-$0.6
-$0.6
-$0.6
-$8.5
$5.0
Figures are in 2005 dollars.
9-85
-------
Final Regulatory Impact Analysis
Table 9 A- 13
: Impact on Small SI Equipment Market: Class II Tractors UL 250 (Average
Price per Equipment = $l,937)a
Small SI Equipment (Class II Tractors UL
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$1
$6
$6
$7
$7
$7
$7
$7
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
$6
Absolute
Change in
Price
$3
$9
$8
$21
$21
$21
$21
$19
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
$18
Change
in Price
(%)
0.1%
0.4%
0.4%
1.1%
1.1%
1.1%
1.1%
1.0%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
Change in
Quantity
(%)
-0.1%
-0.4%
-0.4%
-1.1%
-1.1%
-1.1%
-1.1%
-1.0%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
-0.9%
250)
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$2.9
$12.0
$11.6
$14.6
$14.8
$14.8
$15.1
$15.3
$13.7
$13.9
$14.1
$14.3
$14.5
$14.8
$15.0
$15.2
$15.4
$15.6
$15.9
$16.1
$16.3
$16.5
$16.7
$17.0
$17.2
$17.4
$17.6
$17.9
$18.1
$18.3
$284.9
$171.7
-$0.5
-$1.7
-$1.7
-$4.4
-$4.5
-$4.5
-$4.6
-$4.2
-$4.1
-$4.1
-$4.2
-$4.3
-$4.3
-$4.4
-$4.5
-$4.5
-$4.6
-$4.6
-$4.7
-$4.8
-$4.8
-$4.9
-$5.0
-$5.0
-$5.1
-$5.2
-$5.2
-$5.3
-$5.4
-$5.4
-$81.0
-$47.6
Figures are in 2005 dollars.
9-86
-------
Economic Impact Analysis
Table 9A-14: Impact on Small SI Equipment Market: Class II Other Lawn and Garden Equipment
UL 250 (Average Price per Equipment = $3 12)a
Small SI Equipment (Class II Other Lawn and Garden Equipment
UL 250)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$1
$4
$4
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
$5
Absolute
Change in
Price
$2
$7
$7
$20
$20
$20
$20
$18
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
$17
Change
in Price
(%)
0.6%
2.2%
2.1%
6.4%
6.4%
6.4%
6.4%
5.6%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
Change in
Quantity
(%)
-0.6%
-2.0%
-1.9%
-5.8%
-5.7%
-5.7%
-5.7%
-5.1%
-5.0%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
-4.9%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.1
$0.6
$0.6
$0.8
$0.9
$0.9
$0.9
$0.9
$0.8
$0.8
$0.8
$0.8
$0.8
$0.8
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
$16.2
$9.7
$0.0
-$0.1
-$0.1
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$0.3
-$5.0
-$2.9
Figures are in 2005 dollars.
9-87
-------
Final Regulatory Impact Analysis
Table 9A-15: Impact on Small SI Equipment Market: Class II Gensets/Welders UL 250
(Average Price per Equipment = $666)a
Small SI Equipment
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$2
$6
$6
$10
$10
$10
$10
$10
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
$9
Absolute
Change in
Price
$3
$8
$8
$24
$23
$23
$23
$21
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
$20
(Class II Gensets/Welders UL 250)
Change
in Price
(%)
0.4%
1.2%
1.2%
3.5%
3.5%
3.5%
3.5%
3.2%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
Change in
Quantity
(%)
-0.4%
-1.4%
-1.3%
-3.9%
-3.9%
-3.9%
-3.9%
-3.5%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
-3.3%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$1.1
$4.2
$4.0
$7.4
$7.5
$7.6
$7.7
$7.8
$6.8
$7.0
$7.1
$7.2
$7.3
$7.4
$7.5
$7.6
$7.7
$7.8
$8.0
$8.1
$8.2
$8.3
$8.4
$8.5
$8.6
(TO H
4>0. /
$8.8
$9.0
$9.1
$9.2
$139.8
$83.1
-$0.2
-$0.7
-$0.7
-$2.1
-$2.2
-$2.2
-$2.2
-$2.1
-$2.0
-$2.0
-$2.0
-$2.1
-$2.1
-$2.1
-$2.1
-$2.2
-$2.2
-$2.2
-$2.3
-$2.3
-$2.3
-$2.4
-$2.4
-$2.4
-$2.5
-$2.5
-$2.5
-$2.6
-$2.6
-$2.6
-$38.9
-$22.8
Figures are in 2005 dollars.
9-8
-------
Economic Impact Analysis
Table 9A-16: Impact on Small SI Equipment Market: Class II Pumps/Compressors/ Pressure
Washers UL 250 (Average Price per Equipment = $349)a
Small SI Equipment (Class II Pumps/Compressors/Pressure
Washers UL 250)
Year
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
2037
NPV (3%)
NPV (7%)
Engineering
Cost/Unit
$1
$5
$5
$9
$9
$8
$8
$8
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
$7
Absolute
Change in
Price
$2
$8
$8
$23
$23
$22
$22
$20
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
$19
Change
in Price
(%)
0.7%
2.3%
2.2%
6.5%
6.5%
6.4%
6.4%
5.8%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
5.5%
Change in
Quantity
(%)
-0.7%
-2.3%
-2.2%
-6.6%
-6.5%
-6.4%
-6.4%
-5.9%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
-5.5%
Total
Engineering
Costs
(million $)
$0.4
$1.6
$1.5
$2.6
$2.7
$2.7
$2.7
$2.8
$2.4
$2.5
$2.5
$2.5
$2.6
$2.6
$2.7
$2.7
$2.7
$2.8
$2.8
$2.9
$2.9
$2.9
$3.0
$3.0
$3.1
$3.1
$3.1
$3.2
$3.2
$3.2
$49.6
$29.5
Change in Equipment
Manufacturers Surplus
(million $)
-$0.1
-$0.2
-$0.2
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.6
-$0.6
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$12.5
-$7.3
Figures are in 2005 dollars.
9-89
-------
Final Regulatory Impact Analysis
Table 9A-17: Impact on Small SI Equipment Market:
(Average Price per Equipment :
Small
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$1
$1
$1
$4
$4
$4
$4
$4
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
$3
Class II Snowblowers UL 250
= $665)a
SI Equipment (Class II Snowblowers UL 250)
Absolute
Change in
Price
$1
$1
$1
$4
$4
$4
$4
$4
$3
$o
3
$o
3
$3
$3
$3
$3
$3
$o
3
$o
3
$o
3
$3
$3
$3
$3
$3
$o
3
$o
3
$o
3
$3
$3
$3
Change
in Price
(%)
0.1%
0.1%
0.1%
0.5%
0.5%
0.5%
0.5%
0.5%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
0.4%
Change in
Quantity
(%)
-0.1%
-0.1%
-0.1%
-0.5%
-0.5%
-0.5%
-0.5%
-0.5%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.4
$0.5
$0.4
$2.3
$2.3
$2.3
$2.3
$2.4
$2.0
$2.1
$2.1
$2.1
$2.2
$2.2
$2.2
$2.3
$2.3
$2.3
$2.4
$2.4
$2.4
$2.5
$2.5
$2.5
$2.6
$2.6
$2.6
$2.6
$2.7
$2.7
$40.4
$23.7
$0.0
-$0.1
$0.0
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.2
-$0.3
-$3.7
-$2.2
Figures are in 2005 dollars.
9-90
-------
Economic Impact Analysis
Appendix 9B: Impacts on Marine SI Markets
This appendix provides the time series of impacts from 2008 through 2037 for the
following Small SI engines and equipment markets; a complete set of results for all markets can be
found in the docket for this rule (Li, 2008). For engine markets, Results are presented for the
aggregated categories by power. For the vessel markets, results are presented for the categories
with the highest sales.
Marine SI engines: <25 hp; 26-50 hp; 51-100 hp; 101-175 hp; 176-300 hp; >300 hp
SD/I, 175-300 hp and >300 hp
OB recreational, 50-100 hp
OB luxury, 175-300 hp
PWC 100-175 hp
Table 9B-1 through Table 9B-11 provide the time series of impacts for each engine class
market and each selected equipment market, respectively, includes the following:
• average engine or equipment price
• average engineering costs (variable and fixed) per engine or equipment
• absolute change in the market price ($)
• relative change in market price (%)
relative change in market quantity (%)
total engineering costs associated with each engine or equipment market
changes in producer surplus associated with each engine or equipment market
All prices and costs are presented in 2005 dollars and real engine or equipment prices are
assumed to be constant during the period of analysis. Net present values were estimated using
social discount rates of 3 percent and 7 percent over the period of analysis.
9-91
-------
Final Regulatory Impact Analysis
Table 9B-1 : Impact on Marine SI Engine Market:
<25hp (Average Price per Engine = $2,500)a'b
Marine SI Engine (<25hp)
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Absolute Change in
Engineering Change in Price
Cost/Unit Price (%)
$12
$12
$69
$69
$69
$69
$69
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
$55
2005 dollars.
per engine is
$10
$8
$54
$55
$55
$55
$55
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
$44
a weighted
0.4%
0.3%
2.2%
2.2%
2.2%
2.2%
2.2%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
Change in
Quantity
(%)
-0.8%
-1.6%
-5.5%
-5.2%
-5.2%
-5.2%
-5.2%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
-4.3%
Total
Engineering
Costs
(million $)
$0.8
$0.8
$4.7
$4.8
$4.8
$4.8
$4.9
$3.9
$3.9
$4.0
$4.0
$4.0
$4.1
$4.1
$4.1
$4.1
$4.2
$4.2
$4.2
$4.3
$4.3
$4.3
$4.3
$4.4
$4.4
$4.4
$4.4
$4.5
$4.5
$4.5
$78.2
$47.7
Change in Engine
Manufacturers
Surplus
(million $)
-$0.2
-$0.3
-$1.0
-$1.0
-$1.0
-$1.0
-$1.0
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$16.1
-$9.8
average price of engine by equipment type.
9-92
-------
Economic Impact Analysis
Table 9B-2: Impact on Marine SI Engine Market:
26-50hp (Average Price per Engine = $5,700)a'b
Marine SI Engine (26-50hp)
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Engineering
Cost/Unit
$14
$14
$216
$216
$216
$216
$216
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
$173
2005 dollars.
per engine is a
Absolute
Change in
Price
$12
$12
$198
$197
$196
$196
$196
$156
$156
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
$157
Change in
Price
(%)
0.2%
0.2%
3.5%
3.5%
3.4%
3.4%
3.4%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
Change in
Quantity
(%)
-0.2%
-0.3%
-3.0%
-3.2%
-3.4%
-3.4%
-3.4%
-2.8%
-2.8%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
-2.7%
Total
Engineering
Costs
(million $)
$0.7
$0.7
$11.0
$11.1
$11.2
$11.2
$11.3
$9.1
$9.2
$9.2
$9.3
$9.4
$9.4
$9.5
$9.6
$9.6
$9.7
$9.8
$9.8
$9.9
$9.9
$10.0
$10.1
$10.1
$10.2
$10.3
$10.3
$10.4
$10.5
$10.5
$179.6
$108.8
Change in Engine
Manufacturers
Surplus
(million $)
-$0.1
-$0.1
-$0.9
-$1.0
-$1.0
-$1.1
-$1.1
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$0.9
-$1.0
-$1.0
-$1.0
-$1.0
-$1.0
-$1.0
-$16.6
-$10.0
weighted average price of engine by equipment type.
9-93
-------
Final Regulatory Impact Analysis
Table 9B-3: Impact on Marine SI Engine Market:
51-100hp (Average Price per Engine = $9,100)a'b
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Engineering
Cost/Unit
$20
$20
$203
$203
$203
$203
$203
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
$162
2005 dollars.
per engine is a
Marine
Absolute
Change in
Price
$18
$18
$188
$187
$185
$185
$185
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
$148
SI Engine (51-100hp)
Change in
Price
(%)
0.2%
0.2%
2.1%
2.1%
2.0%
2.0%
2.0%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
Change in
Quantity
(%)
-0.2%
-0.2%
-1.5%
-1.7%
-1.8%
-1.8%
-1.8%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
Total
Engineering
Costs
(million $)
$1.5
$1.5
$15.5
$15.6
$15.7
$15.8
$15.9
$12.8
$12.9
$13.0
$13.1
$13.2
$13.3
$13.4
$13.4
$13.5
$13.6
$13.7
$13.8
$13.9
$14.0
$14.1
$14.2
$14.3
$14.3
$14.4
$14.5
$14.6
$14.7
$14.8
$253.5
$153.8
Change in Engine
Manufacturers
Surplus
(million $)
$1.5
$1.5
$15.5
$15.6
$15.7
$15.8
$15.9
$12.8
$12.9
$13.0
$13.1
$13.2
$13.3
$13.4
$13.4
$13.5
$13.6
$13.7
$13.8
$13.9
$14.0
$14.1
$14.2
$14.3
$14.3
$14.4
$14.5
$14.6
$14.7
$14.8
-$21.6
-$13.0
weighted average price of engine by equipment type.
9-94
-------
Economic Impact Analysis
Table 9B-4: Impact on Marine SI Engine Market:
101-175hp (Average Price per Engine =$12,700)a'b
Marine SI Engine (101-175hp)
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Absolute Change in Change in
Engineering Change in Price Quantity
Cost/Unit Price (%) (%)
$39
$39
$365
$365
$365
$365
$365
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
$292
2005 dollars.
per engine is
$36
$36
$338
$336
$333
$333
$333
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
$266
a weighted
0.3% -0.2%
0.3% -0.3%
2.7% -2.0%
2.6% -2.2%
2.6% -2.4%
2.6% -2.4%
2.6% -2.4%
2.1% -2.0%
2.1% -2.0%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
.9%
Total
Engineering
Costs
(million $)
$2.6
$2.6
$24.7
$24.9
$25.1
$25.3
$25.5
$20.5
$20.7
$20.8
$20.9
$21.1
$21.2
$21.4
$21.5
$21.7
$21.8
$21.9
$22.1
$22.2
$22.4
$22.5
$22.7
$22.8
$23.0
$23.1
$23.2
$23.4
$23.5
$23.7
$406.1
$246.7
Change in Engine
Manufacturers
Surplus
(million $)
-$0.2
-$0.2
-$1.8
-$2.0
-$2.2
-$2.2
-$2.2
-$1.9
-$1.8
-$1.8
-$1.8
-$1.8
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.1
-$34.9
-$21.1
average price of engine by equipment type.
9-95
-------
Final Regulatory Impact Analysis
Table 9B-5: Impact on Marine SI Engine Market:
176-300hp (Average Price per Engine =$17,600)a'b
Marine SI Engine (176-300hp)
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Absolute Change in Change in
Engineering Change in Price Quantity
Cost/Unit Price (%) (%)
$59
$59
$517
$517
$517
$517
$517
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
$414
2005 dollars.
per engine is
$55
$54
$480
$477
$474
$474
$474
$378
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
$379
a weighted
0.3% -0.2%
0.3% -0.2%
2.7% -1.7%
2.7% -1.9%
2.7% -2.0%
2.7% -2.0%
2.7% -2.0%
2.1%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
.7%
.7%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
.6%
Total
Engineering
Costs
(million $)
$5.4
$5.5
$48.3
$48.6
$49.0
$49.3
$49.7
$40.0
$40.3
$40.6
$40.9
$41.2
$41.4
$41.7
$42.0
$42.3
$42.6
$42.8
$43.1
$43.4
$43.7
$44.0
$44.2
$44.5
$44.8
$45.1
$45.4
$45.7
$45.9
$46.2
$793.5
$482.1
Change in Engine
Manufacturers
Surplus
(million $)
-$0.4
-$0.4
-$3.4
-$3.7
-$4.0
-$4.1
-$4.1
-$3.4
$O A
J.4
$O A
J.4
$O A
J.4
-$3.4
-$3.5
-$3.5
-$3.5
-$3.5
-$3.5
$o s:
3.6
$o s:
3.6
-$3.6
-$3.6
-$3.7
-$3.7
-$3.7
$O *1
3.7
$o o
3.8
$o o
3.8
-$3.8
-$3.8
-$3.9
-$65.0
-$39.2
average price of engine by equipment type.
9-96
-------
Economic Impact Analysis
Table 9B-6: Impact on Marine SI Engine Market:
300+ hp (Average Price per Engine = $22,000)a'b
Marine SI Engine (300+ hp)
Year
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
2037
NPV (3%)
NPV (7%)
a Figures are in
b Average price
Engineering
Cost/Unit
$57
$57
$297
$297
$297
$297
$297
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
$238
2005 dollars.
per engine is a
Absolute
Change in
Price
$54
$54
$283
$280
$276
$276
$276
$220
$220
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
$221
Change in
Price
(%)
0.2%
0.2%
1.3%
1.3%
1.3%
1.3%
1.3%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
Change in
Quantity
(%)
-0.1%
-0.1%
-0.6%
-0.8%
-1.0%
-1.0%
-1.0%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
-0.8%
Total
Engineering
Costs
(million $)
$2.1
$2.1
$11.1
$11.2
$11.3
$11.4
$11.5
$9.2
$9.3
$9.4
$9.4
$9.5
$9.6
$9.6
$9.7
$9.8
$9.8
$9.9
$10.0
$10.0
$10.1
$10.1
$10.2
$10.3
$10.3
$10.4
$10.5
$10.5
$10.6
$10.7
$184.7
$112.8
Change in Engine
Manufacturers
Surplus
(million $)
-$0.1
-$0.1
-$0.5
-$0.6
-$0.8
-$0.8
-$0.8
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.7
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$0.8
-$12.8
-$7.7
weighted average price of engine by equipment type.
9-97
-------
Final Regulatory Impact Analysis
Table 9B-7: Impact on Marine Vessels Market:
SD/I Recreational 175-300 hp (Average Price per Equipment =
$32,356)a
Marine Vessel (SD/I Recreational 175-300 hp)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$2
$6
$5
$44
$94
$94
$94
$94
$87
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
$81
Absolute
Change in
Price
$33
$35
$205
$230
$261
$261
$261
$220
$216
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
$212
Change
in Price
(%)
0.1%
0.1%
0.6%
0.7%
0.8%
0.8%
0.8%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
Change in
Quantity
(%)
-0.2%
-0.2%
-1.3%
-1.4%
-1.6%
-1.6%
-1.6%
-1.4%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
-1.3%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.1
$0.2
$0.2
$1.5
$3.2
$3.2
$3.3
$3.3
$3.0
$2.8
$2.9
$2.9
$2.9
$2.9
$2.9
$3.0
$3.0
$3.0
$3.0
$3.0
$3.
$3.
$3.
$3.
$3.
$3.2
$3.2
$3.2
$3.2
$3.2
$50.5
$29.2
-$0.6
-$0.6
-$3.6
-$4.1
-$4.7
-$4.7
-$4.8
-$4.0
-$4.0
-$3.9
-$4.0
-$4.0
-$4.0
-$4.1
-$4.1
-$4.1
-$4.1
-$4.2
-$4.2
-$4.2
-$4.2
-$4.3
-$4.3
-$4.3
-$4.4
-$4.4
-$4.4
-$4.4
-$4.5
-$4.5
-$75.6
-$45.6
Figures are in 2005 dollars.
9-98
-------
Economic Impact Analysis
Table 9B-8: Impact on Marine Vessels Market:
SD/I Luxury 300+ hp (Average Price per Equipment = $205,658)a
Marine Vessel (SD/I Luxury 300+ hp)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$2
$10
$8
$108
$236
$236
$236
$236
$216
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
$206
Absolute
Change in
Price
$56
$61
$291
$354
$435
$435
$435
$378
$365
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
$359
Change
in Price
(%)
0.0%
0.0%
0.1%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
0.2%
Change in
Quantity
(%)
-0.1%
-0.1%
-0.3%
-0.3%
-0.4%
-0.4%
-0.4%
-0.4%
-0.4%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
-0.3%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.1
$0.1
$0.9
$2.0
$2.0
$2.0
$2.0
$1.9
$1.8
$1.8
$1.8
$1.8
$1.8
$1.8
$1.9
$1.9
$1.9
$1.9
$1.9
$1.9
$1.9
$1.9
$2.0
$2.0
$2.0
$2.0
$2.0
$2.0
$2.0
$31.3
$18.0
-$0.2
-$0.3
-$1.3
-$1.6
-$1.9
-$1.9
-$2.0
-$1.7
-$1.7
-$1.6
-$1.7
-$1.7
-$1.7
-$1.7
-$1.7
-$1.7
-$1.7
-$1.7
-$1.7
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.9
-$1.9
-$1.9
-$31.1
-$18.6
Figures are in 2005 dollars.
9-99
-------
Final Regulatory Impact Analysis
Table 9B-9: Impact on Marine Vessels Market:
OB Recreational 50-100 hp (Average Price per Equipment = $21,561)a
Marine Vessel
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$2
$10
$9
$35
$62
$62
$62
$62
$56
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
$50
Absolute
Change in
Price
$16
$21
$165
$181
$197
$197
$197
$165
$162
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
$158
(OB Recreational 50-100 hp)
Change
in Price
(%)
0.1%
0.1%
0.8%
0.8%
0.9%
0.9%
0.9%
0.8%
0.8%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
Change in
Quantity
(%)
-0.2%
-0.2%
-1.5%
-1.7%
-1.8%
-1.8%
-1.8%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
-1.5%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.1
$0.6
$0.5
$2.1
$3.7
$3.7
$3.7
$3.8
$3.5
$3.1
$3.1
$3.1
$3.2
$3.2
$3.2
$3.2
$3.2
$3.3
$3.3
$3.3
$3.3
$3.3
$3.4
$3.4
$3.4
$3.4
$3.5
$3.5
$3.5
$3.5
$56.6
$33.3
$0.0
-$0.2
-$0.2
-$0.7
-$1.3
-$1.3
-$1.3
-$1.3
-$1.2
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.1
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$1.2
-$19.5
-$11.4
Figures are in 2005 dollars.
9-100
-------
Economic Impact Analysis
Table 9B-10: Impact on Marine Vessels Market:
OB Luxury 175-300 hp (Average Price per Equipment = $104,562)a
Marine Vessel (OB Luxury 175-300 hp)
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$o
3
$26
$23
$145
$267
$267
$267
$267
$243
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
$227
Absolute
Change in
Price
$98
$111
$1,009
$1,079
$1,150
$1,150
$1,150
$951
$937
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
$928
Change
in Price
(%)
0.1%
0.1%
1.0%
1.0%
1.1%
1.1%
1.1%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
Change in
Quantity
(%)
-0.2%
-0.2%
-1.9%
-2.1%
-2.2%
-2.2%
-2.2%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
-1.8%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$0.0
$0.1
$0.1
$0.5
$0.9
$1.0
$1.0
$1.0
$0.9
$0.8
$0.8
$0.8
$0.8
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$14.9
$8.7
-$0.2
-$0.2
-$1.9
-$2.0
-$2.1
-$2.2
-$2.2
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.8
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$1.9
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$2.0
-$34.4
-$20.8
Figures are in 2005 dollars.
9-101
-------
Final Regulatory Impact Analysis
Table 9B-1 1 : Impact on Marine Vessels Market:
PWC 100-175 hp (Average Price per Equipment = $9,982)a
Marine Vessel (PWC
Engineering
Year Cost/Unit
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
2037
NPV (3%)
NPV (7%)
$35
$53
$103
$96
$96
$96
$96
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
$79
Absolute
Change in
Price
$25
$38
$74
$69
$69
$69
$69
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
$57
Change
in Price
(%)
0.3%
0.4%
0.7%
0.7%
0.7%
0.7%
0.7%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
100-175 hp)
Change in
Quantity
(%)
-0.5%
-0.8%
-1.5%
-1.4%
-1.4%
-1.4%
-1.4%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
-1.1%
Total
Engineering Change in Equipment
Costs Manufacturers Surplus
(million $) (million $)
$1.9
$2.9
$5.7
$5.3
$5.4
$5.4
$5.5
$4.5
$4.6
$4.6
$4.6
$4.6
$4.7
$4.7
$4.7
$4.8
$4.8
$4.8
$4.9
$4.9
$4.9
$5.0
$5.0
$5.0
$5.1
$5.1
$5.1
$5.2
$5.2
$5.2
$92.7
$57.4
-$0.5
-$0.8
-$1.6
-$1.5
-$1.5
-$1.5
-$1.5
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.3
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$1.4
-$25.6
-$15.8
Figures are in 2005 dollars.
