United States Air and Radiation EPA420-R-02-022
Environmental Protection September 2002
Agency
<>EPA Final Regulatory Support
Document: Control of
Emissions from Unregulated
Nonroad Engines
) Printed on Recycled Paper
-------
EPA420-R-02-022
September 2002
Final Regulatory Support Document:
Control of Emissions from
Unregulated Nonroad Engines
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
-------
-------
Table of Contents
Executive Summary
Chapter 1: Health and Welfare Concerns
1.1 Inventory Contributions 1-1
1.1.1 Inventory Contribution 1-1
1.1.2 Baseline Inventory Adjustment 1-5
1.1.2 Inventory Impacts on a Per Vehicle Basis 1-7
1.2 Ozone 1-8
1.2.1 General Background 1-8
1.2.2 Health and Welfare Effects of Ozone and Its Precursors 1-9
1.2.3 Ozone Nonattainment and Contribution to Ozone Nonattainment 1-11
1.2.4 Public Health and Welfare Concerns from Prolonged and Repeated Exposures to
Ozone 1-14
1.2.5 Additional Health and Welfare Effects of NOx Emissions 1-15
1.3 Carbon Monoxide 1-17
1.3.1 General Background 1-17
1.3.2 Health Effects of CO 1-18
1.3.3 CO Nonattainment 1-18
1.4 Particulate Matter 1-23
1.4.1 General Background 1-23
1.4.2 Health and Welfare Effects of PM 1-25
1.4.3 PM Nonattainment 1-26
1.5 Visibility Degradation 1-31
1.5.1 General Background 1-31
1.5.2 Visibility Impairment Where People Live, Work and Recreate 1-32
1.5.3 Visibility Impairment in Class I Areas 1-44
1.5.4 Recreational Vehicles and Visibility Impairment in Class I Areas 1-47
1.6 Gaseous Air Toxics 1-54
1.6.1 Benzene 1-54
1.6.2 1,3-Butadiene 1-55
1.6.3 Formaldehyde 1-56
1.6.4 Acetaldehyde 1-57
1.6.5 Acrolein 1-57
1.6.6 Toluene 1-57
1.7 Exposure to CO and Air Toxics Associated with Nonroad Engines and Vehicles . . 1-58
1.7.1 Large SI Engines 1-58
1.7.2 Snowmobiles 1-59
Chapter 2: Industry Characterization
2.1 CI Marine Engines and Recreational Boats 2-1
2.1.1 The Supply Side 2-1
2.1.2 The Demand Side 2-6
2.1.3 Industry Organization 2-7
-------
2.1.4 Markets 2-11
2.2 Large SI Engines and Industrial Equipment 2-19
2.2.1 The Supply Side 2-19
2.2.2 The Demand Side 2-23
2.2.3 Industry Organization 2-26
2.2.4 Markets 2-28
2.3 Snowmobile Market 2-31
2.3.1 The Supply Side 2-31
2.3.2 The Demand Side 2-33
2.3.3 Industry Organization 2-35
2.3.4 Snowmobile Retailers and Rental Firms 2-37
2.3.5 Markets 2-37
2.4 All-Terrain Vehicles 2-39
2.4.1 The Supply Side 2-39
2.4.2 The Demand Side 2-41
2.4.3 Industry Organization 2-42
2.4.4 Markets 2-46
2.5 Off-Highway Motorcycles 2-48
2.5.1 The Supply Side 2-49
2.5.2 The Demand Side 2-56
2.5.3 Industry Organization 2-57
2.5.4 Markets 2-60
Chapter 3: Technology
3.1 Introduction to Spark-Ignition Engine Technology 3-1
3.1.1 Four-Stroke Engines 3-1
3.1.2 Two-Stroke Engines 3-2
3.1.3 - Engine Calibration 3-3
3.2 - Exhaust Emissions and Control Technologies 3-6
3.2.1 - Current Two-Stroke Engines 3-6
3.2.2 - Clean Two-Stroke Technologies 3-10
3.2.3 - Current Four-Stroke Engines 3-15
3.2.4 - Clean Four-Stroke Technologies 3-16
3.2.5 - Advanced Emission Controls 3-19
3.3 - Evaporative Emissions 3-22
3.3.1 Sources of Evaporative Emissions 3-22
3.3.2 Evaporative Emission Controls 3-24
3.4 CI Recreational Marine Engines 3-28
3.4.1 Background on Emissions Formation from Diesel Engines 3-28
3.4.2 MarinizationProcess 3-30
3.4.3 General Description of Technology for Recreational Marine Diesel Engines
3-31
Chapter 4: Feasibility of Standards
4.1 CI Recreational Marine 4-1
4.1.1 Baseline Technology for CI Recreational Marine Engines 4-1
-------
4.1.2 Anticipated Technology for CI Recreational Marine Engines 4-4
4.1.3 Emission Measurement Procedures for CI Recreational Marine Engines 4-5
4.1.4 Impacts on Noise, Energy, and Safety 4-14
4.2 Large Industrial SI Engines 4-16
4.2.1 2004 Standards 4-16
4.2.2 2007 Standards 4-16
4.2.3 Impacts on Noise, Energy, and Safety 4-37
4.3 Snowmobile Engines 4-39
4.3.1 Baseline Technology and Emissions 4-39
4.3.2 Potentially Available Snowmobile Technologies 4-39
4.3.3 Test and Measurement Issues 4-44
4.3.4 Impacts on Noise, Energy, and Safety 4-48
4.3.5 Conclusions 4-48
4.4 All-Terrain Vehicles/Engines 4-53
4.4.1 Baseline Technology and Emissions 4-53
4.4.2 Potentially Available ATV Technologies 4-56
4.4.3 Test Cycle/Procedure 4-64
4.4.4 Small Displacement Engines 4-66
4.4.5 Impacts on Noise, Energy, and Safety 4-67
4.4.6 Conclusion 4-67
4.5 Off-Highway Motorcycles 4-68
4.5.1 Baseline Technology and Emissions 4-68
4.5.2 Potentially Available Off-Highway Motorcycle Technologies 4-70
4.5.3 Test Procedure 4-75
4.5.4 Impacts on Noise, Energy, and Safety 4-76
4.5.5 Conclusion 4-76
4.6 Permeation Control from Recreational Vehicles 4-78
4.6.1 Baseline Technology and Emissions 4-78
4.6.2 Permeation Reduction Technologies 4-81
4.6.3 Test Procedures 4-97
4.6.4 Conclusion 4-99
4.6.5 Impacts on Noise, Energy, and Safety 4-99
Appendix to Chapter 4: Emission Index For Recreational Vehicle Hangtags 4-101
Chapter 5: Costs of Control
5.1 Methodology 5-1
5.2 Cost of Emission Controls by Engine/Vehicle Type 5-2
5.2.1 Recreational Marine Diesel Engines 5-2
5.2.2 Large Industrial Spark-Ignition Engines 5-7
5.2.3 Recreational Vehicles 5-17
Chapter 6: Emissions Inventory
6.1 Methodology 6-1
6.1.1 Off-highway Exhaust Emissions 6-1
6.1.2 Off-highway Evaporative Emissions 6-2
6.2 Effect of Emission Controls by Engine/Vehicle Type 6-6
-------
6.2.1 Compression-Ignition Recreational Marine 6-6
6.2.2 Large Spark-Ignition Equipment 6-10
6.2.3 Snowmobile Exhaust Emissions 6-23
6.2.4 All-Terrain Vehicle Exhaust Emissions 6-29
6.2.5 Off-highway Motorcycle Exhaust Emissions 6-34
6.2.6 Evaporative Emissions from Recreational Vehicles 6-40
Appendix to Chapter 6: ATV and Off-highway Motorcycle Usage Rates 6-45
Chapter 7: Cost Per Ton
7.1 Cost Per Ton by Engine Type 7-1
7.1.1 Introduction 7-1
7.1.2 Compression-Ignition Recreational Marine 7-1
7.1.3 Large Industrial SI Equipment 7-3
7.1.4 Recreational Vehicle Exhaust Emissions 7-7
7.1.5 Recreational Vehicle Permeation Emissions 7-10
7.2 Cost Per Ton for Other Mobile Source Control Programs 7-12
7.3 20-Year Cost and Benefit Analysis 7-13
Chapter 8: Small Business Flexibility Analysis
8.1 Requirements of the Regulatory Flexibility Act 8-1
8.2 Need For and Objectives of the Rule 8-3
8.3 Issues Raised by Public Comments 8-3
8.4 Description of Affected Entities 8-5
8.4.1 Recreational Vehicles (ATVs, off-highway motorcycles, and snowmobiles) . 8-6
8.4.2 Large Spark Ignition Engines 8-9
8.4.3 Marine Vessels 8-9
8.4.4 Results for All Small entities 8-11
8.5 Projected Reporting, Recordkeeping, and Other Compliance Requirements of the
Regulation 8-11
8.6 Steps to Minimize Significant Economic Impact on Small Entities 8-11
8.6.1 General Provisions 8-13
8.6.2 Nonroad recreational vehicles 8-14
8.6.3 Nonroad industrial engines 8-19
8.6.4 Recreational marine diesel engines 8-21
8.7 Conclusion 8-26
Chapter 9: Economic Impact Analysis
9.1 Summary of Economic Impact Results 9-2
9.1.1 Summary Results for Marine 9-2
9.1.2 Summary Results for Large SI 9-5
9.1.3 Summary Results for Snowmobiles 9-7
9.1.4 Summary Results for ATVs 9-9
9.1.5 Summary Results for Off-Highway Motorcycles 9-10
9.1.6 Net Present Value of Surplus Loss, Fuel Cost Savings, and Social Costs/Gains
9-12
-------
9.2 Economic Theory 9-14
9.2.1 Partial vs. General Equilibrium Model Scope 9-14
9.2.2 Length-of-Run Considerations 9-14
9.3 Fuel Efficiency Gains 9-18
9.4 Potential Product Attribute Changes 9-22
9.4.1 Technology Changes for Snowmobiles 9-24
9.4.2 Statistical Analysis of Snowmobile Product Attributes 9-24
9.4.3 Anecdotal Pricing Information For Snowmobiles 9-34
9.4.4 Uncertainties and Limitations of the Attribute Study 9-35
9.4.5 Conclusions 9-35
9.5 Methodology 9-36
9.5.1 Conceptual Model 9-36
9.5.2 Price Elasticity Estimation 9-38
9.6 Marine 9-58
9.6.1 Marine Baseline Market Characterization 9-58
9.6.2 Marine Control Costs 9-59
9.6.3 Marine Economic Impact Results 9-60
9.6.4 Marine Engineering Cost and Surplus Loss Comparison 9-62
9.6.5 Marine Economic Impact Results with Fuel Cost Savings 9-62
9.7 Large SI Engines 9-64
9.7.1 Forklift Baseline Market Characterization 9-65
9.7.2 Forklift Control Costs 9-66
9.7.3 Forklift Economic Impact Results 9-67
9.7.4 Forklift Engineering Cost and Surplus Loss Comparison 9-69
9.7.5 Forklift Economic Impact Results with Fuel Cost Savings 9-69
9.7.6 Economic Impacts - Other Large SI Engines 9-73
9.8 Snowmobiles 9-75
9.8.1 Snowmobile Baseline Market Characterization 9-75
9.8.2 Snowmobile Control Costs 9-76
9.8.3 Snowmobile Economic Impact Results 9-77
9.8.4 Snowmobile Engineering Cost and Surplus Loss Comparison 9-79
9.8.5 Snowmobile Economic Impact Results with Fuel Cost Savings 9-80
9.8.6 Economic Impacts on Individual Engine Manufacturers, Snowmobile Retailers
and Snowmobile Rental Firms 9-82
9.9 All-Terrain Vehicles (ATVs) 9-83
9.9.1 ATV Baseline Market Characterization 9-83
9.9.2 ATV Control Costs 9-84
9.9.3 ATV Economic Impact Results 9-85
9.9.4 ATV Engineering Cost and Surplus Loss Comparison 9-87
9.7.5 ATV Economic Impact Results with Fuel Cost Savings 9-88
9.10 Off-Highway Motorcycles 9-90
9.10.1 Off-Highway Motorcycle Baseline Market Characterization 9-90
9.10.2 Off- Highway Motorcycle Control Costs 9-91
9.10.3 Off-Highway Motorcycles Economic Impact Results 9-91
9.10.4 Off-Highway Motorcycle Engineering Cost and Surplus Loss Comparison
9-93
-------
9.10.5 Off-Highway Motorcycle Economic Impact Results with Fuel Cost Savings
9-94
Appendix to Chapter 9: Sensitivity Analyses 9-97
Chapter 10: Benefit-Cost Analysis
10.1 Introduction 10-1
10.2 General Methodology 10-1
10.2.1 PM Methodology - Benefits Transfer 10-1
10.2.2 CO and Air Toxics Methodology : WTP 10-2
10.2.3 Benefits Quantification 10-3
10.3 PM-Related Health Benefits Estimation 10-7
10.3.1 Emissions Inventory Implications 10-7
10.3.2 Benefits Transfer Methodology 10-8
10.3.3 Overview of Heavy Duty Engine/Diesel Fuel Benefits Analysis and
Development of Benefits Transfer Technique 10-11
10.3.4. Quantifying and Valuing Individual Health Endpoints 10-14
10.3.5. Estimating Monetized Benefits Anticipated in Each Year 10-19
10.3.6. Methods for Describing Uncertainty 10-20
10.3.7. Estimated Reductions in Incidences of Health Endpoints and Associated
Monetary Values 10-23
10.3.8 Alternative Calculations of Estimated Reductions in Incidences of Health
Endpoints and Associated Monetary Values 10-28
10.3.9 Alternative Calculations of PM Mortality Risk Estimates and Associated
Monetary Values 10-28
10.4 CO and Air Toxics Health Benefits Estimation 10-34
10.4.1 Direct Valuation of "Clean" Snowmobiles 10-34
10.4.2 Overview of Benefits Estimation for CO and Air Toxics from the Final Rule
10-36
10.5 Total Benefits 10-37
10.6 Comparison of Costs to Benefits 10-39
Chapter 11: Regulatory Alternatives
11.1 Recreational Marine Diesel Engines 11-1
11.1.1 Harmonization with Draft EC Standards 11-1
11.1.2 Earlier Implementation Dates Consistent with Commercial Marine 11-2
11.2 Large Industrial Spark-Ignition Engines 11-3
11.3 Recreational Vehicle Exhaust Emission Standards 11-4
11.3.1 Off-highway Motorcycles 11-4
11.3.2 All-terrain Vehicles 11-12
11.3.3 Snowmobiles 11-20
11.4 Recreational Vehicle Permeation Emission Standards 11-34
11.4.1 Fuel Tanks 11-34
11.4.2 Hoses 11-38
11.5 Incremental Cost Per Ton Analysis 11-42
-------
Executive Summary
Executive Summary
EPA is adopting new standards for emissions of oxides of nitrogen, hydrocarbons, and
carbon monoxide from several categories of engines. This Final Regulatory Support Document
provides technical, economic, and environmental analyses of the new emission standards for the
affected engines. The anticipated emission reductions will translate into significant, long-term
improvements in air quality in many areas of the U.S. Overall, the requirements will
dramatically reduce individual exposure to dangerous pollutants and provide much needed
assistance to states and regions facing ozone and particulate air quality problems that are causing
a range of adverse health effects, especially in terms of respiratory impairment and related
illnesses.
Chapter 1 reviews information related to the health and welfare effects of the pollutants of
concern. Chapter 2 contains an overview of the affected manufacturers, including some
description of the range of engines involved and their place in the market. Chapter 3 covers a
broad description of engine technologies, including a wide variety of approaches to reducing
emissions. Chapter 4 summarizes the available information supporting the specific standards we
are adopting, providing a technical justification for the feasibility of the standards. Chapter 5
applies cost estimates to the projected technologies. Chapter 6 presents the calculated
contribution of these engines to the nationwide emission inventory with and without the
standards. Chapter 7 compares the costs and the emission reductions for an estimate of the cost-
effectiveness of the rulemaking. Chapter 8 presents our Final Regulatory Flexibility Analysis, as
called for in the Regulatory Flexibility Act. Chapters 9 and 10 describe the societal costs and
benefits of the rulemaking. Chapter 11 presents a range of regulatory alternative we considered
in developing the final rule.
There are three sets of engines and vehicles covered by the new standards. The following
paragraphs describe the different types of engines and vehicles and the standards that apply.
Emission Standards
Large industrial spark-ignition engines
These are spark-ignition nonroad engines rated over 19 kW used in commercial applications.
These include engines used in forklifts, electric generators, airport ground service equipment, and
a variety of other construction, farm, and industrial equipment. Many Large SI engines, such as
those used in farm and construction equipment, are operated outdoors, predominantly during
warmer weather and often in or near heavily populated urban areas where they contribute to
ozone formation and ambient CO and PM levels. These engines are also often operated in
factories, warehouses, and large retail outlets throughout the year, where they contribute to high
exposure levels to personnel who work with or near this equipment as well as to ozone formation
and ambient CO and PM levels. In this rulemaking, we call these "Large SI" engines.
-------
Draft Regulatory Support Document
We are adopting two tiers of emission standards for Large SI engines. The first tier,
scheduled to start in 2004, sets standards of 4 g/kW-hr (3 g/hp-hr) for HC+NOx and 50 g/kW-hr
(37 g/hp-hr) for CO. These standards are the same as those adopted earlier by the California Air
Resources Board.
Starting in 2007, the Tier 2 emission standards fall to 2.7 g/kW-hr (2.0 g/hp-hr) for
HC+NOx emissions and 4.4 g/kW-hr (3.3 g/hp-hr) for CO emissions. However, we are
including an option for manufacturers to certify their engines to different emission levels to
reflect the inherent tradeoff of NOx and CO emissions and to add an incentive for HC+NOx
emission reductions below the standard. Generally this involves meeting a less stringent CO
standard if a manufacturer certifies an engine with lower HC+NOx emissions. Table 1 shows
several examples of possible combinations of HC+NOx and CO emission standards. The highest
allowable CO standard for duty-cycle testing is 20.6 g/kW-hr (15.4 g/hp-hr), which corresponds
with HC+NOx emissions below 0.8 g/kW-hr (0.6 g/hp-hr).
Table 1
Samples of Possible Alternative
Emission Standards for Large SI Engines(g/kW-hr)*
Duty-cycle testing
Field testing
HC+NOx
2.70
2.20
1.70
1.30
1.00
0.80
3.80
3.10
2.40
1.80
1.40
1.10
CO
4.4
5.6
7.9
11.1
15.5
20.6
6.5
8.5
11.7
16.8
23.1
31.0
*As described in the Final Regulatory Support Document and the regulations, the values in the table are related by
the following formula: (HC+NOx) x CO0 ™ = 8.57. These values follow directly from the logarithmic relationship
presented with the proposal in the Draft Regulatory Impact Analysis. The analogous formula for field-testing
standards is (HC+NOx) x CO0791 = 16.78.
In addition, Tier 2 engines must have engine diagnostic capabilities that alert the operator to
11
-------
Executive Summary
malfunctions in the engine's emission-control system. Gasoline-fueled Tier 2 engines will also
be required to reduce evaporative emissions. The field-testing procedures and standards in this
final rule make it possible for the manufacturer to easily test engines to meet the requirements of
the in-use testing program for showing that engines undergoing several years of normal operation
in the field continue to meet emission standards.
Nonroad recreational engines and vehicles
These are spark-ignition nonroad engines used primarily in recreational applications. These
include off-highway motorcycles, all-terrain-vehicles (ATVs), and snowmobiles. Some of these
engines, particularly those used on ATVs, are increasingly used for commercial purposes within
urban areas, especially for hauling loads and other utility purposes. These vehicles are typically
used in suburban and rural areas, where they can contribute to ozone formation and ambient CO
and PM levels. They can also contribute to regional haze problems in our national and state
parks. Tables 2 and 3 show the exhaust and permeation emission standards that apply to
recreational vehicles.
Table 2
Recreational Vehicle Exhaust Emission Standards
Vehicle
Model Year
Snowmobile
2006
2007 through 2009
2010
2012*
Off-highway
Motorcycle
ATV
2006
2007 and later
2006
2007 and later
Emission standards
HC
g/kW-hr
100
100
75
75
CO
g/kW-hr
275
275
275
200
HC+NOx
g/km
2.0
2.0
1.5
1.5
CO
g/km
25.0
25.0
35.0
35.0
Phase-in
50%
100%
50%
100%
50%
100
* or equivalent per Section 1051.103; the long term program includes a
provision which acts to cap NOx emission rates
in
-------
Draft Regulatory Support Document
Table 3
Permeation Standards for Recreational Vehicles
Emission Component
Fuel Tank Permeation
Hose Permeation
Implementation Date
2008
2008
Standard
1.5 g/nf/day
15 g/nf/day
Test Temperature
28°C (82°F)
23 °C (73 °F)
Recreational marine diesel engines
These are marine diesel engines used on recreational vessels such as yachts, cruisers, and
other types of pleasure craft. Recreational marine engines are primarily used in warm weather
and therefore contribute to ozone formation and PM levels, especially in marinas, which are
often located in nonattainment areas.
Table 4
Recreational Marine Diesel Emission Limits and Implementation Dates
Displacement
[liters per cylinder]
power > 37 kW
0.5 < disp < 0.9
0.9 < disp<1.2
1.2
-------
Executive Summary
Table 5
2020 HC and NOx Projected Emissions Inventories (thousand short tons)
Category
Industrial SI >19kW
Snowmobiles
ATVs
Off-highway motorcycles
Recreational Marine diesel
Total
Exhaust HC*
base case
318
358
374
232
2.0
1,284
with
standards
34
149
53
117
1.5
355
percent
reduction
89
58
86
50
28
72
Exhaust NOx
base case
472
5
8
1.3
61
547
with
standards
43
10
6
1.5
48
109
percent
reduction
91
(101)
25
(19)
21
80
* The estimate for Industrial SI >19kW includes both exhaust and evaporative emissions. The estimates for
snowmobiles, ATVs and Off-highway motorcycles includes both exhaust and permeation emissions.
Table 6
2020 Projected CO and PM Emissions Inventories (thousand short tons)
Category
Industrial SI >19kW
Snowmobiles
ATVs
Off-highway motorcycles
Recreational Marine diesel
Total
base case
2,336
950
1,250
321
9
4,866
Exhaust CO
with
standards
277
508
1,085
236
9
2,115
percent
reduction
88
46
13
26
0
56
base case
2.3
8.4
13.1
8.7
1.6
34.2
Exhaust PM
with
standards
2.3
4.9
1.9
4.4
1.3
14.8
percent
reduction
0
42
86
50
18
57
Table 7 summarizes the projected costs to meet the emission standards. This is our best
of the cost associated with adopting new technologies to meet the emission standards.
ysis also considers total operating costs, including maintenance and fuel consumption.
estimate or the cost associated with adopting new technologies to meet the emission standan
The analysis also considers total operating costs, including maintenance and fuel consumptk
In many cases, the fuel savings from new technology are greater than the cost to upgrade the
engines. All costs are presented in 2001 dollars.
-------
Draft Regulatory Support Document
Table 7
Estimated Average Cost Impacts of Emission Standards
Standards
Large SI exhaust
Large SI exhaust
Large SI evaporative
Snowmobile exhaust
Snowmobile exhaust
Snowmobile exhaust
Snowmobile permeation
ATV exhaust
ATV permeation
Off -highway motorcycle exhaust
Off -highway motorcycle permeation
Recreational marine diesel
Dates
2004
2007
2007
2006
2010
2012
2008
2006
2008
2006
2008
2006
Increased Production
Cost per Vehicle*
$611
$55
$13
$73
$131
$89
$7
$84
$3
$155
$3
$346
Lifetime Operating Costs
per Vehicle (NPV)
$-3,981
$0
$-56
$-57
$-286
$-191
$-11
$-24
$-6
$-48
$-5
—
*The estimated long-term costs decrease by about 35 percent. Costs presented for the Large SI and snowmobile second-
phase standards are incremental to the first-phase standards.
We also calculated the cost per ton of emission reductions for the standards. For
snowmobiles, this calculation is on the basis of HC plus NOx emissions and CO emissions. For
all other engines, we attributed the entire cost of the program to the control of ozone precursor
emissions (HC or NOx or both). A separate calculation could apply to reduced CO or PM
emissions in some cases. Assigning the full compliance costs to a narrow emissions basis leads
to cost-per-ton values that underestimate of the value of the program.
Table 8 presents the discounted cost-per-ton estimates for the various engine categories and
standards being adopted. Reduced operating costs more than offset the increased cost of
producing the cleaner engines for Large SI and snowmobile engines. The overall fuel savings
associated with the standards being adopted are greater than the total projected costs to comply
with the emission standards.
VI
-------
Executive Summary
Table 8
Estimated Cost-per-Ton of Emission Standards
Standards
Large SI exhaust (Composite of all
fuels)
Large SI exhaust (Composite of all
fuels)
Large SI evaporative
Snowmobile exhaust
Snowmobile exhaust
Snowmobile exhaust
Snowmobile permeation
ATV exhaust
ATV permeation
Off-highway motorcycle exhaust
Off-highway motorcycle permeation
Recreational marine diesel
Aggregate
Dates
2004
2007
2007
2006
2010
2012
2008
2006
2008
2006
2008
2006
—
Discounted
Reductions
per Vehicle
(short tons)*
3.07
0.80
0.13
HC: 0.40
CO: 1.02
HC: 0.10
CO: 0.25
0.03
0.21
0.02
0.38
0.01
0.44
—
Discounted Cost per Ton
ofHC+NOx
Without
Fuel Savings
$240
$80
$80
$90
$1,370
—
$210
$400
$180
$410
$230
$670
$240
With
Fuel Savings
($1,150)
$80
($280)
$20
$0
—
($150)
$290
($180)
$280
($140)
$670
($280)
Discounted Cost per Ton
of CO
Without
Fuel Savings
—
—
—
$40
—
$360
—
—
—
—
—
—
$80
With
Fuel Savings
—
—
—
$10
—
$0
—
—
—
—
—
—
($20)
* HC reductions for evaporative and permeation, and HC+NOx reductions for exhaust (except snowmobiles where CO
reductions are also presented).
Economic Impact Analysis
We performed an analysis to estimate the economic impacts of this final rule on producers
and consumers of recreational marine diesel vessels (specifically, diesel inboard cruisers),
forklifts, snowmobiles, ATVs, off-highway motorcycles, and society as a whole. This economic
impact analysis focuses on market-level changes in price, quantity, and economic welfare (social
gains or costs) associated with the regulation. A description of the methodology used can be
found in Chapter 9 of this document.
We did not perform an economic impact analysis for categories of Large SI nonroad engines
other than forklifts, even though those other Large SI engines are also subject to the standards
contained in this final rule. This was due to the large number of different types of equipment that
use Large SI engines and data availability constraints for those market segments. For the sake of
completeness, the following analysis reports separate estimates for Large SI engines other than
vn
-------
Draft Regulatory Support Document
forklifts. Engineering costs are assumed to be equal to economic costs for those engines. This
approach slightly overestimates the social costs associated with the relevant standards.
Based on the estimated regulatory costs associated with this rule and the predicted changes
in prices and quantity produced in the affected industries, the total estimated annual social gains
of the rule in the year 2030 is projected to be $553.3 million (in 2000 and 2001 dollars). The net
present value of the social gains for the 2002 to 2030 time frame is equal to $4.9 billion. The
social gains are equal to the fuel savings minus the combined loss in consumer and producer
surplus (see Table 9), taking into account producers' and consumers' changes in behavior
resulting from the costs associated with the rule.1 Social gains do not account for the social
benefits (the monetized health and environmental effects of the rule).
Table 9
Surplus Losses, Fuel Efficiency Gains, and Social Gains/Costs in 203Oa
Vehicle Category
Recreational marine diesel
vessels
Forklifts
Other Large SP
Snowmobiles
ATVs
Off-highway motorcycles
All vehicles total
NPV of all vehicles totald
Surplus Losses in
2030 ($millions)
$6.6
$47.8
$48.1
$41.9
$47.2
$25.0
$216.6
$3,231.4
Fuel Efficiency Gains in
2030 ($millions)
$0
$420.1
$138.4
$135.0
$51.4
$25.2
$770.1
$8,130.3
Social Gains/Costs
in 2030b ($millions)
($6.6)
$372.3
$90.3
$93.1
$4.2
$0.2
$553.3
$4,898.9
a Figures are in 2000 and 2001 dollars.
b Figures in this column exclude estimated social benefits. Numbers in parentheses denote social costs.
0 Figure is engineering costs; see Section 9.7.6 of Chapter 9 for explanation.
d Net Present Value is calculated over the 2002 to 2030 time frame using a 3 percent discount rate.
For most of the engine categories contained in this rule, we expect there will be a fuel
savings as manufacturers redesign their engines to comply with emission standards. For ATVs
and off-highway motorcycles, the fuel savings will be realized as manufacturers switch from
Consumer and producer surplus losses are measures of the economic welfare loss
consumers and producers, respectively are likely to experience as a result of the regulations.
Combined these losses represent an estimate of the economic or social costs of the rule. Note
that for the Large SI and recreational vehicle rules, fuel efficiency gains must be netted from
surplus losses to estimate the social costs or social gains (in cases where fuel efficiency gains
exceed surplus losses) attributable to the rules.
Vlll
-------
Executive Summary
two-stroke to four-stroke technologies. For snowmobiles, the fuel savings will be realized as
manufacturers switch some of their engines to more fuel efficient two-stroke technologies and
some of their engines to four-stroke technologies. For Large SI engines, the fuel savings will be
realized as manufacturers adopt more sophisticated and more efficient fuel systems; this is true
for all fuels used by Large SI engines. Overall, we project the fuel savings associated with the
anticipated changes in technology to be about 800 million gallons per year once the program is
fully phased in. These savings are factored into the calculated costs and costs per ton of reduced
emissions, as described above.
IX
-------
Draft Regulatory Support Document
10
-------
Chapter 1: Health and Welfare Concerns
Chapter 1: Health and Welfare Concerns
The engines and vehicles that would be subject to the standards in this final rule generate
emissions of HC, NOx, CO, PM and air toxics. They contribute to ozone and CO nonattainment
and to adverse health effects associated with ambient concentrations of PM and air toxics. They
also contribute to visibility impairment in Class I areas and in other areas where people live,
work, and recreate. This chapter presents our estimates of the contribution these engines make to
our national air inventory. We include in this chapter estimates of pre- and post-control
contributions. These estimates are described in greater detail in Chapter 6.
This chapter also describes the health and environmental effects related to these emissions.
These pollutants cause a range of adverse health and welfare effects, especially in terms of
respiratory impairment and related illnesses and visibility impairment both in Class I areas and in
areas where people live, work and recreate. Air quality modeling and monitoring data presented
in this chapter indicate that a large number of our citizens continue to be affected by these
emissions.
1.1 Inventory Contributions
1.1.1 Inventory Contribution
The contribution of emissions from the nonroad engines and vehicles that would be subject
to the standards to the national inventories of pollutants that are associated with the health and
public welfare effects described in this chapter are considerable. To estimate nonroad engine and
vehicle emission contributions, we used the latest version of our NONROAD emissions model.
This model computes nationwide, state, and county emission levels for a wide variety of nonroad
engines, and uses information on emission rates, operating data, and population to determine
annual emission levels of various pollutants. A more detailed description of the model and our
estimation methodology can be found in the Chapter 6 of this document.
Baseline emission inventory estimates for the year 2000 for the categories of engines and
vehicles covered by this rulemaking are summarized in Table 1.1-1. This table show the relative
contributions of the different mobile-source categories to the overall national mobile-source
inventory. Of the total emissions from mobile sources, the categories of engines and vehicles
covered by this rulemaking contribute about 9 percent, 3 percent, 4 percent, and 2 percent of HC,
NOx, CO, and PM emissions, respectively, in the year 2000. The results for large SI engines
indicate they contribute approximately 2 to 3 percent to HC, NOx, and CO emissions from
mobile sources. The results for land-based recreational engines reflect the impact of the
significantly different emissions characteristics of two-stroke engines. These engines are
estimated to contribute about 6 percent of HC emissions and 2 percent of CO from mobile
sources. Recreational CI marine contribute less than 1 percent to NOx mobile source
inventories. When only nonroad emissions are considered, the engines and vehicles that would
1-1
-------
Draft Regulatory Support Document
be subject to the standards would account for a larger share.
Our emission projections for 2020 and 2030 for the nonroad engines and vehicles subject to
this rulemaking show that emissions from these categories are expected to increase over time if
left uncontrolled. The projections for 2020 and 2030 are summarized in Tables 1.1-2 and 1.1-3,
respectively. The projections for 2020 and 2030 indicate that the categories of engines and
vehicles covered by this rulemaking are expected to contribute approximately 25 percent, 10
percent, 5 percent, and 5 percent of HC, NOx, CO, and PM emissions, respectively. Population
growth and the effects of other regulatory control programs are factored into these projections.
The relative importance of uncontrolled nonroad engines is higher than the projections for 2000
because there are already emission control programs in place for the other categories of mobile
sources which are expected to reduce their emission levels. The effectiveness of all control
programs is offset by the anticipated growth in engine populations.
1-2
-------
Chapter 1: Health and Welfare Concerns
Table 1.1-1
Modeled Annual Emission Levels for
Mobile-Source Categories in 2000 (thousand short tons)
Category
Total for engines subject to
today's standards*
Highway Motorcycles
Nonroad Industrial SI > 19 kW*
Recreational SI*
Recreational Marine CI*
Marine SI Evap
Marine SI Exhaust
Nonroad SI< 19 kW
Nonroad CI
Commercial Marine CI
Locomotive
Total Nonroad
Total Highway
Aircraft
Total Mobile Sources
Total Man-Made Sources
Mobile Source percent of Total
Man-Made Sources
NOx
1000
tons
351
8
308
5
38
0
32
106
2,625
963
1,192
5,269
7,981
178
13,428
24,532
55%
percent
of mobile
source
2.6%
0.1%
2.3%
0.0%
0.3%
0.0%
0.2%
0.8%
19.5%
7.2%
8.9%
39%
59%
1%
100%
-
-
HC
1000
tons
645
84
226
418
1
100
708
1,460
316
30
47
3,305
3,811
183
7,300
18,246
40%
percent of
mobile
source
8.8%
1.2%
3.1%
5.7%
0.0%
1.4%
9.7%
20.0%
4.3%
0.4%
0.6%
45%
52%
3%
100%
-
-
CO
1000
tons
2,860
331
1,734
1,120
6
0
2,144
18,359
1,217
127
119
24,826
49,813
1,017
75,656
97,735
77%
percent of
mobile
source
3.8%
0.4%
2.3%
1.5%
0.0%
0.0%
2.8%
24.3%
1.6%
0.2%
0.2%
33%
66%
1%
100%
-
-
PM
1000
tons
14.6
0.4
1.6
12.0
1
0
38
50
253
41
30
427
240
39
706
3,102
23%
percent
of
mobile
source
2.1%
0.1%
0.2%
1.7%
0.1%
0.0%
5.4%
7.1%
35.9%
5.8%
4.3%
60%
34%
6%
100%
-
-
1-3
-------
Draft Regulatory Support Document
Table 1.1-2
Modeled Annual Emission Levels for
Mobile-Source Categories in 2020 (thousand short tons)
Category
Total for engines subject to
today's standards*
Highway Motorcycles
Nonroad Industrial SI > 19 kW*
Recreational SI*
Recreational Marine CI*
Marine SI Evap
Marine SI Exhaust
Nonroad SI< 19 kW
Nonroad CI
Commercial Marine CI
Locomotive
Total Nonroad
Total Highway
Aircraft
Total Mobile Sources
Total Man-Made Sources
Mobile Source percent of Total
Man-Made Sources
NOx
1000
tons
547
14
472
14
61
0
58
106
1,791
819
611
3,932
2,050
232
6,214
16,190
38%
percent
of mobile
source
8.8%
0.2%
7.6%
0.2%
1.0%
0.0%
0.9%
1.7%
28.8%
13.2%
9.8%
63%
33%
4%
100%
-
-
HC
1000
tons
1,305
142
318
985
2
114
284
986
142
35
35
2,901
2,276
238
5,415
15,475
35%
percent of
mobile
source
24.1%
2.6%
5.9%
18.2%
0.0%
2. 1%
5.2%
18.2%
2.6%
0.6%
0.6%
54%
42%
4%
100%
-
-
CO
1000
tons
4,866
572
2,336
2,521
9
0
1,985
27,352
1,462
160
119
35,944
48,906
1,387
86,237
109,905
79%
percent of
mobile
source
5.6%
0.7%
2.7%
2.9%
0.0%
0.0%
2.3%
31.7%
1.7%
0.2%
0.1%
42%
56%
2%
100%
-
-
PM
1000
tons
34.1
0.8
2.3
30.2
1.6
0
28
77
261
46
21
467
145
43
655
3,039
22%
percent
of
mobile
source
5.2%
0. 1%
0.4%
4.6%
0.2%
0.0%
4.3%
11.8%
40.0%
7.0%
3.2%
71%
22%
7%
100%
-
-
1-4
-------
Chapter 1: Health and Welfare Concerns
Table 1.1-3
Modeled Annual Emission Levels for
Mobile-Source Categories in 2030 (thousand short tons)
Category
Total for engines subject to
today's standards*
Highway Motorcycles
Nonroad Industrial SI > 19 kW*
Recreational SI*
Recreational Marine CI*
Marine SI Evap
Marine SI Exhaust
Nonroad SI< 19 kW
Nonroad CI
Commercial Marine CI
Locomotive
Total Nonroad
Total Highway
Aircraft
Total Mobile Sources
Total Man-Made Sources
Mobile Source percent of Total
Man-Made Sources
NOx
1000
tons
640
17
553
15
72
0
64
126
1,994
1,166
531
4,521
1,648
262
6,431
16,639
39%
percent
of mobile
source
10.0%
0.3%
8.6%
0.2%
1.1%
0.0%
1.0%
2.0%
31.0%
18.1%
8.3%
70%
26%
4%
100%
—
—
HC
1000
tons
1,411
172
371
1,038
2
122
269
1,200
158
52
30
3,242
2,496
262
6,000
17,020
35%
percent of
mobile
source
23.5%
2.9%
6.2%
17.3%
0.0%
2.0%
4.5%
20.0%
2.6%
0.9%
0.5%
54%
42%
4%
100%
—
—
CO
1000
tons
5,363
693
2,703
2,649
11
0
2,083
32,310
1,727
198
119
41,800
56,303
1,502
99,605
123,983
80%
percent of
mobile
source
5.4%
0.7%
2.7%
2.7%
0.0%
0.0%
2.1%
32.4%
1.7%
0.2%
0.1%
42%
56%
2%
100%
—
—
PM
1000
tons
36.5
1.0
2.7
31.9
1.9
0
29
93
306
74
18
557
158
43
758
3,319
23%
percent
of
mobile
source
4.8%
0. 1%
0.4%
4.2%
0.3%
0.0%
3.8%
12.3%
40.4%
9.8%
2.4%
74%
21%
6%
100%
—
-
1.1.2 Baseline Inventory Adjustment
Since we proposed this regulatory program, we revised our baseline inventories for the
covered engines to reflect information we received during the comment period. These inventory
adjustments are discussed in more detail in Chapter 6, and the changes are reflected in the tables
above.
We also revised our national mobile source on-highway and nonroad inventories to reflect
additional information and to incorporate routine updates since we finalized our On-Highway
Heavy-Duty Engine/Diesel Fuel (HD07) rule. The inventory adjustments to our on-highway and
nonroad inventories are of particular importance because the health and visibility results reported
1-5
-------
Draft Regulatory Support Document
in the following sections of this chapter are based on the earlier national mobile source baselines
that were used as inputs to the air quality model. We did not perform new health effects and
visibility modeling for this rule; instead, we relied on the ozone and PM modeling performed for
the HD07 rule. Because our estimates of baseline national mobile source inventories have
increased since the HD07 rule, relying on the earlier inventories would underestimate future PM
levels that we would expect if we conducted new modeling with the revised inventory inputs.
Thus, the health effects and visibility information would underestimate the size of populations
living in counties with air quality above certain levels compared to new modeling.
Table 1.1-4 contains a summary of the changes to the on-highway and nonroad inventories
since the HD07 rule, and reports the percent change in the inventory for each pollutant. This
table shows that the HD07 inventories used in the health and visibility modeling underestimate
2020 direct PM emissions by 0.3 percent for highway engines and 9.4 percent for nonroad
engines. The HD07 inventories underestimate 2030 direct PM emissions by 0.1 percent for on-
highway and 11.9 percent for nonroad engines. HC and NOx emissions could also affect
predicted ambient PM concentrations via secondary formation in the atmosphere.
While the health effects and visibility analyses in the following section may thus
underestimate the extent of health effects and visibility impairment we would predict if we were
to model the information with our updated inventories, the HD07 analysis still supports our
determination that these engines cause or contribute to such health and welfare concerns.
1-6
-------
Chapter 1: Health and Welfare Concerns
Table 1.1-4
Comparison of Inventory Projections to Projections Used for Air Quality Modeling
in the 2007 Highway Heavy-Duty Engine/Diesel Fuel Rule (thousand short tons)
Category
2020 Highway
2020 Nonroad
(including
aircraft)
2030 Highway
2030 Nonroad
(including
aircraft)
Comparison
HD07 Modeling Inventories
Current Estimates
Difference
Difference as a percent of total
mobile inventory
HD07 Modeling Inventories
Current Estimates
Difference
Difference as a percent of total
mobile inventory
HD07 Modeling Inventories
Current Estimates
Difference
Difference as a percent of total
mobile inventory
HD07 Modeling Inventories
Current Estimates
Difference
Difference as a percent of total
mobile inventory
NOx
2,022
2,050
28
0.5%
4,040
4,164
124
2.0%
2,181
2,496
315
4.9%
2,228
3,504
1,276
19.8%
HC
2,019
2,276
257
4.7%
1,995
3,139
1,144
21.1%
1,624
1,648
24
0.4%
4,325
4,783
458
7.6%
CO
48,334
48,906
572
0.7%
33,938
37,331
3,393
3.9%
55,610
56,303
693
0.7%
39,223
43,302
4,079
4.1%
Direct PM
143
145
2
0.3%
449
510
61
9.4%
157
158
1
0.1%
509
600
91
11.9%
1.1.2 Inventory Impacts on a Per Vehicle Basis
In addition to the general inventory contributions described above, the engines that would be
subject to the standards are more potent polluters than their highway counterparts in that they
have much higher emissions on a per vehicle basis. This is illustrated in Table 1.1-5, which
equates the emissions produced in one hour of operation from the different categories of
equipment covered by the rulemaking to the equivalent miles of operation it would take for a car
produced today to emit the same amount of emissions.
1-7
-------
Draft Regulatory Support Document
Table 1.1-5
Per-Vehicle Emissions Comparison
Equipment Category
Recreational Marine CI
Large SI
Snowmobiles
Snowmobiles
2-Stroke ATVs
4-Stroke ATVs
2-Stroke off-road motorcycles
4-Stroke off-road motorcycles
Emission Comparison
HC+NOx
HC+NOx
HC
CO
HC
HC
HC
HC
Miles a Current Passenger Car
Would Need to Drive to Emit the
Same Amount of Pollution as the
Equipment Category Emits in One
Hour of Operation
2,400
1,340
24,300
1,520
6,470
290
9,580
430
The per engine emissions are important because they mean that operators of these engines
and vehicles, as well as those who work in their vicinity, are exposed to high levels of emissions,
many of which are air toxics. These effects are of particular concern for people who operate
forklifts in enclosed areas and for snowmobile riders following a lead rider. These effects are
described in more detail in the next sections.
1.2 Ozone
1.2.1 General Background
Ground-level ozone, the main ingredient in smog, is formed by complex chemical reactions
of volatile organic compounds (VOC) and NOx in the presence of heat and sunlight. Ozone
forms readily in the lower atmosphere, usually during hot summer weather. Volatile organic
compounds are emitted from a variety of sources, including motor vehicles, chemical plants,
refineries, factories, consumer and commercial products, and other industrial sources. Volatile
organic compounds also are emitted by natural sources such as vegetation. Oxides of nitrogen
are emitted largely from motor vehicles, off-highway equipment, power plants, and other sources
of combustion. Hydrocarbons (HC) are a large subset of VOC, and to reduce mobile source
VOC levels we set maximum emissions limits for hydrocarbon as well as particulate matter
emissions.
The science of ozone formation, transport, and accumulation is complex. Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions involving NOx, VOC,
-------
Chapter 1: Health and Welfare Concerns
heat, and sunlight.1 As a result, differences in weather patterns, as well as NOx and VOC levels,
contribute to daily, seasonal, and yearly differences in ozone concentrations and differences from
city to city. Many of the chemical reactions that are part of the ozone-forming cycle are sensitive
to temperature and sunlight. When ambient temperatures and sunlight levels remain high for
several days and the air is relatively stagnant, ozone and its precursors can build up, resulting in
higher ambient ozone levels than typically would occur on a single high temperature day.
Further complicating matters, ozone also can be transported into an area from pollution sources
found hundreds of miles upwind, resulting in elevated ozone levels even in areas with low local
VOC or NOx emissions.
On the chemical level, NOx and VOC are the principal precursors to ozone formation. 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.
When NOx levels are relatively high and VOC levels relatively low, NOx forms inorganic
nitrates but relatively little ozone. Such conditions are called "VOC limited." Under these
conditions, VOC reductions are effective in reducing ozone, but NOx reductions can actually
increase local ozone under certain circumstances. Even in VOC limited urban areas, NOx
reductions are not expected to increase ozone levels if the NOx reductions are sufficiently large.
Rural areas are almost always NOx limited, due to the relatively large amounts of biogenic
VOC emissions in such areas. Urban areas can be either VOC or NOx limited, or a mixture of
both.
Ozone concentrations in an area also can be lowered by the reaction of nitric oxide with
ozone, forming nitrogen dioxide (NO2); as the air moves downwind and the cycle continues, the
NO2 forms additional ozone. The importance of this reaction depends, in part, on the relative
concentrations of NOx, VOC, and ozone, all of which change with time and location.
1.2.2 Health and Welfare Effects of Ozone and Its Precursors
Based on a large number of recent studies, EPA has identified several key health effects
caused when people are exposed to levels of ozone found today in many areas of the country.2'3
Short-term exposures (1-3 hours) to high ambient ozone concentrations have been linked to
increased hospital admissions and emergency room visits for respiratory problems. For example,
studies conducted in the northeastern U.S. and Canada show that ozone air pollution is associated
with 10-20 percent of all of the summertime respiratory-related hospital admissions. Repeated
exposure to ozone can make people more susceptible to respiratory infection and lung
1-9
-------
Draft Regulatory Support Document
inflammation and can aggravate preexisting respiratory diseases, such as asthma. Prolonged (6
to 8 hours), repeated exposure to ozone can cause inflammation of the lung, impairment of lung
defense mechanisms, and possibly irreversible changes in lung structure, which over time could
lead to premature aging of the lungs and/or chronic respiratory illnesses such as emphysema and
chronic bronchitis.
Children and outdoor workers are most at risk from ozone exposure because they typically
are active outside during the summer when ozone levels are highest. For example, summer camp
studies in the eastern U.S. and southeastern Canada have reported significant reductions in lung
function in children who are active outdoors. Further, children are more at risk than adults from
ozone exposure because their respiratory systems are still developing. Adults who are outdoors
and are moderately active during the summer months, such as construction workers and other
outdoor workers, also are among those most at risk. These individuals, as well as people with
respiratory illnesses such as asthma, especially asthmatic children, can experience reduced lung
function and increased respiratory symptoms, such as chest pain and cough, when exposed to
relatively low ozone levels during prolonged periods of moderate exertion.
Evidence also exists of a possible relationship between daily increases in ozone levels and
increases in daily mortality levels. While the magnitude of this relationship is too uncertain to
allow for direct quantification, the full body of evidence indicates the possibility of a positive
relationship between ozone exposure and premature mortality.
In addition to human health effects, ozone adversely affects crop yield, vegetation and forest
growth, and the durability of materials. Because ground-level ozone interferes with the ability of
a plant to produce and store food, plants become more susceptible to disease, insect attack, harsh
weather and other environmental stresses. Ozone causes noticeable foliage damage in many
crops, trees, and ornamental plants (i.e., grass, flowers, shrubs) and causes reduced growth in
plants. Studies indicate that current ambient levels of ozone are responsible for damage to
forests and ecosystems (including habitat for native animal species). Ozone chemically attacks
elastomers (natural rubber and certain synthetic polymers), textile fibers and dyes, and, to a lesser
extent, paints. For example, elastomers become brittle and crack, and dyes fade after exposure to
ozone.
Volatile organic compounds emissions are detrimental not only for their role in forming
ozone, but also for their role as air toxics. Some VOCs emitted from motor vehicles are toxic
compounds. At elevated concentrations and exposures, human health effects from air toxics can
range from respiratory effects to cancer. Other health impacts include neurological
developmental and reproductive effects. The lexicologically significant VOCs emitted in
substantial quantities from the engines that are the subject of this rule are discussed in more
detail in Section 1.6, below.
1-10
-------
Chapter 1: Health and Welfare Concerns
1.2.3 Ozone Nonattainment and Contribution to Ozone Nonattainment
The current primary and secondary ozone National Ambient Air Quality Standard
(NAAQS) is 0.12 ppm daily maximum 1-hour concentration, not to be exceeded more than once
per year on average. The determination that an area is at risk of exceeding the ozone standard in
the future was made for all areas with current design values grater than or equal to 0.125 ppm (or
within a 10 percent margin) and with modeling evidence that exceedances will persist into the
future.
Ground level ozone today remains a pervasive pollution problem in the United States. In
1999, 90.8 million people (1990 census) lived in 31 areas designated nonattainment under the 1-
hour ozone NAAQS.4 This sharp decline from the 101 nonattainment areas originally identified
under the Clean Air Act Amendments of 1990 demonstrates the effectiveness of the last decade's
worth of emission-control programs. However, elevated ozone concentrations remain a serious
public health concern throughout the nation.
Over the last decade, declines in ozone levels were found mostly in urban areas, where
emissions are heavily influenced by controls on mobile sources and their fuels. Twenty-three
metropolitan areas have realized a decline in ozone levels since 1989, but at the same time ozone
levels in 11 metropolitan areas with 7 million people have increased.5 Regionally, California and
the Northeast have recorded significant reductions in peak ozone levels, while four other regions
(the Mid-Atlantic, the Southeast, the Central and Pacific Northwest) have seen ozone levels
increase.
The highest ambient concentrations are currently found in suburban areas, consistent with
downwind transport of emissions from urban centers. Concentrations in rural areas have risen to
the levels previously found only in cities. Particularly relevant to this rulemaking, ozone levels
at 17 of our National Parks have increased, and in 1998, ozone levels in two parks, Shenandoah
National Park and the Great Smoky Mountains National Park, were 30 to 40 percent higher than
the ozone NAAQS over the last decade.6
To estimate future ozone levels, we refer to the modeling performed in conjunction with the
final HD07 rule.7 We performed a series of ozone air quality modeling simulations for nearly the
entire Eastern U.S. covering metropolitan areas from Texas to the Northeast.8 This ozone air
quality model was based upon the same modeling system as was used in the Tier 2 passenger
vehicle air quality analysis,9 with the addition of enhanced inventory estimates for 2007 and 2030
based on the state of knowledge at the time the modeling was performed. Emissions from
nonroad engines, including the engines subject to this final rule, were included as input to the air
quality modeling we describe in this section (as shown in Tables 1.1-2 to 1.1-4 above).
The model simulations were performed for several emission scenarios, and the model
outputs were combined with current air quality data to identify areas expected to exceed the
ozone NAAQS in 2007, 2020, and 2030.10 The results of this modeling are contained in Table
1.2-1. Areas presented in Table 1.2-1 exhibit 1997-99 air quality data indicating violations of the
1-11
-------
Draft Regulatory Support Document
1-hour ozone NAAQS, or are within 10 percent of the standard, are predicted to have exceedance
in 2007, 2020, or 2030. An area was considered likely to have future exceedances if exceedances
were predicted by the model, and the area is currently violating the 1-hour standard, or is within
10 percent of violating the 1-hour standard. Table 1.2-1 shows that 37 areas with a 1999
population of 91 million people are at risk of exceeding the 1-hour ozone standard in 2007.