9-102
-------
Economic Impact Analysis
Appendix 9C: Time Series Projections of Social Cost
This appendix provides a time series of the rule's projected social costs for each year
through 2037. Costs are presented in 2005 dollars. In addition, this appendix includes the net
present values by stakeholder using social discount rates of 3 percent and 7 percent over the period
of analysis. As a result, it illustrates how the choice of the discount rate determines the present
value of the total social costs of the program.
9-103
-------
Table 9C: Time Series Projection of Social Costs: 2008 to 2038 (Million $)a
Consumer Surplus Change, Total
Marine SI
End users (households)
S matt SI
End users (households)
Producer Surplus Change, Total
Marine SI
Engine manufacturers
Equipment manufacturers
Small SI
Engine manufacturers
Equipment manufacturers
Fuel Savings
Consumer savings
Fuel
Tax
Government revenue
Total Surplus Change
2008
-$46.3
-$13.2
-$33.1
-$7.5
-$4.1
-$1.0
-$3.1
-$3.4
-$0.6
-$2.9
$3.2
$3.9
$3.2
$0.7
-$0.7
-$50.6
2009
-$109.2
-$15.8
-$93.4
-$17.0
-$5.7
-$1.3
-$4.4
-$11.3
-$1.7
-$9.6
$8.1
$9.8
$8.1
$1.7
-$1.7
-$118.1
2010
-$224.9
-$113.2
-$111.7
-$42.5
-$30.2
-$8.8
-$21.4
-$12.3
-$2.1
-$10.2
$19.6
$23.8
$19.6
$4.2
-$4.2
-$247.8
2011
-$272.1
-$120.0
-$152.1
-$52.0
-$34.5
-$9.5
-$25.0
-$17.5
-$3.0
-$14.5
$43.9
$53.3
$43.9
$9.4
-$9.4
-$280.2
2012
-$372.8
-$128.2
-$244.6
-$62.7
-$39.6
-$10.3
-$29.3
-$23.1
-$5.3
-$17.8
$70.8
$86.0
$70.8
$15.2
-$15.2
-$364.7
2013
-$376.1
-$129.1
-$247.0
-$63.2
-$39.9
-$10.4
-$29.5
-$23.3
-$5.3
-$18.0
$95.7
$116.3
$95.7
$20.6
-$20.6
-$343.6
2014
-$380.2
-$130.0
-$250.2
-$63.6
-$40.2
-$10.5
-$29.7
-$23.4
-$5.4
-$18.1
$115.9
$140.8
$115.9
$24.9
-$24.9
-$327.9
2015
-$349.8
-$108.5
-$241.4
-$57.0
-$34.7
-$8.8
-$25.9
-$22.3
-$5.2
-$17.2
$134.3
$163.2
$134.3
$28.9
-$28.9
-$272.5
2016
-$348.1
-$107.9
-$240.2
-$56.1
-$34.0
-$8.7
-$25.3
-$22.1
-$5.1
-$16.9
$150.9
$183.3
$150.9
$32.4
-$32.4
-$253.3
2017
-$338.8
-$107.5
-$231.3
-$55.0
-$33.5
-$8.7
-$24.8
-$21.5
-$4.9
-$16.6
$165.3
$200.8
$165.3
$35.5
-$35.5
-$228.5
2018
-$343.2
-$108.2
-$235.0
-$55.6
-$33.7
-$8.7
-$25.0
-$21.9
-$5.0
-$16.9
$178.3
$216.6
$178.3
$38.3
-$38.3
-$220.5
(continued)
-------
Table 9C: Time Series Projection of Social Costs (Million $) (continued)
Consumer Surplus Change, Total
Marine SI
End users (households)
S matt SI
End users (households)
Producer Surplus Change, Total
Marine SI
Engine manufacturers
Equipment manufacturers
Small SI
Engine manufacturers
Equipment manufacturers
Fuel Savings
Consumer savings
Fuel
Tax
Government revenue
Total Surplus Change
2019
-$347.6
-$109.0
-$238.6
-$56.2
-$34.0
-$8.8
-$25.2
-$22.2
-$5.1
-$17.1
$190.4
$231.3
$190.4
$40.9
-$40.9
-$213.4
2020
-$352.0
-$109.7
-$242.3
-$56.7
-$34.2
-$8.8
-$25.4
-$22.5
-$5.1
-$17.4
$201.4
$244.6
$201.4
$43.3
-$43.3
-$207.4
2021
-$356.4
-$110.5
-$245.9
-$57.3
-$34.4
-$8.9
-$25.5
-$22.9
-$5.2
-$17.7
$211.1
$256.5
$211.1
$45.4
-$45.4
-$202.6
2022
-$360.8
-$111.2
-$249.6
-$57.9
-$34.7
-$9.0
-$25.7
-$23.2
-$5.3
-$17.9
$220.5
$267.9
$220.5
$47.4
-$47.4
-$198.2
2023
-$365.2
-$112.0
-$253.2
-$58.5
-$34.9
-$9.0
-$25.9
-$23.6
-$5.4
-$18.2
$229.3
$278.6
$229.3
$49.3
-$49.3
-$194.3
2024
-$369.6
-$112.7
-$256.9
-$59.0
-$35.1
-$9.1
-$26.0
-$23.9
-$5.4
-$18.5
$237.1
$288.1
$237.1
$51.0
-$51.0
-$191.5
2025
-$374.0
-$113.4
-$260.5
-$59.6
-$35.4
-$9.1
-$26.2
-$24.2
-$5.5
-$18.7
$244.2
$296.7
$244.2
$52.5
-$52.5
-$189.3
2026
-$378.4
-$114.2
-$264.2
-$60.2
-$35.6
-$9.2
-$26.4
-$24.6
-$5.6
-$19.0
$250.8
$304.7
$250.8
$53.9
-$53.9
-$187.8
2027
-$382.8
-$114.9
-$267.9
-$60.7
-$35.8
-$9.3
-$26.6
-$24.9
-$5.7
-$19.2
$256.9
$312.2
$256.9
$55.2
-$55.2
-$186.6
2028
-$387.2
-$115.7
-$271.6
-$61.3
-$36.1
-$9.3
-$26.7
-$25.3
-$5.8
-$19.5
$262.7
$319.2
$262.7
$56.5
-$56.5
-$185.8
2029
-$391.7
-$116.4
-$275.2
-$61.9
-$36.3
-$9.4
-$26.9
-$25.6
-$5.8
-$19.8
$268.1
$325.7
$268.1
$57.6
-$57.6
-$185.4
(continued)
-------
Table 9C: Time Series Projection of Social Costs (million $) (continued)
Consumer Surplus Change, Total
Marine SI
End users (households)
Small SI
End users (households)
Producer Surplus Change, Total
Marine SI
Engine manufacturers
Equipment manufacturers
S matt SI
Engine manufacturers
Equipment manufacturers
Fuel Savings
Consumer savings
Fuel
Tax
Government revenue
Total Surplus Change
2030
-$396.1
-$117.2
-$278.9
-$62.5
-$36.5
-$9.4
-$27.1
-$26.0
-$5.9
-$20.0
$273.0
$331.7
$273.0
$58.7
-$58.7
-$185.5
2031
-$400.5
-$117.9
-$282.6
-$63.0
-$36.8
-$9.5
-$27.3
-$26.3
-$6.0
-$20.3
$277.6
$337.3
$277.6
$59.7
-$59.7
-$185.9
2032
-$404.9
-$118.7
-$286.3
-$63.6
-$37.0
-$9.6
-$27.4
-$26.6
-$6.1
-$20.6
$281.9
$342.5
$281.9
$60.6
-$60.6
-$186.7
2033
-$409.4
-$119.4
-$290.0
-$64.2
-$37.2
-$9.6
-$27.6
-$27.0
-$6.2
-$20.8
$285.8
$347.3
$285.8
$61.5
-$61.5
-$187.7
2034
-$413.8
-$120.2
-$293.6
-$64.8
-$37.5
-$9.7
-$27.8
-$27.3
-$6.2
-$21.1
$289.6
$351.9
$289.6
$62.3
-$62.3
-$188.9
2035
-$418.2
-$120.9
-$297.3
-$65.3
-$37.7
-$9.7
-$27.9
-$27.7
-$6.3
-$21.4
$293.3
$356.3
$293.3
$63.1
-$63.1
-$190.3
2036
-$422.6
-$121.7
-$301.0
-$65.9
-$37.9
-$9.8
-$28.1
-$28.0
-$6.4
-$21.6
$296.7
$360.5
$296.7
$63.8
-$63.8
-$191.8
2037
-$427.1
-$122.4
-$304.7
-$66.5
-$38.2
-$9.9
-$28.3
-$28.4
-$6.5
-$21.9
$300.1
$364.6
$300.1
$64.5
-$64.5
-$193.5
NPV (3%)
-$6,551.
-$2,079.
-$4,472.
-$1,065.
-$641.
-$167.
-$474.
-$424.
-$94.
-$329.
$3,374.
$4,100.
$3,374.
$725.
-$725.
-$4,242.
1
,0
1
,5
,5
,0
5
,0
1
,9
,6
,2
,6
5
5
,0
NPV (7%)
-$3,869.9
-$1,257.1
-$2,612.8
-$636.3
-$386.1
-$100.8
-$285.2
-$250.2
-$54.8
-$195.4
$1,774.7
$2,156.3
$1,774.7
$381.6
-$381.6
-$2,731.4
Figures are in 2005 dollars.
-------
Economic Impact Analysis
Appendix 9D: Overview of Model Equations and Calculation
To develop the economic impact model, we use set of nonlinear supply and demand
equations for the affected markets and transform them into a set of linear supply and demand
equations. These resulting equations describe stakeholder production and consumption responses
to policy-induced cost and price changes in each market. They also are used to specify the
conditions for a new with-policy equilibrium. We describe these equations in more detail below.
9D.1 Economic Model Equations
Supply Equations
First, we consider the formal definition of the elasticity of supply with respect to changes
in own price:
dQJQ.
£, =
dpi p
(9D.1)
Next, we can use "hat" notation to transform Eq. (C. 1) to proportional changes and
rearrange terms:
4 = e.P (9D.la)
where
Qs = percentage change in the quantity of market supply,
es = market elasticity of supply, and
p = percentage change in market price.
As Fullerton and Metcalf (2002) note, this approach takes the elasticity definition and turns it into
a linear behavioral equation for each market.
To introduce the direct impact of the regulatory program, we assume the direct per-unit
compliance cost (c) leads to a proportional shift in the marginal cost of production. Under the
assumption of competitive market (price equals marginal cost), we can approximate this shift at
the initial equilibrium point as follows:
MC = -?— = — • (9D.2)
MC0 p0
The with-regulation supply response to price and cost changes can now be written as:
(9D3)
9-107
-------
Final Regulatory Impact Analysis
For equipment producers, the supply response should also simultaneously accounts for
changes in equilibrium input prices (engines). To do this, we modify Eq. (9D.2) as follows:
- c +
MC =
Po (9D.3a)
where Apengine is the equilibrium change in the engine price and a is the ratio of engines used per
unit of equipment. For example, if one piece of equipment uses only one engine, then a = 1 . This
equation can accommodate other engine to equipment ratios by multiplying Apeng by the
appropriate engine-to-equipment ratio (a).
Demand Equations
Similar to supply, we can characterize equipment demand responses to price changes as:
Qd = r\dp (9D.4)
where
Qd = percentage change in the quantity of market demand,
r|d = market elasticity of demand, and
p = percentage change in market price.
In contrast to equipment demand, the demand for engines is a derived demand and is related to
equipment supply decisions. In order to maintain a constant engine-to-equipment ratio, the
demand for engines is specified as:
Qdengines = Qsequipment (9D.5)
Market Equilibrium Conditions
In response to the exogenous increase in equipment and engine production costs,
stakeholder responses are completely characterized by represented in Eq. (9D.3)(equipment and
engine supply), Eq. (9D.4) (equipment demand), and Eq. (9D.5)(engine demand). Next, we
specify the relationship that must hold for markets to "clear", that is, supply in each market equals
demand. Given the equations specified above, the new equilibrium satisfies the condition that for
each market, the proportional change in supply equals the proportional change in demand:
& = & (9D.6)
9-108
-------
Economic Impact Analysis
9D.2 Computing With-Regulation Equilibrium Conditions
The choice of efficient model solution algorithms depends on several factors such as the
number of markets included in the economic model, complexity of interactions between consumers
and producers within these markets, and the software used to construct the model. To
find the new market equilibrium prices and quantities, we used a solution algorithm that has
proven very useful in "searching" for the equilibrium prices and quantities for partial equilibrium
spreadsheet simulations with complicated relationships. We describe this approach in more detail
below.
9D.2.1 Conceptual Description of RTFs Spreadsheet Model Solution Algorithm:
PE_Walrasian_Auctioneer©2005
The French economist Leon Walras proposed one early model of market price adjustment
by using the following thought experiment. Suppose there is a hypothetical agent that facilitates
market adjustment by playing the role of an "auctioneer." He announces prices, collects
information about supply and demand responses (without transactions actually taking place), and
continues this process until market equilibrium is achieved.
For example, consider the with-regulation supply and demand conditions at the without-
regulation equilibrium price (P) (see Figure 9D-la). The auctioneer determines that the quantity
demanded (A) exceeds the quantity supplied (B) at this price and calls out a new (higher) price
(P') based on the amount of excess demand. Consumers and producers make new consumption and
production choices at this new price (i.e., they move along their respective demand and supply
functions), and the auctioneer checks again to see if excess demand or supply exists. This process
continues until P = P* (point C in Figure 9D-la) is reached (i.e., excess demand is zero in the
market). A similar analysis takes place when excess supply exists. The auctioneer calls out lower
prices when the price is higher than the equilibrium price.
Figure 9D-la. Computing with Regulation Equilibrium
$/Q
Price
Increase
P'
P
S,: With Regulation
Q/t
9-109
-------
Final Regulatory Impact Analysis
The model uses a similar type of algorithm for determining with-regulation equilibria, and
the process can be summarized by six recursive steps:
1. Impose the control costs on affected supply segments, thereby affecting their supply
decisions.
2. Recalculate the market supply in each market. Excess demand currently exists.
3. Determine the new prices via a price revision rule. We used a rule similar to the factor
price revision rule described by Kimbell and Harrison (1986). P; is the market price at
iteration i, qd is the quantity demanded, and qs is the quantity supplied. The parameter
z influences the magnitude of the price revision and the speed of convergence. The
revision rule increases the price when excess demand exists, lowers the price when
excess supply exists, and leaves the price unchanged when market demand equals
market supply. The price adjustment is expressed as follows:
(9D.7)
4. Recalculate market supply with new prices.
5. Compute market demand in each market.
6. Compare supply and demand in each market. If equilibrium conditions are not
satisfied, go to Step 3, resulting in a new set of market prices. Repeat until equilibrium
conditions are satisfied (i.e., the ratio of supply and demand is arbitrarily close to one).
When the ratio is appropriately close to one, the market-clearing condition of supply
equals demand is satisfied.
9D.2.2 Consumer and Producer Welfare Calculations
The change in consumer surplus in the affected markets can be estimated using the
following linear approximation method:
ACS=-Q1«Ap + 0.5«AQ«Ap. (9D.8)
As shown, higher market prices and reduced consumption lead to welfare losses for consumers. A
geometric representation of this calculation is illustrated in Figure 9D-lb.
For affected supply, the change in producer surplus can be estimated with the following
equation:
APS = Qj • (Ap - c) - 0.5 • AQ • (Ap - c). (9D.9)
Increased regulatory costs and output declines have a negative effect on producer surplus, because
the net price change (Ap - c) is negative. However, these losses are mitigated, to some degree, as a
result of higher market prices. A geometric representation of this calculation is illustrated in Figure
9D-lb.
9-110
-------
Economic Impact Analysis
A consumer surplus =-[fghd + dhc]
A producer surplus =[fghd - aehb] - bdc
A total surplus =-[aehb + dhc + bdc]
Figure 9D-lb. Welfare Calculations
Price
Increase
S-,: With Regulation
Unit Cost Increase
S0: Without Regulation
Output
9-111
-------
Final Regulatory Impact Analysis
Appendix 9E: Elasticity Parameters for Economic Impact Modeling
The Economic Impact Model (EIM) relies on elasticity parameters to estimate the
behavioral response of consumers and producers to the regulation and its associated social costs.
To operationalize the market model, supply and demand elasticities are needed to represent the
behavioral adjustments that are likely to be made by market participants. The following parameters
are needed:
supply and demand elasticities for Marine SI equipment markets
• supply and demand elasticities for Small SI equipment markets
• supply elasticities for Marine SI engine markets
• supply elasticities for Small SI engine markets
Note that demand elasticities for the Marine SI and Small SI engine markets are not
estimated because they are derived internally in the model. They are a function of changes in
output levels in the equipment markets.
Tables 9E-1 and 9E-2 contain the demand and supply elasticities used to estimate the
economic impact of the rule. Two methods were used to obtain the supply and demand elasticities
used in the EIM. First, the professional literature was surveyed to identify elasticity estimates used
in published studies. Second, when literature estimates were not available for specific markets,
established econometric techniques were used to estimate supply and demand elasticity parameters
directly. Since very few studies have been identified to quantify elasticities for Small SI and
Marine SI markets in the literature survey, the supply and demand elasticities for all of the
equipment and engine markets were estimated econometrically.
This appendix describes the methods used to estimate demand and supply elasticities for
Marine SI and Small SI engines and equipment markets and presents the data sources and the
regression results obtained from applying those methods.
Finally, it should be noted that these elasticities reflect intermediate run behavioral
changes. In the long run, supply and demand are expected to be more elastic since more substitutes
may become available.
9-112
-------
Economic Impact Analysis
Table 9E-1: Summary of Market Supply Elasticities Used in the Market Model
Markets Estimate Source Method Input Data Summary
Recreational
Marine
All vessel types
except PWC
PWC 5.2
Small SI
All lawn and 10.0
garden
equipment
Gensets/welders
All Engines 9.5
EPA econometric
estimate
Table 9E-4
EPA econometric
estimate
Table 9E-5
EPA econometric
estimate
Table 9E-6
EPA econometric
estimate
Table 9E-7
EPA econometric
estimate
Table 9E-3
Cobb-Douglas production
function
Cobb-Douglas production
function
Cobb-Douglas production
function
Cobb-Douglas production
function
Cobb-Douglas production
function
Census of Manufacture,
US Census Bureau; five
years between 1972 and
1997; NAICS 336612
Census of Manufacture,
US Census Bureau; five
years between 1972 and
1997; NAICS 336999
Census of Manufacture,
US Census Bureau; five
years between 1972 and
1997; NAICS 333112
Census of Manufacture,
US Census Bureau; five
years between 1972 and
1997; NAICS 335312
Census of Manufacture,
US Census Bureau; five
years between 1972 and
1997; NAICS 333618
9-113
-------
Final Regulatory Impact Analysis
Table 9E-2: Summary of Market Demand Elasticities Used in the Market Model
Primary Input Data
Market
Equipment
All recreational
marine (including
PWC)
Lawnmowers
Estimate
-2.0
-0.2
Source
EPA econometric
estimate
Table 9E-8
EPA econometric
Table 9E-9,
Method
Simultaneous
equation (3 SLS)
estimate Simultaneous
equation (3 SLS)
Summary
Bartlesman et al.(2000);
Manufacturing Industry Data
from US Census Bureau:;
1958-1996; SIC 3732
AIR/NERA (2003);
1973-2002
Lawn and garden -1.0
tractors
Pumps/compressors/ -1.0a
pressure washers,
snowblowers
Agriculture, -1.0a
construction, general
industrial
Other lawn and -0.9b
garden
Column 2
EPA econometric estimate Simultaneous
Table 9E-9, equation (2SLS)
Columns
EPA econometric estimate Simultaneous
Table 9E-9, equation (2SLS)
Columns
EPA econometric estimate Simultaneous
Table 9E-9, equation (2SLS)
Columns
EPA econometric estimate Simultaneous
Table 9E-9, equation (2SLS)
Columns
U.S. Census Bureau, Current
Industrial Reports, MA333A
2000 and selected previous
years; 1980-1997
U.S. Census Bureau, Current
Industrial Reports, MA333A
2000 and selected previous
years; 1980-1997
U.S. Census Bureau, Current
Industrial Reports, MA333A
2000 and selected previous
years; 1980-1997
U.S. Census Bureau, Current
Industrial Reports, MA333A
2000 and selected previous
years; 1980-1997
All handheld lawn
and garden
equipment
-1.9 EPA econometric estimate Simultaneous
Table 9E-9, equation (2SLS)
Column 4
U.S. Census Bureau, Current
Industrial Reports, MA333A
2000 and selected years;
1980-1997
Gensets/welders
Class 1
-1.4 EPA econometric
estimate
Table 9E-10,
Column 2
Simultaneous
equation (3 SLS)
U.S. Census Bureau, Current
Industrial Reports, MA335H
2000 and selected years;
1980-1997
Gensets/welders
Class 2
All Engines
-1.1 EPA econometric
estimate
Table 9E-10,
Columns
Derived demand
Simultaneous
equation (3 SLS)
NA
U.S. Census Bureau, Current
Industrial Reports, MA335H
2000 and selected years;
1980-1997
a Uses econometric estimate for lawn and garden tractors.
b Uses econometric estimate for commercial mowers.
9-114
-------
Economic Impact Analysis
9E.1 Supply Elasticities
We use a two-step approach to estimate the price elasticity of supply14. In the first step, we
estimate an industry production function by using the regression model. In the second step, we
calculate the supply elasticity by the parameters estimated in the estimated production function.
This section discusses the regression model used to estimate the industry production function, data
sources used for the regression, and estimated results for supply elasticities. The economics theory
on the relationship between the supply elasticity and the production function is discussed in
Appendix 9F.
In economics, the production function is defined as the relationship between inputs and
outputs of the production process. In this case, we assume that Small SI and Marine SI
manufacturers follow the Cobb-Douglas production function with a stochastic error term Uit ~N(0,
a2), recognizing that we have observations on plant i at time t
Vi t - ^t V^i t) U-i tJ Vlvli tJ c; U
where
Qit = At (Kit)K (Lit)L (Mit)M eit
Qi t = total value of shipment on plant i at time t,
Kit = total capital stock, including both structure and equipment, on plant i at time t,
Lit = total plant hours on plant i at time t, and
Mit = cost of materials on plant i at time t.
This equation can be written in linear form by taking the natural logarithms of each side of the
equation. The parameters of this model, aK, aL, aM, can then be estimated using linear regression
techniques:
In Qit = In At + aK In Kit + aL In Lit + aM In Mit + Uit (E9.2)
Under the assumptions of a competitive market, the elasticity of supply with respect to the
price of the final product can be expressed in terms of the parameters of the production function:15
Supply Elasticity = (aL + aM) / (1 - aL - aM). (9E.3)
Our main regressions were carried out imposing the constant returns to scale assumption
(aK+aL+aM=l). We also tested regressions that did not constrain the three a parameters, and
obtained very similar estimates for the parameters. The estimated returns to scale, given by the
sum of the three a parameters, ranges from 1.01 to 1.03, supporting the assumption of constant
returns. We estimate these regressions with and without including the dummies for single-plant
firm and large plant. Table 9E-3 to Table 9E-7 present the estimated production function
14 Please refer to Supply Elasticity Estimation Report, Li, Chi. May 19, 2008. Memorandum to Docket
EPA-HQ-OAR-2004-0008.
15 Appendix 9F provides the derivation of this result.
9-115
-------
Final Regulatory Impact Analysis
coefficients for the five industries including the dummies for single-plant firm and large plant.
9E.1.1 Data Sets
The data used to estimate these elasticities comes from Census of Manufactures, conducted
by the Census Bureau every five years between 1972 and 1997. We had access to the plant-level
data at the Boston Census Research Data Center, and gathered data for all plants that were
identified as belonging to the five industries being studied. The Census data provided us with
information on output as measured by TVS (Plant's Total Value of Shipments), employment as
measured by PH (Total Plant Hours), materials as measured by CM (Cost of Materials), and
capital stock as measured by CAP (Total Capital Stock, including both Structure and Equipment) -
the capital stock data are only available through 1997, which is why we do not include 2002
Census of Manufactures data in our analysis.
Based on comments from reviewers of an earlier supply elasticity analysis done for three
other industries, we added two additional variables to the regression analysis, to control for
possible differences across plants in their productivity levels. SINGLE is a dummy variable,
indicating that the plant's firm owns no other manufacturing plants (a single-plant firm). BIG is a
dummy variable, indicating that this plant has a number of employees larger than the median value
for all other plants in this industry (approximately 50% of the plants in an industry should have
BIG=1).
The data were examined in detail to identify outliers, measured in terms of unusual ratios
between the values (e.g. unusually high or low shipments per worker hour, relative to the other
plants in the industry) or unusual swings from one observation to the next. Those cases were then
adjusted, based on their values in surrounding years or on the ratios between variables for other
plants in the same industry, to avoid biasing the results while retaining all observations for the
analysis.
One potential complication in working with the Census of Manufactures over this time
period is the considerable change in industry definitions in 1997, when the Census Bureau shifted
from SIC (Standard Industry Classification) codes to NAICS (North American Industry
Classification System) codes. For these five industries there were some definitional concerns, but
we were able to solve them reasonably accurately using a combination of SIC and NAICS industry
codes and detailed product codes.
9E.1.2 Results of Supply Elasticity Estimation
By pooling data at establishment or plant level on selected years, we applied ordinary least
square (OLS) procedure to estimate Eq. (9E.2) with two additional dummy variables. As shown in
Tables 9E-3 through 9E-7, supply elasticity estimates for Small SI products range from 3.76 (Boat
Building) to 9.96 (Lawn & Garden Equipment).
9-116
-------
Economic Impact Analysis
Table 9E-3: Gasoline Engines: NAICS 333618 (SIC 3519) Internal Combustion Engines, Not
Elsewhere Classified.
Number of Observations = 1454
Root Mean Square Error = 0.316
Supply Elasticity = 9.46
Variable
intercept
SINGLE
BIG
InK
InL
InM
Estimated Coefficients
1.930
-0.120
0.059
0.096
0.284
0.621
t-statistic
22.3
-4.4
2.1
4.6
13.1
27.0
Table 9E-4: Gasoline-Powered Boats: NAICS 336612 (SIC 3732) Boat Building and Repairing.
Number of Observations = 10521
Root Mean Square Error = 0.239
Supply Elasticity = 3.76
Variable
intercept
SINGLE
BIG
InK
InL
InM
Estimated Coefficients
1.201
-0.045
0.113
0.210
0.141
0.650
t-statistic
43.7
-4.0
18.6
14.4
15.8
45.3
9-117
-------
Final Regulatory Impact Analysis
Table 9E-5: PWCs, ATVs, Snowmobiles: NAICS 336999 (SIC 3799) Transportation Equipment,
Not Elsewhere Classified.
Number of Observations = 2326
Total R-square = 0.237
Supply Elasticity = 5.20
Variable
Estimated Coefficients
t-statistic
intercept
SINGLE
BIG
InK
InL
InM
1.323
-0.069
0.072
0.161
0.184
0.654
13.9
-3.1
7.0
7.0
9.9
19.2
Table 9E-6: Small Handheld/Nonhandheld: NAICS 333112 (SIC 3524) Lawn and Garden Tractors
and Home Lawn and Garden Equipment.
Number of Observations = 839
Root Mean Square Error = 0.232
Supply Elasticity = 9.96
Variable
Estimated Coefficients
t-statistic
intercept
SINGLE
BIG
InK
InL
InM
1.162
-0.012
0.033
0.091
0.156
0.753
12.3
-0.5
1.3
4.5
7.3
23.4
9-118
-------
Economic Impact Analysis
Table 9E-7: Gensets and Marine Generators: NAICS 335312 (SIC 3621) Motors and Generators.
Number of Observations = 2681
Root Mean Square Error = 0.288
Supply Elasticity = 8.79
Variable Estimated Coefficients t-statistic
intercept 1.812 30.8
SINGLE -0.096 -5.3
BIG -0.041 -2.3
InK 0.102 7.3
InL 0.242 15.4
InM 0.655 33.7
9-119
-------
Final Regulatory Impact Analysis
9E.2 Demand Elasticities
To obtain demand elasticity parameters, we estimated a simultaneous system of demand
and supply equations using instrumental variables methodology by either two-stage least squares
(2SLS) or three-stage least squares (3SLS) regression. This type of partial equilibrium market
supply/demand model is specified as a system of interdependent equations in which the price and
output of a product are simultaneously determined by the interaction of producers and consumers
in the market. In simultaneous equation models, where variables in one equation feed back into
variables in another equation, the error terms are correlated with the endogenous variables (price
and output). Use of a single-equation ordinary least squares (OLS) estimation of individual
equations will lead to biased and inconsistent parameter estimates because it does not account for
the correlation of the error term with the endogenous variables. In 2SLS or 3SLS, however, each
equation is identified through the inclusion of exogenous variables as instruments that control for
shifts in the supply and demand curves over time.
Exogenous variables influencing the demand for gasoline-powered boats and Small SI
equipment include measures of general economic activity (per capita household or disposable
income, number of households or housing starts). Exogenous variables influencing the cost of
production and supply of boats and Small SI equipment include changes in prices of key inputs
like labor and raw materials.
The supply/demand system for gasoline powered equipment can be defined as follows:
Qtd = f(Pt,Zt) + ut (9E.4)
Qts = g(Pt,Wt) + vt (9E.5)
Q,d = Q,s (9E.6)
Eq. (9E.4) shows quantity demanded as a function of price, Pt; a vector of demand shifters, Zt (e.g.,
measures of economic activity); and an error term, ut. Eq. (9E.5) represents quantity supplied as a
function of price and a vector of supply shifters, Wt (e.g., input prices), and an error term, vt, while
Eq. (9E.6) specifies the equilibrium condition that quantity supplied equals quantity demanded,
creating a system of three equations with three endogenous variables. The interaction of the
specified market forces solves this system, generating equilibrium values for the variables Pt* and
Q; = Q(d* = Q/.
To generate demand and supply elasticity estimates simultaneously, we used 2SLS and/or
3SLS procedures. For the 2SLS estimates, observed price is regressed against the exogenous
instruments (i.e., the supply and demand "shifter" variables). The fitted (or predicted) values for
the price variable are then employed as observations of the right-hand side price variable in the
supply and demand equations. In the second stage, the 2SLS estimators are generated by running
OLS on these calculated instrumental variables. Also, the 2SLS estimates are used to estimate
errors in the structural equations, which then can be used to estimate the variance-covariance
matrix of the structural equations' errors. For the 3SLS estimates, this information is used at the
third stage to perform a generalized least squares (GLS) estimation of a single large equation
9-120
-------
Economic Impact Analysis
composed from the individual structural equations. If this process is done with all variables
expressed in natural logarithms, the coefficient on the price variable in the demand equation yields
an estimate of the constant elasticity of demand.
9E.2.1 Demand Equation Estimation
Demand equations were estimated using a general specification where the quantity of boats
or Small SI equipment consumed is expressed as a function of price, number of households or
housing starts, per capita household or disposable income, and a time trend. Trends were included
as a general way to model the effects of changes in tastes and preferences. All price and income
variables were deflated by the implicit gross domestic product (GDP) deflator. The endogenous
variables in the equations are unit sales and own-price. The exogenous variables include the
household and income variables and the time trend. The list of instruments includes these
exogenous variables and supply factors influencing the price of the product: wages and a producer
price index for material inputs.
9E.2.2 Data Sets
The National Bureau of Economic Research (NBER) data discussed in the supply elasticity
section of the analysis plan ( RTI, 2005) contain data on production quantities, price indices, and
suitable instruments to inform a demand analysis for recreational boats (SIC 3732). In its Current
Industrial Reports (CIR) series, the U.S. Census Bureau produces an annual summary of the
production of motors and generators and a summary of production of several types of lawn and
garden equipment; both of these reports include the number of units manufactured and the value of
production (U.S. Census Bureau, 1998; 2000). For the walk-behind lawnmowers regression, we
used several data series reported in a study by Air Improvement Resource, Inc., and National
Economic Research Associates (AIR/NERA, 2003). The U.S. Census Bureau publishes historical
data on household income and housing starts (U.S. Census Bureau, 2002; 2004), and we collected
price, wage, and material cost indexes from the Bureau of Labor Statistics (BLS) (BLS,
2004a,b,c,d,e). Lastly, we obtained an implicit GDP price deflator from the U.S. Bureau of
Economic Analysis (BEA) (BEA, 2004). The following variables from these sources were used in
the regression:
unit sales of boats (Bartlesman et al., 2000),
price index for boats (Bartlesman et al., 2000),
lawn and garden equipment units produced (U.S. Census Bureau, AIR/NERA),
lawn and garden equipment value of production (U.S. Census Bureau),
producer price index for walk-behind lawnmowers (BLS),
households (U.S. Census Bureau),
housing starts (U.S. Census Bureau),
per capita income and population (U.S. Census Bureau, 2002; BEA, 2004),
average hourly earnings for production workers (BLS; Bartlesman et al., 2000),
price index for plastic and other materials and engines (BLS; Bartlesman et al., 2000),
and GDP deflator (BEA).
Some care was needed in using the time series from the CIR data set. Occasional changes
9-121
-------
Final Regulatory Impact Analysis
in category definition and the Census Bureau's need to suppress some data to maintain
confidentiality created difficulties in constructing consistent data series over the 2-decade time
period. Nonetheless, we were able to assemble the following series: commercial nonriding
mowers, commercial riding mowers, consumer lawn mowers, tillers and two-wheel tractors, snow
throwers, edgers and trimmers, vacuums and blowers, and lawn and garden tractors. Statistically
significant parameter estimates were obtained for commercial nonriding mowers, tillers/two-wheel
tractors, edgers/trimmers, and lawn and garden tractors.
We were not able to obtain a useful elasticity estimate for consumer lawn mowers using
CIR data, perhaps because of aggregation biases in that category of the CIR data set. Because
consumer lawn mowers are a critical segment of the entire Small SI sector, we used an alternate
data set for our demand elasticity estimate. The data AIR/NERA used in their recent study proved
very useful in this regard (AIR/NERA, 2003). In that study, the authors used a single-equation
OLS regression to obtain a demand elasticity parameter, a procedure that RTI believes to be
inadequate because the market process simultaneously determines price and quantity in the
demand equation. However, using the same data series cited by AIR/NERA supplemented by data
collected by RTI, we were able to obtain a reasonable estimate using the 3SLS regression
described above.
9E.2.3 Results of Demand Elasticity Estimation
In this section, we present regression results used in the EIA. Table 9E-8 shows the
parameter estimate for the marine sector, which is -2.0. Although the methodology and data sets
are quite different, this result is consistent with the ones obtained by Raboy (1987) in his study
almost 20 years ago.
9-122
-------
Economic Impact Analysis
Table 9E-8: Results of Econometric Estimation of Boat Demand Equation: 1958 to 1996
Recreational Boats—SIC 3732
Dependent Variable—Regression Unit Sales per Capita
Intercept -27.9
(-10.3)
Price -2.0
(-2.04)
Disposable income per capita 1.83
(5.85)
Trend -0.19
(-2.15)
Adjusted R2 0.81
Observations (years) 39
(1958-1996)
Notes: 1. Numbers in parentheses are t-ratios (coefficient estimate divided by its standard error)
(except for the year ranges in the last row of the table).
2. All exogenous and endogenous variables are in natural log.
In Table 9E-9, we present demand elasticity results for Small SI equipment. Our estimate
for walk-behind lawnmowers is -0.2 (inelastic). The value obtained for other nonhandheld
categories such as commercial nonriding mowers and lawn and garden tractors is higher at (-0.9,
-1.0). In contrast, the demand estimate for edgers/trimmers is elastic (-1.9), suggesting that
consumers are more willing to forego purchases of these items at higher prices. The
edgers/trimmers' value was used for all handheld equipment. Results for generators, which range
from -1.1 to -1.4, are shown in Table 9E-10.
9-123
-------
Final Regulatory Impact Analysis
Table 9E-9: Results of Econometric Estimation of Small SI Demand Equations:
1980 to 1997 (1973-2002 for Consumer Mowers)
Dependent
Variable —
Regression
Method
Intercept
Price
Per capita income
Housing starts per HH (1
lag)
Trend
Adjusted or system
weighted R2
Observations (years)
Consumer Walk-
Behind Mowers
Units Sold per
Household
3SLS
-0.64
(-2.71)
-0.2
(-3.73)
—
0.23
(4.71)
—
0.547
29
(1973-2002)
Commercial
Mowers
Units Produced
2SLS
-35.19
(-4.41)
-0.9
(-2.74)
4.8
(5.76)
—
-0.20
(-1.58)
0.663
18
(1980-97)
Edgers and
Trimmers
Units Produced
2SLS
-4.69
(-0.63)
-1.9
(-6.05)
1.47
(1.79)
—
0.32
(2.52)
0.877
18
(1980-97)
Lawn and
Garden Tractors
Units Produced
2SLS
-7.22
(-1.46)
-1.0
(-2.29)
2.2
(4.36)
—
0.02
(0.26)
0.939
18
(1980-97)
Notes: 1. Numbers in parentheses are t-ratios (coefficient estimate divided by its standard error) (except for the
year ranges in the last row of the table).
2. All exogenous and endogenous variables are in natural log.
3. For lawnmowers, the income variable is actually per capita disposable income.
9-124
-------
Economic Impact Analysis
Table 9E-10: Results of Econometric Estimation of Gasoline-Powered Generators
Demand Equations: 1973 to 1998
Units Produced
Dependent Variable-Regression Small Generators (<5kW) Large Generators (>15kW)
Intercept
Price
Per capita income
Trend
16.4
(2.64)
-1.4
(-3.64)
-0.46
(-0.71)
-0.02
(-0.51)
-14.3
(-2.48)
-1.1
(-8.59)
2.7
(4.34)
-0.16
(-1.53)
Adjusted R2 0.609 0.723
Observations (years) 26 26
(1973-1998) (1973-1998)
Notes: 1. Numbers in parentheses are t-ratios (coefficient estimate divided by its standard error) (except for the
year ranges in the last row of the table).
2. All exogenous and endogenous variables are in natural log.
9-125
-------
Final Regulatory Impact Analysis
Appendix 9F: Derivation of Supply Elasticity
In economics, a production function is used to describe the relationship between inputs and
outputs of the production process. The production function in general is defined as follows
Q =f(L,K,M,t)
Qs = the quantity of the outputs supplied
L = the labor input or the number of labor hours
K = real capital stock or real capital consumed in the production
M= the material inputs
t = a time trend variable to reflect technology changes
In the competitive market, market forces constrain firms to produce at the cost minimizing
output level. Cost minimization allows for the duality mapping of a firm's technology
(summarized by the firm's production function) to the firm's economic behavior (summarized by
the firm's cost function). The total cost function of an industry in the short term follows:
where TC is the total cost of production, C is the variable cost of production (such as the cost of
materials and labor), and the other variables have previous defined. This approach assumes that
capital stock is fixed, or a sunk cost of production. This assumption is consistent with the goal of
the modeling post-control market changes likely to occur. Firms facing final regulatory emission
controls will consider embedded capital stock as a fixed or sunk cost in economic decision
making. Differentiating the total cost function with respect to Qs derives the marginal cost
function:
MC= h'(C,K,t,Qs)
where MC is the marginal cost of production and all other variables have been previously defined.
Profit maximizing competitive firms will choose to produce the quantity of output that
equate the market price (P) to the marginal cost of the production (MC). Setting the price equal to
the preceding marginal cost function and solving for Qs yields the following implied supply
function:
Q= S(P,PLPM,K,t,)
where P is the market price of the products, PL is the price of the labor, PM is the price of
materials, and all other variables have been previously defined.
To illustrates how the supply elasticity used in Appendix 9E can be expressed in terms of
the parameters of the production function (Equation 9E.3), we assume that production function is
represented by a Cobb-Douglas function with only two inputs (capital [K] and labor [L]) with a
9-126
-------
Economic Impact Analysis
constant return to scale,
where Q = output, L = labor input, and K = capital input. The cost function is written as
TC = wL + rK (9F.2)
where w = wage rate or unit labor cost, r = interest cost or unit capital cost. From equation (9F. 1),
L can be written as,
3)
Substituting L in the cost function with equation (9F.3),
TC = wL + rK = w {Q 1/aK la}tt/ ('-a>
We have
Q = (a/w)a/ ('-a> P a/ ('-K> K (9F.4)
Taking log function on both sides,
lnQ = a/(l-a) In ( a/w)+ a/(l-a) lnP+ InK (9F.5)
The price elasticity of supply can be written as
Supply elasticity = din Q/ din P = a/(I- a) (9F.6)
9-127
-------
Final Regulatory Impact Analysis
Appendix 9G: Initial Market Equilibrium - Price Forecasts
The EIM analysis begins with current market conditions: equilibrium supply and demand.
To estimate the economic impact of a regulation, standard practice uses projected market
equilibrium (time series of prices and quantities) as the baseline and evaluates market changes
from this projected baseline. Consequently, it is necessary to forecast equilibrium prices and
quantities for future years.
Equilibrium price forecasts typically use one of two approaches (EPA 1999, p 5-25). The
first assumes a constant (real) price of goods and services over time. The second models a specific
time series where prices may change over time due to exogenous factors.
In the absence of shocks to the economy or the supply of raw materials, economic theory
suggests that the equilibrium market price for goods and services should remain constant over
time. As shown in Figure 7G-1, demand grows over time, in the long run, capacity will also grow
as existing firms expand or new firms enter the market and eliminate any excess profits. This
produces a flat long run supply curve. Note that in the short to medium run time frame the supply
curve has a positive slope due to limitations in how quickly firms can react.
S/Q
Short Run
Supptyi
Short Run
Supplyj
Short Run
Supp lyj
Long Run
Supply
Figure 9G-1. Prices and Quantities in Long Run Market Equilibrium
9-128
-------
Economic Impact Analysis
If capacity is constrained (preventing the outward shift of the baseline supply curve) or if
the price of production inputs increase (shifting the baseline supply curve upward over time), then
prices may trend upward reflecting that either the growth in demand is exceeding supply or the
commodity is becoming more expensive to produce.
It is very difficult to develop forecasts events (such as those mentioned above) that
influence long run prices. As a result, the approach used in this analysis is to use a constant 2005
observed price.
9-129
-------
Final Regulatory Impact Analysis
Appendix 9H: Sensitivity Analysis
The Economic Impact Analysis presented in this Chapter 9 is based on the Economic
Impact Model (EIM) developed for this analysis. The EIM reflects certain assumptions about
behavioral responses (modeled by supply and demand elasticities), and what the baseline
equipment prices are used in the model. This appendix presents a sensitivity analysis for
alternatives in the model. Three scenarios are examined:
• Scenario 1: alternative market supply and demand elasticity parameters
• Scenario 2: alternative baseline prices for lawn mower and tractor
• Scenario 3: alternative gasoline price for social costs
• Scenario 4: change in consumer's behavior due this rule
The results of these sensitivity analyses are presented below. The results from Scenario 1
are presented for 2014 (the highest cost year) only with 2005$. The results for the Small SI and
Marine SI engine and equipment markets do not include the fuel savings. Instead, fuel savings are
added into the total social costs as a separate item.
In general, varying the elasticity parameters does not significantly change the results of the
economic impact assessment analysis presented above. The expected price increase remains
relative stable across the scenarios in comparing with the primary case for the Small SI and Marine
SI engine and equipment. The difference in expected price change between alternative and
primary scenarios is less than 0.5 percent. Total social costs are about the same across all
sensitivity analysis scenarios, $444 million. In addition, varying these model parameters does not
significantly affect the way the social costs are borne. In all cases, the end user (households) bear
the majority of the burden (over 76 percent), although there are differences in the way the costs are
borne among the scenarios between the change in either demand or supply elasticity. The share of
social costs end users (households) bear, for example, ranges from 66 to 98 percent.
With regard to the scenario of alternative baseline prices, although the difference in prices
is about 27%and 52% for lawn mower and tractors, respectively, the estimates on absolute price
change and social cost for each market are approximately the same as in the base case. However,
given that the baseline prices are different in these scenarios, there is some variation in projected
relative price and quantity change across the scenarios. The expected changes in relative prices
and quantity increase under the lower alternative baseline market price scenarios.
A recent higher gasoline price will have the impacts to our analyses. A higher gasoline
price is expected to increase the fuel savings estimated from primary analysis thus to reduce the
net social cost of the final emission standards. In addition, we will describe qualitatively how the
consumers would response to the fuel efficiency gains that result from applying technologies to
achieve the evaporative emission standards being finalized in this rule in Scenario 4.
9-130
-------
Economic Impact Analysis
9H.1 Model Elasticity Parameters
Consumer demand and producer supply responsiveness to changes in the commodity prices
are referred to by economists as "elasticity." The measure is typically expressed as the
percentage change in quantity (demanded or supplied) brought about by a percent change in own
price. A detailed discussion regarding the estimation and selection of the elasticities used in the
ELM are discussed in Appendix 9E. This component of the sensitivity analysis examines
the impact of changes in selected elasticity values, holding other parameters constant. The goal
is to determine whether alternative elasticity values significantly alter conclusions in this report.
9H.1.1 Alternative Supply and Demand Elasticity Parameters
The choice of supply and demand elasticities for the engine and equipment market is
important because changes in quantities in the equipment markets are the key drivers in the
derived demand functions used to link impacts in the engine and equipment markets. In addition,
the distribution of regulatory costs depends on the relative supply and demand elasticities used in
the analysis. For example, consumers will bear less of the regulatory burden if they are more
responsive to price changes than producers.
Table 9H-1 reports the upper- and lower-bound values of the engine and equipment market
elasticity parameters (supply and demand) used in the sensitivity analysis. The engine and
equipment market supply elasticities are derived econometrically. Therefore, the upper and lower
bound values were computed using the coefficient and standard error values associated with the
econometric analysis and reflected 95 percent confidence interval.
9-131
-------
Final Regulatory Impact Analysis
Table 9H-1: Alternative Supply and Demand Elasticities Used in Sensitivity Analysisa'b
Parameter/Market Upper Bound Primary Case Lower Bound
Supply Elasticities
Engines
Marine and Small SI
5.0
9.5
13.9
Equipment
Marine SI
All other vessel types
PWC
Small SI
Small SI (handheld/nonhandheld)
Gensets/welders
3.1
3.5
5.1
6.2
3.8
5.2
10.0
8.8
4.4
6.9
14.8
11.4
Demand Elasticities
Engines
Marine and Small SI
Derived Demand Derived Demand Derived Demand
Equipment
Marine SI
All vessel types
Small SI
Handheld
Lawn mowers
Other lawn and garden
Gensets/welders — Class I
Gensets/welders — Class II
All other handheld
-3.9
-2.5
-0.3
-1.5
-2.2
-1.4
-1.9
-2.0
-1.9
-0.2
-0.9
-1.4
-1.1
-1.0
-0.1
-1.3
-0.1
-0.3
-0.6
-0.8
-0.1
a For the demand elasticity, EPA computed upper- and lower-bound estimates using the coefficient and standard error
values associated with its econometric analysis and reflect a 95 percent confidence interval.
b For the supply elasticity, see Li, May 19, 2008, memorandum prepared for this Docket, "Supply Elasticity Estimation
Report", for the interval estimates.
9H.1.2 Engines and Equipment Market (Supply Elasticity Parameters)
The results of the EIM using these alternative supply elasticity values for the Small SI and
Marine SI engine and equipment markets are reported in Tables 9H-2. As can be seen in the table,
projected changes in market prices are stable across the upper- and lower-bound sensitivity
scenarios. The relative change in price is around the primary case by 0.3 percent. Absolute
9-132
-------
Economic Impact Analysis
quantities vary but the percentage changes in output are negligible for the two scenarios. The
change in total social surplus for 2014 also remains nearly unchanged across all scenarios and is
approximately the same as for the rule ($444 million).