These estimates include contributions from the engines subject to this rule.2
Additional information about the Regulatory Model System for Aerosols and Deposition
(REMSAD) and our modeling protocols can be found in our Regulatory Impact Analysis: Heavy-
Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements,
document EPA420-R-00-026, December 2000. Docket No. A-2000-01, Document No. A-II-13.
This document is also available at http://www.epa.gov/otaq/disel.htm#documents.
1-12
-------
Chapter 1: Health and Welfare Concerns
Table 1.2-1
Eastern Metropolitan Areas with Modeled Exceedances of the 1-Hour
2007, 2020, or 2030 (Includes all national emission controls through
Ozone Standard in
HD07 standards)
MSA or CMSA / State
Atlanta, GA MSA
Barnstable-Yarmouth, MA MSA*
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA*
Biloxi-Gulrport-Pascagoula, MS MSA*
Birmingham, AL MSA
Boston- Worcester-Lawrence, MA CMSA
Charleston, WV MSA*
Charlotte-Gastonia-Rock Hill, NC MSA
Chicago-Gary-Kenosha, IL CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA*
Cleveland-Akron, OH CMSA*
Detroit- Ann Arbor-Flint, MI CMSA
Grand Rapids-Muskegon-Holland, MI MSA*
Hartford, CT MSA
Houma, LA MSA*
Houston-Galveston-Brazoria, TX CMSA
Huntington- Ashland, WV-KY-OH MSA
Lake Charles, LA MSA*
Louisville, KY-IN MSA
Macon, GA MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
New London-Norwich, CT-RI MSA
New Orleans, LA MSA*
New York-Northern NJ-Long Island, NY-NJ-CT-PA
CMSA
Norfolk- Virginia Beach-Newport News, VA-NC MSA*
Orlando, FL MSA*
Pensacola, FL MSA
Philadelphia- Wilmington-Atlantic City, PA-NJ-DE-MD
CMSA
Providence-Fall River- Warwick,RI-MAMSA*
Richmond-Petersburg, VA MSA
St. Louis, MO-IL MSA
Tampa-St. Petersburg, FL MSA*
Washington-Baltimore
Total number of areas
Population
2007
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
37
91 2
2020
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
32
885
2030
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
32
878
pop (1999)
3.9
0.2
0.6
0.4
0.2
0.3
0.9
5.7
0.3
1.4
8.9
1.9
2.9
5.4
1.1
1.1
0.2
4.5
0.3
0.2
1
0.3
1.1
1.7
1.2
0.3
1.3
20.2
1.6
1.5
0.4
6
1.1
1
2.6
2.3
7.4
91 4
These areas have registered 1997-1999 ozone concentrations within 10 percent of standard.
1-13
-------
Draft Regulatory Support Document
With regard to future ozone levels, our air quality ozone modeling for 2020 predicts
exceedances of the 1-hour ozone standard in 32 areas with a total of 89 million people (1999
census; see Table 1.2-1). We expect that the control strategies contained in this rulemaking will
further assist state efforts already underway to attain and maintain the 1-hour ozone standard.
The inventories that underlie this predictive modeling for 2020 and 2030 include reductions
from all current and committed to federal control programs, including the recently promulgated
NOx and PM standards for heavy-duty vehicles and low sulfur diesel fuel (HD07 rule). The
geographic scope of these areas at risk of future exceedances underscores the need for additional,
nationwide controls of ozone precursors.
It should be noted that this modeling did not attempt to examine the prospect of areas
attaining or maintaining the ozone standard with possible future controls (i.e., controls beyond
current or committed controls). Therefore, this information should be interpreted as indicating
what areas are at risk of ozone violations in 2007, 2020 or 2030 without federal, State, or local
measures that may be adopted and implemented in the future. We expect many of these areas to
adopt additional emission reduction programs, but we are unable to quantify or rely upon future
reductions from additional State or local programs since they have not yet been adopted.
1.2.4 Public Health and Welfare Concerns from Prolonged and Repeated Exposures to
Ozone
In addition to the health effects described above, there exists a large body of scientific
literature that shows that harmful effects can occur from sustained levels of ozone exposure
much lower than 0.125 ppm. Studies of prolonged exposures, those lasting about 7 hours,
showed health effects from exposures to ozone concentrations as low as 0.08 ppm. Prolonged
and repeated exposures to ozone at these levels are common in areas that do not attain the 1-hour
NAAQS, and also occur in areas where ambient concentrations of ozone are in compliance with
the 1-hour NAAQS.
Prolonged exposure to levels of ozone below the NAAQS have been reported to cause or be
statistically associated with transient pulmonary function responses, transient respiratory
symptoms, effects on exercise performance, increased airway responsiveness, increased
susceptibility to respiratory infection, increased hospital and emergency room visits, and transient
pulmonary respiratory inflamation. Such acute health effects have been observed following
prolonged exposures at moderate levels of exertion at concentrations of ozone as low as 0.08
ppm, the lowest concentration tested. The effects are more pronounced as concentrations
increase, affecting more subjects or having a greater effect on a given subject in terms of
functional changes or symptoms. A detailed summary and discussion of the large body of ozone
health effects research may be found in Chapters 6 through 9 (Volume 3) of the 1996 Criteria
Document for ozone.11 Monitoring data indicates that 333 counties in 33 states exceed these
levels in 1997-99.12
1-14
-------
Chapter 1: Health and Welfare Concerns
To provide a quantitative estimate of the projected number of people anticipated to reside in
areas in which ozone concentrations are predicted to exceed the 8-hour level of 0.08 to 0.12 ppm
or higher for multiple days, we performed regional modeling using the variable-grid Urban
Airshed Model (UAM-V) for the HD07 rule.13 UAM-V is a photochemical grid model that
numerically simulates the effects of emissions, advection, diffusion, chemistry, and surface
removal processes on pollutant concentrations within a 3-dimensional grid. As with the previous
modeling analysis, the inventories that underlie this predictive modeling include reductions from
all current and committed to control programs, including the HD07 NOx and PM reductions.
This HD07 ozone modeling forecast that 111 million people are predicted to live in areas
that areas at risk of exceeding these moderate ozone levels for prolonged periods of time in 2020
after accounting for expected inventory reductions due to controls on light- and heavy-duty on-
highway vehicles; that number is expected to increase to 125 million in 2030.14 Prolonged and
repeated ozone concentrations at these levels are common in areas throughout the country. These
concentrations are found both in areas that are exceeding, and areas that are not exceeding, the 1-
hour ozone standard. Areas with these high concentrations are more widespread than those in
nonattainment for that 1-hour ozone standard.
Ozone at these levels can have other welfare effects, with damage to plants and ecosystems
being of most concern. Plant damage affects crop yields, forestry production, and ornamentals.
The adverse effect of ozone on forests and other natural vegetation can in turn cause damage to
associated ecosystems, with additional resulting economic losses. Prolonged ozone
concentrations of 0.10 ppm can be phytotoxic to a large number of plant species, and can produce
acute injury and reduced crop yield and biomass production. Ozone concentrations within the
range of 0.05 to 0.10 ppm have the potential over a longer duration of creating chronic stress on
vegetation that can result in reduced plant growth and yield, shifts in competitive advantages in
mixed populations, decreased vigor, and injury. Ozone effects on vegetation are presented in
more detail in Chapter 5, Volume II of the 1996 Criteria Document.
1.2.5 Additional Health and Welfare Effects of NOx Emissions
In addition to their role as an ozone precursor, NOx emissions are associated with a wide
variety of other health and welfare effects.15'16 Nitrogen dioxide can irritate the lungs and reduce
resistance to respiratory infection (such as influenza). Nitrogen dioxide and airborne nitrate also
contribute to pollutant haze, which impairs visibility and can reduce residential property values
and the value placed on scenic views. Elevated levels of nitrates in drinking water pose
significant health risks, especially to infants. NOx emissions are an important precursor to acid
rain that may affect both terrestrial and aquatic ecosystems. Atmospheric deposition of nitrogen
leads to excess nutrient enrichment problems ("eutrophication"). Deposition of nitrogen-
containing compounds also affects terrestrial ecosystems.
1-15
-------
Draft Regulatory Support Document
1.2.3.1 Acid Deposition
Acid deposition, or acid rain as it is commonly known, occurs when SO2 and NOx react in
the atmosphere with water, oxygen, and oxidants to form various acidic compounds that later fall
to earth in the form of precipitation or dry deposition of acidic particles.17 It contributes to
damage of trees at high elevations and in extreme cases may cause lakes and streams to become
so acidic that they cannot support aquatic life. In addition, acid deposition accelerates the decay
of building materials and paints, including irreplaceable buildings, statues, and sculptures that are
part of our nation's cultural heritage. To reduce damage to automotive paint caused by acid rain
and acidic dry deposition, some manufacturers use acid-resistant paints, at an average cost of $5
per vehicle—a total of $61 million per year if applied to all new cars and trucks sold in the U.S.
Acid deposition primarily affects bodies of water that rest atop soil with a limited ability to
neutralize acidic compounds. The National Surface Water Survey (NSWS) investigated the
effects of acidic deposition in over 1,000 lakes larger than 10 acres and in thousands of miles of
streams. It found that acid deposition was the primary cause of acidity in 75 percent of the acidic
lakes and about 50 percent of the acidic streams, and that the areas most sensitive to acid rain
were the Adirondacks, the mid-Appalachian highlands, the upper Midwest and the high elevation
West. The NSWS found that approximately 580 streams in the Mid-Atlantic Coastal Plain are
acidic primarily due to acidic deposition. Hundreds of the lakes in the Adirondacks surveyed in
the NSWS have acidity levels incompatible with the survival of sensitive fish species. Many of
the over 1,350 acidic streams in the Mid-Atlantic Highlands (mid-Appalachia) region have
already experienced trout losses due to increased stream acidity. Emissions from U.S. sources
contribute to acidic deposition in eastern Canada, where the Canadian government has estimated
that 14,000 lakes are acidic. Acid deposition also has been implicated in contributing to
degradation of high-elevation spruce forests that populate the ridges of the Appalachian
Mountains from Maine to Georgia. This area includes national parks such as the Shenandoah
and Great Smoky Mountain National Parks.
1.2.3.2 Eutrophication and Nitrification
Nitrogen deposition into bodies of water can cause problems beyond those associated with
acid rain. The Ecological Society of America has included discussion of the contribution of air
emissions to increasing nitrogen levels in surface waters in a recent major review of causes and
consequences of human alteration of the global nitrogen cycle in its Issues in Ecology series.18
Long-term monitoring in the United States, Europe, and other developed regions of the world
shows a substantial rise of nitrogen levels in surface waters, which are highly correlated with
human-generated inputs of nitrogen to their watersheds. These nitrogen inputs are dominated by
fertilizers and atmospheric deposition.
Human activity can increase the flow of nutrients into those waters and result in excess algae
and plant growth. This increased growth can cause numerous adverse ecological effects and
economic impacts, including nuisance algal blooms, dieback of underwater plants due to reduced
light penetration, and toxic plankton blooms. Algal and plankton blooms can also reduce the
1-16
-------
Chapter 1: Health and Welfare Concerns
level of dissolved oxygen, which can also adversely affect fish and shellfish populations. This
problem is of particular concern in coastal areas with poor or stratified circulation patterns, such
as the Chesapeake Bay, Long Island Sound, or the Gulf of Mexico. In such areas, the
"overproduced" algae tends to sink to the bottom and decay, using all or most of the available
oxygen and thereby reducing or eliminating populations of bottom-feeder fish and shellfish,
distorting the normal population balance between different aquatic organisms, and in extreme
cases causing dramatic fish kills.
Collectively, these effects are referred to as eutrophication, which the National Research
Council recently identified as the most serious pollution problem facing the estuarine waters of
the United States.19 Nitrogen is the primary cause of eutrophication in most coastal waters and
estuaries.20 On the New England coast, for example, the number of red and brown tides and
shellfish problems from nuisance and toxic plankton blooms have increased over the past two
decades, a development thought to be linked to increased nitrogen loadings in coastal waters.
We believe that airborne NOx contributes from 12 to 44 percent of the total nitrogen loadings to
United States coastal water bodies. For example, some estimates assert that approximately one-
quarter of the nitrogen in the Chesapeake Bay comes from atmospheric deposition.
Excessive fertilization with nitrogen-containing compounds can also affect terrestrial
ecosystems.21 Research suggests that nitrogen fertilization can alter growth patterns and change
the balance of species in an ecosystem, providing beneficial nutrients to plant growth in areas
that do not suffer from nitrogen over-saturation. In extreme cases, this process can result in
nitrogen saturation when additions of nitrogen to soil over time exceed the capacity of the plants
and microorganisms to utilize and retain the nitrogen. This phenomenon has already occurred in
some areas of the U.S.
1.3 Carbon Monoxide
1.3.1 General Background
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). 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.
1-17
-------
Draft Regulatory Support Document
1.3.2 Health Effects of CO
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.22
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 declines23. The subsequent hypoxia in brain tissue then produces behavioral
effects, including decrements in continuous performance and reaction time.24
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. Persons
with heart disease are especially sensitive to carbon monoxide poisoning and may experience
chest pain if they breathe the gas while exercising. In Ontario, 18 deaths of snowmobilers
involved myocardial infarction and 14 involved sudden cardiac death25. It is unknown if these
deaths are linked to CO exposures.
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. More
importantly to many individuals is the frequent exposure of individuals to exhaust emissions
from engines operating indoors. The Occupational Safety and Health Administration sets
standards regulating the concentration of indoor pollutants, but high local CO levels are still
commonplace.
Several recent 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 of ambient CO
exposures with frequency of cardiovascular hospital admissions, especially for congestive heart
failure (CHF). An association of ambient CO exposure with 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 Criteria Document review process.26 There is emerging
evidence suggesting that CO is linked with asthma exacerbations.
1.3.3 CO Nonattainment
The current primary NAAQS for CO are 35 parts per million for the one-hour average and 9
parts per million for the eight-hour average. These values are not to be exceeded more than once
per year. Air quality carbon monoxide value is estimated using EPA guidance for calculating
design values. Over 22.4 million people currently live in the 13 non-attainment areas for the CO
NAAQS.27 As described in Section 1.1, the engines subject to this rule currently account for
about 3.8 percent of the mobile source CO inventory; this is expected to increase to 8.8 percent
by 2020 without the emission controls in this action.
1-18
-------
Chapter 1: Health and Welfare Concerns
Emissions from the engines and vehicles covered by this rule contribute to the national CO
inventory and to CO levels in several nonattainment areas. Large SI engines are used in forklifts
and many types of construction, industrial, and lawn care equipment that are used in urban areas,
including nonattainment areas.
ATVs and off-highway motorcycles are also used in counties and cities within CO-
nonattainment areas, and are operated on private land and in and around non-attainment areas.
This is illustrated by information about ATV use provided by Honda in public comments, which
included recent warranty claims for ATVs in three serious CO non-attainment areas: Fairbanks,
AK, in 1998 and 2001, in Phoenix, AZ in 2001, and in Las Vegas, NV in 2000.28 In our
December 7, 2000 notice finding that recreational vehicles cause or contribute to CO
nonattainment, we provided information showing CO emissions in six nonattainment areas in
2000. Five of these areas remain in nonattainment.
In addition, Western state studies of off-highway vehicle use in Colorado and Utah both
indicate that ATVs and off-highway motorcycles are operated on private land about 20 to 30
percent of the time (22.4 percent for off-highway motorcycles and 27.8 percent for ATVs in
Utah, and combined vehicles 22.4 percent of off-highway vehicles are operated on the survey
respondent's own private land or ranch).29 In addition, operation of these vehicles is not limited
to established trails. Half of the off-highway motorcyclists and 40 percent of the ATV owners in
Utah reported riding off established trails or roads.30 Furthermore, according to the U.S.
Consumer Product Safety Commission, almost three quarters of ATV drivers use ATVs for at
least one non-recreational activity; half use ATVs for farming or ranching; 63 percent use ATVs
for household chores (e.g., yard work); and about 8 percent use ATVs for occupational or
commercial tasks.31 Another CO nonattainment area, Anchorage, AK, estimates ATVs and
motorcycles (on- and off-road) contribute 0.19 tons per day in 2000.32
Several states that contain CO nonattainment areas also have large populations of registered
off-highway motorcycles, as shown in Table 1.3-1 (similar information was not available for
ATVs).
1-19
-------
Draft Regulatory Support Document
Table 1.3-1
Off-Highway Motorcycle Use in Selected CO Nonattainment Areas
City and State
Anchorage, AK
Fairbanks, AK
Las Vegas, NV
Los Angeles, CA
Phoenix, AZ
Spokane, WA
New York/New Jersey /Long Island, NY,
NJ, CT
Provo, UT
El Paso, TX
Fort Collins, CO
Medford, OR
Missoula, MT
Reno, NV
CO Nonattainment
Classification
Serious
Serious
Serious
Serious
Serious
Serious
Moderate > 12.7 ppm
Moderate > 12.7 ppm
Moderate
Moderate
Moderate
Moderate
Moderate
2001 State off-highway
motorcycle population3
5,100b
15,800
175,100
20,400
44,800
81,300
16,600
61,600
30,200
28,800
96,00
15,800b
'Source: Motorcycle Industry Council, 2001 Motorcycle Statistical Annual, Docket A-2000-01, Document No. II-G.
b State has more than one CO nonattainment area.
Snowmobiles, which have relatively high per engine CO emissions, can also be an important
source of ambient CO levels in CO nonattainment areas. While some of these areas have
experienced improved CO air quality in recent years, an area cannot be redesignated to
attainment until it can show EPA that it has had air quality levels within the level required for
attainment and that it has a plan in place to maintain such levels. Until areas have been
redesignated, they remain non-attainment areas.33 Snowmobiles contribute to CO nonattainment
in more than one of these areas.
The state of Alaska estimated (and a National Research Council study confirmed) that
snowmobiles contributed 0.3 tons/day in 2001 to Fairbanks' CO nonattainment area or 1.2
percent of a total inventory of 23.3 tons per day in 2001.3'4 There is some indication that
3 Draft Anchorage Carbon Monoxide Emission Inventory and Year 2000 Attainment
Projections, Air Quality Program, May 2001, Docket Number A-2000-01, Document II-A-40;
Draft Fairbanks 1995-2001 Carbon Monoxide Emissions Inventory, June 1, 2001, Docket
1-20
-------
Chapter 1: Health and Welfare Concerns
Fairbanks' snowmobile population is significantly higher than EPA's estimates.34 While
Fairbanks has made significant progress in reducing ambient CO concentrations, existing climate
conditions make achieving and maintaining attainment challenging. Anchorage, AK, reports a
similar contribution of snowmobiles to their emissions inventories (0.34 tons per day in 2000).
Furthermore, a recent National Academy of Sciences report concludes that "Fairbanks will be
susceptible to violating the CO health standards for many years because of its severe
meteorological conditions. That point is underscored by a December 2001 exceedance of the
standard in Anchorage which had no violations over the last 3 years."5 There is also a
snowmobile trail within the Spokane, WA, CO nonattainment area.
Several states that contain CO nonattainment areas also have large populations of registered
snowmobiles. This is shown in Table 1.3-2. A review of snowmobile trail maps and public
comments indicate that snowmobiles are used in counties containing these CO nonattainment
areas or in adjoining counties.35 These include the Mt. Spokane and Riverside trails near the
Spokane, Washington, CO nonattainment area; the Larimer trails near the Fort Collins, Colorado
CO nonattainment area; and the Hyatt Lake, Lake of the Woods, and Cold Springs trails near the
Klamath Falls and Medford, Oregon CO nonattainment area. There are also trails in Missoula
County, Montana that demonstrate snowmobile use in the Missoula, Montana CO nonattainment
area. While Colorado has a large snowmobile population, the snowmobile trails are fairly distant
from the Colorado Springs CO nonattainment area.36
Number A-2000-01, Document H-A-39.
4National Research Council. The Ongoing Challenge of Managing Carbon Monoxide
Pollution in Fairbanks, AK. May 2002. Docket A-2000-01, Document No. IV-A-115.
5National Research Council. The Ongoing Challenge of Managing Carbon Monoxide
Pollution in Fairbanks, AK. May 2002. Docket A-2000-01, Document IV-A-115.
1-21
-------
Draft Regulatory Support Document
Table 1.3-2
Snowmobile Use in Selected CO Nonattainment Areas
City and State
Anchorage, AK
Fairbanks, AK
Spokane, WA
Fort Collins, CO
Medford, OR
Missoula, MT
CO Nonattainment
Classification
Serious
Serious
Serious
Moderate
Moderate
Moderate
2001 State snowmobile
population*
35,576
31,532
32,500
16,809
23,440
! Source: Letter from International Snowmobile Manufacturers Association to US-EPA, March 14, 2002, Docket A-2000-01, Document No. II-G
While snowmobile trails are often located in rural areas and many are located outside CO
nonattainment areas, it is nevertheless the case that snowmobiles are used in urban areas within
nonattainment areas. In some northeast cities, "snowmobiles are a common sight in downtown
areas [and] are driven in large numbers along streets and recreational paths ... in close proximity
to pedestrians, motorists, and those using public parks such as cross-country skiers."37 A search
of the available literature indicates that snowmobiles are ridden in areas other than trails. For
example, a report by the Michigan Department of Natural Resources indicates that from 1993 to
1997, of the 146 snowmobile fatalities studied, 46 percent occurred on a state or county roadway
(another 2 percent on roadway shoulders) and 27 percent occurred on private lands.38
Furthermore, accident reports in the CO nonattainment area Fairbanks, AK, document that
snowmobiles driven on streets have collided with motor vehicles.39 On certain days there may be
concentrations of snowmobiles operated in non-attainment areas due to public events such as
snowmachine races (such as the Iron Dog Gold Rush Classic, which finishes in Fairbanks, AK),
during which snowmobiles will be present and operated. There is some indication that
Fairbanks snowmobile population is significantly higher than EPA's estimates.40
While the operation of snowmobiles alone in an area would not necessarily result in CO
nonattainment, emissions from regulated categories need only contribute to, not themselves
cause, nonattainment. Concentrations of NAAQS-related pollutants are by definition a result of
multiple sources of pollution. The above discussion shows that snowmobiles are operated on
snowmobile trails and some are within CO nonattainment areas (e.g., Spokane). Snowmobiles
are also used for maintenance operations and other uses in CO nonattainment areas (e.g.,
Fairbanks and Anchorage), and there is evidence that snowmobiles are operated in town along
streets in these and other CO nonattainment areas.
While CO air quality is improving in several northern areas, further reductions may still be
required. Exceedances of the 8-hour CO standard were recorded in three of the six CO
1-22
-------
Chapter 1: Health and Welfare Concerns
nonattainment areas located in the northern portion of the country over the five year period from
1994 to 1999: Fairbanks, AK; Medford, OR; and Spokane, WA.41 Given the variability in CO
ambient concentrations due to weather patterns such as inversions, the absence of recent
exceedances for some of these nonattainment areas should not be viewed as eliminating the need
for further reductions to consistently attain and maintain the standard. A review of CO monitor
data in Fairbanks from 1986 to 1995 shows that while median concentrations have declined
steadily, unusual combinations of weather and emissions have resulted in elevated ambient CO
concentrations well above the 8-hour standard of 9 ppm. Specifically, a Fairbanks monitor
recorded average 8-hour ambient concentrations at 16 ppm in 1988, around 9 ppm from 1990 to
1992, and then a steady increase in CO ambient concentrations at 12, 14 and 16 ppm during some
extreme cases in 1993, 1994 and 1995, respectively.42 Furthermore, a recent National Academy
of Sciences report concludes that "Fairbanks will be susceptible to violating the CO health
standards for many years because of its severe meteorological conditions. That point is
underscored by a December 2001 exceedance of the standard in Anchorage which had no
violations over the last 3 years."43 Fairbanks is located in a mountain valley with a much higher
potential for air stagnation than cities within the contiguous United States. Nocturnal inversions
that give rise to elevated CO concentrations can persist 24-hours a day due to the low solar
elevation, particularly in December and January. These inversions typically last from 2 to 4 days,
and thus inversions may continue during hours of maximum CO emissions from mobile sources.
While Fairbanks has made significant progress in reducing ambient CO concentrations, existing
climate conditions make achieving and maintaining attainment challenging.
In addition to the CO nonattainment areas, there are 6 areas that have not been classified as
non-attainment where air quality monitoring indicated a need for CO control. For example, CO
monitors in northern locations such as Des Moines, IA, and Weirton, WV/Steubenville, OH,
registered levels above the level of the CO standards in 1998.44
1.4 Particulate Matter
1.4.1 General Background
Particulate pollution is a problem affecting urban and non-urban localities in all regions of
the United States. Nonroad engines and vehicles that would be subject to the standards
contribute to ambient particulate matter (PM) levels in two ways. First, they contribute through
direct emissions of particulate matter. Second, they contribute to indirect formation of PM
through their emissions of organic carbon, especially HC. As shown in Table 1.4-1, organic
carbon accounts for between 27 and 36 percent of ambient fine particle mass depending on the
area of the country.
1-23
-------
Draft Regulatory Support Document
Table 1.4-1
Percent Contribution to PM2 5 by Component, 1998
Sulfate
Elemental Carbon
Organic Carbon
Nitrate
Crustal Material
East
56
5
27
5
7
West
33
6
36
8
17
Source: National Air Quality and Emissions Trends Report, 1998, March, 2000, at 28. This document is available at
http ://www.epa. gov/oar/aqtrnd98/. Relevant pages of this report can be found in Memorandum to Air Docket A-2000-01 from Jean Marie Revelt,
September 5, 2001, Document No. II-A-63.
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. All particles equal to and less than 10 microns are
called PM10. Fine particles can be generally defined as those particles with an aerodynamic
diameter of 2.5 microns or less (also known as PM25), and coarse fraction particles are those
particles with an aerodynamic diameter greater than 2.5 microns, but equal to or less than a
nominal 10 microns.