However, varying the supply elasticity changes the social impacts (how the burden is
shared across markets). Manufacturers bear a smaller share of the social costs when they are more
responsive to price changes (supply upper bound scenario). As shown for the Small SI market,
engine and equipment manufacturers bear approximately 1.4 and 4.7 percent, respectively, in the
supply upper bound scenario compared to 2.0 and 6.6 percent in the base case. In contrast, they
bear a higher share of social cost when they are less responsive to price changes relative to the
base case (the supply lower bound scenario). For the Marine SI market, engine and equipment
manufacturers bear approximately 4.5 and 15.4 percent, respectively, in supply upper bound
scenario compared to 6.2 and 17.5 percent in the base case. In contrast, they bear a higher share
when they are less responsive to price changes relative to the base case (supply lower bound
scenario).
9-133
-------
Final Regulatory Impact Analysis
Table 9H-2: Sensitivity Analysis for Engine and Equipment Market Supply Elasticities
for 2013 a b
Scenario
Marine
Market-Level Impacts
Price
Engines
Equipment
Quantity
Engines
Equipment
Welfare Impacts (million $)
Change in engine manufacturers surplus
Change in equipment manufacturers
surplus
Change in end user (households) surplus
Small SI
Market-Level Impacts
Price
Engines
Equipment
Class I
Class II
HH
Quantity
Engines
Equipment
Class I
Class II
HH
Welfare Impacts (million $)
Change in engine manufacturers surplus
Change in equipment manufacturers
surplus
Change in end user (households) surplus
Subtotal Social Costs (million $)
Fuel Savings (million $)
Total Social Costs (million $)
Primary
Absolute
$266.5
$285.4
-10,883
-12,229
$10.5
$29.7
$130.0
$13.7
$9.5
$16.6
$23.7
$0.3
-303.992
-360.310
-209,284
-101,104
^9,992
$5.4
$18.1
$250.2
$443.8
$115.9
$327.9
Case
Relative"
2.4%
1.6%
-2.7%
-3.2%
6.2%
17.5%
76.4%
8.3%
2.6%
6.2%
2.6%
0.2%
-1.9%
-1.4%
-2.1%
-2.8%
-0.3%
2.0%
6.6%
91.4%
Supply Lower Bound
Absolute
$249.3
$252.1
-9,385
-10,684
$17.4
$34.1
$119.2
$13.5
$8.9
$15.7
$22.1
$0.3
-279,592
-326,161
-194,465
-92,979
-38,717
$9.31
$30.4
$234.1
$444.4
$115.9
$328.6
Relative"
2.2%
1.4%
-2.3%
-2.8%
10.2%
20.0%
69.8%
8.1%
2.4%
5.9%
2.4%
0.1%
-1.7%
-1.2%
-2.0%
-2.6%
-0.2%
3.4%
11.1%
85.5%
Supply Upper Bound
Absolute
$273.5
$307.0
-11,838
-13,203
$7.7
$26.3
$136.0
$13.8
$9.8
$16.9
$24.5
$0.4
-314,636
-367,250
-215,513
-104,638
^7,100
$3.76
$12.9
$256.9
$443.4
$115.9
$327.6
Relative"
2.5%
1.7%
-2.9%
-3.5%
4.5%
15.4%
80.0%
8.3%
2.6%
6.3%
2.7%
0.2%
-1.9%
-1.4%
-2.2%
-2.9%
-0.3%
1.4%
4.7%
93.9%
a Figures are in 2005 dollars.
b For "prices" rows the "relative" column refers to the relative change in price (with regulation) from the baseline
price. For "Surplus" rows, the "relative" column contains the distribution of total surplus changes among
stakeholders (consumers and producers).
9-134
-------
Economic Impact Analysis
9H.1.3 Equipment Market (Demand Elasticity Parameters)
Sensitivity analysis was also conducted for the equipment market demand elasticities. The
range of demand elasticity values evaluated for each market is provided in Table 9H-1.
The demand elasticities for the engine markets are derived as part of the model, and therefore
sensitivity analysis was not conducted on those parameters.16 In other words, the change in the
equipment market quantities determines the demand responsiveness in the engine market. As a
result, the demand sensitivity analysis for engine markets is indirectly shown in Table 9H-2.
16For a discussion of the concept of derived demand, see Section 9.2.3.2 Incorporating Multimarket
Interactions.
9-135
-------
Final Regulatory Impact Analysis
Table 9H-3: Sensitivity Analysis for Equipment Market Demand Elasticities for 2013
a,b
Scenario
Marine
Market-Level Impacts
Price
Engines
Equipment
Quantity
Engines
Equipment
Welfare Impacts (million S)
Change in engine manufacturers surplus
Change in equipment manufacturers
surplus
Change in end user (households) surplus
Small SI
Market-Level Impacts
Price
Engines
Equipment
Class I
Class II
HH
Quantity
Engines
Equipment
Class I
Class II
HH
Welfare Impacts (million $)
Change in engine manufacturers surplus
Change in equipment manufacturers
surplus
Change in end user (households) surplus
Subtotal Social Costs (million S)
Fuel Savings (million S)
Total Social Costs (million S)
Primary
Absolute
$266.5
$285.4
-10,883
-12,229
$10.5
$29.7
$130.0
$13.7
$9.5
$16.6
$23.7
$0.3
-303.992
-360.310
-209,284
-101,104
^9,992
$5.4
$18.1
$250.2
$443.8
$115.9
$327.9
Case
Relative"
2.4%
1.6%
-2.7%
-3.2%
6.2%
17.5%
76.4%
8.4%
2.6%
6.2%
2.6%
0.2%
-1.9%
-1.4%
-2.1%
-2.8%
-0.3%
2.0%
6.6%
91.4%
Demand Lower Bound
Absolute
$290.5
$443.4
-905
-934
$0.9
$2.3
$170.4
$14.0
$10.1
$17.2
$25.1
$0.3
-99,098
-146,272
-66,388
-35,948
^3,939
$1.8
$6.2
$267.3
$448.8
$115.9
$333.0
Relative"
2.7%
2.5%
-0.2%
-0.2%
0.5%
1.3%
98.2%
8.4%
2.7%
6.4%
2.7%
0.2%
-0.6%
-0.6%
-0.8%
-1.9%
-0.3%
0.7%
2.3%
97.1%
Demand Upper Bound
Absolute
$255.7
$211.2
-15,198
-17,854
$14.7
$42.8
$110.9
$13.5
$9.0
$15.9
$22.5
$0.3
^86,671
-558,525
-340,554
-156,182
-61,788
$8.4
$28.3
$235.4
$440.6
$115.9
$324.7
Relative"
2.3%
1.2%
-3.7%
-4.7%
8.7%
25.4%
65.8%
8.2%
2.5%
6.0%
2.5%
0.2%
-3.0%
-2.1%
-3.3%
-3.6%
-0.4%
3.1%
10.4%
86.5%
a Figures are in 2005 dollars.
b For "prices" rows the "relative" column refers to the relative change in price (with regulation) from the baseline
price. For "Surplus" rows, the "relative" column contains the distribution of total surplus changes among
stakeholders (consumers and producers).
9-136
-------
Economic Impact Analysis
As shown in Tables 9H-3, market prices are relative stable across the upper- and lower-
bound sensitivity scenarios. The relative change in price is around the primary case by 0.5
percent. Absolute quantities vary and the percentage changes in output are small for the two
scenarios. There is also a small change in total social surplus for 2014 compared to the primary
case ($444 million) but this is negligible in terms of the percentage change.
In comparing Table 9H-3 with Table 9H-2 , all quantitative estimates for the market
impacts (price and quantity changes) by the ELM model are a little more sensitive to the alternative
demand elasticities than the alternative supply elasticities. However, theses changes remain in a
reasonable range when compared with the rule, across both the upper and lower bound demand
elasticity scenarios for the equipment markets.
It should be noted, varying the demand elasticity changes the social impacts (how the
burden is shared across markets) as in the case of changing the supply elasticity. Manufacturers
bear a smaller share of the social costs when consumers are less responsive to price changes
(demand lower bound scenario). As shown for the Small SI market, engine and equipment
manufacturers bear approximately 0.7 and 2.3 percent, respectively, in the demand lower bound
scenario compared to 2.0 and 6.6 percent in the base case. In contrast, they bear a higher share of
social cost when consumers are more responsive to price changes relative to the base case (the
demand upper bound scenario). For the Marine SI market, engine and equipment manufacturers
bear approximately 0.5 and 1.3 percent, respectively, in demand lower bound scenario compared
to 6.2 and 17.5 percent in the base case. In contrast, they bear a higher share when consumers are
more responsive to price changes relative to the base case (demand upper bound scenario).
9-137
-------
Final Regulatory Impact Analysis
9H.2 Alternative Baseline Prices for Lawn Mower & Tractor
As discussed in Section 9.3.2, the starting point for the economic impact analysis is initial
market equilibrium conditions (prices and quantities) that exist prior to the implementation of new
standards. At the pre-control market equilibrium conditions, consumers are willing to purchase the
same amount of a product that producers are willing to produce at the market price. Since the
lawn mower and tractor equipment are the most popular equipment in the Small SI market and
their prices range widely, a sensitivity analysis was performed to examine how alternative baseline
prices for lawn mower and tractor influence the EIM results.
Table 9H-4: Market Sensitivity Analysis for Alternative Baseline for Lawnmower & Tractor
Prices in 2014 a
Scenario
Lawn Mowers
(UL 125)
Primary
scenario
Low price
scenario
Tractors
(UL 250)
Primary
scenario
Low price
scenario
Average
Baseline
Price
$218
$159
$1,937
$928
Market Results
Change in
Price
(Absolute)
$12.07
$12.03
$21.11
$20.98
Change
in Price
5.5%
7.6%
1.1%
2.3%
Change
Change in in
Quantity Quantity
(Absolute) (%)
-76,121 -1.1%
-104,068 -1.4%
-23,690 -1.1%
-49,134 -2.3%
Welfare Results
Change in
End Users
(Households)
Surplus
(Million $)
-$87.0
-$86.5
-$45.6
-$45.1
Change in
Equipment
Manufacture
r Surplus
(Million $)
-$1.7
-$1.7
-$4.6
-$4.5
Change
in Total
Surplus
(Million
$)
-$88.7
-$88.2
-$50.2
-$49.6
a Figures are in 2005 dollars.
We selected the lower end market prices as the alternative baseline prices for lawn mower
and tractor in this sensitivity analysis. As shown in Table 9H-4, when these pre-control baseline
prices are allowed to vary, the absolute change in market prices remains nearly unchanged when
compared with the rule, although the relative price change and absolute quantity change are
expected to be higher in the alternative baseline price case. This is because the change in absolute
price is ultimately determined by the per unit compliance cost and market supply and demand
elasticities. In contrast, the change in relative price is determined by the ratio between the per-unit
compliance cost and the baseline price. The lower the initial baseline price, the higher the ratio is
for a given per unit compliance cost. Therefore, the change in the relative price is higher. In this
market, consumers are expected to response to the higher relative price change by purchasing less
equipment. As a result, the expected change for quantity is higher in the lower baseline prices
case. Also as seen in Table 9H-4, varying the baseline prices are not expected to substantially
change the social cost estimates in these markets or alter the distribution of the social costs across
9-138
-------
Economic Impact Analysis
the stakeholders.
9H.3 Alternative Gasoline Price for Social Costs
As discussed in 9.2.4.2, there are fuel savings attributed to the final emission control
programs, reflecting the use of more fuel efficient technology to meet evaporative and exhaust
emission requirements. These fuel savings are included in the social cost analysis because they are
savings that accrue to society.
For the social costs analysis, EPA calculated fuel savings using the 2005 pre-tax price of
gasoline of $1.81 per gallon and this value is held constant for each future years (see section
9.3.5). Because of the recent trend of increasing gasoline prices, we may be understating the fuel
savings in our cost analysis. This is reflected in recent fuel price projections from the EIA's 2008
Annual Energy Outlook.18 To investigate the sensitivity of the net social cost to future fuel prices,
we used the AEO "reference case" and "high price case" scenarios, as described in section 6.7.1.
As indicated in Table 9H-5, comparing the AEO 2008 reference case with the primary
case, the annual net social cost of the final standards is lower because of a higher fuel savings.
The net present value of social cost decreases from $4.2 billion to $3.9 billion using a 3 percent
discount rate. The net present value of social costs for the period of analysis falls from $2.7 billion
to $2.6 billion using a 7 percent discount rate.
As shown in Table 9H-6, fuel savings are even higher in the AEO 2008 high price case.
Based on these fuel price projections, the increased fuel savings estimates would actually be higher
than the projected costs, once the new equipment is fully phased in. In comparison with the
primary case, the net present value of social costs decreases from $4.2 billion to $2.0 billion using
a 3 percent discount rate. Using a 7 percent discount rate, the net present value of social costs for
the period of analysis falls from $2.7 billion to $1.6 billion.
18 Energy Information Administration, "Annual Energy Outlook 2008; with Projections to 2030,'
DOE/EIA-0383(2008), June 2008.
9-139
-------
Final Regulatory Impact Analysis
Table 9.H-5: Sensitivity of Gasoline Price to Net Social Cost -AEO 2008 Reference Case
Year
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
2037
NPVat3%
NPVat7%
Total
Engineering
Costs
$54
$127
$271
$329
$442
$445
$450
$412
$409
$398
$403
$408
$413
$418
$423
$428
$433
$438
$443
$449
$454
$459
$464
$469
$474
$479
$484
$489
$494
$499
$7,705
$4,559
Total Social
Costs
$54
$126
$267
$324
$435
$439
$444
$407
$404
$394
$399
$404
$409
$414
$419
$424
$429
$434
$439
$444
$449
$454
$459
$464
$469
$474
$479
$484
$489
$494
$7,617
$4,506
Fuel Savings
AEO 2008
Reference Case
Projected
Gasoline Price
$5
$11
$24
$52
$80
$105
$126
$140
$153
$168
$184
$204
$221
$228
$241
$251
$259
$269
$278
$287
$297
$306
$314
$319
$324
$329
$333
$337
$341
$345
$3,743
$1,956
Net Social Costs
AEO 2008
Reference case
Projected
Gasoline Price
$49
$116
$243
$272
$355
$334
$318
$266
$251
$225
$215
$200
$187
$185
$178
$172
$169
$164
$160
$157
$152
$147
$144
$144
$144
$145
$146
$147
$148
$149
$3,873
$2,550
Net Social Costs
2005 Constant
Gasoline Price
(Primary Analysis)
$51
$118
$248
$280
$365
$344
$328
$273
$253
$229
$221
$213
$207
$203
$198
$194
$191
$189
$188
$187
$186
$185
$186
$186
$187
$188
$189
$190
$192
$193
$4,242
$2,731
9-140
-------
Economic Impact Analysis
Table 9.H-6: Sensitivity of Gasoline Price to Net Social Cost - AEO 2008 Higher Case
Year
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
2037
NPVat3%
NPVat7%
Total Engineering
Costs
$54
$127
$271
$329
$442
$445
$450
$412
$409
$398
$403
$408
$413
$418
$423
$428
$433
$438
$443
$449
$454
$459
$464
$469
$474
$479
$484
$489
$494
$499
$7,705
$4,559
Total Social
Costs
$54
$126
$267
$324
$435
$439
$444
$407
$404
$394
$399
$404
$409
$414
$419
$424
$429
$434
$439
$444
$449
$454
$459
$464
$469
$474
$479
$484
$489
$494
$7,617
$4,506
Fuel Savings
AEO 2008
Higher Case
Projected
Gasoline Price
$5
$11
$27
$61
$100
$137
$169
$198
$226
$253
$277
$301
$325
$351
$371
$386
$402
$411
$423
$437
$450
$464
$479
$487
$494
$501
$507
$514
$520
$525
$5,585
$2,886
Net Social Costs
AEO 2008
Higher Case
Projected
Gasoline Price
$49
$115
$240
$263
$336
$302
$274
$209
$179
$141
$122
$102
$84
$63
$47
$37
$27
$23
$16
$7
($2)
($11)
($20)
($23)
($25)
($27)
($29)
($30)
($31)
($32)
$2,032
$1,620
Net Social Costs
2005 Constant
Gasoline Price
(Primary Analysis)
$51
$118
$248
$280
$365
$344
$328
$273
$253
$229
$221
$213
$207
$203
$198
$194
$191
$189
$188
$187
$186
$185
$186
$186
$187
$188
$189
$190
$192
$193
$4,242
$2,731
9-141
-------
Final Regulatory Impact Analysis
9H.4 Discussion of Fuel Savings Effects on EIA Analysis
The final evaporative and exhaust controls are expected to result in fuel savings for
consumers due to the use of more fuel efficient technology to meet evaporative and exhaust
emission requirements. Although these fuel savings represent a cost savings for consumers, our
ELM model assumes that consumers do not consider these fuel savings in their purchasing
decisions. As a result, the ELM model does not include a shift in the demand curve due to more
fuel efficient products causing the demand curve in each small SI or marine SI equipment market
to remain unchanged. This has been explained in Section 9.2.4.2. Although the fuel savings are
not directly incorporated in the market model, the benefits to society are captured as an added line
item to offset the total social costs of the program.
Figure 9H. 1- (a) summarizes the EIA primary analysis that was presented earlier. The
market equilibrium of a representative equipment market is at point A prior to the regulation (the
baseline case). With the regulation, compliance costs lead to an upward shift in the market supply
curve. The new market equilibrium is at point B (the primary case). In this case, the demand curve
for the equipment remains unchanged; the regulatory program only influences the supply side of
the market (e.g. additional costs of the program [S0 to St ]); and fuel savings are added as a line
item to the social costs.
This section explores an alternative way to treat the fuel savings in the analysis by
assuming consumers are fully aware of the fuel savings and realize that these improvements will
lower the future cost of using equipment. Since many consumers may prefer more fuel efficient
equipment, they may be willing to initially pay more for Small SI and Marine SI equipment
because of these future operation savings. This leads to a change in demand and shifts the demand
curve upward.
This concept may be described using illustrations of supply and demand curves. As
indicated in Figure 9H.l-(b), if the fuel savings (lower costs of operating equipment) are
incorporated in the model as a demand shift (DA to Dc) that ultimately raises the quantity that
buyers wish to purchase at a higher market price. The new equilibrium that includes this demand
shift is at point C (the alternative case). Incorporating this demand shift in the model would lead
to the following effects:
• Equipment price and quantity: Including the demand shift due to fuel savings in the
model will increase the projected price and quantity in the equipment market
relative to the primary case. Consumers are willing to purchase more Small SI and
Marine SI equipment at a higher price because of fuel efficiency characteristics.
Since the equipment has become more desirable, producers could sell these
equipment at a higher price. In comparing with the primary case (PB and QB), the
new market equilibrium price and quantity (Pc and Qc) in the alternative case are
higher.
• Fuel Savings: At a given Q, consumers are willing to pay more for the equipment
with fuel efficiency technology. The fuel savings are measured by difference in
9-142
-------
Economic Impact Analysis
consumer's willingness to pay as indicated in the shaded area in Figure 9.H. l-(b)
• Social costs: Including the demand shift due to fuel savings will increase total
social welfare as indicated in Figure 9.H.2. by the area GFBC. This is because
consumers are willing to purchase more at a higher price and producers are
charging more and selling more. Both consumer surplus and producer surplus are
increased because the fuel efficient technology. The social cost is measured as the
change in total social welfare and is now lower in the alternative case.
• Distribution of social costs: the primary case assumes all the fuel savings benefits
are attributed to the consumer. In the alternative case, consumers and producers
share these benefits since the demand shift leads to high equipment prices and
higher sells. If producers are more responsive to equipment price changes than
consumers (e.g. their supply elasticity is higher than the consumer's demand
elasticity), they may receive a higher share of the fuel savings benefits in the form
of higher profits from equipment sales.
The likelihood of consumers considering fuel savings in their purchasing decisions is
dependent on the magnitude of the fuel savings relative to their income. For the small handheld
equipment and lawnmowers used for personal lawn maintenance, operation cost savings would
likely have a small influence on equipment purchase decisions. In contrast, for operators of large
recreational boats or lawnmowers used by lawn care businesses, the additional gasoline cost
savings could influence purchase decisions. While the directional effects of these decisions are
discussed above, EPA does not have the sufficient data and information to quantify these effects.
9-143
-------
Final Regulatory Impact Analysis
Figure 9.H.1: Demand Side Effect of Fuel Savings in the Equipment Market
$/
equipment
I I
equipment
QQ QA Equipment
A} 2005 Gasoline Price
EIA Primary Scenario
J I
Equipment
B) Include Demand Shift from
Increased Fuel Efficiency
Figure 9.H.2: Social Welfare Change
equipment
'DA
QEQ,;QA
Welfare Change
Equipment
9-144
-------
-------
Small-Business Flexibility Analysis
CHAPTER 10: Small-Business Flexibility Analysis
This chapter presents our Small Business Flexibility Analysis (SBFA) which evaluates
the impacts of the rule on small businesses. Prior to issuing our proposed rule, we analyzed the
potential impacts of our program on small businesses. As a part of this analysis, we convened a
Small Business Advocacy Review Panel (SBAR Panel or 'the Panel'), under the requirements of
the Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory
Enforcement Fairness Act (SBREFA) of 1996, 5 USC 601 et seq. Through the Panel process,
we gathered advice and recommendations from small entity representatives (SERs) who would
be affected by the regulation. The Panel issued a report recommending that EPA consider and
seek comment on a wide range of regulatory alternatives to mitigate the impacts of the
rulemaking on small businesses. The Panel report has been placed in the rulemaking record.
In the proposal, EPA proposed provisions consistent with each of the Panel's
recommendations and sought comments on all of the small business provisions. We received a
number of comments during the comment period after we issued the proposal. A summary of all
comments pertaining to the small business provisions can be found in our Summary and
Analysis of Comments document contained in the public docket for this rulemaking. A list of
the small business provisions being adopted with the final rule is presented in section 10.7
below.
10.1 Overview of the Regulatory Flexibility Act
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 72 FR 28098, May 18, 20007). 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 (72
FR 28098).
Section 609(b) of the Regulatory Flexibility Act further directs the Panel to report on the
comments of small entity representatives and make findings on issues related to identified
elements of the IRFA under section 603 of the Regulatory Flexibility Act. Key elements of an
IRFA are:
- A description of and, where feasible, an estimate of the number of small entities to
which the proposed rule will apply;
- Projected reporting, record keeping, and other compliance requirements of the proposed
rule, including an estimate of the classes of small entities which will be subject to the
requirements and the type of professional skills necessary for preparation of the report or
record;
- An identification to the extent practicable, of all other relevant Federal rules which may
10-1
-------
Final Regulatory Impact Analysis
duplicate, overlap, or conflict with the proposed rule;
- Any significant alternatives to the proposed rule which accomplish the stated objectives
of applicable statutes and which minimize any significant economic impact of the
proposed rule on small entities.
The Regulatory Flexibility Act was amended by SBREFA to ensure that concerns
regarding small entities are adequately considered during the development of new regulations
that affect those entities. Although we are not required by the Clean Air Act to provide special
treatment to small businesses, the Regulatory Flexibility Act requires us to carefully consider the
economic impacts that our rules will have on small entities. The recommendations made by the
Panel may serve to help lessen these economic impacts on small entities when consistent with
Clean Air Act requirements.
For purposes of assessing the impacts of this action on small entities, a small entity is
defined as: (1) a small business as defined by the Small Business Administration's (SBA)
regulations at 13 CFR 121.201; (2) a small governmental jurisdiction that is a government of a
city, county, town, school district or special district with a population of smaller than 50,000;
and (3) a small organization that is any not-for-profit enterprise which is independently owned
and operated and is not dominant in its field.
After considering the economic impacts of today's final rule on small entities, we believe
that this action will not have a significant economic impact on a substantial number of small
entities. The small entities directly regulated by this final rule cover a wide range of small
businesses including engine manufacturers, equipment manufacturers, boat manufacturers, fuel
tank manufacturers, and fuel hose manufacturers. No small governmental jurisdictions or small
non-profits are impacted by this final rule. We have determined that 61 small businesses will
experience an impact of greater than 1 percent. These 61 companies represent less than 5
percent of the small business identified by EPA.
Despite the determination that this rule will not have a significant economic impact on a
substantial number of small entities, we have prepared a Small Business Flexibility Analysis that
has all the components of a final regulatory flexibility analysis (FRFA). An FRFA examines the
impact of the final rule on small businesses along with regulatory alternatives that could reduce
that impact. The Small Business Flexibility Analysis is presented in this chapter.
10.2 Need for and Objective of the Rulemaking
A detailed discussion on the need for and objectives of this final rule are located in the
preamble to the final rule. As presented in Chapter 8, controlling exhaust and evaporative
emissions from Small SI engines and equipment and Marine SI engines and vessel has important
public health and welfare benefits.
Section 213(a) of the CAA directs EPA to: (1) conduct a study of emissions from
nonroad engines and vehicles; (2) determine whether emissions of CO, NOx, and VOCs from
nonroad engines and vehicles are significant contributors to ozone or CO in more than one area
10-2
-------
Small-Business Flexibility Analysis
which has failed to attain the National Ambient Air Quality Standard (NAAQS) for ozone or
CO; and (3) if nonroad emissions are determined to be significant, regulate those categories or
classes of new nonroad engines and vehicles that cause or contribute to such air pollution.
Section 213(a)(3) states that the emission standards "shall achieve the greatest degree of
emission reduction achievable through the application of technology" giving appropriate
consideration to cost, noise, energy, safety, and lead time.
The Nonroad Engine and Vehicle Emission Study required by section 213(a)(l) was
completed in November 1991. The determination of the significance of emissions from nonroad
engines and vehicles in more than one NAAQS nonattainment area was published on June 17,
1994. At the same time, the first set of regulations for new land-based nonroad compression-
ignition (CI) engines at or above 37 kW was promulgated. EPA has also issued proposed or
final rules for most other categories of nonroad engines, including engines used in lawn and
garden equipment, recreational marine vessels, forklifts, recreational vehicles, locomotives, and
ships. In addition, EPA has revised the emission standards for many of these categories of
nonroad engines one or more times to achieve further emission reductions.
In addition to the general authority to regulate nonroad engines under the CAA, section
428 of the Omnibus Appropriations Bill for 2004 requires EPA to propose and finalize new
regulations for nonroad spark-ignition engines less than 50 horsepower (hp). The Bill directs
EPA to propose regulations by December 1, 2004 and finalize them by December 31, 2005.
EPA's assessment of new standards is to be carried out under section 213 of the CAA.
Finally, section 205 of Public Law 109-54 included an additional requirement that EPA
complete a technical study, to look at safety issues related to the potential standards called for
under the Omnibus Appropriations Bill for 2004. The law directed EPA to complete the study
prior to issuing the proposal called for in the Omnibus Appropriations Bill for 2004. In response
to this requirement, EPA prepared a technical study on safety in coordination with the Consumer
Product Safety Commission (CPSC). The study analyzes the incremental risk of fire and burn to
consumers that could result from the new standards. EPA published the study in March 2006.
In response to these requirements, today's action adopts controls on exhaust and
evaporative emissions from Small SI engines and equipment and Marine SI engines and vessels.
10.3 Summary of Significant Public Comments
In the proposal, EPA proposed provisions consistent with each of the Panel's
recommendations and sought comments on all of the small business provisions (see 72 FR
28245, May 18, 2007). As noted earlier, we received a number of comments during the
comment period after we issued the proposal. A summary of all comments pertaining to the
small business provisions can be found in our Summary and Analysis of Comments document
contained in the public docket for this rulemaking. A few changes have been made to some of
the proposed flexibilities in response to the comments as well as other changes made in the
rulemaking. Those changes are noted in section 10.7.1 below.
10-3
-------
Final Regulatory Impact Analysis
10.4 Definition and Description of Small Entities
Small entities include small businesses, small organizations, and small governmental
jurisdictions. As noted earlier, for the purposes of assessing the impacts of a rule on small
entities, a small entity is defined as: (1) a small business that meets the definition for business
based on the Small Business Administration's (SBA) size standards (see Table 10.4-1); (2) a
small governmental jurisdiction that is a government of a city, county, town, school district or
special district with a population of less than 50,000; and (3) a small organization that is any not-
for-profit enterprise which is independently owned and operated and is not dominant in its field.
Table 10.4-1 provides an overview of the primary SBA small business categories that will be
affected by this regulation.
Table 10.4-1: Small Business Definitions for Entities Affected by this Rule
Industry
Nonroad SI Engine Manufacturers
Equipment Manufacturers:
Farm Machinery
Lawn and Garden
Construction
Sawmill and Woodworking
Pumps
Air and Gas Compressors
Generators
Boat Builders
Fuel Tank Manufacturers:
Other Plastic Products
Metal Stamping
Metal Tank (Heavy Gauge)
Fuel Hose Manufacturers:
Rubber and Plastics Hoses
NAICS Codes3
333618
333111
333112
333120
333210
333911
333912
335312
336612
326199
332116
332420
326220
Defined as small entity by
SBA if less than or equal to:b
1,000 employees
500 employees
500 employees
750 employees
500 employees
500 employees
500 employees
1,000 employees
500 employees
500 employees
500 employees
500 employees
500 employees
a North American Industry Classification System
b As defined in SBA's regulations at 13 CFR part 121.
10.4.1 Small SI Engines and Equipment
For Small SI engines and equipment, the SBA small business size standards are 1,000
employees for engine manufacturers, 1,000 employees for generator manufacturers, 750
employees for construction equipment manufacturers, and 500 employees for manufacturers of
other types of equipment. To identify companies that meet these criteria, we compiled a list of
engine manufacturers and equipment manufacturers using information from a database prepared
by Power Systems Research (PSR) that contains data on Small SI engines and equipment sold in
the United States. EPA augmented this information with the list of engine manufacturers
10-4
-------
Small-Business Flexibility Analysis
currently certifying with EPA under the Small SI engine regulations. We then found
employment data for each company (or parent company if an individual company is part of a
larger group) using databases such as the Thomas Register and Dunn and Bradstreet.
The SB A small business size standard for manufacturers that produce fuel tanks or fuel
hose is 500 employees. To identify companies that meet this criterion, we compiled a list of
manufacturers that produce fuel tanks and fuel hoses for the Small SI equipment market. The list
was based on information from the California Air Resources Board, who has recently adopted
requirements for Small SI engine fuel tank and fuel hose manufacturers, and additional
information from Small SI equipment manufacturers and the Association of Rotational Molders
International. We then found employment data for each of the companies (or parent company if
an individual company is part of a larger group) using databases such as Thomas Register and
onesourceexpress.com and discussions with some of the manufacturers.
10.4.2 Marine SI Engines and Vessels
For Marine SI engines and vessels, the SB A small business size standards are 1,000
employees for engine manufacturers and 500 employees for boat builders. To identify
companies that meet these criteria, we used a number of different sources. For engine
manufacturers, we compiled a list based on the engine manufacturers currently certifying with
EPA and the California Air Resources Board (CARB) under the existing Marine SI engine
regulations and augmented the list with additional information on SD/I manufacturers, who do
not currently have to certify with EPA. We gathered additional information from boat shows,
the Internet, trade magazines, the National Marine Manufacturers Association (NMMA), and
discussions with individual manufacturers. For vessel manufacturers, we used information from
a database of boat builders maintained by the U.S. Coast Guard.
The SB A small business size standard for manufacturers that produce fuel tanks or fuel
hose is 500 employees. For fuel tank and fuel hose manufacturers, we compiled a list based on
information gathered from the NMMA, trade shows, the Internet and discussions with
manufacturers. We then found employment data for these companies (or parent company if an
individual company is part of a larger group) using databases such as Thomas Register and
discussions with trade groups and individual manufacturers.
10.5 Type and Numbers of Small Entities Affected
As noted above, for each sector impacted by this final rule, SBA defines small entities by
number of employees. This section gives an overview of the Small SI engine and equipment
industries and the Marine SI engine and vessel industries, specifically related to small
businesses.
10.5.1 Small SI Engines and Equipment
10-5
-------
Final Regulatory Impact Analysis
Based on EPA certification records, the Small SI nonhandheld engine industry is made
up primarily of large manufacturers including Briggs and Stratton, Tecumseh, Honda, Kohler
and Kawasaki. The Small SI handheld engine industry is also made up primarily of large
manufacturers including Electrolux Home Products, MTD, Homelite, Stihl and Husqvarna. EPA
has identified 10 Small SI engine manufacturers that qualify as a small business under SB A
definitions. Half of these small manufacturers certify gasoline engines and the other half certify
liquefied petroleum gas (LPG) engines.
The Small SI equipment market is dominated by a few large businesses including Toro,
John Deere, MTD, Briggs and Stratton, and Electrolux Home Products. While the Small SI
equipment market may be dominated by just a handful of companies, there are many small
businesses in the market; however these small businesses account for less than 10 percent of
equipment sales. We have identified over three hundred equipment manufacturers that qualify as
a small business under the SB A definitions. More than 90 percent of these small companies
manufacture less than 5,000 pieces of equipment per year. The median employment level is 65
employees for nonhandheld equipment manufacturers and 200 employees for handheld
equipment manufacturers. The median sales revenue is approximately $9 million for
nonhandheld equipment manufacturers and $20 million for handheld equipment manufacturers.
EPA has identified 25 manufacturers that produce fuel tanks for the Small SI equipment
market that meet the SB A definition of a small business. Fuel tank manufacturers rely on three
different processes for manufacturing plastic tanks - rotational molding, blow molding and
injection molding. EPA has identified small business fuel tank manufacturers using the
rotational molding and blow molding processes but has not identified any small business
manufacturers using injection molding. In addition, EPA has identified two manufacturers that
produce fuel hose for the Small SI equipment market that meet the SBA definition of a small
business. The majority of fuel hose in the Small SI market is made by large manufacturers
including Avon Automotive and Dana Corporation.
10.5.2 Marine SI Engines and Vessels
Based on EPA certification records, the OB/PWC market is made up primarily of large
manufacturers including, Brunswick (Mercury), Bombardier Recreational Products, Yamaha,
Honda, Kawasaki, Polaris, Briggs & Stratton, and Nissan. Two companies qualify as a small
business under the SBA definitions. Tohatsu makes outboard engines. The other small business
is Surfango which makes a small number of motorized surfboards and has certified their product
asaPWC.
The SD/I market is made up mostly of small businesses; however, these businesses
account for less than 20 percent of engine sales. Two large manufacturers, Brunswick
(Mercruiser) and Volvo Penta, dominate the market. We have identified 28 small entities
manufacturing SD/I marine engines. The third largest company is Indmar, which qualifies as a
small business based on the SBA threshold of 1,000 employees. Based on sales estimates,
number of employees reported by Thomas Register, and typical engine prices, we estimate that
10-6
-------
Small-Business Flexibility Analysis
the average revenue for the larger small SD/I manufacturers is about $50-60 million per year.
However, the vast majority of the SD/I engine manufacturers produce low production volumes
of engines and typically have less than 50 employees.
The two largest boat building companies are Brunswick and Genmar. Brunswick owns
approximately 25 boat companies and Genmar owns approximately 12 boat companies. Based
on a manufacturer list maintained by the U.S. Coast Guard, there are over 1,600 boat builders in
the United States. We estimate that, based on manufacturer identification codes, more than
1,000 of these companies produce boats using gasoline marine engines. According to the
National Marine Manufacturers Association (NMMA), most of these boat builders are small
businesses. These small businesses range from individuals building one boat per year to
businesses near the SBA small business threshold of 500 employees.
We have identified 14 marine fuel tank manufacturers in the United States that qualify as
small businesses under the SBA definition. These manufacturers include five rotational molders,
two blow molders, six aluminum fuel tank manufacturers, and one specialty fuel tank
manufacturer. The small rotational molders average less than 50 employees while the small
blow-molders average over 100 employees.
We have only identified one small hose manufacturer that produces for the Marine SI
market. Novaflex primarily distributes hoses made by other manufacturers, but does produce its
own fill neck hose. Because we expect vessel manufacturers will design their fuel systems such
that there will not be standing liquid fuel in the fill neck (and therefore the low permeation fuel
hose requirements will not apply to the fill neck), we have not included this manufacturer in our
analysis. The majority of fuel hose in the Marine SI market is made by large manufacturers
including Goodyear and Parker-Hannifin.
10.6 Reporting, Recordkeeping, and Compliance Requirements
For any emission control program, EPA must have assurances that the regulated products
will meet the standards. Historically, EPA programs for Small SI engines and Marine SI
engines have included provisions placing engine manufacturers responsible for providing these
assurances. The program that EPA is adopting for manufacturers subject to this final rule will
include testing, reporting, and record keeping requirements for manufacturers of engines,
equipment, and vessels, and will also include fuel system component manufacturers if they
choose to certify their fuel tank, fuel hose, and fuel cap products.
For Small SI engine manufacturers and OB/PWC engine manufacturers, EPA is generally
continuing the same reporting, record keeping, and compliance requirements prescribed in the
current regulations. For SD/I engine manufacturers, which are not currently subject to EPA
regulation, EPA is planning to apply similar reporting, record keeping, and compliance
requirements to those for OB/PWC engine manufacturers. Testing requirements for engine
manufacturers will include certification emission (including deterioration factor) testing and
production line testing. Reporting requirements will include emission test data and technical
data on the engines. Manufacturers will also need to keep records of this information.
10-7
-------
Final Regulatory Impact Analysis
Because of the new evaporative emission requirements, there will be new reporting,
record keeping and compliance requirements for Small SI equipment manufacturers. Small SI
equipment manufacturers participating in the transition program will also be subject to reporting,
record keeping and compliance requirements. Depending on who chooses to certify fuel system
components, there may also be new reporting, record keeping and compliance requirements for
fuel tank manufacturers, fuel hose manufacturers, fuel cap manufacturers, and marine vessel
manufacturers. Testing requirements for these manufacturers will include certification emission
testing. Reporting requirements will include emission test data and technical data on the designs.
Manufacturers will also need to keep records of this information.
10.7 Steps Taken to Minimize the Impact on Small Entities
The Panel developed a wide range of regulatory alternatives to mitigate the impacts of
the rulemaking on small businesses, and recommended that we propose and seek comment on
the flexibilities. The Panel's findings and discussions were based on the information that was
available during the term of the Panel and issues that were raised by the SERs during the
outreach meetings and in their written comments. It was agreed that EPA should consider the
issues raised by the SERs (and issues raised in the course of the Panel) and that EPA should
consider the comments on flexibility alternatives that would help to mitigate any negative
impacts on small businesses. Alternatives discussed throughout the Panel process included those
offered in the development of the upcoming rule. Though some of the recommended flexibilities
may be appropriate to apply to all entities affected by the rulemaking, the Panel's discussions
and recommendations were focused mainly on the impacts, and ways to mitigate adverse
impacts, on small businesses. A summary of the Panel's recommendations can be found in the
SBREFA Final Panel Report.1
A list of the small business provisions being adopted with the final rule are presented in
the following section.
10.7.1 Small SI Exhaust Emission Standards
Described below are the regulatory alternatives being adopted with the final rule related
to the Small SI nonhandheld engine exhaust emission standards.
10.7.1.1 Regulatory Flexibility Options for Nonhandheld Engine Manufacturers
The following section contains a discussion of the provisions in the final rule for small
business nonhandheld engine manufacturers.
Additional Lead Time for Nonhandheld Engine Manufacturers - Small-volume engine
manufacturers can delay implementation of the Phase 3 exhaust emission standards for two years
(see §1045.145). Small-volume engine manufacturers will be required to comply with the Phase
3 exhaust emission standards beginning in model year 2014 for Class I engines and model year
2013 for Class II engines. Under this approach, small-volume engine manufacturers can apply
this delay to all their nonhandheld engines or to just a portion of their production. They could
10-8
-------
Small-Business Flexibility Analysis
therefore sell engines that meet the Phase 3 standards on some product lines while delaying
introduction of emission control technology on more challenging product lines. This option
provides more time for small-volume engine manufacturers to redesign their products. They
would also be able to learn from some of the hurdles overcome by larger manufacturers.
Assigned Deterioration Factors - Small-volume engine manufacturers will be able to
rely on an assigned deterioration factor to demonstrate compliance with the standards rather than
doing service accumulation and additional testing to measure deteriorated emission levels at the
end of the regulatory useful life (see §1054.240). EPA is not adopting actual levels for the
assigned deterioration factors with this final rule. EPA intends to analyze emissions
deterioration information that becomes available over the next few years to determine what
deterioration factors would be appropriate for nonhandheld engines. This data is likely to
include deterioration data for engines certified to comply with CARB's Tier 3 standards and
engines certified early to EPA's Phase 3 standards. Prior to the implementation date for the
Phase 3 standards, EPA expects to provide guidance to engine manufacturers specifying the
levels of the assigned deterioration factors for small-volume engine manufacturers.
Production Line Testing Exemption - Small-volume engine manufacturers will be
exempt from the production-line testing requirements for all of their nonhandheld engine
families (see §1054.301).
Broader Definition of Engine Family - Small-volume engine manufacturers may use a
broader definition of engine family than generally applies for certification purposes. Under the
existing engine family criteria specified in the regulations, manufacturers group their various
engine lines into engine families that have similar design characteristics including the
combustion cycle, cooling system, cylinder configuration, number of cylinders, engine class,
valve location, fuel type, aftertreatment design, and useful life category. With this final rule, we
are allowing small-volume engine manufacturers to group all of their nonhandheld engines into a
single engine family for certification by engine class and useful life category, subject to good
engineering judgment (see §1054.230).
Eligibility for the Small Business Flexibilities - We are retaining the current criteria
(i.e., 10,000 units per year of nonhandheld engines) for determining who is a small-volume
engine manufacturer and, as a result, eligible for the Phase 3 flexibilities described above (see
§1054.801). Based on confidential sales data provided to EPA by engine manufacturers, the
10,000 unit cut-off for engine manufacturers would include all of the small business engine
manufacturers using SBA's employee-based definition. However to ensure all small businesses
that meet SBA's employee-based definition have access to the flexibilities described above, EPA
is also allowing engine manufacturers which exceed the production cut-off level of 10,000 units
but have fewer than 1,000 employees, to request treatment as a small volume engine
manufacturer (see §1054.635). In such a case, the manufacturer would need to provide
information to EPA demonstrating that the manufacturer has fewer employees than the 1,000
cut-off level established by SB A.
10-9
-------
Final Regulatory Impact Analysis
10.7.1.2 Regulatory Flexibility Options for Nonhandheld Equipment Manufacturers
The following section contains a discussion of the provisions in the final rule for small
business nonhandheld equipment manufacturers.
Additional Lead Time for Small SI Equipment Manufacturers - Small-volume
equipment manufacturers will have two extra years beyond the implementation dates for the
Phase 3 standards to continue using Phase 2 engines in their Class II equipment. Alternatively,
the manufacturer can use Phase 3 engines without the catalysts, provided the engine
manufacturer submitted data at the time of certification showing that the engine without the
catalyst complied with EPA's Phase 2 standards. As described in Section V.E.3 of the preamble,
EPA is adopting a flexibility program for all equipment manufacturers that produce Class II
equipment. Under that program, equipment manufacturers can install Phase 2 engines in limited
numbers of Class II equipment over the first four years the Phase 3 standards apply (i.e., 2011
through 2015). The number of equipment that can use Phase 2 engines is based on 30 percent of
an average annual production level of Class II equipment. In an effort to provide additional
flexibility to small-volume equipment manufacturers within the context of the flexibility
program, EPA is adopting provision that allow small-volume equipment manufacturers to use
Phase 2 engines at a level of 200 percent of an average annual production level of Class II
equipment over the four year period (see §1054.625). Therefore, a small-volume equipment
manufacturer can potentially use Phase 2 engines on all their Class II equipment for two years
(consistent with the SBAR Panel's recommendation) or they may, for example, sell half their
Class II equipment with Phase 2 engines for four years.
Simplified Engine Certification for Equipment Manufacturers - We are adopting a
simplified engine certification procedure for small-volume equipment manufacturers. (As
discussed in Section V.E.4 of the preamble, we are also adopting this provision for other
manufacturers, regardless of the company's size.) Generally, it has been engine manufacturers
who certify with EPA for the exhaust emission standards because the standards are engine-based
standards. However, because the Phase 3 standards under consideration are expected to result in
the use of catalysts, a number of equipment manufacturers, especially those that make low-
volume models, believe it may be necessary to certify their own unique engine/muffler designs
with EPA, but using the same catalyst substrate already used in a muffler certified by the engine
manufacturer. In order to allow the possibility of an equipment manufacturer certifying an
engine/muffler design with EPA, we are adopting a simplified engine certification process for
small-volume equipment manufacturers (see §1054.612). Under such a simplified certification
process, the equipment manufacturer will need to demonstrate that it is using the same catalyst
substrate as the approved engine manufacturer's family, provide information on the differences
between their engine/exhaust system and the engine/exhaust system certified by the engine
manufacturer, and explain why the emissions deterioration data generated by the engine
manufacturer is representative for the equipment manufacturer's configuration.
Eligibility for the Small Business Flexibilities - EPA is retaining the current criteria (i.e.,
5,000 units per year of nonhandheld equipment) for determining who is a small-volume
equipment manufacturer and, as a result, eligible for the Phase 3 flexibilities described above
10-10
-------
Small-Business Flexibility Analysis
(see §1054.801). Based on sales data, the 5,000 unit cut-off for equipment manufacturers would
include the vast majority of the small business equipment manufacturers using SBA's employee-
based definition. However to ensure all small businesses that meet SBA's employee-based
definition have access to the flexibilities described above, EPA will also allow equipment
manufacturers which exceed the production cut-off level noted above but have fewer employees
than the SB A definition of small business (i.e., 500 employees for manufacturers of most types
of equipment), to request treatment as a small-volume equipment manufacturer (see §1054.635).
In such a case, the manufacturer will need to provide information to EPA demonstrating that the
manufacturer has fewer employees than the applicable employee cut-off level established by
SBA.
10.7.2 Marine SI Exhaust Emission Standards—Regulatory Flexibility Options for SD/I
Engine Manufacturers
Described below are the regulatory alternatives being adopted with the final rule related
to the SD/I engine exhaust emission standards.
Additional Lead Time for SD/I Engine Manufacturers - We are adopting an
implementation date of 2011 for <373 kW SD/I engines produced by small business marine
engine manufacturers and a date of 2013 for small business manufacturers of high-performance
(>373 kW) marine engines (see §1045.145). These dates provide 1 year of additional leadtime
for small businesses producing <373 kW SD/I engines and 3 years of additional leadtime for
small businesse producing >373 kW SD/I engines compared the implementation dates for large
manufacturers.
Exhaust Emission ABT - We are adopting an averaging, banking, and trading (ABT)
credit program for exhaust emissions from <373 kW SD/I marine engines (see part 1045, subpart
H). Under the proposal, the ABT program would have applied to >373 kW SD/I engines as well.
However, as described in section 3.4 of the Summary and Analysis of Comments document for
the Final Rule, we are adopting different standards for high performance SD/I engines than
originally proposed. High performance (>373 kW) SD/I engines are required to meet the new
standards without the use of an ABT program.
Early Credit Generation for ABT - We are adopting an early banking program in which
bonus credits can be earned for certifying early (see §1045.145). This program, combined with
the additional lead time for small businesses noted above, give small-volume manufacturers of
SD/I engines <373 kW ample opportunity to bank emission credits prior to the implementation
date of the standards and provide greater incentive for more small business engine manufacturers
to introduce advanced technology earlier than would otherwise occur. Because the ABT
program being adopted with the final rule only applies to SD/I engines <373 kW, the early credit
provisions will not apply to high-performance (>373 kW) SD/I engines.