Manmade emissions that contribute to airborne particulate matter result principally from
combustion sources (stationary and mobile sources) and fugitive emissions from industrial
processes and non-industrial processes (such as roadway dust from paved and unpaved roads,
wind erosion from crop land, construction, etc.). Human-generated sources of particles include a
variety of stationary sources (including power generating plants, industrial operations,
manufacturing plants, waste disposal) and mobile sources (light- and heavy-duty on-road
vehicles, and off-highway vehicles such as construction, farming, industrial, locomotives, marine
vessels and other sources). Natural sources also contribute to particulate matter in the
atmosphere and include sources such as wind erosion of geological material, sea spray, volcanic
emissions, biogenic emanation (e.g., pollen from plants, fungal spores), and wild fires.
The chemical and physical properties of PM vary greatly with time, region, meteorology,
and source category. Particles may be emitted directly to the atmosphere (primary particles) or
may be formed by transformations of gaseous emissions of sulfur dioxide, nitrogen oxides or
volatile organic compounds (secondary particles). Secondary PM is dominated by sulfate in the
eastern U.S. and nitrate in the western U.S.45 The vast majority (>90 percent) of the direct
mobile source PM emissions and their secondary formation products are in the fine PM size
range. Mobile sources can reasonably be estimated to contribute to ambient secondary nitrate
and sulfate PM in proportion to their contribution to total NOx and SOx emissions.
1-24
-------
Chapter 1: Health and Welfare Concerns
1.4.2 Health and Welfare Effects of PM
Particulate matter can adversely affect human health and welfare. Discussions of the health
and welfare effects associated with ambient PM can be found in the Air Quality Criteria for
Particulate Matter.46
Key EPA findings regarding the health risks posed by ambient PM are summarized as
follows:
a. Health risks posed by inhaled particles are affected both by the penetration and deposition of
particles in the various regions of the respiratory tract, and by the biological responses to
these deposited materials.
b. The risks of adverse effects associated with deposition of ambient particles in the thorax
(tracheobronchial and alveolar regions of the respiratory tract) are markedly greater than for
deposition in the extrathoracic (head) region. Maximum particle penetration to the thoracic
regions occurs during oronasal or mouth breathing.
c. Published, peer-reviewed studies have reported statistical associations between PM and
several key health effects, including premature death; aggravation of respiratory and
cardiovascular disease, as indicated by increased hospital admissions and emergency room
visits, school absences, work loss days, and restricted activity days; changes in lung function
and increased respiratory symptoms; changes to lung tissues and structure; and altered
respiratory defense mechanisms. Most of these effects have been consistently associated
with ambient PM concentrations, which have been used as a measure of population
exposure, in a large number of community epidemiological studies. Additional information
and insights on these effects are provided by studies of animal toxicology and controlled
human exposures to various constituents of PM conducted at higher than ambient
concentrations. Although mechanisms by which particles cause effects are not well known,
there is general agreement that the cardio-respiratory system is the major target of PM
effects.
d. Based on a qualitative assessment of the epidemiological evidence of effects associated with
PM for populations that appear to be at greatest risk with respect to particular health
endpoints, we have concluded the following with respect to sensitive populations:
1. Individuals with respiratory disease (e.g., chronic obstructive pulmonary disease, acute
bronchitis) and cardiovascular disease (e.g., ischemic heart disease) are at greater risk of
premature mortality and hospitalization due to exposure to ambient PM.
2. Individuals with infectious respiratory disease (e.g., pneumonia) are at greater risk of
premature mortality and morbidity (e.g., hospitalization, aggravation of respiratory
symptoms) due to exposure to ambient PM. Also, exposure to PM may increase
individuals' susceptibility to respiratory infections.
1-25
-------
Draft Regulatory Support Document
3. Elderly individuals are also at greater risk of premature mortality and hospitalization for
cardiopulmonary problems due to exposure to ambient PM.
4. Children are at greater risk of increased respiratory symptoms and decreased lung
function due to exposure to ambient PM.
5. Asthmatic individuals are at risk of exacerbation of symptoms associated with asthma,
and increased need for medical attention, due to exposure to PM.
e. There are fundamental physical and chemical differences between fine and coarse fraction
particles. The fine fraction contains acid aerosols, sulfates, nitrates, transition metals, diesel
exhaust particles, and ultra fine particles; the coarse fraction typically contains high mineral
concentrations, silica and resuspended dust. It is reasonable to expect that differences may
exist in both the nature of potential effects elicited by coarse and fine PM and the relative
concentrations required to produce such effects. Both fine and coarse particles can
accumulate in the respiratory system. Exposure to coarse fraction particles is primarily
associated with the aggravation of respiratory conditions such as asthma. Fine particles are
closely associated with health effects such as premature death or hospital admissions, and
for cardiopulmonary diseases.
With respect to welfare or secondary effects, fine particles have been clearly associated with
the impairment of visibility over urban areas and large multi-State regions. Particles also
contribute to soiling and materials damage. Components of particulate matter (e.g., sulfuric or
nitric acid) also contribute to acid deposition, nitrification of surface soils and water
eutrophication of surface water.
1.4.3 PM Nonattainment
1.4.3.1 PM10 Concentrations and Nonattainment
The NAAQS for PM10 was established in 1987. According to these standards, the short term
(24-hour) standard of 150 |ig/m3 is not to be exceeded more than once per year on average over
three years. The long-term standard specifies an expected annual arithmetic mean not to exceed
50 |ig/m3 over three years.
PM10 monitoring data indicate that 14 designated PM10 nonattainment areas with a projected
population of 23 million violated the PM10 NAAQS in the period 1997-1999. Table 1.4-2 lists
the 14 areas, and also indicates the PM10 nonattainment classification, and 1999 projected
population for each PM10 nonattainment area. The projected population in 1999 was based on
1990 population figures which were then increased by the amount of population growth in the
county from 1990 to 1999.
1-26
-------
Chapter 1: Health and Welfare Concerns
Table 1.4-2
PM,n Nonattainment Areas Violating the PM,n NAAQS in 1997-1999
Nonattainment Area or County
\nthony, NM (Moderate)"
:iark Co [Las Vegas], NV (Serious)
^oachella Valley, CA (Serious)
il Paso Co, TX (Moderate)3
layden/Miami, AZ (Moderate)
mperial Valley, CA (Moderate)
^os Angeles South Coast Air Basin, CA (Serious)
Regales, AZ (Moderate)
)wens Valley, CA (Serious)
3hoenix, AZ (Serious)
san Joaquin Valley, CA (Serious)
searles Valley, CA (Moderate)
\Vallula, WA (Moderate)15
\Vashoe Co [Reno], NV (Moderate)
Total Areas: 14
1999 Population
(projected, in thousands)
3
1,200
239
611
4
122
14,352
25
18
2,977
3,214
29
52
320
23.167
' EPA has determined that continuing PM10 nonattainment in El Paso, TX is attributable to transport under section 179(B).
The violation in this area has been determined to be attributable to natural events under section 188(f) of the Act.
In addition to the 14 PM10 nonattainment areas that are currently violating the PM10 NAAQS
listed in Table 1.4-2, there are 25 unclassifiable areas that have recently recorded ambient
concentrations of PM10 above the PM10 NAAQS. EPA adopted a policy in 1996 that allows areas
with PM10 exceedances that are attributable to natural events to retain their designation as
unclassifiable if the State is taking all reasonable measures to safeguard public health regardless
of the sources of PM10 emissions. Areas that remain unclassifiable areas are not required under
the Clean Air Act to submit attainment plans, but we work with each of these areas to understand
the nature of the PM10 problem and to determine what best can be done to reduce it. With respect
to the monitored violations reported in 1997-99 in the 25 areas designated as unclassifiable, we
have not yet excluded the possibility that factors such as a one-time monitoring upset or natural
events, which ordinarily would not result in an area being designated as nonattainment for PM10,
may be responsible for the problem. Emission reductions from today's action will assist these
currently unclassifiable areas to achieve ambient PM10 concentrations below the current PM10
NAAQS.
1.4.3.2 PM2 5 Concentrations
Fine particle concentrations contribute to both health effects and visibility impairment.
This section presents our assessment of current and future PM2.5 levels. Because monitoring
data are not available for all areas, we have modeled PM2.5 levels for those areas using the
EPA's Regulatory Model System for Aerosols and Deposition (REMSAD) model. These
1-27
-------
Draft Regulatory Support Document
concentrations are related to both health effects and visibility impairment. After a brief
description of the PM air quality model, we present current PM2.5 data, both modeled and
estimated. Then we present projections of PM2.5 levels that were estimated using REMSAD.
1.4.3.2.1 Description ofPMAir Quality Modeling
To estimate both current PM2.5 levels in areas for which no monitoring data are available
and future PM2 5 levels for all areas, we refer to the PM air quality modeling performed in
conjunction with EPA's on-highway Heavy Duty Engine/Diesel Fuel (HD07) final rule. This
modeling was performed using EPA's Regulatory Model System for Aerosols and Deposition
(REMSAD) model.47 We describe the REMSAD modeling because we use the modeling
examine visibility impairment and population exposures related to the PM health effects we
would anticipate would occur without the emissions reductions from this rulemaking. The
REMSAD modeling was also a key input for the economic benefits transfer technique described
in Chapter 10 related to selected PM health effects.
REMSAD simulates every hour of every day of the year and, thus, requires a variety of input
files that contain information pertaining to the modeling domain and simulation period. These
include gridded, 3-hour average emissions estimates and meteorological fields, initial and
boundary conditions, and land-use information. As applied to the contiguous U.S., the model
segments the area within the region into square blocks called grids (roughly equal in size to
counties), each of which has several layers of air conditions. Using this data, REMSAD
generates predictions of 3-hour average PM concentrations for every grid. We then calculated
daily and seasonal PM air quality metrics.
REMSAD was peer-reviewed in 1999 for EPA as reported in "Scientific Peer-Review of the
Regulatory Modeling System for Aerosols and Deposition." Earlier versions of REMSAD have
been employed for the EPA's Prospective CAA Section 812 Report to Congress and for EPA's
Analysis of the Acid Deposition and Ozone Control Act (Senate Bill 172). Version 4.1 of
REMSAD was employed for the HD07 final rule analysis and is fully described in the air quality
technical support documents for that HD07 rulemaking. We focus on the HD07 modeling
because it is the most current modeling for mobile sources.
For the FID07 rulemaking, EPA modeled PM air quality in 1996 and in 2030 after those
requirements were to take effect using REMSAD. Although we did not undertake new air
quality modeling for this rulemaking, the modeling from the FID07 rulemaking can be
considered a baseline for this rulemaking. As explained in Section 1.1.2, the emissions
inventories that were used in the FID07 REMSAD modeling have been updated and that the
FID07 modeling may underestimate the PM2.5 levels that we would expect with revised
emissions inventories.
1.4.3.2.2 Current PMAir Quality
The 1999-2000 PM2 5 monitored values, which cover about a third of the nation's counties,
1-28
-------
Chapter 1: Health and Welfare Concerns
indicate that at least 82 million people live in areas where long-term ambient fine particulate
matter levels are at or above 15 jig/m3.48
To estimate the current number of people who live in areas where long-term ambient fine
particulate matter levels are at or above 16 |ig/m3 but for which there are no monitors, we can use
the HD07 REMSAD modeling described above. At the time the HD07 modeling was performed,
1999 PM monitoring data were not yet available, so we conducted 1996 base year modeling to
reproduce the atmospheric processes resulting in formation and dispersion of PM25 across the
U.S. and to evaluate operational model performance for PM25 and its related speciated
components (e.g., sulfate, nitrate, elemental carbon) which are important to visibility impairment.
This 1996 modeling included emissions from the engines subject to this final rule (although
earlier emissions estimates were used). According to our national modeled predictions, there
were a total of 76 million people (1996 population) living in areas with modeled annual average
PM25 concentrations at or above 16 |ig/m3 (29 percent of the population).49
1.4.3.2.3 Future PMAir Quality
To estimate future year concentrations, we can use the air quality model to predict changes
between current and future states. The most reliable information would be to compare future
levels in counties for which we have monitoring data. Thus, we estimated future conditions for
the areas with current PM25 monitored data (which covered about a third of the nation's counties
at that time).50 For these counties, REMSAD predicts the current level of 37 percent of the
population living in areas where fine PM levels above 15 |ig/m3 to increase to 49 percent in
2030.51 Again, this 2030 modeling included emissions from the engines subject to this final rule
(although earlier emissions estimates were used). These emissions are contributing to air quality
levels that may result in future PM nonattainment. Nonattainment status is related to both health
impacts described above and welfare impacts, such as visibility impairment, soiling, and material
damage. Thus, for areas with levels above the NAAQS, unacceptable health and welfare effects
are anticipated to be occurring, and emissions from the engines subject to this rulemaking are
contributing to these anticipated adverse effects. In Table 1.4-3, we summarize the national PM
air quality based on the HD07 REMSAD modeling.
1-29
-------
Draft Regulatory Support Document
Table 1.4-3
Summary of Anticipated 2030 National PM Baseline Air Quality (ug/m3)
Statistic
PM10
Minimum Annual Meanb
Maximum Annual Meanb
Average Annual Mean
Median Annual Mean
Population- Weighted Average Annual Mean0
PM25
Minimum Annual Meanb
Maximum Annual Meanb
Average Annual Mean
Median Annual Mean
Population- Weighted Average Annual Mean0
2030 Air Quality Value
(ug/m3)a
1.49
64.29
10.03
7.97
21.04
1.16
38.2
7.6
5.79
14.2
* Based on public comment received on the proposed Large Si/Recreational Vehicle rule and other updated
information, we revised our emissions estimates in some categories downwards and other categories upwards;
however, on net, we believe this modeling would underestimate the baseline PM emissions without regulation.
b The minimum (maximum) is the value for the populated grid-cell with the lowest (highest) annual average.
c Calculated by summing the product of the projected 2030 grid-cell population and the estimated 2030 PM
concentration, for that grid-cell and then dividing by the total population in the 48 contiguous States.
Nonroad engines and vehicles that are subject these standards contribute to ambient fine PM
levels in two ways. First, they contribute through direct emissions of fine PM. As shown in
Table 1.1-1, these engines emitted 14,600 tons of PM (about 2.1 percent of all mobile source
PM) in 2000. As shown in Table 1.1-3, they are modeled to emit 36,500 tons of PM (about 4.8
percent of all mobile source PM) in 2030. Second, these engines contribute to indirect formation
of PM through their emissions of gaseous precursors which are then transformed in the
atmosphere into particles. For example, these engines emitted about 1,411,000 tons of HC or
23.5 percent of the HC emitted from mobile sources in 2030. Furthermore, recreational vehicles,
such as snowmobiles and ATVs emit high levels of organic carbon (as HC) on a per engine basis.
Some organic emissions are transformed into particles in the atmosphere and other volatile
organics can condense if emitted in cold temperatures, as is the case for emissions from
snowmobiles, for example. Organic carbon accounts for between 27 and 36 percent of ambient
fine particle mass depending on the area of the country. The relationship between HC and PM
have implications for the most efficient controls of ambient PM as discussed in Chapter 4.
1-30
-------
Chapter 1: Health and Welfare Concerns
Further, as discussed below, the nonroad engines we are regulating contribute to PM levels
in areas with PM levels above 15 |ig/m3.
1.5 Visibility Degradation
1.5.1 General Background
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.52 Visibility impairment has been considered the "best understood and most easily
measured effect of air pollution."53 Visibility degradation is often directly proportional to
decreases in light transmittal in the atmosphere. Scattering and absorption by both gases and
particles decrease light transmittance. It is an easily noticeable effect of fine PM present in the
atmosphere, and fine PM is the major cause of reduced visibility in parts of the United States,
including many of our national parks and in places where people live, work, and recreate. Fine
particles with significant light-extinction efficiencies include organic matter, sulfates, nitrates,
elemental carbon (soot), and soil.
Visibility is an important effect 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, both in where they live and work, and in places where they enjoy
recreational opportunities. Visibility is highly valued in significant natural areas such as national
parks and wilderness areas, because of the special emphasis given to protecting these lands now
and for future generations.
To quantify changes in visibility, the analysis presented in this chapter computes a light-
extinction coefficient, based on the work of Sisler, which shows the total fraction of light that is
decreased per unit distance.54 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. Visibility can be described in terms of visual range, light extinction
or deciview.6
In addition to limiting the distance that one can see, the scattering and absorption of light
caused by air pollution can also degrade the color, clarity, and contrast of scenes. Visibility
impairment also has a temporal dimension in that impairment might relate to a short-term
6Visual range can be defined as the maximum distance at which one can identify a black
object against the horizon sky. It is typically described in miles or kilometers. Light extinction is
the sum of light scattering and absorption by particles and gases in the atmosphere. It is typically
expressed in terms of inverse megameters (Mm"1), with larger values representing worse
visibility. The deciview metric describes perceived visual changes in a linear fashion over its
entire range, analogous to the decibel scale for sound. A deciview of 0 represents pristine
conditions. Under many scenic conditions, a change of 1 deciview is considered perceptible by
the average person.
1-31
-------
Draft Regulatory Support Document
excursion or to longer periods (e.g., worst 20 percent of days or annual average levels). More
detailed discussions of visibility effects are contained in the EPA Criteria Document for PM.
Visibility effects are manifest in two principal ways: (1) as local impairment (e.g., localized
hazes and plumes) and (2) as regional haze. The emissions from engines covered by this rule
contribute to both types of visibility impairment.
Local-scale visibility degradation is commonly in the form of either a plume resulting from
the emissions of a specific source or small group of sources, or it is in the form of a localized
haze such as an urban "brown cloud." Plumes are comprised of smoke, dust, or colored gas that
obscure the sky or horizon relatively near sources. Impairment caused by a specific source or
small group of sources has been generally termed as "reasonably attributable."
The second type of impairment, regional haze, results from pollutant emissions from a
multitude of sources located across a broad geographic region. It impairs visibility in every
direction over a large area, in some cases over multi-state regions. Regional haze masks objects
on the horizon and reduces the contrast of nearby objects. The formation, extent, and intensity of
regional haze is a function of meteorological and chemical processes, which sometimes cause
fine particulate loadings to remain suspended in the atmosphere for several days and to be
transported hundreds of kilometers from their sources.55
On an annual average basis, the concentrations of non-anthropogenic fine PM are generally
small when compared with concentrations of fine particles from anthropogenic sources.56
Anthropogenic contributions account for about one-third of the average extinction coefficient in
the rural West and more than 80 percent in the rural East.57 Because of significant differences
related to visibility conditions in the eastern and western U.S., we present information about
visibility by region. Furthermore, it is important to note that even in those areas with relatively
low concentrations of anthropogenic fine particles, such as the Colorado plateau, small increases
in anthropogenic fine particle concentrations can lead to significant decreases in visual range.
This is one of the reasons Class I areas have been given special consideration under the Clean Air
Act.
1.5.2 Visibility Impairment Where People Live, Work and Recreate
Visibility impairment occurs in many areas throughout the country, where people live, work,
and recreate. In this section, in order to estimate the magnitude of the problem, we use
monitored PM2.5 data and modeled air quality using emissions inventories from the engines
subject to this rule. The engines covered by this rule contribute to PM2.5 levels in areas across
the country with unacceptable visibility conditions.
1.5.2.1 Areas Affected by Visibility Impairment
The secondary PM NAAQS is designed to protect against adverse welfare effects such as
visibility impairment. In 1997, the secondary PM NAAQS was set as equal to the primary
1-32
-------
Chapter 1: Health and Welfare Concerns
(health-based) PMNAAQS (62 Federal Register No. 138, July 18, 1997). EPA concluded that
PM can and does produce adverse effects on visibility in various locations, depending on PM
concentrations and factors such as chemical composition and average relative humidity. In 1997,
EPA demonstrated that visibility impairment is an important effect on public welfare and that
visibility impairment is experienced throughout the U.S., in multi-state regions, urban areas, and
remote Federal Class I areas.
In many cities having annual mean PM2.5 concentrations exceeding 17 ug/m3,
improvements in annual average visibility resulting from the attainment of the annual PM2.5
standard are expected to be perceptible to the general population (e.g., to exceed 1 deciveiew).
Based on annual mean monitored PM2.5 data, many cities in the Northeast, Midwest, and
Southeast as well as Los Angeles would be expected to experience perceptible improvements in
visibility if the PM2.5 annual standard were attained. For example, in Washington, DC, where
the IMPROVE monitoring network shows annual mean PM2.5 concentrations at about 19 ug/m3
during the period of 1992 to!995, approximate annual average visibility would be expected to
improve from 21 km (29 deciview) to 27 km (27 deciview). The PM2.5 annual average in
Washington, DC, was 18.9 ug/m3 in 2000.
The updated monitored data and air quality modeling presented below confirm that the
visibility situation identified during the NAAQS review in 1997 is still likely to exist.
Specifically, there will still likely be a broad number of areas that are above the annual PM2.5
NAAQS in the Northeast, Midwest, Southeast and California , such that the determination in the
NAAQS rulemaking about broad visibility impairment and related benefits from NAAQS
compliance are still relevant. Thus, levels above the fine PM NAAQS cause adverse welfare
impacts, such as visibility impairment (both regional and localized impairment).
In addition, in setting the PM NAAQS, EPA acknowledged that levels of fine particles
below the NAAQS may also contribute to unacceptable visibility impairment and regional haze
problems in some areas, and Clean Air Act Section 169 provides additional authorities to remedy
existing impairment and prevent future impairment in the 156 national parks, forests and
wilderness areas labeled as Class I areas.
In making determinations about the level of protection afforded by the secondary PM
NAAQS, EPA considered how the Section 169 regional haze program and the secondary
NAAQS would function together. Regional strategies are expected to improve visibility in many
urban and non-Class I areas as well. The following recommendation for the National Research
Council, Protecting Visibility in National Parks and Wilderness Areas (1993), addresses this
point: "Efforts to improve visibility in Class I areas also would benefit visibility outside these
areas. Because most visibility impairment is regional in scale, the same haze that degrades
visibility within or looking out from a national park also degrade visibility outside it. Class I
areas cannot be regarded as potential islands of clean air in a polluted sea."
Visibility impairment (localized and regional haze) in Class I areas is discussed in the next
section.
1-33
-------
Draft Regulatory Support Document
1.5.2.1.1 Areas Affected by Visibility Impairment: Monitored Data
The 1999-2000 PM2 5 monitored values, which cover only a portion of the nation's
counties, indicate that at least 82 million people live in areas where long-term ambient fine
particulate matter levels are at or above 15 jig/m3.58 Thus, these populations (plus others who
travel to these areas) would be experiencing visibility impairment that is unacceptable, and based
on our modeling, emissions of PM and its precursors from engines in these categories contribute
to this unacceptable impairment.
Another way to consider this information is to compare the values directly to the PM
NAAQS in the format required by regulation. EPA regulations require 3 consecutive years of
PM2.5 data in order to make comparisons with the National Ambient Air Quality Standards; see
Part 50, Appendix N. In Table 1.5-1, we list areas with 1999 and 2000 monitored annual average
PM2.5 levels above 15 ug/m3 in 2000, as represented by design values that can be compared to
the PM2.5 NAAQS. There were a total of 129 counties representing 65 million people with
levels above the design value for the annual PM2.5 NAAQS based on 1999 and 2000 monitored
data. The table also notes areas which have made a note of "exceptional events" in their
reporting of the monitored data.
1-34
-------
Table 1.5-1.
Areas with Monitored Annual Average PM2.5 Concentrations Above 15 ug/m3.
EPA regulations require 3 consecutive years of PM2.5 data in order to make comparisons with the National Ambient
Air Quality Standards; see Part 50, Appendix N. The data represented in this table reflect air quality monitoring from
1999 to 2001, although not all data have been verified by the monitoring agency.
State
ALABAMA
ALABAMA
ALABAMA
ALABAMA
ALABAMA
County
CLAY
COLBERT
DEKALB
HOUSTON
JEFFERSON*
Population 2000
14,254
54,984
64,452
88,787
662,047
Annual PM2.5
Standard
Desian Value
15.5
15.3
16.8
16.3
20.8*
Design
Value Data
Flacmed for
Exceptional
Events? 1
* Two sites in Jefferson County are encompassed in a Community Monitoring Zone (i.e. utilize spatial
averaging); the spatially averaged design value for the CMZ is 20.8, which is the maximum for the county.
ALABAMA
ALABAMA
ALABAMA
ALABAMA
ALABAMA
ALABAMA
ALABAMA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA
CALIFORNIA 2
CALIFORNIA 2
CALIFORNIA
CALIFORNIA
CALIFORNIA
MADISON
MOBILE
MONTGOMERY
MORGAN
RUSSELL
SHELBY
TALLADEGA
BUTTE
FRESNO
MPERIAL
KERN
KINGS
LOS ANGELES
MERCED
ORANGE
RIVERSIDE
SAN BERNARDINO
SAN DIEGO
SAN JOAQUIN
STANISLAUS
276,700
399,843
223,510
111,064
49,756
143,293
80,321
203,171
799,407
142,361
661,645
129,461
9,519,338
210,554
2,846,289
1,545,387
1,709,434
2,813,833
563,598
446,997
15.5
15.3
16.8
19.1
18.4
17.2
17.8
15.4
24.0
15.7
23.7
16.6
25.9
18.9
22.4
29.8
25.8
17.1
16.4
19.7
yes
yes
yes
yes
yes
yes
1-35
-------
State
CALIFORNIA
CONNECTICUT
DELAWARE
DISTRICT OF
COLUMBIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
GEORGIA
ILLINOIS
ILLINOIS
ILLINOIS
ILLINOIS
ILLINOIS
INDIANA
INDIANA
INDIANA
INDIANA
KENTUCKY
KENTUCKY
KENTUCKY
KENTUCKY
KENTUCKY
KENTUCKY
Countv
TULARE
NEW HAVEN
NEW CASTLE
WASHINGTON
BIBB
CHATHAM
CLARKE
CLAYTON
COBB
DEKALB
DOUGHERTY
FLOYD
FULTON
HALL
MUSCOGEE
PAULDING
RICHMOND
WASHINGTON
WILKINSON
COOK
DU PAGE
MADISON
ST CLAIR
WILL
CLARK
FLOYD
LAKE
MARION
BOYD
BULLITT
CAMPBELL
FAYETTE
JEFFERSON
KENTON
Pooulation 2000
368,021
824,008
500,265
572,059
153,887
232,048
101,489
236,517
607,751
665,865
96,065
90,565
816,006
139,277
186,291
81,678
199,775
21,176
10,220
5,376,741
904,161
258,941
256,082
502,266
96,472
70,823
484,564
860,454
49,752
61,236
88,616
260,512
693,604
151,464
Annual Std
Desian Value
24.7
16.8
16.6
16.6
17.6
16.5
18.6
19.2
18.6
19.6
16.6
18.5
21.2
17.2
18.0
16.8
17.4
16.5
18.1
18.8
15.4
17.3
17.4
15.9
17.3
15.6
16.3
17.0
15.5
16.0
15.5
16.8
17.1
15.9
DataFlacmed
for Exc.