Assigned Emission Rates for High Performance (>373 kW) SD/I Engines - In the
proposal, we noted that in the case where an engine manufacturer is using emission credits to
comply with the standard, the manufacturer will still need to test engines to calculate how many
10-11
-------
Final Regulatory Impact Analysis
emission credits are needed. In order to minimize this testing burden, we proposed to allow
manufacturers to use assigned baseline emission rates for certification based on previously
generated emission data. As discussed in section 3.4 of the Summary and Analysis of Comments
document for the Final Rule, we are adopting less stringent standards for high-performance
(>373 kW) SD/I engines that do not allow for the use of the ABT program for demonstrating
compliance with the standards. Therefore, we are not adopting baseline HC+NOx and CO
emission rates for high-performance SD/I engines since the proposed levels were higher than the
standards being adopted and therefore, are of no use without an ABT program.
Alternative Standards for High Performance (>373 kW) SD/I Engines - In the
proposal, EPA cited concerns raised by small businesses that catalysts had not been
demonstrated on high-performance engines and that they may not be practicable for this
application and therefore requested comments on the need for and level of alternative standards
for high-performance marine engines. As described in section 3.4 of the Summary and Analysis
of Comments document for the Final Rule, we are adopting a less stringent set of exhaust
emission standards for high performance (>373 kW) SD/I engines than originally proposed.
These standards are not expected to result in the use of catalysts on high performance (>373 kW)
SD/I engines.
Furthermore, we are not adopting NTE standards for high-performance SD/I engines (See
§1045.105). This is consistent with the SBAR Panel recommendation that NTE standards not
apply to any high-performance SD/I engines.
Broad Engine Families for High Performance (>373 kW) SD/IEngines - Typically in
EPA engine and equipment programs, manufacturers are able to group their engine lines into
engine families for certification to the standards. Engines in a given family must have many
similar characteristics including the combustion cycle, cooling system, fuel system, air
aspiration, fuel type, aftertreatment design, number of cylinders and cylinder bore sizes. A
manufacturer would then only perform emission tests on the engine in that family that would be
most likely to exceed an emission standard. We are adopting provisions that allow small
businesses to group all of their high performance (>373 kW) SD/I engines into a single engine
family for certification, subject to good engineering judgment (see §1045.230). A manufacturer
will need to perform emission tests only on the engine design that will be most likely to exceed
the emission standard.
Simplified Test Procedures for High Performance (>373 kW) SD/I Engines - Existing
testing requirements include detailed specifications for the calibration and maintenance of testing
equipment and tolerances for performing the actual tests. For high performance (>373 kW) SD/I
engines, it may be difficult to hold the engine at idle or high power within the tolerances
currently specified by EPA in the test procedures. Therefore, we are adopting less restrictive
specifications and tolerances, for small businesses testing high performance (>373 kW) SD/I
engines, which would allow the use of portable emission measurement equipment (see
§1065.901(b)). This will facilitate less expensive testing for these small businesses without
having a negative effect on the environment.
10-12
-------
Small-Business Flexibility Analysis
Reduced Testing Requirements for SD/I Engine Manufacturers - We are adopting
provisions to allow small-volume engine manufacturers to use an assigned deterioration factor to
demonstrate compliance with the standards for the purposes of certification rather than doing
service accumulation and additional testing to measure deteriorated emission levels at the end of
the regulatory useful life (see §1045.240). EPA is not specifying actual levels for the assigned
deterioration factors in this final rule. EPA intends to analyze available emission deterioration
information to determine appropriate deterioration factors for SD/I engines. The data will likely
include durability information from engines certified to California ARB's standards and may also
include engines certified early to EPA's standards. Prior to the implementation date for the SD/I
standards, EPA will provide guidance to engine manufacturers specifying the levels of the
assigned deterioration factors for small-volume engine manufacturers.
We proposed to exempt small-volume manufacturers of SD/I engines from the
production -line testing requirements. As noted in section 3.10 of the Summary and Analysis of
Comments document for the Final Rule, we are dropping the production-line testing
requirements for all engine manufacturers including large manufacturers. Therefore, no
production-line testing will be required of any SD/I engine manufacturer (see §1045.301).
Eligibility for the Small Business Flexibilities - For purposes of determining which
engine manufacturers are eligible for the small business flexibilities described above for SD/I
engine manufacturers, we are adopting criteria based on the number of employees. SD/I engine
manufacturers that have no more than 250 employees will be considered a small business for the
purposes of the flexibilities being adopted with the final rule. We originally proposed criteria
based on a production cut-off of 5,000 SD/I engines per year. However, engine manufacturers
commented that it was more appropriate to use an employee level than a production level for
defining which companies are small businesses. We believe a 250 employee limit should be
roughly consistent with the production level we targeted in our proposal, although some
manufacturers would likely be able to produce more than 5,000 units. Under the small-volume
engine manufacturer definition being adopted for the final rule, there will be no option to
consider the production volume instead of the 250 employee count.
10.7.3 Small SI and Marine SI Evaporative Emission Standards— Flexibility Alternatives
for Equipment, Vessel, and Fuel Tank Manufacturers
Described below are the regulatory alternatives being adopted with the final rule related
to the evaporative emission standards for Small SI engines and equipment and Marine SI engines
and vessels. The provisions discussed below applied to Small SI equipment and to SD/I marine
vessels, except where noted. Because the majority of fuel tanks produced for the Small SI
equipment and the SD/I marine vessel market are made by small businesses, the flexibility
provisions being adopted for fuel tank manufacturers apply regardless of whether the
manufacturer was a small business or not.
Consideration of Appropriate Lead Time - We are adopting an implementation schedule
that we believe provides sufficient lead time for blow-molded and marine rotational molded fuel
tanks. For Small SI equipment, we are establishing tank permeation implementation dates of
10-13
-------
Final Regulatory Impact Analysis
2011 for Class II equipment and 2012 for Class I equipment. For marine fuel tanks, we are
implementing the tank permeation standards in 2011 with an additional year (2012) for installed
fuel tanks which are typically rotational-molded marine fuel tanks (see §1054.110 and
§1045.107).
With regard to the diurnal requirements for marine vessels, we are providing an
additional year of lead time, compared to the proposal. This means that the diurnal standard
willll apply beginning with the 2011 model year. In addition, we are adopting an interim
allowance program that will give additional time for a limited number of boats. Under this
program, each boat builder will be allowed to sell these boats without the diurnal emission
controls that would otherwise be required. Specifically, each boat builder will have a total of
1,200 allowances that may be used, at the manufacturer's discretion, for boats produced before
December 31, 2012. These allowances are intended to help boat builders engage in an orderly
transition to the new standards and will only be available for boats produced in 2011 and 2012.
This allowance program applies only to boats with installed fuel tanks that are expected to use
carbon canisters to meet the diurnal emission standards. Therefore, it does not apply to portable
fuel tanks or personal watercraft. This provision will apply to both small and large businesses
because we believe that even large businesses may have specific, small-volume models where
additional lead time may be especially helpful due to atypical design constraints. For very small
companies, we expect that this allowance program will result in an additional year, or even two,
of lead time for them to address potential installation issues related to carbon canisters.
Fuel Tank ABT and Early Incentive Program - We are adopting an ABT program for
fuel tank permeation and an early-allowance program for fuel tank permeation. In the proposal,
we requested comment on including service tanks in the ABT program. (Service tanks are fuel
tanks sold as replacement parts for in-use equipment.) Based on comments received, we do not
believe it is appropriate to include such tanks in the ABT program. Equipment manufacturers
will be required to demonstrate that their equipment models meet the evaporative emission
standards. If the certified equipment uses a fuel tank included in the ABT program, the credits
generated were based on a useful life of five years. Therefore, if the tank being replaced is less
than five years old, the replacement tank would result in double counting of some of the credits.
While manufacturers could potentially gather information to account for the age of the fuel tank
being replaced, we do not want to complicate the provisions of the ABT program and therefore
we are not including replacement tanks in the fuel tank ABT program.
Broad Definition of Evaporative Emission Family for Fuel Tanks - We are adopting
provisions that allow fuel tank emission families to be based on type of material (including
additives such as pigments, plasticizers, and UV inhibitors), emission control strategy, and
production methods. This would allow fuel tanks of different sizes, shapes, and wall thicknesses
can be grouped into the same emission family (see §1045.230 and §1054.230). In addition,
Small SI and Marine SI fuel tanks could be allowed in the same emission family if the tanks
meet these criteria. Manufacturers therefore will be able to broadly group similar fuel tanks into
the same emission family and then only test the configuration most likely to exceed the emission
standard.
10-14
-------
Small-Business Flexibility Analysis
Compliance Progress Review for Marine Fuel Tanks - We believe the 2012 fuel
permeation standards are technologically feasible for rotationally-molded marine fuel tanks.
This conclusion is supported by data presented in the Regulatory Impact Analysis. In addition,
several rotationally-molded tank manufacturers support EPA's proposed standards and
implementation dates and have provided information to support their positions. However,
several other rotationally-molded tank manufacturers are not as far along in their technological
progress toward meeting the standards and are not certain about their ability meet the EPA
requirements in 2012. To address this situation, these manufactures requested that EPA perform
a technical review in 2010 to determine whether the compliance dates should be adjusted.
However, for the reasons discussed above, we believe that the tank permeation standards have
been demonstrated to be technologically feasible in the 2012 time frame and do not look
favorably upon the request for a technology review of the permeation standard.
Nevertheless, we are concerned about the potential long-term impacts on the small
businesses that have not yet developed technology that meets the requirements. During the next
few years, EPA intends to hold periodic progress reviews with small businesses that rotationally
mold fuel tanks. The purpose of these progress reviews will be to monitor the progress of
individual companies towards compliance with the tank permeation standards and to provide
feedback as needed. Rather than conducting a broad program with the entire industry, we will
conduct separate, voluntary reviews with each interested company. These sessions will be
instrumental to EPA in following the progress for these companies and assessing their efforts
and potential problems.
To help address small business concerns, we expect we would rely on the small volume
manufacturer hardship relief provisions contained in 40 CFR 1068.250, and described in the
following section. In the event that a small business is unsuccessful in the 2012 model year and
seeks hardship relief, the progress reviews described above would provide an important
foundation in determining whether a manufacturer has taken all steps to comply with the
permeation standards in a timely and orderly manner.
Design-Based Certification - We are adopting design-based certification for carbon
canisters for boats. For the carbon canisters, the design requirement call for a ratio of carbon
volume (liters) to fuel tank capacity (gallons) of 0.04 liter/gallon for boats less than 26 feet in
length, and 0.016 liter/gallon for larger boats. We are also adopting design-based certification
for certain fuel tanks. For fuel tanks, we will allow design-based certification for metal tanks as
well as plastic fuel tanks with a continuous EVOH barrier.
The National Marine Manufacturers Association (NMMA) the American Boat and Yacht
Council (ABYC) and the Society of Automotive Engineers (SAE) have industry recommended
practices for boat designs that must be met as a condition of NMMA membership. We will
allow this data to be used as part of EPA certification as long as it is collected consistent with the
test procedures and other requirements described in this final rule.
Additional Lead Time for Small SI Fuel Hose Requirement - We proposed an
implementation date of 2008 for Small SI hose permeation standards for non-handheld
10-15
-------
Final Regulatory Impact Analysis
equipment (see §90.127). Given that we are not adopting the final rule until mid-2008, we have
delayed the implementation of the low permeation fuel line requirement until 2009 for
nonhandheld equipment. However, we are keeping the 2009 implementation date for
low-permeation fuel line for small businesses producing Small SI nonhandheld equipment. We
believe the 2009 date is feasible for all equipment manufacturers, given that fuel line meeting the
low permeation standards is already widely available and manufacturers selling most types of
nonhandheld equipment in California were required to use such hose starting in 2007 or 2008.
10.7.4 Hardship Provisions—Regulatory Flexibility Options for Engine, Equipment,
Vessel, and Fuel System Component Manufacturers
The following section summarizes the hardship provisions we are adopting which would
be available to engine manufacturers, equipment manufacturers, vessel manufacturers, and fuel
system component manufacturers (i.e., fuel tank, fuel hose, and fuel cap manufacturers).
Unusual Circumstances Hardship - Under the unusual circumstances hardship
provision, manufacturers can apply for hardship relief if circumstances outside their control
cause the failure to comply and if failure to sell the subject engines or equipment will jeopardize
the company's solvency (see §1068.245). The terms and time frame of the relief will depend on
the specific circumstances of the company and the situation involved. As part of its application
for hardship, a company will be required to provide a compliance plan detailing when and how it
will achieve compliance with the standards. This hardship provision will be available to all
business engine manufacturers, equipment manufacturers, vessel manufacturers, and fuel system
component manufacturers, regardless of size.
Economic Hardship - Under the economic hardship provision, small business
manufacturers can petition EPA for limited additional lead time to comply with the standards
(see §1068.250). A manufacturer will have to make the case that it has taken all possible
business, technical, and economic steps to comply, but the burden of compliance costs will have
a significant impact on the company's solvency. Hardship relief may include requirements for
interim emission reductions and/or purchase and use of emission credits. The length of the
hardship relief will be established during the initial review and will likely need to be reviewed
annually thereafter. As part of its application for hardship, a company will be required to
provide a compliance plan detailing when and how it will achieve compliance with the
standards. This hardship provision will be available only to engine manufacturers, equipment
manufacturers, vessel manufacturers, and fuel system component manufacturers that are small
businesses.
10.8 Projected Economic Effects of the Rulemaking
The following section summarizes the economic impact on small businesses of the new
exhaust and evaporative emission standards for both Small SI engines and equipment and Marine
SI engines and vessels. As noted earlier, the types of companies that will be affected by the new
Marine SI standards include OB/PWC engine manufacturers, SD/I engine manufacturers, boat
builders, and marine fuel system component manufacturers (e.g., fuel tank and fuel hose
10-16
-------
Small-Business Flexibility Analysis
manufacturers). Similarly, the types of companies that will be affected by the Small SI standards
include nonhandheld engine manufacturers, equipment manufacturers, and Small SI fuel system
component manufacturers (e.g., fuel tank and fuel hose manufacturers). For the purposes of this
analysis, it is assumed that engine manufacturers will bear the cost of complying with the
exhaust emission standards, whereas equipment manufacturers and vessel manufacturers will
bear the cost of complying with the evaporative emission standards.
To gauge the impact of the new standards on small businesses, EPA employed a cost-to-
sales ratio test to estimate the number of small businesses that would be impacted by less than
one percent, between one and three percent, and above three percent. The costs used in this
analysis are based on the cost estimates developed in Chapter 6 of this Final RIA with the
exception of the costs used for Small SI engine and equipment manufacturers. A description of
the inputs used for each affected industry sector (except small SI engine and equipment
manufacturers) and the methodology used to develop the estimated impact on small businesses in
each industry sector is presented in the docket for this rulemaking.2
For small SI engine and equipment manufacturers, we relied on the costs from the
proposal instead of the final rule. The basic cost inputs for the final rule (e.g., the cost of the
various technologies, the number of engine and equipment models, etc.) have not changed from
the proposal. However, recent certification data suggests that a number of Class II engines may
be able to comply with the standards without the use of a catalyst. Our cost analysis for the final
rule reflects this change and results in significantly lower costs for Class II. Because we project
that more than half of the engines in Class II will use catalysts, but we do not know which
engines small business equipment manufacturers will purchase for their equipment, we believe it
is appropriate to continue using the higher costs associated with the proposal rather than the final
rule cost numbers in gauging the potential costs of the new standards on small manufacturers.
We believe this approach will result in an overestimation of the impacts (i.e., a conservative
estimate) of the new standards on small SI engine and equipment manufacturers.
For OB/PWC engine manufacturers, EPA identified two small businesses. One of the
small businesses identified by EPA manufactures personal watercraft today using four-stroke
engines with certified emission levels below the new standards, so we project negligible
incremental costs resulting from our rule. The other small business manufactures outboard
engines. Several of their currently certified engines already comply with the new standards,
while the remaining engines would need to be recalibrated, which we project would cost on the
order of less than $10 per engine. Given the cost of personal watercraft and outboard engines,
we therefore believe the impact of the rule is well below one percent of revenues for both of
these OB/PWC engine manufacturers.
For <373 kW SD/I engine manufacturers, EPA identified nine small businesses. Of these
companies, eight produce conventional SD/I engines and the remaining one company produces
SD/I engines for airboats. Of the conventional SD/I small business engine manufacturers, five of
the small businesses may incur compliance costs between one and three percent of their annual
revenues. Three of the small businesses that produce <373 kW SD/I engines as part of a much
broader line of work (such as engine rebuilding or selling land-based engines) will be impacted
10-17
-------
Final Regulatory Impact Analysis
by less than one percent of annual revenues.
Using available information for the airboat engine manufacturer, we project that the
manufacturer will have compliance costs between one and three percent of annual revenues.
Some of this company's engines are >373 kW, so their estimated compliance burden reflects a
combination of costs for conventional SD/I engines and for high-performance >373 kW engines.
(They are included in the conventional SD/I category for this impact analysis.) This company is
unique in that it manufacturers many of its engines for sale to other airboat manufacturers,
resulting in a concentrated cost impact relative to their revenues.
We also identified a number of other airboat manufacturers. These small businesses
making engines for airboats are less reliant on selling engines to other boat builders, instead
making engines for the boats they build themselves. Most of these businesses are very small,
with little ability to marshal the technical resources needed to comply with emission standards.
If these companies would take on the effort to design and certify compliant engines, they would
likely experience compliance costs exceeding three percent of their revenues. However, given
their place in the market and the fact that they are primarily boat builders with the
resourcefulness to make their own engines, we believe the most likely approach for these
companies is to buy a certified engine from manufacturers of conventional SD/I engines. As
such, these companies would be treated with other boat builders, in which case their main
compliance cost is related to evaporative emissions (as described below). We therefore do not
consider any of these companies as engine manufacturers for the purposes of analyzing the
impact of the new standards on engine manufacturers.
For >373 kW SD/I engine manufacturers, EPA identified 19 small businesses. Of the
>373 kW SD/I small business engine manufacturers, all of the small businesses are projected to
be impacted by less than one percent of annual revenues.
For boat builders, EPA believes there are over 1,000 small business manufacturers.
Many of these companies make small numbers of vessels for certain segments of the marine
market. Given the high cost of most boats, EPA believes the cost impact will be below one
percent for all small business boat builders, including those that manufacture SD/I vessels, and
OB/PWC boat manufacturers as well.
While boat builders have the primary responsibility under the new regulations for
complying with evaporative emission standards, fuel hose and fuel tank manufacturers will have
to certify their product with EPA. EPA has identified one small business that manufactures fuel
hose for marine applications and 14 small businesses that manufacturer fuel tanks for marine
applications. The company producing fuel hose primarily distributes hoses made by other
manufacturers but does produce its own fill neck hose. Because we expect vessel manufacturers
will design their fuel systems such that there will not be standing liquid fuel in the fill neck (and
therefore the new low permeation fuel hose requirements will not apply to the fill neck), we have
not included this manufacturer in our analysis. Of the 14 fuel tank manufacturers, EPA has
estimated that all of them will incur costs below one percent of annual revenues.
10-18
-------
Small-Business Flexibility Analysis
For Small SI engine and equipment manufacturers, EPA has identified 370 small
businesses.3 Ten of the small businesses are engine manufacturers and the remaining companies
are equipment manufacturers. Based on an analysis of sales revenues by company, EPA projects
that 314 of the small businesses are estimated to incur compliance costs representing less than 1
percent of their annual revenues. EPA projects that 38 companies will incur compliance costs
between 1 and 3 percent of their annual revenues, and 18 companies will incur compliance costs
representing more than 3 percent of their annual revenues.
Similar to the requirements noted above for boat manufacturers under the Marine SI
evaporative emission regulations, equipment manufacturers will have the primary responsibility
under the regulations for complying with the Small SI evaporative emission standards.
However, fuel hose and fuel tank manufacturers will have to certify their product with EPA.
EPA has identified two small businesses that manufactures fuel hose for Small SI applications
and 25 small businesses that manufacturer fuel tanks for Small SI applications. Of these
companies, EPA has estimated that all of these companies will incur costs below one percent of
annual revenues.
Table 10.8-1 summarizes the impacts of the new regulations on small businesses
impacted by the exhaust and evaporative emission standards for Small SI engines and equipment
and Marine SI engines and vessels.
Table 10.8-1: Summary of Impacts on Small Businesses
Market Sector
Manufacturers of Marine OB/PWC engines
Manufacturers of Marine SD/I engines < 373 kW
Manufacturers of Marine SD/I engines > 373 kW
(high-performance)
Boat Builders
Manufacturers of Fuel Hose and Fuel Tanks for
Marine SI Vessels
Small SI engines and equipment
Manufacturers of Fuel Hose and Fuel Tanks for
Small SI Applications
Total
0-1 percent
2
4
19
>1,000
14
314
27
380 plus
>l,000boat
builders
1 -3
percent
0
5
0
0
0
38
0
43
> 3 percent
0
0
0
0
0
18
0
18
After considering the economic impacts of today's final rule on small entities, we believe
10-19
-------
Final Regulatory Impact Analysis
this action will not have a significant economic impact on a substantial number of small entities.
The small entities directly regulated by this final rule cover a wide range of small businesses
including engine manufacturers, equipment manufacturers, boat manufacturers, fuel tank
manufacturers, and fuel hose manufacturers. Small governmental jurisdictions and small
organizations as described above will not be impacted. We have determined that the estimated
effect of the rule is to impact 43 companies with costs between one and three percent of
revenues, and 18 additional companies with costs over three percent of revenues. These 61
companies represent less than 5 percent of the total number of small businesses impacted by the
new regulations. All remaining companies (over 1,000 of them) would be impacted with costs
by less than one percent of revenues. It should be noted that this estimate is based on the highest
level of estimated cost in the first years of the program. We estimate substantially lower
long-term costs as manufacturers learn to produce compliant products at a lower cost over time.
For a complete discussion of the economic impacts of the final rulemaking, see Chapter
9, the Economic Impact Analysis chapter, of this Final Regulatory Impact Analysis.
10-20
-------
Small-Business Flexibility Analysis
Chapter 10 References
1. Final Panel Report of the Small Business Advocacy Review Panel on EPA's Planned Proposed Rule—Control of
Emissions from Nonroad Spark-Ignition Engines and Equipment, October 17, 2006. (A copy has been placed in
docket EPA-HQ-OAR-2004-0008.)
2. "Small Business Impact Memo, Control of Emissions from Nonroad Spark-Ignition Engines and Equipment -
Determination of No SISNOSE," EPA memorandum from Phil Carlson to Alex Cristofaro, March 13, 2008.
(Docket Identification EPA-HQ-OAR-2004-0008- .)
3. "Small Entity Analysis of Small Spark Ignition Nonroad Engine and Equipment Manufacturers," memorandum
from Alex Rogozhin and Brooks Depro, RTI Interational, to Phil Carlson, U.S. EPA, December 15, 2006. (Docket
Identification EPA-HQ-OAR-2004-0008-0541.)
10-21
-------
Regulatory Alternatives
CHAPTER 11: Regulatory Alternatives
Our program represents a blend of exhaust and evaporative emission standards for small
nonroad spark-ignition (SI) engines used in land-based or auxiliary marine applications, and also
recreational Marine SI engines. We believe that the combination of emission standards and their
associated timing are superior to the alternative program options we considered given their
feasibility, cost, and environmental impact. In this chapter we present and discuss the options
that we evaluated in order to make this determination.
Section 11.1 presents each element of our requirements and discusses a variety of
specific alternatives that are either less and more stringent. After this initial assessment, options
that merit a more rigorous examination are identified for analysis in subsequent sections.
Section 11.2 describes the cost of the selected options for each affected engine or system.
Section 11.3 presents the emissions inventory impacts associated with each option. Section 11.4
describes the cost effectiveness ($/ton of emission reduced) of the selected options. Finally, we
present our assessment of the rationale, feasibility, and issues associated with each alternative in
Section 11.5.
The costs, emission reductions, and cost effectiveness of the options analyzed in Sections
11.2 through 11.5 are incremental to the base case (i.e., current requirements) ignoring this rule,
unless otherwise specified. For example, the more stringent recreational marine exhaust
standards for OB/PWC are evaluated as follow-on requirements to the new requirements and
would begin in a later year. Therefore, the analysis for that option reflects only the more
stringent subsequent standards.