Events?1
yes
yes
yes
yes
yes
yes
yes
1-36
-------
State
KENTUCKY
KENTUCKY
KENTUCKY
MARYLAND
MICHIGAN
MISSISSIPPI
MISSISSIPPI
MISSOURI
MONTANA
NEW JERSEY
NEW JERSEY
NEW YORK
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
NORTH CAROLINA
OHIO
OHIO
Countv
MC CRACKEN
PIKE
WARREN
BALTIMORE (CITY)
WAYNE
HINDS
JONES
ST LOUIS (CITY)
LINCOLN
HUDSON
UNION
NEW YORK
ALAMANCE
CABARRUS
CATAWBA
CUMBERLAND
DAVIDSON
DURHAM
FORSYTH
GASTON
GUILFORD
HAYWOOD
MC DOWELL
MECKLENBURG
MITCHELL
WAKE
BUTLER
CUYAHOGA
Pooulation 2000
65,514
68,736
92,522
651,154
2,061,162
250,800
64,958
348,189
18,837
608,975
522,541
1,537,195
130,800
131,063
141,685
302,963
147,246
223,314
306,067
190,365
421,048
54,033
42,151
695,454
15,687
627,846
332,807
1,393,978
Annual Std
Desian Value
15.1
16.1
15.4
17.8
18.9
15.1
16.6
16.3
16.4
17.5
16.3
17.8
15.3
15.7
17.1
15.4
17.3
15.3
16.2
15.3
16.3
15.4
16.2
16.8
15.5
15.3
17.4
20.3
DataFlacmed
for Exc.
Events?1
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
1-37
-------
State
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
OHIO
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
PENNSYLVANIA
SOUTH CAROLINA
SOUTH CAROLINA
SOUTH CAROLINA
SOUTH CAROLINA
TENNESSEE
TENNESSEE
TENNESSEE
TENNESSEE
TENNESSEE
TENNESSEE
Countv
FRANKLIN
HAMILTON
JEFFERSON
LORAIN
MAHONING
MONTGOMERY
PORTAGE
SCIOTO
STARK
SUMMIT
TRUMBULL
ALLEGHENY
BERKS
CAMBRIA
DAUPHIN
LANCASTER
PHILADELPHIA
WASHINGTON
WESTMORELAND
YORK
GREENVILLE
LEXINGTON
RICHLAND
SPARTANBURG
DAVIDSON
HAMILTON
KNOX
ROANE
SHELBY
SULLIVAN
Pooulation 2000
1,068,978
845,303
73,894
284,664
257,555
559,062
152,061
79,195
378,098
542,899
225,116
1,281,666
373,638
152,598
251,798
470,658
1,517,550
202,897
369,993
381,751
379,616
216,014
320,677
253,791
569,891
307,896
382,032
51,910
897,472
153,048
Annual Std
Desian Value
18.1
19.3
18.9
15.1
16.4
17.6
15.3
20.0
18.3
17.3
16.2
21.0
15.6
15.3
15.5
16.9
16.6
15.5
15.6
16.3
17.0
15.6
15.4
15.4
17.0
18.9
20.4
17.0
15.6
17.0
DataFlacmed
for Exc.
Events?1
yes
yes
yes
yes
yes
yes
yes
DataFlaqqed
1-38
-------
County
State
Population 2000 Annual Std for Exc.
Design Value Events?1
TENNESSEE
VIRGINIA
VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
WEST VIRGINIA
TOTAL
SUMNER
BRISTOL
ROANOKE (CITY)
BERKELEY
BROOKE
CABELL
HANCOCK
KANAWHA
MARSHALL
OHIO
WOOD
129 Counties
130,449
17,367
94,911
75,905
25,447
96,784
32,667
200,073
35,519
47,427
87,986
65,185,812
15.7
16.0
15.2
16.0
17.4
17.8
17.4
18.4
16.5
15.7
17.6
yes
yes
yes
yes
yes
1. Design Values include all valid data. Some valid data were impacted by exceptional events.
These special situations are being reviewed by EPA.
2. Sacramento County CA does not exceed the PM2.5 annual standard but does exceed
the daily standard.
Source: EPA Trends Reports
1-39
-------
Draft Regulatory Support Document
1.5.2.1.2 Areas Affected by Visibility Impairment: Modeled Future PM Levels and
Visibility Index Estimates
Because the chemical composition of the PM affects visibility impairment, we used
REMSAD air quality model to project visibility conditions in 2030 accounting for the chemical
composition of the particles and to estimate visibility impairment directly as changes in
deciview. Our projections included anticipated emissions from the engines subject to this rule,
and although our emission predictions reflected our best estimates of emissions projections at the
time the modeling was conducted, we now have new estimates, as discussed above in Table 1.1-
4. Based on public comment for this rule and new information, we have revised our emissions
estimates in some categories downwards and other categories upwards; however, on net, we
believe the HD07 modeling underestimates the PM air quality levels that would be predicted if
new inventories were used.
The most reliable information about the future visibility levels would be in areas for which
monitoring data are available to evaluate model performance for a base year (e.g., 1996).
Accordingly, we predicted that in 2030, 49 percent of the population will be living in areas where
fine PM levels are above 15 |ig/m3 and monitors are available.59 This can be compared with the
1996 level of 37 percent of the population living in areas where fine PM levels are above 15
|ig/m3 and monitors are available.
Based upon the light-extinction coefficient, we also calculated a unitless visibility index,
called a "deciview," which is used in the valuation of visibility. The deciview metric provides a
linear scale for perceived visual changes over the entire range of conditions, from clear to hazy.
Under many scenic conditions, the average person can generally perceive a change of one
deciview. The higher the deciview value, the worse the visibility. Thus, an improvement in
visibility is a decrease in deciview value.
As shown in Table 1.5-2, in 2030 we estimate visibility in the East to be about 19 deciviews
(or visual range of 60 kilometers) on average, with poorer visibility in urban areas, compared to
the visibility conditions without man-made pollution of 9.5 deciviews (or visual range of 150
kilometers). Likewise, in we estimate visibility in the West to be about 9.5 deciviews (or visual
range of 150 kilometers) in 2030, compared to the visibility conditions without man-made
pollution of 5.3 deciviews (or visual range of 230 kilometers). Thus, in the future, a substantial
percent of the population may experience unacceptable visibility impairment in areas where they
live, work and recreate.
1-40
-------
Chapter 1: Health and Welfare Concerns
Table 1.5-2
Summary of 2030 National Visibility Conditions Based on
REMSAD Modeling (Deciviews)
Regions'5
Eastern U.S.
Urban
Rural
Western U.S.
Urban
Rural
Predicted 2030 Visibility"
(annual average)
18.98
20.48
18.38
9.54
10.21
9.39
Natural
Background Visibility
9.5
5.3
a The results incorporate earlier emissions estimates from the engines subject to this rule. We have revised our estimates both
upwards for some categories and downwards for others based on public comment and updated information; however, on net, we
believe that the results would underestimate future PM emissions.
Eastern and Western Regions are separated by 100 degrees north longitude. Background visibility conditions differ by region.
The emissions from nonroad engines generally, and in particular the engines subject to this
rule, contribute to this visibility impairment shown in Table 1.5-2. Nonroad engines emissions
contribute a large portion of the total PM emissions from mobile sources and anthropogenic
sources, in general. These emissions occur in and around areas with PM levels above the annual
PM2.5 NAAQS. The engines subject to the final rule will contribute to these effects. They are
estimated to emit 36,500 tons of direct PM in 2030, which is 1.1 percent of the total
anthropogenic PM emissions in 2030. Similarly, for PM precursors, the engines subject to this
rule will emit 640,000 tons of NOx and 1,411,000 tons HC in 2030, which are 3.8 and 8.3
percent of the total anthropogenic NOx and HC emissions, respectively, in 2030. Recreational
vehicles in particular contribute to these levels. In Table I.E-1 through I.E-3, we show that
recreational vehicles emitted about 1.7 percent of mobile source PM emissions in 2000.
Similarly, recreational vehicles are modeled to emit over 4 percent of mobile source PM in 2020
and 2030. Thus, the emissions from these sources contribute to the visibility impairment
modeled for 2030 summarized in the table.
Snowmobiles are operated in and around areas with PM2.5 levels above the level of the
secondary NAAQS. For 20 counties across nine states, snowmobile trails are found within or
near counties that registered ambient PM2 5 concentrations at or above 15 |ig/m3, the level of the
PM25 NAAQS.7 These counties are listed in Table 1.5-3. To obtain the information about
7 Memo to file from Terence Fitz-Simons, OAQPS, Scott Mathias, OAQPS, Mike Rizzo,
Region 5, "Analyses of 1999 PM Data for the PM NAAQS Review," November 17, 2000, with
attachment B, 1999 PM2.5 Annual Mean and 98th Percentile 24-Hour Average Concentrations.
1-41
-------
Draft Regulatory Support Document
snowmobile trails contained in the table, we consulted snowmobile trail maps that were supplied
by various states.60 Fine particles may remain suspended for days or weeks and travel hundreds to
thousands of kilometers, and thus fine particles emitted or created in one county may contribute
to ambient concentrations in a neighboring county.8
Docket No. A-2000-01, Document No. II-B-17.
8Review of the National Ambient Air Quality Standards for Particulate Matter: Policy
Assessment for Scientific and Technical Information, OAQPS Staff Paper, EPA-452VR-96-013,
July, 1996, at IV-7. This document is available from Docket A-99-06, Document U-A-23.
1-42
-------
Chapter 1: Health and Welfare Concerns
Table 1.5-3
Counties with Annual PM2 5 Levels Above 16 ^ig/m3 and Snowmobile Trails
State
Ohio
Montana
California
Minnesota
Wisconsin
Oregon
Pennsylvania
Illinois
Iowa
PM2 5 Exceedances
County
Machining
Trumbull
Summit
Montgomery
Portage
Franklin
Marshall/Ohio (WV)
Lincoln
Tulane
Butte
Fresno
Kern
Washington
Wright
Waukesha
Milwaukee
Jackson
Klammath
Washington
Rock Island
Rock Island (IL)
County with
Snowmobile Trails
Machining
Trumbull
Summit
Montgomery
Portage
Delaware
Belmont
Lincoln
Tulane
Butte
Fresno
Kern
Washington
Wright
Waukesha
Milwaukee
Douglas
Douglas
Layette
Somerset
Rock Island
Henry
Dubuque
Proximity to PM2 5
Exceedances County
Same County
Same County
Same County
Same County
Same County
Borders North
Borders West
Same County
Same County
Same County
Same County
Same County
Same County
Same County
Same County
Same County
Borders NNE
Borders North
Borders East
—
Same County
Borders East
Borders West
1-43
-------
Draft Regulatory Support Document
Achieving the annual PM2 5 NAAQS will help improve visibility across the country, but it
will not be sufficient (64 FR 35722 July 1, 1999 and 62 FR July 18, 1997 PM NAAQS). In
setting the NAAQS, EPA discussed how the NAAQS in combination with the regional haze
program, is deemed to improve visibility consistent with the goals of the CAA. In the East, there
are wide areas above 15 ug/m3 and light extinction is significantly above natural background.
Thus, large areas of the Eastern United States have air pollution that is causing unacceptable
visibility problems. In the West, scenic vistas are especially important to public welfare.
Although the annual PM25 NAAQS is met in most areas outside of California, virtually the entire
West is in close proximity to a scenic Class I area protected by 169A and 169B of the CAA.
1.5.3 Visibility Impairment in Class I Areas
The Clean Air Act establishes special goals for improving visibility in many national parks,
wilderness areas, and international parks. In the 1977 amendments to the Clean Air Act,
Congress set as a national goal for visibility the "prevention of any future, and the remedying of
any existing, impairment of visibility in mandatory class I Federal areas which impairment
results from manmade air pollution" (CAA section 169A(a)(l)). The Amendments called for
EPA to issue regulations requiring States to develop implementation plans that assure
"reasonable progress" toward meeting the national goal (CAA Section 169A(a)(4)). EPA issued
regulations in 1980 to address visibility problems that are "reasonably attributable" to a single
source or small group of sources, but deferred action on regulations related to regional haze, a
type of visibility impairment that is caused by the emission of air pollutants by numerous
emission sources located across a broad geographic region. At that time, EPA acknowledged that
the regulations were only the first phase for addressing visibility impairment. Regulations
dealing with regional haze were deferred until improved techniques were developed for
monitoring, for air quality modeling, and for understanding the specific pollutants contributing to
regional haze.
In the 1990 Clean Air Act amendments, Congress provided additional emphasis on regional
haze issues (see CAA section 169B). In 1999 EPA finalized a rule that calls for States to
establish goals and emission reduction strategies for improving visibility in all 156 mandatory
Class I national parks and wilderness areas. In this rule, EPA established a "natural visibility"
goal.61 In that rule, EPA also encouraged the States to work together in developing and
implementing their air quality plans. The regional haze program is focused on long-term
emissions decreases from the entire regional emissions inventory comprised of major and minor
stationary sources, area sources and mobile sources. The regional haze program is designed to
improve visibility and air quality in our most treasured natural areas from these broad sources.
At the same time, control strategies designed to improve visibility in the national parks and
wilderness areas will improve visibility over broad geographic areas. In the PM NAAQS
rulemaking, EPA also anticipated the need in addition to the NAAQS and Section 169 regional
haze program to continue to address localized impairment that may relate to unique
circumstances in some Western areas. For mobile sources, there may also be a need for a Federal
role in reduction of those emissions, in particular, because mobile sources are regulated primarily
1-44
-------
Chapter 1: Health and Welfare Concerns
at the federal level.
As described above, regional haze is caused by the emission from numerous sources located
over a wide geographic area.62 Visibility impairment is caused by pollutants (mostly fine
particles and precursor gases) directly emitted to the atmosphere by several activities (such as
electric power generation, various industry and manufacturing processes, truck and auto
emissions, construction activities, etc.). These gases and particles scatter and absorb light,
removing it from the sight path and creating a hazy condition. Visibility impairment is caused by
both regional haze and localized impairment.
Because of evidence that fine particles are frequently transported hundreds of miles, all 50
states, including those that do not have Class I areas, participate in planning, analysis and, in
many cases, emission control programs under the regional haze regulations. Even though a given
State may not have any Class I areas, pollution that occurs in that State may contribute to
impairment in Class I areas elsewhere. The rule encourages states to work together to determine
whether or how much emissions from sources in a given state affect visibility in a downwind
Class I area.
The regional haze program calls for states to establish goals for improving visibility in
national parks and wilderness areas to improve visibility on the haziest 20 percent of days and to
ensure that no degradation occurs on the clearest 20 percent of days. The rule requires states to
develop long-term strategies including enforceable measures designed to meet reasonable
progress goals toward natural visibility conditions. Under the regional haze program, States can
take credit for improvements in air quality achieved as a result of other Clean Air Act programs,
including national mobile-source programs.9
As noted above, EPA issued regulations in 1980 to address Class I area localized visibility
impairment that is "reasonably attributable" to a single source or small group of sources. In 40
CFR Part 51.301 of the visibility regulations, visibility impairment is defined as "any humanly
perceptible change in visibility (light extinction, visual range, contrast, coloration) from that
which would have existed under natural conditions." States are required to develop
implementation plans that include long-term strategies for improving visibility in each Class I
area. The long-term strategies under the 1980 regulations should consist of measures to reduce
impacts from local sources and groups of sources that contribute to poor air quality days in the
9 Though a recent case, American Corn Growers Association v. EPA, 291F.3d 1(D.C .Cir
2002) vacated the BART provisions of the Regional Haze rule, the court denied industry's
challenge to EPA's requirement that state's SIPS provide for reasonable progress towards
achieving natural visibility conditions in national parks and wilderness areas and the "no
degradation" requirement. Industry did not challenge requirements to improve visibility on the
haziest 20 percent of days. The court recognized that mobile source emission reductions would
need to be a part of a long-term emission strategy for reducing regional haze. A copy of this
decision can be found in Docket A-2000-01, Document IV- A-l 13.
1-45
-------
Draft Regulatory Support Document
class I area. Types of impairment covered by these regulations includes layered hazes and visible
plumes. While these kinds of visibility impairment can be caused by the same pollutants and
processes as those that cause regional haze, they generally are attributed to a smaller number of
sources located across a smaller area. The Clean Air Act and associated regulations call for
protection of visibility impairment in Class I areas from localized impacts as well as broader
impacts associated with regional haze.
As part of the HD07 PM air quality modeling described above, we modeled visibility
conditions in the Class I areas nationally. The results by region are summarized in Table 1.5-4.
In Figure 1.5-1, we define the regions used in this analysis based on a visibility study.63 These
results show that visibility is impaired in most Class I areas and additional reductions from
behicles subject to this rule are needed to achieve the goals of the Clean Air Act of preserving
natural conditions in Class I areas.
Table 1.5-4
Summary of 2030 Visibility Conditions in Class I
Areas Based on REMSAD Modeling (Annual Average Deciview)
Region
Eastern
Southeast
Northeast/Midwest
Western
Southwest
California
Rocky Mountain
Northwest
National Class I Area Average
Predicted 2030
Visibility
25.02
21.00
8.69
11.61
12.30
15.44
14.04
Natural
Background
Visibility
9.5
5.3
a Regions are depicted in Figure 1-5.1. Background visibility conditions differ by region
based on differences in relative humidity and other factors: Eastern natural background is 9.5
deciviews (or visual range of 150 kilometers) and in the West natural background is 5.3
deciviews (or visual range of 230 kilometers).
b The results incorporate earlier emissions estimates from the engines subject to this rule. We have
revised our estimates both upwards for some categories and downwards for others based on public
comment and updated information; however, on net, we believe that the HD07 analyses underestimate
future PM emissions.
1-46
-------
Chapter 1: Health and Welfare Concerns
Figure 1.5-1. Visibility Regions for Continental U.S.
Note: Study regions were represented in the Chestnut and Rowe (1990a, 1990b) studies used in evaluating the
benefits of visibility improvements.
The overall goal of the regional haze program is to prevent future and remedy existing
visibility impairment in Class I areas. As shown by the future deciview estimates in Table 1.5-4,
additional emissions reductions will be needed from the broad set of sources that contribute,
including the emissions from engines subject to this rule.
1.5.4 Recreational Vehicles and Visibility Impairment in Class I Areas
This section presents information about the contribution of recreational vehicles to visibility
impairment in Class I areas. Although this discussion focuses primarily on snowmobiles, we
present information on other recreational vehicles as well. We use monitoring data to show that
many of the worst 20 percent of days in terms of visibility levels occur in the wintertime, when
snowmobiles are used. We also summarize air quality modeling information of future visibility
for Class I areas where snowmobiles are operated and a case study of localized impairment in a
national park.
1.5.4.1 Snowmobiles Emissions in Class I Areas
Emissions of HC from snowmobiles contribute to direct and secondary formation of fine
particulate matter which can cause a variety of adverse health and welfare effects, including
visibility impairment discussed above. This section presents snowmobile-related emissions
information for Class I areas where snowmobiles are operated as further evidence of their
1-47
-------
Draft Regulatory Support Document
contribution in Class I areas.
Ambient concentrations of fine particles are the primary pollutant responsible for visibility
impairment. The classes of fine particles principally responsible for visibility impairment are
sulfates, nitrates, organic carbon particles, elemental carbon, and crustal material. Hydrocarbon
emissions from automobiles, trucks, snowmobiles, and other industrial processes are common
sources of organic carbon. The organic carbon fraction of fine particles ranges from 47 percent
in western Class I areas such as Denali National Park, to 28 percent in Rocky Mountain National
Park, to 13 percent in Acadia National Park.64
The contribution of snowmobiles to elemental carbon and nitrates is relatively small. Their
contribution to sulfates is a function of fuel sulfur and is small and will decrease even more as
the sulfur content of their fuel decreases due to our recently finalized fuel sulfur requirements. In
the winter months, however, hydrocarbon emissions from snowmobiles can be significant, as
indicated in Table 1.5-5 and these HC emissions can contribute significantly to the organic
carbon fraction of fine particles which are largely responsible for visibility impairment. This is
because snowmobiles are typically powered by two-stroke engines that emit large amounts of
hydrocarbons. In Yellowstone, a park with high snowmobile usage during the winter months,
snowmobile hydrocarbon emissions can exceed 500 tons per year, as much as several large
stationary sources. Other parks with less snowmobile traffic are also impacted, though to a lesser
extent, by these hydrocarbon emissions.65
Table 1.5-5
1999 Winter Season Snowmobile Emissions in Selected Class I Areas (tons)
Class I area
Denali NP and Preserve
Grand TetonNP
Rocky Mountain NP
Voyager NP
Yellowstone NP
HC
>9.8
13.7
106.7
138.5
492
CO
>26.1
36.6
284.7
369.4
1311.9
NOx
>0.08
0.1
0.8
1.1
3.8
PM
>0.24
0.3
2.6
3.4
12
Source: Letter from Aaron J. Worstell, Environmental Engineer, National Park Service, Air Resources Division, to Drew Kodjak, August 21, 2001,
particularly Table 1. Docket No. A-2000-01, Document No. II-G-178.
The national park areas outside of Denali in Alaska are open to snowmobile operation in
accordance with special regulations (36 CFR Part 7). Denali National Park permits snowmobile
operation by local rural residents engaged in subsistence uses (36 CFR Part 13). Emission
calculations are based on an assumed 2 hours of use per snowmobile visit at 16 hp with the
exception of Yellowstone where 4 hours of use at 16 hp was assumed. The emission factors used
to estimate these emissions are identical to those used by the NONROAD model. Two-stroke
snowmobile emission factors are: 111 g/hp-hr HC, 296 g/hp-hr CO, 0.86 g/hp-hr NOx, and 2.7
1-48
-------
Chapter 1: Health and Welfare Concerns
g/hp-hr PM. These emission factors are based on a number of engine tests performed by the
International Snowmobile Manufacturers Association (ISMA) and the Southwest Research
Institute (SwRI).
1.5.4.2 Air Quality Monitoring Information
To explore whether recreational vehicles, such as snowmobiles, contribute to visibility
impairment in Class I areas, we examine current monitored PM levels. Visibility and particulate
monitoring data are available for 8 Class I areas where snowmobiles are commonly used. These
are Acadia, Boundary Waters, Denali, Mount Ranier, Rocky Mountain, Sequoia and Kings
Ganyon, Voyager, and Yellowstone. Monitored fine particle data for these parks are set out in
Table 1.5-6. This table shows the number of monitored days in the winter that fell within the 20-
percent haziest days for each of these eight parks. Monitors collect data two days a week for a
total of about 104 days of monitored values. Thus, for a particular site, a maximum of 21 worst
possible days of these 104 days with monitored values constitute the set of 20-percent haziest
days during a year which are tracked as the primary focus of regulatory efforts.66 With the
exception of Denali in Alaska, we defined the snowmobile season as January 1 through March 15
and December 15 through December 31 of the same calendar year, consistent with the
methodology used in the Regional Haze Rule, which is calendar-year based. For Denali, Alaska,
the snowmobile season is October 1 to April 30.
1-49
-------
Draft Regulatory Support Document
Table 1.5-6
Winter Days That Fall Within the 20 Percent Worst Visibility Days
At National Parks Where Snowmobiles Are Operated
Class I Area
Acadia NP
Denali NP and Preserve
Mount Rainier NP
Rocky Mountain NP
Sequoia and Kings Canyon NP
Voyager NP
(1989-1992)
- Boundary Waters USFS
Wilderness Area (close to
Voyaguers with recent data)
Yellowstone NP
State(s)
ME
AK
WA
CO
CA
MN
MN
ID, MT, WY
Number of Sampled Wintertime Days
Within 20 Percent Worst Visibility Days
(maximum of 21 out of 104 monitored days)
1996
4
10
1
2
4
1989
o
5
2
0
1997
4
10
3
1
9
1990
4
5
2
1998
2
12
1
2
1
1991
6
1
0
1999
1
9
1
1
8
1992
8
5
0
Source: Letter from Debra C. Miller, Data Analyst, National Park Service, to Drew Kodjak, August 22, 2001. Docket No. A-2000-01.
1.5.4.3 Future Visibility Impairment in Class I Areas: Regional Haze
We also examined future air quality information to whether the emissions from recreational
vehicles, such as snowmobiles, contribute to regional visibility impairment in Class I areas. We
present results from the HD07 future air quality modeling described above for these Class I areas
in addition to inventory and air quality measurements. Specifically, in Table 1-5.7, we
summarize the expected future visibility conditions in these areas without these regulations.
1-50
-------
Chapter 1: Health and Welfare Concerns
Table 1.5-7
Estimated 2030 Visibility in Selected Class I Areas
Class I Area
Eastern areas
Acadia
Boundary Waters
Voyager
Western areas
Grand TetonNP
Kings Canyon
Mount Rainier
Rocky Mountain
Sequoia-Kings
Yellowstone
County
Hancock Co
St. Louis Co
St. Louis Co
Teton Co
Fresno Co
Lewis Co
Larimer Co
Tulare Co
Teton Co
State
ME
MN
MN
WY
CA
WA
CO
CA
WY
Predicted 2030 Visibility
(annual average deciview)
23.42
22.07
22.07
11.97
10.39
16.19
8.11
9.36
11.97
Natural Background
Visibility
(annual average
deciview)
9.5
5.3
a Natural background visibility conditions differ by region because of differences in factors such as relative humidity: Eastern natural
background is 9.5 deciviews (or visual range of 150 kilometers) and in the West natural background is 5.3 deciviews (or visual range
of 230 kilometers).
The results incorporate earlier emissions estimates from the engines subject to this rule. We have revised our estimates both
upwards for some categories and downwards for others based on public comment and updated information; however, on net, we
believe that HD07 analysis would underestimate future PM emissions from these categories.
In these areas, snowmobiles represent a signficant part of wintertime visibility-impairing
emissions. In fact, as the following discussion shows, snowmobile emissions can even be a
sizable percentage of annual emissions in some Class I areas. The snowmobiles thus are a
significant contributor to visibility impairment in these areas during the winter. As indicated,
winter days can often be among the worst visibility impairment. In addition, as the CAA
specifically states a goal of prevention and of remedying of any impairment of visibility in Class
I areas, the contribution of snowmobiles to visibility impairment even on winter days that are not
among the days of greatest impairment is a contribution to pollution that may reasonably be
anticipated to endanger public welfare and is properly regulated in this rule.
The information presented in Table 1.5-6 shows that visibility data supports a conclusion
that there are at least 8 Class I areas frequented by snowmobiles with one or more wintertime
days within the 20-percent worst visibility days of the year. For example, Rocky Mountain
National Park in Colorado was frequented by about 27,000 snowmobiles during the 1998-1999
winter. Of the monitored days characterized as within the 20-percent worst visibility monitored
days, 2 of those days occurred during the wintertime when snowmobile emissions such as HC
contributed to visibility impairment. The information in Table 1.5-7 shows that these areas also
1-51
-------
Draft Regulatory Support Document
have high predicted annual average deciview levels in the future. According to the National Park
Service, "[significant differences in haziness occur at all eight sites between the averages of the
clearest and haziest days. Differences in mean standard visual range on the clearest and haziest
days fall in the approximate range of 115-170 km."67
1.5.4.4 Localized Visibility Impairment in Class I Areas: Yellowstone National Park
The Class I are with the most detailed analysis of snowmobile contribution is Yellowstone
National Park. This provides an example of the extent to which snowmobiles can contribute to
emissions that can cause visibility impairment in Class I areas. Annual and particularly
wintertime hydrocarbon emissions from snowmobiles are high in the five parks considered in
Table 1.5-7, with two parks having HC emissions nearly as high as Yellowstone (Rocky
Mountain and Voyageurs). The proportion of snowmobile emissions to emissions from other
sources affecting air quality in these parks is likely to be similar to that in Yellowstone.
Inventory analysis performed by the National Park Service for Yellowstone National Park
suggests that snowmobile emissions can be a significant source of total annual mobile source
emissions for the park year round. Table 1.5-8 shows that in the 1998 winter season
snowmobiles contributed 64 percent, 39 percent, and 30 percent of HC, CO, and PM emissions.68
When the emission factors used by EPA in its NONROAD model are used, the contribution of
snowmobiles to total emissions in Yellowstone is still high: 59 percent, 33 percent, and 45
percent of HC, CO and PM emissions. The University of Denver used remote-sensing
equipment to estimate snowmobile HC emissions at Yellowstone during the winter of 1998-
1999, and estimated that snowmobiles contribute 77 percent of annual HC emissions at the
park.69 The portion of wintertime emissions attributable to snowmobiles is even higher, since all
snowmobile emissions occur during the winter months.
Table 1.5-8
1998 Annual HC Emissions (tons per year), Yellowstone National Park
Source
Coaches
Autos
RVs
Snowmobiles
Buses
TOTAL
HC
2.69
307.17
15.37
596.22
4.96
926.4
0%
33%
2%
64%
1%
CO
24.29
2,242.12
269.61
1,636.44
18.00
4190.46
1%
54%
6%
39%
0%
NOx
0.42
285.51
24.33
1.79
13.03
325.08
0%
88%
7%
1%
4%
PM
0.01
12.20
0.90
6.07
1.07
20.25
0%
60%
4%
30%
5%
Source: National Park Service, February 2000. Air Quality Concerns Related to Snowmobile Usage in National Parks. Air Docket A-2000-01,
Document No. II-A-44.
1-52
-------
Chapter 1: Health and Welfare Concerns
As part of public comments, Sierra Research conducted modeling of local impairment using
EPA's SCREENS Model Version 96043. This methodology consists of a single source Gaussian
plume model, which provides maximum ground-level concentrations for point, area, flare, and
volume sources, as well as concentrations in the cavity zone and concentrations due to inversion
break-up and shoreline fumigation.
The Sierra Research modeling demonstrated that there is up to an 8 percent contribution to
visibility degradation from snowmobile exhaust based on worst case conditions in Yellowstone
national park. It should be noted that SCREENS is not an EPA-approved model for conducting
visibility modeling. In interpreting the results of this modeling, the International Snowmobile
Manufacturers Association (ISMA) notes that the conversion factors used by SCREENS are
"conservatively high" and meant for worst case conditions, where there is a "pronounced [wind]
polarity... such as where a sea breeze exists."70 Consequently, ISMA appears to believe that data
gathered away from a coastline would actually have a lower demonstrated visual impact than the
impact determined by the model. Even using this modeling, ISMA presents modeling results that
support an 8 percent contribution to visibility impairment. ISMA reasons that by using the same
model for automobiles, the impairment contribution is double of what was expected, and
therefore, the 8 percent is most likely double of what it should be. As a result, ISMA concludes
an up to 4% contribution to visibility impairment from snowmobile emissions in national parks
"on best visibility days."71 Though the contribution levels in this industry-sponsored study are
lower than those discussed above, and though we have some concerns with this study, as
discussed in the Summary and Analysis of Comments, they still confirm that snowmobiles are
indeed a significant contributor to visibility degradation in Yellowstone.
In addition to the national modeling presented in Tables 1.4-3, 1.5-1, and 1.5-6, we also
conducted local-scale modeling using an EPA-approved visibility model, VISCREEN Version
1.01, to evaluate whether current emissions from recreational vehicles, such as snowmobiles,
contribute to localized visibility impairment in Class I areas. This analysis focused on localized
visibility impairments in Yellowstone National Park.72 The VISCREEN model is a visibility
screening level-I and -II model that characterizes point source plumes and visibility effects at 34
lines of sight. Thus, in this modeling, EPA treated snowmobiles as a synthetic point source in
order to determine plume perceptibility effects in a national park.
Using VISCREEN Version 1.01, we determined plume perceptibility from snowmobile
usage at four entrances (North, South, East, and West) in Yellowstone National Park as a case
study of visibility impairment from recreational vehicles. We conclude that plume perceptibility
would be noticeable at all entrances, even at the North entrance where the smallest numbers of
snowmobiles enter. Variations in the parameters concluded that perceptibility increased as the
observer neared the plume and at smaller plume-offset angles. As well, a sensitivity analysis was
conducted in order to demonstrate visibility impairment when the source is located within the
Class I boundaries and concluded that visibility impairment increases if the source is located
within the boundary. This provides further proof that snowmobile usage can lead to visibility
impairment at Yellowstone.
1-53
-------
Draft Regulatory Support Document
These results all indicate that snowmobiles contribute to visibility impairment concerns in
Yellowstone National Park, a Class I area.
1.6 Gaseous Air Toxics
In addition to the human health and welfare impacts described above, emissions from the
engines covered by this rulemaking also contain several other substances that are known or
suspected human or animal carcinogens, or have serious non-cancer health effects. These
include benzene, 1,3-butadiene, formaldehyde, acetaldehyde, acrolein, and toluene. The health
effects of these air toxics are highlighted below. Additional information can also be found in the
Technical Support Document four our final Mobile Source Air Toxics rule.73
1.6.1 Benzene
Benzene is an aromatic hydrocarbon which is present as a gas in both exhaust and
evaporative emissions from motor vehicles. Benzene in the exhaust, expressed as a percentage
of total organic gases (TOG), varies depending on control technology (e.g., type of catalyst) and
the levels of benzene and other aromatics in the fuel, but is generally about three to five percent.
The benzene fraction of evaporative emissions depends on control technology and fuel
composition and characteristics (e.g., benzene level and the evaporation rate), and is generally
about one percent.74
EPA has recently reconfirmed that benzene is a known human carcinogen by all routes of
exposure.75 Respiration is the major source of human exposure. Long-term respiratory exposure
to high levels of ambient benzene concentrations has been shown to cause cancer of the tissues
that form white blood cells. Among these are acute nonlymphocytic leukemia,76 chronic
lymphocytic leukemia and possibly multiple myeloma (primary malignant tumors in the bone
marrow), although the evidence for the latter has decreased with more recent studies.77'78
Leukemias, lymphomas, and other tumor types have been observed in experimental animals
exposed to benzene by inhalation or oral administration. Exposure to benzene and/or its
metabolites has also been linked with genetic changes in humans and animals79 and increased
proliferation of mouse bone marrow cells.80 The occurrence of certain chromosomal changes in
individuals with known exposure to benzene may serve as a marker for those at risk for
contracting leukemia.81
A number of adverse non-cancer health effects, blood disorders such as preleukemia and
aplastic anemia, have also been associated with low-dose, long-term exposure to benzene.82
People with long-term exposure to benzene may experience harmful effects on the blood-forming
tissues, especially the bone marrow. These effects can disrupt normal blood production and
cause a decrease in important blood components, such as red blood cells and blood platelets,
leading to anemia (a reduction in the number of red blood cells), leukopenia (a reduction in the
number of white blood cells), or thrombocytopenia (a reduction in the number of blood platelets,
thus reducing the ability for blood to clot). Chronic inhalation exposure to benzene in humans
1-54
-------
Chapter 1: Health and Welfare Concerns
and animals results in pancytopenia,83 a condition characterized by decreased numbers of
circulating erythrocytes (red blood cells), leukocytes (white blood cells), and thrombocytes
(blood platelets).84'85 Individuals that develop pancytopenia and have continued exposure to
benzene may develop aplastic anemia,86 whereas others exhibit both pancytopenia and bone
marrow hyperplasia (excessive cell formation), a condition that may indicate a preleukemic
state.87 88 The most sensitive non-cancer effect observed in humans is the depression of absolute
lymphocyte counts in the circulating blood.89
1.6.2 1,3-Butadiene
1,3-Butadiene is formed in vehicle exhaust by the incomplete combustion of fuel. It is not
present in vehicle evaporative emissions, because it is not present in any appreciable amount in
fuel. 1,3-Butadiene accounts for 0.4 to 1.0 percent of total organic gas exhaust, depending on
control technology and fuel composition.90
1,3-Butadiene was classified by EPA as a Group B2 (probable human) carcinogen in 1985.91
This classification was based on evidence from two species of rodents and epidemiologic data.
In the EPA1998 draft Health Risk Assessment of 1,3-Butadiene, that was reviewed by the
Science Advisory Board (SAB), the EPA proposed that 1,3-butadiene is a known human
carcinogen based on human epidemiologic, laboratory animal data, and supporting data such as
the genotoxicity of 1,3-butadiene metabolites.92 The Environmental Health Committee of EPA's
Scientific Advisory Board (SAB) reviewed the draft document in August 1998 and recommended
that 1,3-butadiene be classified as a probable human carcinogen, stating that designation of 1,3-
butadiene as a known human carcinogen should be based on observational studies in humans,
without regard to mechanistic or other information.93 In applying the 1996 Guidelines for
Carcinogen Risk Assessment, the Agency relies on both observational studies in humans as well
as experimental evidence demonstrating causality, and therefore the designation of 1,3-butadiene
as a known human carcinogen remains applicable.94 The Agency has revised the draft Health
Risk Assessment of 1,3-Butadiene based on the SAB and public comments. The draft Health
Risk Assessment of 1,3-Butadiene will undergo the Agency consensus review, during which time
additional changes may be made prior to its public release and placement on the Integrated Risk
Information System (IRIS).
1,3-Butadiene also causes a variety of non-cancer reproductive and developmental effects in
mice and rats (no human data) when exposed to long-term, low doses of butadiene.95 The most
sensitive effect was reduced litter size at birth and at weaning. These effects were observed in
studies in which male mice exposed to 1,3-butadiene were mated with unexposed females. In
humans, such an effect might manifest itself as an increased risk of spontaneous abortions,
miscarriages, still births, or very early deaths. Long-term exposures to 1,3-butadiene should be
kept below its reference concentration of 4.0 microgram/m3 to avoid appreciable risks of these
reproductive and developmental effects.96 EPA has developed a draft chronic, subchronic, and
acute RfC values for 1,3-butadiene exposure as part of the draft risk characterization mentioned
above. The RfC values will be reported on IRIS.
1-55
-------
Draft Regulatory Support Document
1.6.3 Formaldehyde
Formaldehyde is the most prevalent aldehyde in vehicle exhaust. It is formed from
incomplete combustion of both gasoline and diesel fuel and accounts for one to four percent of
total organic gaseous emissions, depending on control technology and fuel composition. It is not
found in evaporative emissions.
Formaldehyde exhibits extremely complex atmospheric behavior.97 It is formed by the
atmospheric oxidation of virtually all organic species, including biogenic (produced by a living
organism) hydrocarbons. Mobile sources contribute both primary formaldehyde (emitted directly
from motor vehicles) and secondary formaldehyde (formed from photooxidation of other VOCs
emitted from vehicles).
EPA has classified formaldehyde as a probable human carcinogen based on limited evidence
for carcinogen!city in humans and sufficient evidence of carcinogen!city in animal studies, rats,
mice, hamsters, and monkeys.98 Epidemiological studies in occupationally exposed workers
suggest that long-term inhalation of formaldehyde may be associated with tumors of the
nasopharyngeal cavity (generally the area at the back of the mouth near the nose), nasal cavity,
and sinus. Studies in experimental animals provide sufficient evidence that long-term inhalation
exposure to formaldehyde causes an increase in the incidence of squamous (epithelial) cell
carcinomas (tumors) of the nasal cavity. The distribution of nasal tumors in rats suggests that not
only regional exposure but also local tissue susceptibility may be important for the distribution of
formaldehyde-induced tumors.99 Research has demonstrated that formaldehyde produces
mutagenic activity in cell cultures.100
Formaldehyde exposure also causes a range of non-cancer health effects. At low
concentrations (0.05-2.0 ppm), irritation of the eyes (tearing of the eyes and increased blinking)
and mucous membranes is the principal effect observed in humans. At exposure to 1-11 ppm,
other human upper respiratory effects associated with acute formaldehyde exposure include a dry
or sore throat, and a tingling sensation of the nose. Sensitive individuals may experience these
effects at lower concentrations. Forty percent of formaldehyde-producing factory workers
reported nasal symptoms such as rhinitis (inflammation of the nasal membrane), nasal
obstruction, and nasal discharge following chronic exposure.101 In persons with bronchial
asthma, the upper respiratory irritation caused by formaldehyde can precipitate an acute
asthmatic attack, sometimes at concentrations below 5 ppm.102 Formaldehyde exposure may also
cause bronchial asthma-like symptoms in non-asthmatics.103104
Immune stimulation may occur following formaldehyde exposure, although conclusive
evidence is not available. Also, little is known about formaldehyde's effect on the central
nervous system. Several animal inhalation studies have been conducted to assess the
developmental toxicity of formaldehyde. The only exposure-related effect noted in these studies
was decreased maternal body weight gain at the high-exposure level. No adverse effects on
reproductive outcome of the fetuses that could be attributed to treatment were noted. An
1-56
-------
Chapter 1: Health and Welfare Concerns
inhalation reference concentration (RfC), below which long-term exposures would not pose
appreciable non-cancer health risks, is not available for formaldehyde at this time.
1.6.4 Acetaldehyde
Acetaldehyde is a saturated aldehyde that is found in vehicle exhaust and is formed as a
result of incomplete combustion of both gasoline and diesel fuel. It is not a component of
evaporative emissions. Acetaldehyde comprises 0.4 to 1.0 percent of total organic gas exhaust,
depending on control technology and fuel composition.105
The atmospheric chemistry of acetaldehyde is similar in many respects to that of
formaldehyde.106 Like formaldehyde, it is produced and destroyed by atmospheric chemical
transformation. Mobile sources contribute to ambient acetaldehyde levels both by their primary
emissions and by secondary formation resulting from their VOC emissions. Acetaldehyde
emissions are classified as a probable human carcinogen. Studies in experimental animals
provide sufficient evidence that long-term inhalation exposure to acetaldehyde causes an increase
in the incidence of nasal squamous cell carcinomas (epithelial tissue) and adenocarcinomas
(glandular tissue).107 108
Non-cancer effects in studies with rats and mice showed acetaldehyde to be moderately toxic
by the inhalation, oral, and intravenous routes.109 no 1U The primary acute effect of exposure to
acetaldehyde vapors is irritation of the eyes, skin, and respiratory tract. At high concentrations,
irritation and pulmonary effects can occur, which could facilitate the uptake of other
contaminants. Little research exists that addresses the effects of inhalation of acetaldehyde on
reproductive and developmental effects. The in vitro and in vivo studies provide evidence to
suggest that acetaldehyde may be the causative factor in birth defects observed in fetal alcohol
syndrome, though evidence is very limited linking these effects to inhalation exposure. Long-
term exposures should be kept below the reference concentration of 9 |j.g/m3 to avoid appreciable
risk of these non-cancer health effects.112
1.6.5 Acrolein
Acrolein is extremely toxic to humans from the inhalation route of exposure, with acute
exposure resulting in upper respiratory tract irritation and congestion. Although no information
is available on its carcinogenic effects in humans, based on laboratory animal data, EPA
considers acrolein a possible human carcinogen.113
1.6.6 Toluene
Toluene is a known respiratory irritant with central nervous system effects. Reproductive
toxicity has been observed in exposed humans and rats.114 Toluene toxicity is most prominent in
the central nervous system after acute and chronic exposure, and that the brain is the principal
target organ for toluene toxicity in humans. Specifically, recent studies indicate that toluene and
other similar solvents alter the function of ion channels in neuronal membranes, including
1-57
-------
Draft Regulatory Support Document
receptors stimulated by y-amino butyric acid (GABA), w-methyl-o-aspartate (NMDA), nicotinic
acetylcholine (nACh), and those sensitive to membrane voltage.115'116> 117'118'119 Anesthetic
agents, ethanol, toluene, and other solvents inhibit the function of receptors that are excitatory in
the nervous system (NMDA, nACh), and enhance the function of inhibitory receptors
(GAB A).120'121 Thus, these compounds tend to suppress the activity of the nervous system,
yielding slowed reaction times, reduced arousal and, at high concentrations, anesthesia,
unconsciousness and respiratory failure.122
1.7 Exposure to CO and Air Toxics Associated with Nonroad Engines and
Vehicles
The previous section describes national-scale adverse public health effects associated with
the nonroad engines and vehicles covered by this rulemaking. This section describes significant
adverse health and welfare effects arising from the usage patterns of snowmobiles, large SI
engines, and gasoline marine engines on the regional and local scale. Studies suggest that
emissions from these engines can be concentrated in specific areas, leading to elevated ambient
concentrations of particular pollutants and associated elevated exposures to operators and by-
standers. This section describes these exposures.
1.7.1 Large SI Engines
Exhaust emissions from applications with significant indoor use can expose individual
operators or bystanders to dangerous levels of pollution. Forklifts, ice-surfacing machines,
sweepers, and carpet cleaning equipment are examples of large industrial spark-ignition engines
that often operate indoors or in other confined spaces. Forklifts alone account for over half of the
engines in this category. Indoor use may include extensive operation in a temperature-controlled
environment where ventilation is kept to a minimum (e.g., for storing, processing, and shipping
produce). Although our standards are not designed to eliminate occupational exposures, the
standards will reduce CO and HC emissions that contribute to those exposures.
The principal concern for human exposure relates to CO emissions. One study showed
several forklifts with measured CO emissions ranging from 10,000 to 90,000 ppm (1 to 9
percent).123 The threshold limit value for a time-weighted average 8-hour workplace exposure set
by the American Conference of Governmental Industrial Hygienists is 25 ppm.
One example of a facility that addressed exposure problems with new technology is in the
apple-processing field.124 Trout Apples in Washington added three-way catalysts to about 60
LPG-fueled forklifts to address multiple reports of employee health complaints related to CO
exposure. The emission standards are based on the same technologies installed on these in-use
engines.
1-58
-------
Chapter 1: Health and Welfare Concerns
Additional exposure concerns occur at ice rinks. Numerous papers have identified ice-
surfacing machines with spark-ignition engines as the source of dangerous levels of CO and NO2,
both for skaters and for spectators.125 This is especially problematic for skaters, who breathe air
in the area where pollutant concentration is highest, with higher respiration rates resulting from
their high level of physical activity. This problem has received significant attention from the
medical community.
In addition to CO emissions, HC emissions from these engines can also lead to increased
exposure to harmful pollutants, particularly air toxics. Since many gasoline or dual-fuel engines
are in forklifts that operate indoors, reducing evaporative emissions could have direct health
benefits to operators and other personnel. Fuel vapors can also cause odor problems.
1.7.2 Snowmobiles
In addition to their contribution to CO concentrations generally and visibility impairment,
snowmobile emissions are of concern because of their potential impacts on riders and on park
attendants, as well as other groups of people who are in contact with these vehicles for extended
periods of time.
Snowmobile users can be exposed to high air toxic and CO emissions, both because they sit
very close to the vehicle's exhaust port and because it is common for them to ride their vehicles
in lines or groups on trails where they travel fairly close behind other snowmobiles. Because of
these riding patterns, snowmobilers breathe exhaust emissions from their own vehicle, the
vehicle directly in front as well as those farther up the trail. This can lead to relatively high
personal exposure levels of harmful pollutants. A study of snowmobile rider CO exposure
conducted at Grand Teton National Park showed that a snowmobiler riding at distances of 25 to
125 feet behind another snowmobiler and traveling at speeds from 10 to 40 mph can be exposed
to average CO levels ranging from 0.5 to 23 ppm, depending on speed and distance. The highest
CO level measured in this study was 45 ppm, as compared to the current 1-hour NAAQS for CO
of 35 ppm.126 While exposure levels can be less if a snowmobile drives 15 feet off the centerline
of the lead snowmobile, the exposure levels are still of concern. This study led to the
development of an empirical model for predicting CO exposures from riding behind
snowmobiles.
Hydrocarbon speciation for snowmobile emissions was performed for the State of Montana
in a 1997 report.127 Using the dispersion model for CO from the Grand Teton exposure study
with air toxic emission rates from the State of Montana's emission study, average benzene
exposures for riders driving at an average speed of 23 mph, 25 feet behind another snowmobile
were predicted to be 0.402 ppm, (95% bootstrap confidence intervals = 0.285-0.555). Average
toluene concentrations in this scenario were modeled at 10.3 ppm (95% bootstrap CI = 8.1-12.8).
With an average speed of 23 mph with a 50 foot space between snowmobiles, average benzene
concentrations were estimated to be 0.210 ppm (95% bootstrap CI = 0.154 - 0.271).
The cancer risk posed to those exposed to benzene emissions from snowmobiles must be
1-59
-------
Draft Regulatory Support Document
viewed within the broader context of expected lifetime benzene exposure. Observed monitoring
data and predicted modeled values demonstrate that a significant cancer risk already exists from
ambient concentrations of benzene for a large portion of the US population. The Agency's 1996
National-Scale Air Toxics Assessment of personal exposure to ambient concentrations of air
toxic compounds emitted by outside sources (e.g., cars and trucks, power plants) found that
benzene was among the five air toxics appear to pose the greatest risk to people nationwide. This
national assessment found that for approximately 50% of the US population in 1996, the
inhalation cancer risks associated with benzene exceeded 10 in one million. Modeled predictions
for ambient benzene from this assessment correlated well with observed monitored
concentrations of benzene ambient concentrations.
Specifically, the draft National-Scale Assessment predicted nationwide annual average
benzene exposures from outdoor sources to be 1.4 |ig/m3.128 In comparison, snowmobile riders
and those directly exposed to snowmobile exhaust emissions had predicted benzene levels two to
three orders of magnitude greater than the 1996 national average benzene concentrations.129
These elevated levels are also known as air toxic "hot spots," which are of particular concern to
the Agency. Thus, total annual average exposures to typical ambient benzene concentrations
combined with elevated short-term exposures to benzene from snowmobiles may pose a
significant risk of adverse public health effects to snowmobile riders and those exposed to
exhaust benzene emissions from snowmobiles.
Toluene concentrations, also elevated in snowmobile plumes, were predicted to be within
the concentrations typically observed in occupational settings. While not considered a human
carcinogenic hazard, toluene at high concentrations can affect the central nervous system,
causing effects similar to intoxication. Weakness, confusion, euphoria, dizziness, and headache
are associated with high exposures to toluene. National Institute of Occupational Safety and
Health. NIOSH Pocket Guide to Chemical Hazards. NIOSH web site.
http://www.cdc.gov/niosh/npg/npgdQ619.html. Exposure to constituents of snowmobile exhaust at the
levels predicted is anticipated to cause such effects in the human central nervous system.
Since snowmobile riders often travel in large groups, the riders towards the back of the
group are exposed to the accumulated exhaust of those riding ahead. This scenario was not
modeled, given the lack of data on snowmobile plume concentrations in trains of several
vehicles. However, snowmobile trains, consisting of multiple riders in a line, are common riding
scenarios. In these conditions, exhaust concentrations are anticipated to be significantly higher
than those predicted here. These exposure levels can continue for hours at a time, depending on
the length of a ride. An additional consideration is that the risk to health from CO exposure
increases with altitude, especially for unacclimated individuals. Therefore, a park visitor who
lives at sea level and then rides his or her snowmobile on trails at high-altitude is more
susceptible to the effects of CO than local residents.
In addition to snowmobilers themselves, people who are active in proximity to the areas
where snowmobilers congregate may also be exposed to high CO levels. An OSHA industrial
hygiene survey reported a peak CO exposure of 268 ppm for a Yellowstone employee working at
1-60
-------
Chapter 1: Health and Welfare Concerns
an entrance kiosk where snowmobiles enter the park. This level is greater than the NIOSH peak
recommended exposure limit of 200 ppm. OSHA's survey also measured employees' exposures
to several air toxics. Benzene exposures in Yellowstone employees ranged from 67-600 |ig/m3,
with the same individual experiencing highest CO and benzene exposures. The highest benzene
exposure concentrations exceeded the NIOSH Recommended Exposure Limit of 0.1 ppm for 8-
hour exposures.