For the more stringent options, it is important to note that the analyses depend on data
supporting them. Generally, a scenario was picked for analysis because there was evidence to
suggest that controls such as those identified in the write-ups could be technically feasible at
some point in the future. However, there is some uncertainty with regard to the technical
feasibility of implementing the standards or requirements across all products, the level of the
potential standards selected for analysis (if applicable), the timing for potential introduction, and
the costs of control. However, while these standards were ultimately not selected as the basis for
this rule, it appears that in some cases they could form the basis for potential future rulemaking
actions.
11.1 Identification of Alternative Program Options
This section provides our description of potential options for each element of our rule.
Options that do not merit further consideration are eliminated and those that warrant additional
analysis in subsequent sections are identified.
ll-l
-------
Final Regulatory Impact Analysis
11.1.1 Alternative Exhaust Emission Requirements
11.1.1.1 Small SI Engine HC+NOx Standards
11.1.1.1.1 Class I
We considered, but rejected, a less stringent HC+NOx emission standard for Class I
spark-ignition engines. The standard of 10 g/kW-hr is readily achievable with reasonably priced
emission control technology. Furthermore, the lead time for implementing the standard in 2012
is adequate for applying the catalyst-based technology that will be used on many of these
engines. A less stringent emission standard would not be consistent with the requirements of
section 213 of the Clean Air Act.
A more stringent standard was also considered. Under this option an 8 g/kW-hr
HC+NOx standard would be implemented. For purposes of this analysis we elected to begin the
requirement in the 2015 model year. Due to the technical design relationship between the engine
and running loss control requirement we modeled running loss controls to start in 2015 as well.
This standard represents about a 50 reduction from the existing Phase 2 standard, rather than the
approximately 38 percent reduction associated with the final standards. As analyzed this option
also provides 3 more years of lead time. We believe that manufacturers of side-valve (SV)
engines would choose to convert these families to overhead-valve (OHV) designs. The
emissions from OHV engine are typically lower and deteriorate less than SV engines and thereby
result in the need for only a slightly more active catalyst and improved cooling relative to the
technology changes needed for the final standards. Cooling for the slightly more active OHV
catalyst would be supplied by the engine improvements anticipated for this rule, such as include
optimized head design for cooling and fan design for cooling air generation. The slightly more
active catalyst can be achieved with either a larger volume and/or a more active mix of precious
metals in the catalyst substrate. It may be possible for SV engines to meet the more stringent
emission standards using catalysts. For SV engines the catalysts would likely need to be larger
and more active. This would result in higher costs and greater catalyst heat generation which
may or may not be able to be handled by the engine's cooling system.
11.1.1.1.2 Class II
For Class II spark-ignition engines, we considered an alternative program option that was
less stringent than the final standards. However, for the same reasons previously stated for Class
I engines, we rejected this alternative from further consideration; the standards are readily
achievable at a reasonable cost within the lead time provided. A less stringent standard, such as
one at a level not depending on catalyst technology, would not have been consistent with section
213 of the Clean Air Act.
An alternative for a more stringent exhaust HC+NOx emission standard would be 4.0
g/kW-hr along with a delay in the corresponding running loss requirement such that engine
changes are made at one time. For analytical purposes we started this requirement in 2015, four
years beyond that for the new standard. Such an exhaust emission standard represents a 67
11-2
-------
Regulatory Alternatives
percent reduction relative to the existing Phase 2 standard, rather than the 34 percent reduction
associated with the new standards. It also provides four more years of lead time; a phase-in
could be needed since implementation would require the equipment manufacturers involvement
for non-integrated products. In order to achieve the 4.0 g/kW-hr HC+NOx emission standard,
we expect manufacturers would need to make widespread use of closed loop control EFI and
three-way catalysts. The EFI systems would keep engine air-to-fuel mixture closer to
stoichiometry and provide an optimum environment for the maximum reduction in HC+NOx by
a three way catalyst. Changes to the catalyst would likely involve a more active mix of precious
metals in the catalyst substrate. In addition, engine upgrades would be required in some of the
Class II engines commonly used in residential lawn care equipment.
11.1.1.2 Marine Auxiliary Engine CO Standard
The standards for marine auxiliary engines include a CO standard that would require the
use of highly efficient catalytic control. This standard would require the use of technology to
meet emission levels demanded by the market. Manufacturers of gasoline marine generators are
equipping their engines with catalysts for the primary purpose of reducing ambient CO
concentrations around boats. Therefore, we do not believe that it would be useful to consider a
less stringent standard which could enable market penetration of new engine offerings which
potentially endanger public health. At the same time, the standard is very stringent and
manufacturers are already designing for reductions which are more than 95 percent below the
current CO emission standard. A more stringent standard would do little more to push
technology. Thus, we do not believe that it would be useful to analyze a more stringent standard.
11.1.1.3 Outboard/Personal Watercraft (OB/PWC) Engine HC+NOx and CO
Emission Standards
The standards for OB/PWC are based on technology that manufacturers are already
certifying and selling nationwide. To meet the new requirements, manufacturers would continue
to sell this technology and discontinue their sale of high-emitting old technology carbureted two-
stroke engines. Because the standards can be met with existing technology, we do not believe
that there is an alternative between the new standards and the current standards which would be
consistent with the CAA section 213 requirement. Therefore, we did not analyze a less stringent
alternative.
For a more stringent alternative, we considered an addition tier of standards beginning in
2012. For OB/PWC engines greater than 40 kW these would be at a level of 10 g/kW-hr. For
engines less than 40 kW, we use an equation of 28 - 0.45 x rated power(kW) to maintain a
continuous curve function. This alternative also considers a lower CO standard of 200 g/kW-hr
for engines greater than 40 kW with an adjusted standard of 500 - 7.5 x rated power(kW) for
engines less than 40 kW to maintain a continuous standard function. Such standards would be
consistent with currently certified emission levels from some four-stroke outboard engines.
Although many four-stroke engines may be able to meet a 10 g/kW-hr standard with improved
engine calibration, it is not clear that all engines could meet this standard without applying yet
unproven catalyst technology in this application. To model this scenario, we evaluated the costs
11-3
-------
Final Regulatory Impact Analysis
and emission reductions that could be achieved through the combined use of calibrated four-
stroke engines and four-stroke engines with catalytic control. This analysis applied catalytic
control to larger OB/PWC engines, which already use or are expected to use electronic fuel
injection.
11.1.1.4 Sterndrive/Inboard (SD/I) Engine HC+NOx and CO Standards
For the purposes of this analysis, we subdivided the SD/I category into traditional and
high-performance engine categories. Based on our definitions, high-performance engines have
a rated power greater than or equal to 373 kW (500 hp).
11.1.1.4.1 SD/K373kW
In developing regulatory alternatives for SD/I engines, we considered both what was
achievable without catalysts and what could be achievable with larger, more efficient catalysts
than those we evaluated in our test programs.
With regard to a less stringent option, we considered non-catalyst based standards to be
implemented in the 2010 model year. Chapter 4 presents data on SD/I engines equipped with
exhaust gas recirculation (EGR). HC+NOx emission levels below 10 g/kW-hr were achieved for
each of the engines. CO emissions ranged from 25 to 185 g/kW-hr. For this less stringent
alternative, we consider standards of 10 g/kW-hr HC+NOx and 150 g/kW-hr CO. The current
California HC+NOx standard for these engines is 160 g/kW-hr.
For a more stringent option, we considered more stringent catalyst-based standards.
Many of the SD/I marine engines with catalysts described in Chapter 4 had HC+NOx emission
rates appreciably below 5 g/kW-hr, even with deteriorated catalysts. In the development testing
for this rulemaking, we did not investigate larger catalysts for SD/I applications. The goal of the
development testing was to demonstrate catalysts that would work within the packaging
constraints associated with water jacketing the exhaust and fitting the engines into engine
compartments on boats. However, we did perform testing on engines equipped with both
catalysts and EGR. These engines showed emission results in the 2-3 g/kW-hr range. We
expect that these same reductions could be achieved more simply through the use of larger
catalysts or catalysts with higher precious metal loading. As a more stringent regulatory
alternative, we considered a standard of 2.5 g/kW-hr HC+NOx, with no change in the CO
standard, based on the use of larger catalysts. To account for additional development work that
would need to be performed by manufacturers to achieve a lower standard than the existing
California standard, we consider a later implementation date of 2012 for this more stringent
alternative with no standard before that time.
11.1.1.4.2 SD/I >373kW
For high-performance SD/I marine engines, we originally proposed a standard based on
the use of catalysts and then considered a less stringent alternative based on engine fuel system
upgrades, calibration, or other minor changes such as an air injection pump rather than catalytic
control. However, manufacturers commented that catalysts are not be practical for these engines
11-4
-------
Regulatory Alternatives
due to the high exhaust flow rates, high emission rates, and low useful life period between
rebuild. In the final rule, we are establishing standards that can be met through the use of engine
controls, similar to the alternative standard that was analyzed in the proposal. Because we do not
consider catalyst-based standards to be feasible for high-performance engines at this time, we are
not modeling a more stringent alternative.
11.1.2 Alternative Evaporative Emission Requirements
11.1.2.1 Small SI Engines
For Small SI engines, we are finalizing both permeation and venting emission standards.
The permeation standards are for fuel tanks and fuel lines. We believe that the standards are
reflective of available technology and represent a step change in emissions performance. Venting
emissions include diurnal breathing losses, diffusion, and running loss emissions. For non-
handheld Small SI engines (i.e., Classes I and II), we are finalizing standards for running loss1
but not for diurnal emissions. We are not finalizing any type of venting emissions control for
handheld equipment.
For a less stringent alternative, we considered not requiring running loss emission control
for non-handheld Small SI engines. These requirements would be deleted from the rule and thus
modeled as being deleted in the years otherwise required.
For a more stringent alternative, we considered applying running loss and diurnal
standards to handheld equipment and setting a diurnal standard for non-handheld (Classes I and
II). In these alternatives, we consider an implementation date of 2012 for handheld and Class I
equipment, and a date of 2011 for Class II equipment.
11.1.2.2 Marine
Similar to the analysis described above for Small SI equipment, we base the less stringent
and more stringent regulatory alternatives on changes in the venting emission standards. For
marine vessels, we are adopting diurnal emission standards for all vessel types. For portable fuel
tanks and PWC fuel tanks, the control technology of a sealed system with pressure relief is fairly
straightforward and commonly used today. However, we anticipate that the diurnal emissions
standards for vessels with installed fuel tanks would be based on the use of passively purged
carbon canisters. For a less stringent alternative, we consider not setting a diurnal emission
standard for marine vessels.2 For a more stringent scenario, we consider a diurnal requirement
wherein boat builders would be required to employ active purge of carbon canister with installed
tanks. This means that, when the engine is operating, it would draw air through the canister to
purge the stored hydrocarbons. These purged gasoline vapors would be used in the engine as
1 We anticipate that running loss control measures will also reduce diffusion emissions.
2Note that PWC already meet the standard and would not be affected differently for the less
stringent standard. PWC use sealed systems with pressure relief to prevent fuel spillage during operation.
11-5
-------
Final Regulatory Impact Analysis
fuel.
11.1.3 Summary of Alternative Standards
Table 11.1-1 and Table 11.1-2 show the alternative program options that were selected
above for further consideration.
Table 11.1-1: Exhaust Alternative Program Options for Quantitative Analysis
Source
Exhaust
Alt
1
2
3
4
5
Target
Class I
Class II
OB/PWC
SD/I
<373 kW
Standard
• lOg/kW-hrHC+NOx
•Begins 20 12
•8g/kW-hrHC+NOx
•Begins 20 11
• Decreases with power
output (P)
• 2008 California
HC+NOx equation
• CO g/kW-hr equation
is 500-5Pfor <40 kW
• 300 g/kW-hr CO for
>40kW
•Begins 20 10
• 5 g/kW-hr HC+NOx
• 75 g/kW-hr CO
•Begins 20 10a
less/
more
more
more
more
less
more
Alternative Description
• 8 g/kW-hr HC+NOx
• Begins 2015 in lieu of standard
•3.5g/kW-hrHC+NOx
• Begins 2015 in lieu of standard
<40kW
• power output (P)
• HC+NOx g/kW-hr equation is 28-0.45P
• COg/kW-hr equation is 500-7. 5P
>40kW
• 10 g/kW-hr HC+NOx
• 200 g/kW-hr CO
•Both begin 2012 in addition to 2010
standards
• 10 g/kW-hr HC+NOx
• 150 g/kW-hr CO
• Same effective dates as standard
• 2.5 g/kW-hr HC+NOx
• 75 g/kW-hr CO
• Begins 201 2 in lieu of standards3
a 2011 for certain engine blocks. Does not include small business flexibilities that will delay the effective date of the
requirements for some companies.
-------
Regulatory Alternatives
Table 11.1-2: Evaporative Alternative Program Options for Quantitative Analysis
1 Source
Alt
Target
Standard
less/
more
Alternative Description 1
Evap
6
7
8
9
10
11
HH
diurnal/
running
loss
Class I
& Class
II
running
loss
Class I
& Class
II diurnal
Installed
marine
fuel tank
diurnal
Portable
marine
fuel tank
diurnal
• None
• Running loss is a
"zero emission" design
standard
•Class I begins 20 12
and Class II begins
2011
• None
• 0.4g/gal/day HC
trailerable boat
•0.16g/gal/dayHC
non-trailerable boat
•Begins 20 11
• Diurnal is a "zero
emission" design
standard
• Begins 20 10
more
less
more
less
more
less
•Begins 20 12
• No running loss
• Requirement would begin in 2012 for Class
I and 20 11 for Class II
•No diurnal for 20 10
• More stringent test procedure. If charcoal
canister is used, active purge required.
•Would begin 20 11
• No diurnal
11.2 Cost per Engine
This section describes the estimated cost of complying with the alternative program
options. We developed the costs for individual technologies using estimates from ICF
Incorporated,1'23 conversations with manufacturers, other information including the published
literature, and our best technical judgment. Also, the cost estimates for the alternatives rely
heavily on the methodology and in some cases the actual cost data, used to characterize the
standards. For ease of presentation, we have not repeated the methodology or those detailed
cost data here. Instead, we focus on presenting information regarding the requirements or
changes that we expect will be needed to comply with the alternative options. The reader is
encouraged to refer to Chapter 6 for more information. Finally, we did not specifically analyze
the incremental costs of setting standards which would not result in technology which would
allow certification in all 50 states (a harmonized program).
The costs of complying with the alternative program options are presented as
incremental to the base case (current requirements) without considering the final standard. The
only exception to this is the second phase of OB/PWC standards where costs are incremental to
the final standard. The alternatives and the requisite technology are described in Section 11.1.
Further, results are provided as the average cost per affected engine and the total net present
value (NPV) for a 30-year period beginning in 2008. The NPV estimates are based on a seven
11-7
-------
Final Regulatory Impact Analysis
percent discount rate. All costs are in 2005 dollars.
11.2.1 Costs for Exhaust Emission Standards
11.2.1.1 More Stringent Small SI Engine HC+NOx Standards
11.2.1.1.1 Class I
Meeting more stringent standards would require OHV engines to use a slightly larger or
more active catalyst than for the final standards. For current SV engines they would need to
utilize larger and more active catalysts than considered in the analysis for the final standards, or
convert to OHV design and use a slightly larger catalyst or more active catalyst than for the
final standards.
The cost for the SV sized catalyst is outlined in Chapter 6. The cost for the conversion
from SV to OHV design is drawn from ICF International's 2006 report "Small SI Engine
Technologies and Costs4," and is listed as $9.42 in variable costs per engine, $2,010,147 in
tooling changes and design and development, as well as $15 million in facility upgrades per
Class I SV engine family. The 2005 EPA certification database lists five SV engine families
certified to Phase 2 of which two engines have OHV engine designs in the same power range
and one engine family is listed as a small volume engine family. The remaining two engine
families have sales estimates in the millions of engines. As a result, fixed costs are applied two
engine families and variable costs are applied to all SV engines.
The cost for improvements in OHV current engine designs includes improved cylinder
head design for improved engine cooling, redesign of the engine flywheel to provide optimum
cooling for the catalyst muffler as well as carburetor improvements. Research and development
and tooling for these changes are estimated at $456,450 per engine family as shown in Chapter
6.
Upgrades in catalysts for OHV engines include additional precious metal for more
active catalysts. The catalyst estimates for the SV engine families, that are replaced by OHV
engine families, are also replaced with the OHV catalyst costs. These costs for improved OHV
engines, upgraded catalysts for OHV engines are included in Table 11.2-1 together with those
for SV engines.
11.2.1.1.2 ClassII
Technologies for the more stringent option include improved engine design (redesign of
cooling fins, fan design, combustion chamber design), closed loop control electronic fuel
injection (EFI), catalysts and pressurized oil lube system for engines intended for residential
use. The fixed costs for improved engine design are $456,000 per engine family and include
R&D and tooling costs, as listed in Chapter 6. The same Chapter lists EFI variable costs at $79
per engine when it includes the credit for the removal of the carburetor. The fixed costs for
closed loop fuel injection design is estimated at $103,000 per engine family. Increased catalyst
11-8
-------
Regulatory Alternatives
efficiency is achieved through use of a larger catalyst and increased precious metal loading at
an estimated increased catalyst cost of $4 (1000 hr engine) - $16 (250 hr engine) per engine. A
pressurized lube oil system is listed by ICF5 to be $15.48 in variable costs and $210,000 in fixed
costs per engine family for the residential engines which often do not use it in today's design.
Overall, fuel savings would be increased due to the application of electronic fuel injection to all
Class II engines.
Table 11.2-1: Small SI Per-Engine Cost Estimates (Without Fuel Savings)
Sales Weighted Averages
Short Term (years 1-5)
Long Term (years 6-10)
Standard
Class I
Class II
$10-$26
$17-$60
$10-$12
$12-$30
More Stringent
Class I
Class II
$17-$23
$110-$149
$12-$18
$76-$89
11.2.1.2 Outboard/Personal Watercraft (OB/PWC) Engine HC+NOx and CO
Emission Standards
We believe that, to meet the more stringent alternative considered here, manufacturers
would need to convert their product lines primarily to a mix of calibrated four-stroke engines
and engines equipped with catalysts. To model this approach, we looked at a technology mix
that would achieve the 10 g/kW-hr HC+NOx limit, with appropriate considerations given to
emissions deterioration rates and compliance margins. This technology mix was developed by
assuming that all carbureted two-stroke engines would be removed from the fleet and replaced
with four-stroke engines. All engines over 75 kW (100 hp) were modeled as using catalytic
control. Detailed costs for converting engines from two-stroke to four-stroke and for equipping
OB/PWC engines with catalysts are presented in Chapter 6. Table 11.2-2 compares the average
per-engine equipment costs for the primary and the more stringent alternatives for OB/PWC
engines.
11-9
-------
Final Regulatory Impact Analysis
Table 11.2-2: OB/PWC Per-Engine Cost Estimates (Without Fuel Savings)
Sales Weighted Averages
Standard
More Stringent
Incremental Cost3
OB
PWC
OB
PWC
OB
PWC
Short Term (years 1-5)
$291
$359
$388
$528
$102
$169
Long Term (years 6-10)
$224
$272
$275
$392
$51
$120
a Incremental cost is presented here because the more stringent alternative for OB/PWC includes the primary
standard in 2010 plus a second, more stringent, standard in 2012.
We did not model differences in fuel savings between the primary and more stringent
alternatives. The fuel savings for all three alternatives primarily come from the replacement of
carbureted two-stroke engines with cleaner engine designs. In both the primary and more
stringent scenarios, we model the discontinuation of sales of carbureted two-stroke engines.
11.2.1.3 Sterndrive/Inboard (SD/I) Engine HC+NOx and CO Emission Standards
With regard to the less stringent alternative, Chapter 4 presents costs for using exhaust
gas recirculation (EGR) on SD/I engines. To estimate the costs for the less stringent alternative,
all SD/I engines less than 373 kW were modeled to be equipped with electronic closed loop
control fuel injection and EGR.
For the more stringent case, we consider a larger catalyst size with a higher precious
metal loading for engines. Specifically, for engines less than 373 kW, we model a 25 percent
larger catalyst and an additional 25 percent precious metal loading. We do not model a
difference in fuel consumption for any of these scenarios because, in each case, all engines are
anticipated to use electronic fuel injection. Table 11.2-3 compares the per-engine cost estimates
for the primary, less stringent, and more stringent alternatives. As discussed above, we do not
including high-performance engines in this analysis.
Table 11.2-3: SD/I <373 kW Per-Engine Cost Estimates (Without Fuel Savings)
Sales Weighted Averages
Standard
Less Stringent
More Stringent
Short Term (years 1-5)
$355
$200
$431
Long Term (years 6-10)
$266
$149
$333
11-10
-------
Regulatory Alternatives
11.2.2 Costs for Evaporative Emission Standards
11.2.2.1 Small SI Engine
For the less stringent case, we simply subtract the costs of running loss controls for non-
handheld equipment. For the more stringent case, we add the incremental costs of diurnal
emission control for all nonhandheld engines and diurnal emission and running loss control for
handheld engines. These technology costs are presented in Chapter 6. Table 11.2-4 compares
the per-equipment cost estimates for the primary, less stringent, and more stringent alternatives.
Table 11.2-4: Evaporative Small SI Per-Equipment Cost Estimates (Without Fuel Savings)
Sales Weighted Averages
Standard Aggregate
Handheld
Class I
Class II
Less Stringent
Aggregate
Handheld
Class I
Class II
More Stringent Aggregate
Handheld
Class I
Class II
Short Term (years 1-5)
$3.27
$0.82a
$3.05
$6.73
$1.86
$0.82a
$1.13
$4.50
$6.76
$4.40
$6.01
$11.08
Long Term (years 6-10)
$2.46
$0.69a
$2.20
$5.16
$1.34
$0.69a
$0.67
$3.38
$5.25
$3.55
$4.57
$8.64
a Values reflect the final permeation standards. These costs are used in the alternative analysis
only to develop aggregate values for comparison purposes.
11-11
-------
Final Regulatory Impact Analysis
Table 11.2-5 presents the fuel savings for the three alternatives, based on the
evaporative emission reductions for each of the scenarios. Because evaporative emissions are
basically gasoline vapor lost to the atmosphere, these hydrocarbon reductions can be directly
translated to gasoline savings using a gasoline cost of $1.81 per gallon. Cost savings are
presented both with a 3 percent and a 7 percent discount factor over the life of the equipment.
Table 11.2-5: Projected Evaporative Fuel Savings for Small SI Equipment
Sales Weighted Averages
Standard Aggregate
Handheld
Class I
Class II
Less Stringent
Aggregate
Handheld
Class I
Class II
More Stringent Aggregate
Handheld
Class I
Class II
Lifetime Gallons Saved
1.4
0.2
0.8
4.7
0.9
0.2a
0.5
3.0
1.5
0.3
0.9
5.3
Discounted Cost Savings
3 percent 7 percent
$2.36
$0.33
$1.41
$6.53
$1.53
$0.33a
$0.92
$4.16
$2.63
$0.49
$1.53
$7.32
$2.17
$0.31
$1.31
$5.96
$1.41
$0.31a
$0.85
$3.80
$2.41
$0.46
$1.41
$6.69
Values reflect the final permeation standards. These costs are used in the alternative analysis only to develop
aggregate values for comparison purposes.
11-12
-------
Regulatory Alternatives
11.2.2.2 Marine
For the less stringent case, we simply subtract the costs of diurnal emission controls
from marine vessels with installed and portable fuel tanks. For the more stringent case, we add
the incremental costs of actively purged diurnal emission control for vessels with installed fuel
tanks. These technology costs are presented in Chapter 6. Table 11.2-6 compares the per-
equipment cost estimates for the primary, less stringent, and more stringent alternatives. Cost
savings are presented both with a 3 percent and a 7 percent discount factor over the life of the
vessel.