1-61
-------
Draft Regulatory Support Document
Notes to Chapter 1
1.Carbon monoxide also participates in the production of ozone, albeit at a much slower rate than
most VOC and NOx compounds.
2.U.S. EPA, 1996, Review of National Ambient Air Quality Standards for Ozone, Assessment of
Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007. A copy of this
document can be obtained from Air Docket A-99-06, Document No. II-A-22.
3.U.S. EPA, 1996, Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. The document is available on the internet at
http://www.epa.gov/ncea/ozone.htm. A copy can also be obtained from Air Docket No. A-99-
06, Documents Nos. II-A-15, II-A-16, U-A-17.
4.National Air Quality and Emissions Trends Report, 1999, EPA, 2001, at Table A-19. This
document is available at http://www.epa.gov/oar/aqtrnd99/. The data from the Trends report are
the most recent EPA air quality data that has been quality assured. A copy of this table can also
be found in Docket No. A-2000-01, Document No. II-A-64.
5 .National Air Quality and Emissions Trends Report, 1998, March, 2000, at 28. This document
is available at http ://www.epa. gov/oar/aqtrnd98/. Relevant pages of this report can be found in
Memorandum to Air Docket A-2000-01 from Jean Marie Revelt, September 5, 2001, Document
No. II-A-63.
6.National Air Quality and Emissions Trends Report, 1998, March, 2000, at 32. This document
is available at http ://www.epa. gov/oar/aqtrnd98/. Relevant pages of this report can be found in
Memorandum to Air Docket A-2000-01 from Jean Marie Revelt, September 5, 2001, Document
No. II-A-63.
7. Additional information about this modeling can be found in our Regulatory Impact Analysis:
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements, document EPA420-R-00-026, December 2000. Docket No. A-2000-01,
Document No. U-A-13. This document is also available at
http://www.epa.gov/otaq/diesel.htmtfdocuments.
8.We also performed ozone air quality modeling for the western United States but, as described
further in the air quality technical support document, model predictions were well below
corresponding ambient concentrations for out heavy-duty engine standards and fuel sulfur control
rulemaking. Because of poor model performance for this region of the country, the results of the
Western ozone modeling were not relied on for that rule.
9.U.S. EPA Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles:
Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements. EPA420-
R-99-023. December 1999. A copy of this document is also available in Docket A-97-10,
Document No. V-B-01.
1-62
-------
Chapter 1: Health and Welfare Concerns
10. Additional information about these studies can be found in Chapter 2 of "Regulatory Impact
Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements," December 2000, EPA420-R-00-026. Docket No. A-2000-01, Document
Number II-A-13. This document is also available at
http://www.epa.gov/otaq/diesel.htmtfdocuments.
11 .Air Quality Criteria Document for Ozone and Related Photochemical Oxidants, EPA National
Center for Environmental Assessment, July 1996, Report No. EPA/600/P-93/004cF. The
document is available on the internet at http://www.epa.gov/ncea/ozone.htm. A copy can also be
obtained from Air Docket No. A-99-06, Documents Nos. II-A-15, II-A-16, II-A-17.
12. A copy of this data can be found in Air Docket A-2000-01, Document No.U-A-80.
13.Memorandum to Docket A-99-06 from Eric Ginsburg, EPA, "Summary of Model-Adjusted
Ambient Concentrations for Certain Levels of Ground-Level Ozone over Prolonged Periods,"
November 22, 2000. Docket A-2000-01, Document Number II-B-13.
14.Memorandum to Docket A-99-06 from Eric Ginsburg, EPA, "Summary of Model-Adjusted
Ambient Concentrations for Certain Levels of Ground-Level Ozone over Prolonged Periods,"
November 22, 2000, at Table C, Control Scenario - 2020 Populations in Eastern Metropolitan
Counties with Predicted Daily 8-Hour Ozone greater than or equal to 0.080 ppm. Docket A-
2000-01, Document Number II-B-13.
15.U.S. EPA, 1995, Review of National Ambient Air Quality Standards for Nitrogen
Dioxide, Assessment of Scientific and Technical Information, OAQPS Staff Paper,
EPA-452/R-95-005.
16.U.S.EPA, 1993, Air Quality Criteria for Oxides of Nitrogen, EP A/600/8-9 l/049aF.
17.Much of the information in this subsection was excerpted from the EPA document, Human
Health Benefits from Sulfate Reduction., written under Title IV of the 1990 Clean Air Act
Amendments, U.S. EPA, Office of Air and Radiation, Acid Rain Division, Washington, DC
20460, November 1995. Air Docket A-2000-01, Document No. U-A-32.
IS.Vitousek, Peter M., John Aber, Robert W. Howarth, Gene E. Likens, et al. 1997. Human
Alteration of the Global Nitrogen Cycle: Causes and Consequences. Issues in Ecology. Published
by Ecological Society of America, Number 1, Spring 1997.
19. 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 document is available on the internet at
http://www.nap.edu/books/0309048443/html/
20.Much of this information was taken from the following EPA document: Deposition of Air
Pollutants to the Great Waters-Second Report to Congress, Office of Air Quality Planning and
1-63
-------
Draft Regulatory Support Document
Standards, June 1997, EPA-453/R-97-011.
21.Terrestrial nitrogen deposition can act as a fertilizer. In some agricultural areas, this effect can
be beneficial.
22. Coburn, R.F. (1979) Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322.
23. Helfaer, M.A., and Traystman, RJ. (1996) Cerebrovascular effects of carbon monoxide. In:
Carbon Monoxide (Penney, D.G., ed). Boca Raton, CRC Press, 69-86.
24. Benignus, V.A. (1994) Behavioral effects of carbon monoxide: meta analyses and
extrapolations. J. Appl. Physiol. 76:1310-1316. Docket A-2000-01, Document IV-A-127.
25. Rowe, B., Milner, R., Johnson, C. Bota, G. Snowmobile-Related Deaths in Ontario: A 5-
Year Review. Canadian Medical Association Journal, Vol. 146, Issue 2, pp 147-152. Docket
A-2000-01, Document IV-A-194.
26.The CO Criteria Document (EPA 600/P-99/001F) contains additional information about the
health effects of CO, human exposure, and air quality. It was published as a final document and
made available to the public in August 2000 (www.epa. gov/ncea/co/). A copy of this document
is also available in Docket A-2000-01, Document A-U-29.
27.National Air Quality and Emissions Trends Report, 1999, EPA, 2001, at Table A-19. This
document is available at http://www.epa.gov/oar/aqtrnd99/. The data from the Trends report are
the most recent EPA air quality data that has been quality assured. A copy of this table can also
be found in Docket No. A-2000-01, Document No. II-A-64.
28. Information attached to written comments, P. Amette, Vice President, Motorcycle Industry
Council, Incorporated. Docket A-2000-01, Document IV-D-214.
29. Economic Contribution of Off-Highway Vehicle Use in Colorado" Prepared for the Colorado
Of-Highway Vehicle Coalition, by Hazen and Sawer Environmental Engineers & Scientists.
July, 2001. Colorado OHV User Survey" Summary of Results: prepared for State of Colorado
OHV Coalition under a contract with the Colorado State Parks OHV Program, prepared by T.
Crimins, Trails Consultant. January 1999. Off Highway Vehicle Uses and Owners Preferences
in Uta", prepared for Utah DNR, Div. Of Parks and recreation, prepared by Institute for Outdoor
recreation and Tourism Department of Forest Resources, Utah State University. July 22, 2001.
These documents are available in Docket A-2000-01, Documents IV-A-02, 03, 05.
30. Off Highway Vehicle Uses and Owners Preferences in Uta", prepared for Utah DNR, Div. Of
Parks and recreation, prepared by Institute for Outdoor recreation and Tourism Department of
Forest Resources, Utah State University. July 22, 2001. This document is available in Docket
A-2000-01, Document IV-A-03.
1-64
-------
Chapter 1: Health and Welfare Concerns
31. All-Terrain Vehicle Exposure, Injury, Death and Risk Studies. U.S. Consumer Product
Safety Commission, April, 1998. Docket A-2000-01, Document IV-A-197.
32. Anchorage Carbon Monoxide Emission Inventory and Year 2000 Attainment Projections"
Air Quality Program, Environmental Services Division, Department of Health and Human
Services [DRAFT]. May, 2001. Docket A-2000-01, Document II-A-40.
33.Areas with a few years of attainment data can and often do have exceedances following such
years of attainment because of several factors including different climatic events during the later
years, increases in inventories, etc. Thus, a plan to maintain the NAAQS is critical to showing
attainment.
34.Dulla, Robert G. Sierra Research, Inc. "A Review of Vehicle Test Programs Conducted in
Alaska in Recent Years and a Summary of the Fairbanks Co. Inventory 1995-2001. June 4,
2001. Docket A-2000-01, Document IV-A-198.
35.St. Paul, Minnesota was recently reclassified as being in attainment but is still considered a
maintenance area. There is also a significant population of snowmobiles in Minnesota, with
snowmobile trails in Washington County.
36.The trail maps consulted for this rulemaking can be found in Docket No. A-2000-01,
Document No. U-A-65.
37. Written comments from J.S. Grumet, Executive Director, Northeast States for Coordinated
Use Management (NESCAUM), Docket A-2000-01, Document IV-D-196.
38. Doss, Howard. Snowmobile Safety. Michigan Agricultural Safety Health Center. A copy of
this document can be found in Docket A-2000-01, Document IV-A-148 (an attachment).
39. Mauer, Richard. "Snowmobile Perils" Anchorage Daily News. Internet search 7/3/02.
Docket A-2000-01, IV-A-184.
40.Dulla, Robert G. Sierra Research, Inc. "A Review of Vehicle Test Programs Conducted in
Alaska in Recent Years and a Summary of the Fairbanks Co. Inventory 1995-2001. June 4,
2001. Docket A-2000-01, Document IV-A-198.
41. Technical Memorandum to Docket A-2000-01 from Drew Kodjak, Attorney-Advisor, Office
of Transportation and Air Quality, "Air Quality Information for Selected CO Nonattainment
Areas," July 27, 2001, Docket Number A-2000-01, Document Number U-B-18.
42. Air Quality Criteria for Carbon Monoxide, US EPA, EPA 600/P-99/001F, June 2000, at 3-
38, Figure 3-32 (Federal Bldg, AIRS Site 020900002). Air Docket A-2000-01, Document
1-65
-------
Draft Regulatory Support Document
Number II-A-29. This document is also available at http://www.epa.gov/ncea/coabstract.htm.
43 .National Research Council. The Ongoing Challenge of Managing Carbon Monoxide
Pollution in Fairbanks, AK. May 2002. Docket A-2000-01, Document IV-A-115.
44. National Air Quality and Emissions Trends Report, 1999, EPA, 2001, at Table A-19. This
document is available at http://www.epa.gov/oar/aqtrnd99/. The data from the Trends report are
the most recent EPA air quality data that have been quality assured. A copy of this table can also
be found in Docket No. A-2000-01, Document No. II-A-64. See also the air quality update,
1998-2000 Ozone and 1999-2000 Carbon Monoxide, available at www.epa.gov/oar/aqtrndOO. A
copy of this document is also available at Docket A-2000-01, Document No. IV-A-141.
45. Air Quality and Emissions Trends Report, 1998, March, 2000. This document is available at
http ://www.epa. gov/oar/aqtrnd98/. Relevant pages of this report can be found in Memorandum
to Air Docket A-2000-01 from Jean Marie Revelt, September 5, 2001, Document No. II-A-63.
46.EPA (1996) Review of the National Ambient Air Quality Standards for Particulate Matter:
Policy Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA-452/R-96-
013. Docket Number A-99-06, Documents Nos. H-A-18, 19, 20, and 23. The particulate matter
air quality criteria documents are also available at http://www.epa.gov/ncea/partmatt.htm.
47. Additional information about the Regulatory Model System for Aerosols and Deposition
(REMSAD) and our modeling protocols can be found in our Regulatory Impact Analysis: Heavy-
Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements,
document EPA420-R-00-026, December 2000. Docket No. A-2000-01, Document No. A-II-13.
This document is also available at http://www.epa.gov/otaq/disel.htm#documents.
48.Memorandum to Docket A-99-06 from Eric O. Ginsburg, Senior Program Advisor,
"Summary of 1999 Ambient Concentrations of Fine Particulate Matter," November 15, 2000.
Air Docket A-2000-01, Document No. II-B-12.
49.Memorandum to Docket A-99-06 from Eric O. Ginsburg, Senior Program Advisor,
"Summary of Absolute Modeled and Model-Adjusted Estimates of Fine Particulate Matter for
Selected Years," December 6, 2000. This memo is also available in the docket for this rule.
Docket A-2000-01, Document Number II-B-14.
50.The fine particle monitoring network was expanding with more monitors being added
between 1996 and 2002.
51. Technical Memorandum, EPA Air Docket A-99-06, Eric O. Ginsburg, Senior Program
Advisor, Emissions Monitoring and Analysis Division, OAQPS, Summary of Absolute Modeled
and Model-Adjusted Estimates of Fine Particulate Matter for Selected Years, December 6, 2000,
Table P-2. Docket Number 2000-01, Document Number II-B-14.
1-66
-------
Chapter 1: Health and Welfare Concerns
52.National Research Council, 1993 (Ibid); U.S. EPA Criteria for Particulate Matter, 8-3; US
EPA Review of the National Ambient Air Quality Standards for Particulate Matter: Policy
Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA-452/R-96-013.
1996. Docket Number A-99-06, Documents Nos. II-A-18, 19, 20, and 23. The particulate matter
air quality criteria documents are also available at http://www.epa.gov/ncea/partmatt.htm.
53. Council on Environmental Quality, 1978. Visibility Protection for Class I Areas, the
Technical Basis. Washington DC. Cited in US EPA, Review of the National Ambient Air
Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical
Information. OAQPS Staff Paper. EPA-452 \ R-96-013. This document is available in Docket
A-99-06, Document U-A-23.
54. Sisler, James F. Spatial and Seasonal Patterns and Long Term Variability of the Composition
of Haze in the United States: An Analysis of Data from the IMPROVE Network. 1996. A copy
of the relevant pages of this document can be found in Docket A-99-06, Document No. U-B-21.
55. National Research Council, 1993 (Ibid).
56.National Research Council, 1993 (Ibid).
57. National Acid Precipitation Assessment Program (NAPAP), 1991. Office of the Director.
Acid Deposition: State of Science and Technology. Report 24, Visibility: Existing and Historical
Conditions - Causes and Effects. Washington, DC. Cited in US EPA, Review of the National
Ambient Air Quality Standards for Particulate Matter: Policy Assessment of Scientific and
Technical Information. OAQPS Staff Paper. EPA-452 \R-96-013. This document is available
in Docket A-99-06, Document II-A-23. Also, US EPA. Review of the National Ambient Air
Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical
Information, OAQPS Staff Paper. Preliminary Draft. June 2001. Docket A-2000-01, Document
IV-A-199.
58.Memorandum to Docket A-99-06 from Eric O. Ginsburg, Senior Program Advisor,
"Summary of 1999 Ambient Concentrations of Fine Particulate Matter," November 15, 2000.
Air Docket A-2000-01, Document No. II-B-12.
59. Technical Memorandum, EPA Air Docket A-99-06, Eric O. Ginsburg, Senior Program
Advisor, Emissions Monitoring and Analysis Division, OAQPS, Summary of Absolute Modeled
and Model-Adjusted Estimates of Fine Particulate Matter for Selected Years, December 6, 2000,
Table P-2. Docket Number 2000-01, Document Number U-B-14.
60. The trail maps consulted for this rulemaking can be found in Docket No. A-2000-01,
Document No. U-A-65.
61 .This goal was recently upheld by the US Court of Appeals. American Corn Growers
Association v. EPA, 291F.3d 1(D.C .Cir 2002). A copy of this decision can be found in Docket
A-2000-01, Document IV- A-113..
1-67
-------
Draft Regulatory Support Document
62.U.S. EPA Review of the National Ambient Air Quality Standards for Particulate Matter:
Policy Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA-452/R-96-
013. 1996. Docket Number A-99-06, DocumentsNos.il-A-18, 19, 20, and 23. The particulate
matter air quality criteria documents are also available at http://www.epa.gov/ncea/partmatt.htm.
63. Chestnut, L.G., and R.D. Rowe. 1990a. Preservation for Visibility Protection at the National
Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards, US
Environmental Protection Agency, and Air Quality Management Division, National Park
Service; Chestnut, L.G., and R.D. This document is available from Docket A-97-10, Document
II-A-33 Rowe. 1990b. A New National Park Visibility Value Estimates. In Visibility and Fine
Particles., Transactions of an AWMA/EPA International Speciality Conference. C.V. Mathai, ed.,
Air and Waste Management Association, Pittsburg. Docket A-2000-01, IV-A-2000.
64.Letter from Debra C. Miller, Data Analyst, National Park Service, to Drew Kodjak, August
22, 2001. Docket No. A-2000-01, Document Number. II-B-28.
65.Technical Memorandum, Aaron Worstell, Environmental Engineer, National Park Service,
Air Resources Division, Denver, Colorado, particularly Table 1. Docket No. A-2000-01,
Document Number U-G-178.
66.Letter from Debra C. Miller, Data Analyst, National Park, to Drew Kodjak, August 22, 2001.
Docket No. A-2000-01, Document Number. II-B-28.
67.Letter from Debra C. Miller, Data Analyst, National Park Service, to Drew Kodjak, August
22, 2001. Docket No. A-2000-01, Document. Number. H-B-28.
68.National Park Service, February 2000. Air Quality Concerns Related to Snowmobile Usage
in National Parks. Air Docket A-2000-01, Document No. II-A-44.
69.G. Bishop, et al., Snowmobile Contributions to Mobile Source Emissions in Yellowstone
National Park, Environmental Science and Technology, Vol. 35, No. 14, at 2873. Docket No. A-
2000-01, Document No. II-A-47.
70.Memorandum to IV-D-204 at 13
71.Ibid, at 14.
72. Julia Rege, Environmental Scientist, EPA. Memorandum to Docket A-2000-0. Predicted
visibility effects from snowmobile exhaust (particulate matter) on or near snowmobile trails in
Yellowstone National Park. July, 12, 2002. Docket A-2000-01, Document IV-A-147.
73.See our Mobile Source Air Toxics final rulemaking, 66 FR 17230, March 29, 2001, and the
Technical Support Document for that rulemaking. Docket No. A-2000-01, Documents Nos. U-
A-42 and U-A-30.
1-68
-------
Chapter 1: Health and Welfare Concerns
74.U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA by Sierra
Research, Inc. and Radian International Corporation/Eastern Research Group, November 30,
1999. Report No. EPA420-R-99-029. http://www.epa.gov/otaq/toxics.htm.
75.U.S. EPA (1998) Environmental Protection Agency, Carcinogenic Effects of Benzene: An
Update, National Center for Environmental Assessment, Washington, DC. 1998. EPA/600/P-
97/001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html.
76.Leukemia is a blood disease in which the white blood cells are abnormal in type or number.
Leukemia may be divided into nonlymphocytic (granulocytic) leukemias and lymphocytic
leukemias. Nonlymphocytic leukemia generally involves the types of white blood cells
(leukocytes) that are involved in engulfing, killing, and digesting bacteria and other parasites
(phagocytosis) as well as releasing chemicals involved in allergic and immune responses. This
type of leukemia may also involve erythroblastic cell types (immature red blood cells).
Lymphocytic leukemia involves the lymphocyte type of white bloods cell that are responsible for
the immune responses. Both nonlymphocytic and lymphocytic leukemia may, in turn, be
separated into acute (rapid and fatal) and chronic (lingering, lasting) forms. For example; in
acute myeloid leukemia (AML) there is diminished production of normal red blood cells
(erythrocytes), granulocytes, and platelets (control clotting) which leads to death by anemia,
infection, or hemorrhage. These events can be rapid. In chronic myeloid leukemia (CML) the
leukemic cells retain the ability to differentiate (i.e., be responsive to stimulatory factors) and
perform function; later there is a loss of the ability to respond.
77.U.S. EPA (1985) Environmental Protection Agency, Interim quantitative cancer unit risk
estimates due to inhalation of benzene, prepared by the Office of Health and Environmental
Assessment, Carcinogen Assessment Group, Washington, DC. for the Office of Air Quality
Planning and Standards, Washington, DC., 1985. Air Docket A-2000-01, Document No. H-A-
74.
78.Clement Associates, Inc. (1991) Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991. Air Docket A-2000-01, Document
No. II-A-49.
79.International Agency for Research on Cancer (IARC) (1982) IARC monographs on the
evaluation of carcinogenic risk of chemicals to humans, Volume 29, Some industrial chemicals
and dyestuffs, International Agency for Research on Cancer, World Health Organization, Lyon,
France, p. 345-389.
80.Irons, R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry (1992) Synergistic action of
the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci. 89:3691-3695.
Sl.Lumley, M., H. Barker, and J.A. Murray (1990) Benzene in petrol, Lancet 336:1318-1319.
1-69
-------
Draft Regulatory Support Document
82.U.S. EPA (1993) Motor Vehicle-Related Air Toxics Study, U.S. Environmental Protection
Agency, Office of Mobile Sources, Ann Arbor, MI, EPA Report No. EPA 420-R-93-005, April
1993.
83.Pancytopenia is the reduction in the number of all three major types of blood cells
(erythrocytes, or red blood cells, thrombocytes, or platelets, and leukocytes, or white blood cells).
In adults, all three major types of blood cells are produced in the bone marrow of the vertebra,
sternum, ribs, and pelvis. The bone marrow contains immature cells, known as multipotent
myeloid stem cells, that later differentiate into the various mature blood cells. Pancytopenia
results from a reduction in the ability of the red bone marrow to produce adequate numbers of
these mature blood cells.
84.Aksoy, M (1991) Hematotoxicity, leukemogenicity and carcinogenicity of chronic exposure
to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds. Molecular Aspects of
Monooxygenases and Bioactivation of Toxic Compounds. New York: Plenum Press, pp. 415-
434.
85.Goldstein, B.D. (1988) Benzene toxicity. Occupational medicine. State of the Art Reviews.
3: 541-554.
86. Aplastic anemia is a more severe blood disease and occurs when the bone marrow ceases to
function, i.e.,these stem cells never reach maturity. The depression in bone marrow function
occurs in two stages - hyperplasia, or increased synthesis of blood cell elements, followed by
hypoplasia, or decreased synthesis. As the disease progresses, the bone marrow decreases
functioning. This myeloplastic dysplasia (formation of abnormal tissue) without acute
leukemiais known as preleukemia. The aplastic anemia can progress to AML (acute mylogenous
leukemia).
87.Aksoy, M., S. Erdem, and G. Dincol. (1974) Leukemia in shoe-workers exposed chronically
to benzene. Blood 44:837.
SS.Aksoy, M. and K. Erdem. (1978) A follow-up study on the mortality and the development of
leukemia in 44 pancytopenic patients associated with long-term exposure to benzene. Blood 52:
285-292.
89.Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E. Marti, Y.Z. Wang, M. Linet, L.Q.
Xi, W. Lu, M.T. Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, and R.B. Hayes
(1996) Hematotoxicity among Chinese workers heavily exposed to benzene. Am. J. Ind. Med.
29: 236-246.
90.U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA by Sierra
Research, Inc. and Radian International Corporation/Eastern Research Group, November 30,
1999. Report No. EPA420-R-99-029. http://www.epa.gov/otaq/toxics.htm.
1-70
-------
Chapter 1: Health and Welfare Concerns
91.U.S. EPA (1985) Mutagenicity and Carcinogenicity Assessment of 1,3-Butadiene.
EPA/600/8-85/004F. U.S. Environmental Protection Agency, Office of Health and
Environmental Assessment. Washington, DC.
92.U.S. EPA (1998) Draft Health Risk Assessment of 1,3-Butadiene, National Center for
Environmental Assessment, Office of Research and Development, U.S. EPA, EPA/600/P-
98/001 A, February 1998.
93.Scientific Advisory Board. 1998. An SAB Report: Review of the Health Risk Assessment of
1,3-Butadiene. EPA-SAB-EHC-98, August, 1998.
94.EPA 1996. guidelines for carcinogen risk assessment. Federal Register 61(79):17960-18011.
95.U.S. EPA (1985) Mutagenicity and carcinogenicity assessment of 1,3-butadiene. EPA/600/8-
85/004F. U.S. Environmental Protection Agency, Office of Health and Environmental
Assessment. Washington, DC. http://www.epa.gov/ngispgm3/iris/subst/0139.htm.
96.U.S. EPA (1985) Mutagenicity and carcinogenicity assessment of 1,3-butadiene. EPA/600/8-
85/004F. U.S. Environmental Protection Agency, Office of Health and Environmental
Assessment. Washington, DC. http://www.epa.gov/ngispgm3/iris/subst/0139.htm.
97.Ligocki, M.P., G.Z. Whitten, R.R. Schulhof, M.C. Causley, and G.M. Smylie (1991)
Atmospheric transformation of air toxics: benzene, 1,3-butadiene, and formaldehyde, Systems
Applications International, San Rafael, CA (SYSAPP-91/106).
98.U.S. EPA (1987) Environmental Protection Agency, Assessment of health risks to garment
workers and certain home residents from exposure to formaldehyde, Office of Pesticides and
Toxic Substances, April 1987. Air Docket A-2000-01, Document No. II-A-48.
99.Clement Associates, Inc. (1991) Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991. Air Docket A-2000-01, Document
No. II-A-49.
100.U.S. EPA (1993) Motor Vehicle-Related Air Toxics Study, U.S. Environmental Protection
Agency, Office of Mobile Sources, Ann Arbor, MI, EPA Report No. EPA 420-R-93-005, April
1993. http://www.epa.gov/otaq/toxics.htm.
lOl.Wilhelmsson, B. and M. Holmstrom. (1987) Positive formaldehyde PAST after prolonged
formaldehyde exposure by inhalation. The Lancet:l64.
102.Burge, P.S., M.G. Harries, W.K. Lam, I.M. O'Brien, and P.A. Patchett. (1985) Occupational
asthma due to formaldehyde. Thorax 40:225-260.