Table 11.2-6: Per-Vessel Cost Estimates (Without Fuel Savings)
Sales Weighted Averages
Standard Aggregate
portable
PWC
installed
Less Stringent Aggregate
portable
PWC
installed
More Stringent Aggregate
portable
PWC
installed
Short Term (years 1-5)
$55
$12
$17
$74
$33
$11
$17a
$42
$69
$12a
$17a
$94
Long Term (years 6-10)
$45
$8
$11
$62
$27
$7
$lla
$36
$56
$8a
$lla
$77
a Values reflect the final permeation and diurnal standards. These costs used in the alternative analysis only to
develop aggregate values for comparison purposes.
11-13
-------
Final Regulatory Impact Analysis
Table 11.2-7 presents the fuel savings for the three alternatives. These fuel savings are
based on the evaporative emission reductions for each of the scenarios. Because evaporative
emissions are basically gasoline vapor lost to the atmosphere, preventing these hydrocarbon
emissions can be directly translated to gasoline savings using a gasoline cost of $1.81 per
gallon.
Table 11.2-7: Projected Evaporative Fuel Savings for Marine Vessels
Sales Weighted Averages
Standard Aggregate
portable
PWC
installed
Less Stringent
Aggregate
portable
PWC
installed
More Stringent Aggregate
portable
PWC
installed
Lifetime Gallons Saved
28
13
9
38
20
11
9a
26
30
13a
9a
39
Discounted Cost Savings
3 percent 7 percent
$42
$20
$14
$54
$30
$17
$14a
$37
$44
$20a
$14a
$57
$33
$17
$12
$42
$24
$14
$12a
$29
$34
$17a
$12a
$44
a Values reflect the final permeation and diurnal standards. These costs used in the alternative analysis only to
develop aggregate values for comparison purposes.
11-14
-------
Regulatory Alternatives
11.2.3 Cost Summary of Regulatory Alternatives
Table 11.2-8 summarizes the average cost per engine for the various alternative program
options described above. The costs presented are for the short term and do not include fuel
savings.
Table 11.2-8: Engine Cost Summary Range for Alternative Program Options ($/Engine)
Sales Weighted Averages of Short-Term Costs without Fuel Savings, 2005$
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
$10-$26
$17-$60
$-
$360
$-
$4.32
$-
$74
$12
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
$17-$23
$110-$149
$70
$216
$435
$3.58
$2.30
$3.45
$42
$94
$11
a Costs are presented incremental to the standards for OB/PWC because, for this alternative, a second stage of
standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
Table 11.2-9 summarizes the 30-year net present value for costs for the standards and
the various alternative program options described in Table 11.2-1. Cost results are provided as
the total net present value (NPV) for a 30-year period. The NPV estimates are based on a 7
percent discount rate. These costs do not include fuel savings. Table 11.2-10 presents the same
information with a 3 percent discount rate.
11-15
-------
Final Regulatory Impact Analysis
Table 11.2-9: 30-Year Net Present Value Cost Summary for Alternative
Program Options with a 7 Percent Discount Rate (Million 2005$)
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
$1,228
$1,146
$-
$343
$-
$718
$-
$250
$8
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
$1,558
$4,040
$347
$194
$388
$318
$394
$570
$144
$310
$7
a Costs are presented incremental to the final standards for OB/PWC because, for this alternative, a second stage of
standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
Table 11.2-10: 30-Year Net Present Value Cost Summary for Alternative
Program Options with a 3 Percent Discount Rate (Million 2005$)
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standards
$2100
$1831
$-
$541
$-
$1,180
$-
$413
$12
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
$2,944
$7,366
$556
$304
$626
$544
$630
$962
$239
$512
$11
a Costs are presented incremental to the standards for OB/PWC because, for this alternative, a second stage of
standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
11.3 Emission Reduction
11-16
-------
Regulatory Alternatives
This section describes the estimated emission reductions associated with each of the
alternative program options. We developed these estimates using the NONROAD emissions
inventory model and methodology described in Chapter 3. The modeling inputs for alternative
options are provided in Appendix 11A and Appendix 1 IB.
The incremental emission reductions of complying with the alternative program options
are presented as incremental to the base case without the final standards. The only exception to
this is the second phase of OB/PWC standards. The alternatives and the requisite technology
are described in Section 11.1. Further, emission inventory results are provided as the total net
present value (NPV) for a 30-year period. The NPV estimates are calculated based on both a 7
percent and a 3 percent discount rate. Small SI and Marine SI emission reductions are
presented separately in Tables 11.3-1 and 11.3-2.
Table 11.3-1: 30-Year Net Present Value
Emission Reduction Summary for Alternative
Program Options with a 7 Percent Discount Rate (Million Tons)
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
0.73
1.05
0
0.33
0
1.04
0
0.36
0.07
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
0.63
1.27
0.26
0.22
0.32
0.04
0.63
0.12
0.26
0.38
0.06
a Tons reduced are presented incremental to the standards for OB/PWC because, for this alternative, a second stage
of standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
11-17
-------
Final Regulatory Impact Analysis
Table 11.3-2: 30-Year Net Present Value
Emission Reduction Summary for Alternative
Program Options with a 3 Percent Discount Rate (Million Tons)
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
1.33
1.90
0
0.64
0
1.83
0
0.70
0.13
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
1.22
2.52
0.50
0.42
0.65
0.07
1.09
0.21
0.50
0.73
0.11
a Tons reduced are presented incremental to the standards for OB/PWC because, for this alternative, a second stage
of standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
11-18
-------
Regulatory Alternatives
11.4 Cost Effectiveness
This section describes the cost effectiveness associated with each of the alternative
program options. The costs are expressed as millions of dollars and the emission reductions are
in terms of short tons. All results are presented as incremental to the base case without the final
standards. The only exception to this is the second phase of OB/PWC standards where the
values are calculated based on costs and emission reductions incremental to the final standards.
Tables 11.4-1 and 11.4-2 present cost per ton estimates, using both a 7 percent and a 3 percent
discount rate, for Small SI engines/equipment and Marine SI engines/vessels as outlined in
Table 11.2-1.
Table 11.4-1: Comparison of Cost Effectiveness for Final Standards and Alternatives
Without Fuel Savings, 7 Percent Discount Rate ($/ton) 2005$
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
$1,680
$1,086
$790
$1,030
NA
$690
NA
$690
$115
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
$2,540
$3,170
$1,340
$880
$1.210
$8,150
$630
$4,900
$550
$820
$120
a Cost effectiveness of more stringent alternative is presented incremental to the standards for OB/PWC because,
for this alternative, a second stage of standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
11-19
-------
Final Regulatory Impact Analysis
Table 11.4-2: Comparison of Cost Effectiveness for Final Standards and Alternatives
Without Fuel Savings, 3 Percent Discount Rate ($/ton) 2005$
Source
Exhaust
Evap
Alt
1
2
3a
4
5
6b
7
8b
9
10
11
Target
Class I
Class II
OB/PWC
SD/I <373 kW
HH
Class I & Class II
Class I & Class II
Installed marine fuel
tank
Portable marine fuel tank
Standard
$1,580
$965
$670
$840
NA
$640
NA
$590
$100
Scenario
more
more
more
less
more
more
less
more
less
more
less
Alternative
$2,410
$2,930
$1100
$720
$970
$7,620
$580
$4560
$500
$700
$100
a Cost effectiveness of more stringent alternative is presented incremental to the standards for OB/PWC because,
for this alternative, a second stage of standards is considered in 2012 beyond the final standards.
b Only considers standards for venting emission control which are not in the final standards. The venting emission
standards considered here are diurnal for Class I and Class II and diurnal/running loss for handheld.
Ideally, this analysis would include an assessment of the monetized benefits which
would potentially accompany each alternative as was provided in Chapter 8. This would
provide further information for decision making and comparison to the final program.
Unfortunately, the emissions data needed to conduct such an analysis, such as the potential PM
benefits for the more stringent exhaust emission scenarios, is not available for this NPRM. This
limits the utility of any comparisons which could be made since monetized benefits are partially
dependent on PM health benefits.
11.5 Summary and Analysis of Alternative Program Options
This section presents a comparative summary of the important aspects related to the
various alternative program options and our rationale for not pursuing an option relative to the
final standards.
11.5.1 Exhaust Emission Standards
11.5.1.1 Small SI Engine HC+NOx Standards
11.5.1.1.1 Class I
This alternative considers a more stringent standard of 50 percent HC+NOx emission
reduction beginning in 2015 for Phase 3 Class I engines instead of a reduction of 38 percent
11-20
-------
Regulatory Alternatives
beginning in 2012 . While these emission standards may be feasible, it is clearly in the in the
longer term relative to the timing of the final standards. For analytical purposes the time line to
begin implementation of the new standards was set at the 2015 model year. This is three model
years past the implementation year for the final standards. For the analytical period we
considered, the final standards provide more emission reductions than the alternative by
202,600 tons between 2012 and 2020. Postponing the exhaust emission standards to 2015 could
likely also lead to postponing controls on running loss emissions with an additional loss of
47,000 tons of control. States with air quality problems would benefit from emission reductions
in an earlier time frame. Thus, while both approaches are cost effective, we elected to go with
the 38 percent reduction in 2012. In the context of section 213(a)(3) of the Clean Air Act, it
represents the most stringent standards feasible within the lead time considered.
11.5.1.1.2 ClassII
This alternative considers a more stringent standard of 4 g/kW-hr HC+NOx , a reduction
of about 67 percent for Class II engines over phase 2. These standards assume the use of closed
loop electronic fuel injection and catalysts on all Class II engines. We are expecting engine
manufacturers to meet the final standards by applying closed loop EFI on a portion of their V-
twin engines and for the engine manufacturers or equipment manufacturers to use catalytic
mufflers on the remaining engines. While these emission standards may be feasible it is clearly
in the in the longer term relative to the timing of the final standards. For analytical purposes the
time line to begin implementation of the new standards was set at the 2015 model year. This is
four model years past the implementation year for the final standards. For the 30 year analytical
period we considered, the final rule provides fewer overall emission reductions than the
alternative, but between 2011 and 2020 the final rule gives 150,300 tons more reduction than
the alternative assuming that running loss control is also postponed to begin in the 2015 model
year. States with air quality problems would benefit from emission reductions in an earlier time
frame. Thus, while both approaches are cost effective, we elected to go with the 34 percent
reduction in 2011. In the context of section 213(a)(3) of the Clean Air Act, it represents the
most stringent standards feasible within the lead time considered.
11.5.1.2 Outboard/Personal Watercraft (OB/PWC) Engine HC+NOx and CO
Emission Standards
We analyzed the costs and emission reductions associated with more stringent standards
for OB/PWC engines. We have concerns with this second tier of OB/PWC standards at this
time. While some four-stroke engines may be able to meet a 10 g/kW-hr standard with
improved calibrations, it is not clear that all engines could meet this standard without applying
catalyst technology. Direct injection two-strokes engines would face additional challenges. At
this time, we believe it is not appropriate to base standards in this rule on the use of catalysts for
OB/PWC engines. Although this technology may be attractive in the longer term, little
development work has been performed on the application of 3-way catalysts to OB/PWC
engines. For this alternative, our modeling assumes all OB/PWC engines which need to can
successfully apply aftertreatment technology.
11-21
-------
Final Regulatory Impact Analysis
11.5.1.3 Sterndrive/Inboard (SD/I) Engine HC+NOx and CO Emission Standards
With regard to less stringent standards, we believe that EGR would be a technologically
feasible and cost-effective approach to reducing emissions from SD/I marine engines.
However, we believe that greater reductions could be achieved through the use of catalysts. We
considered basing an interim standard on EGR, but were concerned that this would divert
manufacturers' resources away from catalyst development and could have the effect of delaying
emission reductions from this sector. Setting a less stringent standard would likely be
inconsistent with the requirements of section 213 of the Clean Air Act because at least one SD/I
engine manufacturer offers a compliant product for sale in the US. In the NPRM we do ask for
comment on a short-phase-in to deal with a change in the engine a supplier's product lines.
With regard to more stringent requirements, we do not believe that they would
necessarily lead to any further significant emission reductions in HC+NOx. Because this is the
first generation of emission standards for this category of recreational marine engines, we
believe that most manufacturers will strive to achieve emission levels below the final standards
to give them certainty that they will pass the standards in-use, especially as catalysts on SD/I
engines are a new technology. Therefore, we do not believe that it is necessary at this time to
consider a lower standard for these engines.
11.5.2 Evaporative Emission Standards
11.5.2.1 Small SI Engine
We analyzed requiring diurnal and running loss control from handheld equipment in
2012. Even though it would be feasible from a strict technical perspective it is not a attractive
option at this time. Fuel tanks from this equipment are very small, most less than one liter, and,
with the exception of commercial equipment, their use is less than 15 hours per year. Adding
hardware to control diurnal and running loss emissions would add weight which could be
problematic on handheld equipment. In addition, it could create the potential for fuel leaks in
equipment which can be used in rotated and inverted positions in the field. In addition, this
option does not appear cost effective. For these reasons we elected not to pursue it.
With regard to controlling running loss emissions control from non-handheld equipment
we believe it is feasible at a relatively low cost. Running loss emissions can be controlled by
sealing the fuel cap and routing vapors from the fuel tank to the engine intake. This emission
control approach is relatively straight-forward and inexpensive and do not have the weight and
in-use position issues such as mentioned above for handheld equipment. Deleting the
requirement does not meaningfully improve the cost effectiveness. Not finalizing these
requirements would be inconsistent with the section 213 of the Clean Air Act.
California requires control diurnal fuel tank emissions from Class I and Class II
equipment as part of its overall fuel evaporative certification requirements. California requires
an active purge of the control system. We evaluated the alternative of adding a diurnal
requirement like that in California. Even though it would be feasible from a strict technical
11-22
-------
Regulatory Alternatives
perspective it is not a attractive option at this time. While workable, there are some important
issues would need to be resolved for diurnal emission control, such as cost, packaging, and
vibration. Also, California requires an active purge, but we believe that a substantial reduction
on the order of 50 percent could be achieved with a less complicated and less expensive passive
purge approach. Finally, the cost and cost effectiveness of this program sub-element are of
concern given the relatively low emissions levels (on a per-equipment basis) from such small
fuel tanks. Overall, we do not consider this to be an attractive option at this time for Small SI
engines as a group.
11.5.2.2 Marine
Although we considered the alternative of not requiring diurnal emission control for
installed fuel tanks, we believe that carbon canisters are feasible for boats at relatively low cost.
Carbon canisters have been installed on fourteen boats by industry in a pilot program intended
to demonstrate the feasibility of this technology. The final standards are achievable through
engineering design-based certification with canisters that are much smaller than the fuel tanks.
In addition, sealed systems, with pressure control strategies would be accepted under the
engineering design-based certification provisions. Eliminating these requirements would not
meaningfully affect the cost effectiveness of the marine evaporative program. Not finalizing
these controls would be inconsistent with the requirements of section 213 of the Clean Air Act.
We also considered the feasibility of requiring the use of carbon canisters with active
purging to control diurnal emissions. However, we are concerned that active purging would
occur infrequently due to the low hours of operation per year seen by many boats. In addition,
active purge adds complexity into the system in that the engine must be integrated into the
control strategy. This could end up involving engine, tank, and vessel manufacturers in
certification processes. Although we did not model it, this approach would undoubtedly require
more lead time to implement because it is more complex and involves more entities. Based on
data presented in Chapter 5, carbon canisters can be used to reduce emissions by more than 50
percent with passive purging. This passive purging occurs during the normal tank breathing
process caused by ambient temperature changes without creating any significant pressure in the
fuel tank. The small additional benefit of an actively purged diurnal control system would
likely not justify the cost and complexity of implementing such a system, even though it
appears to be cost effective.
Portable marine fuel tanks are used in vessels with outboard motors. Many of these
tanks employ self-sealing vents which close the tank to the atmosphere when it is not in-use.
This is quite straightforward, and it can be applied to all such tanks in the future for a
reasonable cost. Not finalizing these controls would be inconsistent with the requirements of
section 213 of the Clean Air Act.
11-23
-------
Final Regulatory Impact Analysis
APPENDIX 11 A: Emission Factors for the Less Stringent Alternative
11 A.I Exhaust Emission Factors and Deterioration Rates
11 A. 1.1 Small SI Exhaust
No less stringent exhaust emission standards were quantitatively analyzed for either
Class I or Class II Small SI engines.
11 A. 1.2 Marine SI Exhaust
In the less stringent alternative, the same standards are considered for OB/PWC engines
as for the primary scenario. However, for SD/I engines, we consider less stringent standards.
As discussed above, these standards are based on the use of EGR for SD/I engines less than 373
kW and engine calibration for larger engines. For engines less than 373 kW we considered less
stringent alternative standards of 10 g/kW-hr HC+NOx and 150 g/kW-hr CO for SD/I engines
less than 373 kW. For high-performance engines, we did not model alternative scenarios, as
discussed above. Because these emission factors are based on engine-out emissions, we use the
same deterioration factors (DF) as for the baseline case. Table A-l presents the zero-hour SD/I
emission factors and the accompanying deterioration factors used to model the less stringent
alternative.
Table 11 A-l: Less Stringent Alternative EFs [g/kW-hr] and DFs for SD/I
Engine Category
<373 kW SD/I
HC
EF
4.05
DF
1.26
NOx
EF
4.00
DF
1.03
CO
EF
96.3
DF
1.35
BSFC
345
11A.2 Evaporative Emission Factors
As discussed above, no changes in the hose and tank permeation standards were
considered in the less stringent alternative. The less stringent scenario was modeled for Small
SI equipment by using the baseline running loss and diffusion rates for Class I and Class II
equipment. For marine, the less stringent alternative was modeled by using the baseline diurnal
emission rates for vessels with installed fuel tanks.
11-24
-------
Regulatory Alternatives
APPENDIX 11B: Emission Factors for the More Stringent Alternative
11B.1 Exhaust Emission Factors and Deterioration Rates
1 IB. 1.1 Small SI Exhaust
For analytical purposes, we identified a more stringent program option of 8 g/kW-hr
HC+NOx standard for Class I engines that would be implemented beginning in 2015. This
standard represents about a 50 reduction from the existing Phase 2 standard, rather than the
approximately 38 percent reduction associated with the final rule. The option also provides 3
more years of lead time. For Class II engines, we identified an alternative for a more stringent
exhaust HC+NOx emission standard of 4.0 g/kW-hr beginning in 2015. (This option also
includes an associated delay in the corresponding running loss requirement such that engine
changes are made simultaneously.) Such an exhaust emission standard represents a 67 percent
reduction relative to the existing Phase 2 standard, rather than the 34 percent reduction
associated with the final rule.
In modeling this more stringent option, we assumed the same phase-in schedule that
reflects a number of flexibilities for engine and equipment manufacturers, and allows them to
sell some Phase 2 compliant engines in the early years of the program. We also assumed that
Class I side-valve technology would be completely replaced with overhead valve designs, and
that all of the Class II engines would require closed loop control electronic fuel injection (EFI).
Since EFI equipped engines enjoy a 10 percent fuel consumption advantage over their
carbureted counterparts, we also revised the brake-specific fuel consumption (BSFC) for Class
II engines. The new BSFC value is 0.666 Ib/hp-hr.
All the modeling inputs were developed using a methodology consistent with that
described in Chapter 3 of this draft RIA. The alternative emission standards and phase-in
assumptions are shown in Table B-l. The emission factors are shown in Table B-2.
Table 11B-1: More Stringent Phase 3 Emission Standards and Implementation Schedule
for Class I and II Small SI Engines (g/kW-hr or Percent)
Engine
Class
Class I
Class II
Requirement
HC+NOx
Required Sales
Percentage
HC+NOx
Required Sales
Percentage
2015
8
95
4
83
2016
8
95
4
83
2017
8
100
4
93
2018
8
100
4
93
2019+
8
100
4
100
11-25
-------
Final Regulatory Impact Analysis
Table 11B-2: More Stringent Phase 3 Modeling Emission Factors
for Small SI Engines (g/KW-hr)
Class/
Technology
Class I - SV
Class I -
Class II
HC
ZML
4.48
4.07
2.13
HC "A"
1.011
1.011
1.011
NOx
ZML
1.12
1.53
0.67
NOx "A"
0.470
0.470
0.470
CO ZML
319.76
325.06
391.13
CO "A"
0.070
0.070
0.080
1 IB. 1.2 Marine SI Exhaust
For OB/PWC engines, the more stringent alternative considers exhaust emissions
standards that are about 40 percent lower for HC+NOx and about 30 percent lower for CO than
the final standard. The more stringent alternative emission standards are modeled as a second
phase of standards, beyond the primary, beginning in 2012. In determining the combined
HC+NOx emission factor, we used the final emission standards with a 10 percent compliance
margin (with deterioration factor applied). To determine the NOx emission factors, we used
certification data and other emissions data presented in Chapter 4, to determine the sales
weighted average NOx for low emission technologies in each power bin. HC was then
determined as the difference between the HC+NOx and the NOx emission factors. Because we
are finalizing the same standards for OB and PWC and because they use similar engines, we use
the same HC+NOx emission factors and deterioration factors for both engine types. Because
the final CO standard primarily acts as a cap on CO for many of the engines, the CO emission
factors differ somewhat for CO based on data in the certification database for low CO engines.
We use the same deterioration rates as in the primary case. Table B-3 presents the zero-hour
OB/PWC emission factors used in analyzing the more stringent alternative.
Table B-3: More Stringent Alternative Emission Factors for OB/PWC [g/kW-hr]
Power Bin
0-2.2 kW
2.3-4.5 kW
4.6-8.2 kW
8. 3-11. 9 kW
12.0-18.6 kW
18.7-29.8 kW
29.9-37.3 kW
37.4-55.9 kW
55.9-74.6 kW
74.7-130.5 kW
130.6+ kW
HC
11.7
10.9
10.5
9.0
9.5
7.5
5.7
5.2
5.2
5.4
6.3
NOx
3.02
2.25
3.50
4.22
2.69
3.55
3.70
3.38
3.38
3.13
2.30
CO
OB PWC
362
238
195
165
137
120
114
115
115
101
93
426
359
162
154
145
137
137
137
137
135
119
BSFC
563
560
555
552
543
528
507
471
471
415
387
11-26
-------
Regulatory Alternatives
For SD/I engines greater than 373 kW, we did not model the use of catalysts for reasons
discussed above. However, for SD/I engines less than 373 kW, we considered a more stringent
HC+NOx standard of 2.5 g/kW-hr. To model this standard, we used zero-hour emission factors
of 0.90 g/kW-hr HC and 0.80 g/kW-hr NOx. No changes were made in other emission factors
for this more stringent alternative. In addition, the same deterioration factors were used here as
in the primary alternative.
11B.2 Evaporative Emission Factors
As discussed above, no changes in the hose and tank permeation standards were
considered in the more stringent alternative. The more stringent scenario modeled for Small SI
equipment by considering diurnal standards beginning in 2011 for Class II and 2012 for
handheld and Class I equipment. This diurnal emission standards was modeled using a 60
percent reduction from baseline. Also, the more aggressive option for Class II exhaust
standards was modeled as also including a corresponding delay in the running loss requirement
such that engine changes are made simultaneously.
For marine, the more stringent alternative was a standard requiring active purging of
canisters for vessels with installed fuel tanks. This was modeled by using a 70 percent
reduction in diurnal emissions compared to the baseline.
11-27
-------
Final Regulatory Impact Analysis
Chapter 11 References
1. "Small SI Engine Technologies and Costs, Final Report," ICF International, August 2006.
2. "Marine Outboard and Personal Watercraft SI Engine Technologies and Costs," ICF Consulting, prepared for the
U.S. Environmental Protection Agency, July 2006, Docket Identification EPA-HQ-OAR-2004-0008-0452.
3. "Sterndrive and Inboard Marine SI Engine Technologies and Costs," ICF Consulting, prepared for the U.S.
Environmental Protection Agency, July 2006, Docket Identification EPA-HQ-OAR-2004-0008-0453.
4. "Small SI Engine Technologies and Costs, Final Report," ICF International, August 2006.
5. "Small SI Engine Technologies and Costs, Final Report," ICF International, August 2006.
11-28
-------
-------
|