103.Hendrick, D.J., RJ. Rando, DJ. Lane, and MJ. Morris (1982) Formaldehyde asthma:
Challenge exposure levels and fate after five years. J. Occup. Med. 893-897.
1-71
-------
Draft Regulatory Support Document
104.Nordman, H., H. Keskinen, and M. Tuppurainen. (1985) Formaldehyde asthma - rare or
overlooked? J. Allergy Clin. Immunol. 75:91-99.
105.U.S. EPA. (1999) Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I. Prepared for EPA by Sierra
Research, Inc. and Radian International Corporation/Eastern Research Group, November 30,
1999. Report No. EPA420-R-99-029. http://www.epa.gov/otaq/toxics.htm
106.Ligocki, M.P., G.Z. Whitten (1991) Atmospheric transformation of air toxics: acetaldehyde
and polycyclic organic matter, Systems Applications International, San Rafael, CA, (SYSAPP-
91/113).
107. Environmental Protection Agency, Health assessment document for acetaldehyde, Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office, Research
Triangle Park, NC, EPA-600/8-86/015A (External Review Draft), 1987. Air Docket A-2000-01,
Document No. II-A-33.
108.Environmental Protection Agency, Integrated Risk Information System (IRIS), Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office,
Cincinnati, OH, 1992. Acetaldehyde. http://www.epa.gov/iris/subst/0290.htm
109.U.S. EPA (1987) Health Assessment Document for Acetaldehyde — External Review
Draft. Office of Health and Environmental Assessment, Research Triangle Park, NC. Report
No. EPA 600/8-86/015A.
110.California Air Resources Board (CARB) (1992) Preliminary Draft: Proposed identification
of acetaldehyde as a toxic air contaminant, Part B Health assessment, California Air Resources
Board, Stationary Source Division, August, 1992. Air Docket A-2000-01, Document No. II-A-
34.
111.U.S. EPA (1997) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office, Cincinnati, OH, 1997. Acetaldehyde. http://www.epa.gov/iris/subst/0290.htm
112.U.S. EPA (1999) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office, Cincinnati, OH. 1,3-Butadiene, http://www.epa.gov/iris/subst/0139.htm .
113.U.S. EPA (1993) Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office, Cincinnati, OH. Acrolein http ://www. epa. gov/iri s/sub st/03 64. htm .
114. U.S. EPA Integrated Risk Information System (IRIS) summary for Toluene.
Www.epa.gov/iris/subst/0118.htm
1-72
-------
Chapter 1: Health and Welfare Concerns
115. Peoples, R.W., Li, C., and Weight, F.F. (1996) Lipid vs. protein theories of alcohol action
in the nervous system. Annu. Rev. Pharmacol. Toxicol. 36:185-201.
116. Beckstead, M.J., Weiner, J.L., Eger, E.I. II, Gong, D.H. and Mihic, S.J. (2000) Glycine and
a-aminobutyric acidA receptor function is enhanced by inhaled drugs of abuse. Mol. Pharm.
57:1199-1205.
117. Cruz, S.L., R.L. Balster, and JJ. Woodward. 2000. Effects of volatile solvents on
recombinant N-methyl-D-aspartate receptors expressed in Xenopus oocytes. Brit. J. Pharmacol.
131:1303-1308. Cruz, S.L., T. Mirshahi, B. Thomas, R.L. Balster and JJ. Woodward. 1998.
Effects of the abused solvent toluene on recombinant N-methyl-D-aspartate and
non-N-methyl-D-aspartate receptors expressed in Xenopus oocytes. J. Pharmacol. Exp. Ther.
286:334-40. Both of these documents are cited in Benignus, V., Bushnell, P. and Boyes, W.
"Acute Behavioral Effects of Exposure to Toluene and Carbon Monoxide from Snowmobile
Exhaust" Docket A-2000-01, Document IV-A-143.
118. Bale, A.S., Cruz, S.L., Balster, R.L., and Woodward, JJ. Effects of Toluene onNicotinic
Acetylcholine Receptors Expressed in Xenopus Oocytes. Presented at the Sixty-First Annual
College on Problems of Drug Dependence, Acapulco, Mexico, 1999. This document is cited in
Benignus, V., Bushnell, P. and Boyes, W. "Acute Behavioral Effects of Exposure to Toluene
and Carbon Monoxide from Snowmobile Exhaust" Docket A-2000-01, Document IV-A-143.
119. Tillar, R., Shafer, TJ. and Woodward, JJ. (2002) Toluene inhibits voltage-sensitive
calcium channels expressed in pheochromocytoma cells. Neurochemistry International, in press.
120. Danysz, W., W. Dyr, E. Jankowska, S. Glazewski, W. Kostowski. The involvement of
NMDA receptors in acute and chronic effects of ethanol. Alcohol. Clin. Exp. Res. 16:499-504,
1992. This document is cited in Benignus, V., Bushnell, P. and Boyes, W. "Acute Behavioral
Effects of Exposure to Toluene and Carbon Monoxide from Snowmobile Exhaust" Docket A-
2000-01, Document IV-A-143.; Tassonyi, E., Charpantier, E., Muller, D., Dumont, L. and
Bertrand, D. (2002 The role of nicotinic acetylcholine receptors in the mechanisms of
anesthesia. Brain Res. Bull. 57:133-150.
121.Mihic, S.J., McQuilkin, S.J., Eger, E.I. H, lonescu, P. and Harris, R.A. (1994) Potentiation
of a-aminobutyric acid type A receptor-mediated chloride currents by novel halogenated
compounds correlates with their abilities to induce general anesthesia. Mol. Pharm. 46:851-857.
Olsen, R.W. (1998) The molecular mechanism of action of general anesthetics: Structural
aspects of interactions with GABAA receptors. Toxicol. Lett. 100-101:193-201.
122. Balster, R.L. (1998) Neural basis of inhalant abuse. Drug Ale. Dependence 51:207-214.
123."Warehouse Workers' Headache, Carbon Monoxide Poisoning from Propane-Fueled
Forklifts," Thomas A. Fawcett, et al, Journal of Occupational Medicine, January 1992, p. 12.
Docket A-2000-01, Document No. II-A-36.
1-73
-------
Draft Regulatory Support Document
124."Terminox System Reduces Emissions from LPG Lift Trucks," Material Handling Product
News. Docket A-2000-01, Document No. II-A-14.
125.Summary of Medical Papers Related to Exhaust Emission Exposure at Ice Rinks," EPA
Memorandum from Alan Stout to Docket A-2000-01. Docket A-2000-01, Document No. II-A-
38.
126. Snook and Davis, 1997, "An Investigation of Driver Exposure to Carbon Monoxide While
Traveling Behind Another Snowmobile." Docket No. A-2000-01, Document Number II-A-35.
127. Emissions from Snowmobile Engines Using Bio-based Fuels and Lubricants, Southwest
Research Institute, August, 1997, at 22. Docket No. A-2000-01, Document Number II-A-50.
128. National-Scale Air Toxics Assessment for 1996, EPA-453/R-01-003, Draft, January 2001.
129. Technical Memorandum, Chad Bailey, Predicted benzene exposures and ambient
concentrations on and near snowmobile trails, August 17, 2001. Air Docket A-2000-01,
Document No. U-B-27.
1-74
-------
Chapter 2: Industry Characterization
Chapter 2: Industry Characterization
To accurately assess the potential impact of this emission control program, it is important to
understand the nature of the affected industries. This chapter describes relevant background
information related to each of the categories of engines and vehicles subject to this proposal. For
each engine category, descriptions of the supply and demand sides of the markets are provided.
Additionally, industry organization and historical market trends data are discussed.
2.1 CI Marine Engines and Recreational Boats
This section gives a general characterization of the segments of the marine industry that may
be affected by the regulation. The emission control program may affect diesel marine engines
and recreational boats that contain these engines. We therefore focus on the compression-
ignition (CI) diesel marine engine manufacturing and recreational boat building industries.
Information is also provided for several spark-ignition vessel categories, even though they are not
directly affected by this rule (spark-ignition engines and vessels are the subject of a separate
proposed rulemaking regarding evaporative emissions; See 67 FR 53050, August 14, 2002).
This industry characterization was developed in part under contract with ICF Consulting1 as well
as independent analyses conducted by EPA through interaction with the industry and other
sources.2'3'4
2.1.1 The Supply Side
This section describes the types of recreational boats that may contain CI marine engines,
the inputs used to manufacture both boats and engines, and the costs associated with boat and
engine production.
2.1.1.1 Product Types
Diesel engines are primarily available in inboard marine configurations and are most
commonly found in inboard cruisers and inboard runabouts. The National Marine Manufacturers
Association estimates that 18 percent of all inboard boats are equipped with diesel engines, with
the dominant application being cruisers.5 Diesel engines are also available in sterndrive
configurations on a limited basis, and in the past, a small number of outboard boats contained
diesel engines as well (currently there are no outboard diesel engines being manufactured).
Descriptions of these boat types, taken from the Economic Impact Analysis of the Proposed Boat
Manufacturing NESHAP, are provided here6:
Inboard runabouts are mid-sized boats powered by an attached engine located inside the
hull at the middle or rear of the boat, with a prop shaft running through the bottom of the
boat. Most inboard runabouts are tournament ski boats.
2-1
-------
Draft Regulatory Support Document
Inboard cruisers are large boats with cabins. Almost all cruisers are equipped with two
inboard engines.
• Sterndrives are mid-sized boats powered by an attached inboard engine combined with a
drive unit that is located on the transom at the stern (rear) of the boat. Sterndrives are also
known as inboard/outboards or I/Os.
• Outboards are small to medium-sized boats powered by a self-contained detachable engine
and propulsion system, which is attached to the transom. This category of boats includes
most runabouts, bass boats, utility boats, offshore fishing boats, and pontoons.
Larger boats are powered exclusively by diesel inboard engines. These boats are generally
40 feet or greater in length. Recreational boats in ports with access to the ocean (e.g. Seattle) can
be 80 to 100 feet or longer. The larger boats typically require twin inboard diesel engines with
2,000 total horsepower or more. Recreational diesel marine engines are generally produced by
domestic companies that have been long-standing players in the marine diesel engine market.
The three companies that tend to dominate the market are Caterpillar, Cummins, and Detroit
Diesel (see Section 2.1.3.2 for details about these companies). Nearly 75 percent of diesel
engines sales for recreational vessels in 2000 can be attributed to these three companies.
Sterndrive boats equipped with diesel engines account for less than 1 to 2 percent of the
market. A minority of mid-sized boat owners insist on diesel powered Sterndrive engines for
their boats. Diesel marine Sterndrive systems generally power the same types of boats as their
gasoline counterparts, which tend to be 15 to 30 feet in length. Customers that choose a diesel
Sterndrive marine engine are generally seeking three main advantages over gasoline Sterndrive
marine engines. First, diesel fumes are much less ignitable and explosive that gasoline fumes.
Second, diesel powered craft have a greater range than gasoline powered craft with similar fuel
capacity. Lastly, diesel engines tend to be more reliable and tend to run more hours between
major overhauls than gasoline engines. This last point is particularly important to boat owners
who operate their boats higher than the average.
One major disadvantage of diesel Sterndrive engines is their cost relative to comparably
powered gasoline Sterndrive engines. For example, a 40 foot twin cabin cruiser with twin
gasoline Sterndrive engines costs $238,000. For twin diesel Sterndrive engines, the price
increases by approximately $50,000. The fact that the diesel engine is more expensive, coupled
with the fact that diesel fuel is often less available than gasoline in the U.S., has resulted in
limited domestic demand for recreational diesel Sterndrive marine engines.
2.1.1.2 Primary Inputs
The primary inputs used to produce marine engines and recreational boats, can be divided
into four major categories: capital, labor, energy, and materials. Capital refers to the type of
equipment used in production where the type of capital depends upon the good being produced.
The same is true for labor, as different skills are required for the production of boats relative to
engines. Energy refers to the electricity, natural gas, or other power sources used to operate
production equipment and plants at which boats and engines are manufactured. Material inputs
2-2
-------
Chapter 2: Industry Characterization
are what differ the most across the production of these end products. The remainder of this
section focuses on the different materials used to produce CI marine engines and recreational
boats.
Some of the main materials used to produce CI marine engines include fluid power pumps,
motors, and transmissions; fluid power cylinders, filters, valves, hoses, and their assemblies;
metal bolts, nuts, screws, washers, and tanks; iron, steel, and nonmetal forgings and castings;
steel bars, plates, piston rings, and other steel shapes and forms; gears, gaskets, and fabricated
plastic products; engine electrical equipment such as spark plugs, generators, and starters; and
rubber and plastic hosing and belting. All of these inputs are used in conjunction with energy,
capital, and skilled labor to manufacture engines.
Main inputs used in the production of recreational boats include marine engines, plastic and
aluminum fuel tanks, and rubber fuel hoses. However, these are but a few of the materials used
in boat manufacturing. Others include marine metal hardware, such as propellers, castings,
screws, washers, and rivets; metal forgings, castings, and other steel forms; aluminum and
aluminum-base alloy sheet, plate, foil, rod, bars, and pipes; fiberglass, lumber, plywood, canvas
products, and carpeting; plastic rods, tubes, and shapes; and paints, varnishes and lacquers.
2.1.1.3 Costs of Production
The historical production costs of marine engines and recreational boats are divided into the
primary input categories of labor, materials, and capital expenditures. Table 2.1-1 presents the
value of shipments (VOS), production costs, and production costs as a share of the VOS for the
other engine equipment manufacturing industry (which includes marine engine manufacturing).
Table 2.1-2 shows the same figures for the boat manufacturing industry. The other engine
equipment manufacturing industry is identified by Standard Industrial Classification (SIC) code
3519 and the North American Industrial Classification System (NAICS) code 333618. The SIC
code and the NAICS code for the boat building industry are 3732 and 336612.
For both engine manufacturing and boat building, the average share of the cost of materials
and total capital expenditures is similar. The cost of materials represents an average of 57 to 58
percent of the VOS for both industries and average share of capital expenditures for both
industries is approximately 2 to 3 percent. Another trend evident for both industries is that the
cost shares of materials and payroll tended to be higher in the earlier part of the 1990s than in the
late 1990s. Payroll, which includes the costs associated with employee wages and benefits,
differs slightly across the industries. For the boat manufacturing industry, payroll represents an
average of 20 percent of VOS while for engine manufacturing, it is equal to an average share of
14 percent of its shipment value.
Also notable in these tables is that the average VOS for the engine manufacturing industry,
over $16 billion, is about three times the average VOS for the boat manufacturing industry. It is
important to keep in mind that the data in Table 2.1-1 include other engine equipment
manufacturing and does not represent marine engine manufacturing exclusively. Likewise, the
2-3
-------
Draft Regulatory Support Document
figures in Table 2.1-2 for boat manufacturing include vessels that are not powered by CI engines,
such as outboards, jet skis, personal water craft, and boats that are not motorized, such as canoes
and kayaks.
Table 2.1-1
Value of Shipments and Production Costs for the SIC and NAICS Codes that
Include Recreational Boat Engine Manufacturers*, 1992 -1999 78910n 1213
Year
1992
1993
1994
1995
1996
1997
1998
1999
Industry
Code
SIC 3519
SIC 3519
SIC 3519
SIC 3519
SIC 3519
NAICS 333618
NAICS 333618
NAICS 333618
Average
Value of
Shipments
($106)
$11,827
$12,600
$15,308
$16,642
$17,286
$19,011
$20,312
$22,389
$16,922
Payroll
($106)
$2,072
$1,900
$2,162
$2,238
$2,237
$2,374
$2,471
$2,652
$2,263
%of
VOS
18%
15%
14%
13%
13%
12%
12%
12%
14%
Cost of Materials
($106)
$6,996
$7,545
$8,977
$9,940
$9,905
$10,539
$11,963
$12,474
$9,792
%of
VOS
59%
60%
59%
60%
57%
55%
59%
56%
58%
Total Capital
Expenditures
($106)
$461
$371
$406
$499
$528
$631
$682
$786
$545
%of
VOS
4%
3%
3%
3%
3%
3%
3%
4%
3%
! Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars
Table 2.1-2
Value of Shipments, and Production Costs for the SIC and NAICS Codes
that Include Recreational Boat Manufacturers*, 1992 -1999 ^15,16,17,18,19,20
Year
1992
1993
1994
1995
1996
1997
1998
1999
Industry
Code
SIC 3732
SIC 3732
SIC 3732
SIC 3732
SIC 3732
NAICS 336612
NAICS 336612
NAICS 336612
Average
Value of
Shipments
($106)
$4,599
$4,975
$5,334
$5,597
$5,823
$5,607
$5,939
$7,463
$5,667
Payroll
($106)
$1,006
$1,033
$1,081
$1,105
$1,177
$1,030
$1,114
$1,361
$1,113
%of
VOS
22%
21%
20%
20%
20%
18%
19%
18%
20%
Cost of Materials
($106)
$2,609
$2,919
$3,075
$3,218
$3,396
$3,237
$3,202
$4,099
$3,219
%of
VOS
57%
59%
58%
57%
58%
58%
54%
55%
57%
Total Capital
Expenditures
($106)
$63
$83
$90
$89
$109
$122
$263
$231
$131
%of
VOS
1%
2%
2%
2%
2%
2%
4%
3%
2%
Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars
2-4
-------
Chapter 2: Industry Characterization
Looking specifically at the engine manufacturing industry, we see that the share of payroll
steadily declined over the 1992 - 1999 time period. In 1992, payroll represented 18 percent of
the VOS but by 1995, it was down to 13 percent. Labor costs fell to 12 percent of the VOS in
1997 and remained at this lower share value through 1999. A declining trend is also evident for
the share of payroll for the boat manufacturing industry, however it was more recently that the
share of labor costs fell. In 1992, labor costs were equal to 22 percent of the boat manufacturing
industry's VOS. It dropped to 20 percent from 1994 to 1996 and most recently was equal to 18
to 19 percent in the late 1990s.
2.1.1.4 Recreational Boat Production Practices
Based on information supplied by a variety of recreational boat builders, the following
discussion provides a description of the general production practices used in this sector of the
marine industry.
Engines are usually purchased from factory authorized distribution centers. The boat
builder provides the specifications to the distributor who helps match an engine for a particular
application. It is the boat builders responsibility to fit the engine into their vessel design. The
reason for this is that sales directly to boat builders are a very small part of engine manufacturers'
total engine sales. These engines are not generally interchangeable from one design to the next.
Each recreational boat builder has their own designs. In general, a boat builder will design one
or two molds that are intended to last 5-8 years. Very few changes are tolerated in the molds
because of the costs of building and retooling these molds.
Recreational vessels are designed for speed and therefore typically operate in a planing
mode. To enable the vessel to be pushed onto the surface of the water where it will subsequently
operate, recreational vessels are constructed of lighter materials and use engines with high power
density (power/weight). The tradeoff on the engine side is less durability, and these engines are
typically warranted for fewer hours of operation. Fortunately, this limitation typically
corresponds with actual recreational vessel use. With regard to design, these vessels are more
likely to be serially produced. They are generally made out of light-weight fiberglass. This
material, however, minimizes the ability to incorporate purchaser preferences, not only because
many features are designed into the fiberglass molds, but also because these vessels are very
sensitive to any changes in their vertical or horizontal centers of gravity. Consequently, optional
features are generally confined to details in the living quarters, and engine choice is very limited
or is not offered at all.
Based on information supplied by a variety of recreational boat builders, fuel tanks for
recreational boats are usually purchased from fuel tank manufacturers. However, some boat
builders construct their own fuel tanks. The boat builder provides the specifications to the fuel
tank manufacturer who helps match the fuel tank for a particular application. It is the boat
builder's responsibility to install the fuel tank and connections into their vessel design. For
vessels designed to be used with small outboard engines, the boat builder may not install a fuel
tank; therefore, the end user would use a portable fuel tank with a connection to the engine.
2-5
-------
Draft Regulatory Support Document
2.1.2 The Demand Side
The information provided in this section addresses the various options consumers have
available regarding recreational marine vessels and the engines used to power them. Some of the
engine-powered recreational boats available to consumers include inboards, sterndrives,
outboards, personal water craft, and jet boats.
2.1.2.1 Uses and Consumers
Recreational boats are used for a number of water-related pastimes including fishing,
waterskiing, cruising, vacationing, relaxing on the water, sunning, and a host of other activities.
Runabouts are commonly used for waterskiing, tubing, and wakeboarding. Larger cruisers and
yachts can be used for extended trips because they may be equipped with cabins for cooking and
sleeping. Fishing boats can vary in size depending on whether they are used for offshore sport
fishing or local lake fishing. Other boats, such as personal water craft, sailboats, canoes, and
rowboats can be used for cruising along the water.
According to the National Marine Manufacturers Association (NMMA), there are currently
close to 70 million people participating in recreational boating. In the late 1990s, this figure was
closer to 80 million, but the recent economic downturn has led consumers to engage in fewer
leisure activities. From Table 2.1-3, we can see that outboard boats are the most common boat
type, followed further behind by inboard and sterndrive boats. The number of inboards and
sterndrives owned in the U.S. are roughly equivalent over the 1997 to 2001 time period.
Table 2.1-3
Recreational Boating Population Estimates (103)*, 1997 - 20012122
People participating in recreational
boating
All boats in use
Outboard boats owned
Inboard boats owned
Sterndrive boats owned
Personal water craft
1997
78,406
16,230
8,125
1,587
1,582
1,000
1998
74,847
16,824
8,300
1,609
1,673
1,100
1999
73,208
16,790
8,211
1,635
1,665
1,096
2000
72,269
16,991
8,288
1,660
1,709
1,078
2001
69,486
16,999
8,342
1,678
1,743
1,631
* These in-use figures are based on the actual state and Coast Guard registrations. Population estimates are rounded to
the nearest thousandths.
The type of boat purchased by a consumer and the type of engine it is equipped with are
affected by the recreational activity the consumer plans to engage in, the size of the boat being
purchased, and other consumer preferences. For example, if a larger inboard cruiser is selected
for purchase, the consumer will likely opt for a diesel engine. Diesel engines are, in general,
2-6
-------
Chapter 2: Industry Characterization
more expensive, but have a longer life span than gasoline engines. In addition, diesel engines are
available at much higher power ratings. However, if the consumer prefers a smaller fishing boat
with an outboard engine configuration, it will be equipped with a gasoline engine.
Generally speaking, recreational boats are considered final goods while the engines that
power them are intermediate goods. As discussed in Section 2.1.1.4, boat builders purchase
engines from distribution centers and then use these engines as inputs to the production of boats.
Boat builders may provide their own engine designs to engine manufacturers so that the engines
will properly fit into the boat builders' specific models.
2.1.2.2 Substitution Possibilities
Consumers can substitute across different boat types but may be limited by the water-related
activities they want to engage in. Runabouts and cruisers are available in different engine
configurations and different engine types. Consumers will first evaluate the purpose for which
they'd like to buy a boat and will then consider the various types of boats that will suit their
preferences. If consumers choose to purchase either sterndrive or inboard boats, they have both
diesel and gasoline engines available to them. Outboards, on the other hand, are only available
with gasoline engines.
Consumers may be interested in engaging in water-related activities, but may instead
consider purchasing non-motorized boats. For example, consumers who are like to float out on
the water or engage in lake fishing may choose to purchase a sailboat, row boat, or canoe. These
non-motorized boating options do not allow the consumer to participate in the same set of water-
related activities as would the purchase of a motorized boat, but they may be considered
substitutes for less intensive water-related past times.
2.1.3 Industry Organization
It is important to gain an understanding of how the recreational marine vessel and CI marine
engine industries may be affected by the emissions control program. One way to determine how
increased costs might affect the market is to examine the organization of each industry. This
section provides data to measure the competitive nature of the boat building and marine engine
industries and lists the manufacturers of recreational boats, marine engines, and marine fuel
tanks.
2.1.3.1 Market Structure
Market structure is of interest because it determines the behavior of producers and
consumers in the industry. In perfectly competitive industries, no producer or consumer is able
to influence the price of the product sold. In addition, producers are unable to affect the price of
inputs purchased for use in production. This condition is most likely to hold if the industry has a
large number of buyers and sellers, the products sold and inputs used are homogeneous, and entry
and exit of firms is unrestricted. Entry and exit of firms are unrestricted for most industries,
2-7
-------
Draft Regulatory Support Document
except in cases where the government regulates who is able to produce output, where one firm
holds a patent on a product, where one firm owns the entire stock of a critical input, or where a
single firm is able to supply the entire market. In industries that are not perfectly competitive,
producer and/or consumer behavior can have an effect on price.
Concentration ratios (CRs) and Herfindahl-Hirschman indices (HHI) can provide some
insight into the competitiveness of an industry. The U.S. Department of Commerce reports these
ratios and indices for the six digit NAICS code level for the year 1997, the most recent year
available. Tables 2.1-4 and 2.1-5 provide the four- and eight-firm concentration ratios (CR4 and
CRS, respectively) and the Herfindahl-Hirschman indices for the other engine equipment
manufacturing and boat building industries (the other engine equipment manufacturing industry
includes manufacturers of marine engines). These industries are represented by NAICS codes
333618 and 336612, respectively. Concentration ratios are provided in percentage terms while
HHI are based on a scale formulated by the Department of Justice.
Table 2.1-4
Measures of Market Concentration for the NAICS Code that
Includes Recreational Boat Engine Manufacturers, 1997 23
Description
NAICS 333618
CR4
55.8
CRS
76.0
HHI
1019.1
VUS
($106)
$19,011.09
Number ol 1
Companies
245 U
Table 2.1-5
Measures of Market Concentration for the NAICS Code that
Includes Recreational Boat Manufacturers, 1997 24
Description
NAICS 336612
CR4
41.4
CRS
48.9
HHI
644.5
VOS
($106)
$5,607.30
Number ot 1
Companies 1
984 |
The criteria for evaluating the HHI are based on the 1992 Department of Justice Horizontal
Merger Guidelines. According to these criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are
considered moderately concentrated (i.e., moderately competitive), and those with HHIs above
1,800 are considered highly concentrated (i.e., less competitive). In general, firms in less
concentrated industries have more ability to influence market prices. Based on these criteria, the
marine vessel industry can be modeled as perfectly competitive for the purposes of the economic
impact analysis. The other engine equipment manufacturing industry is slightly more
concentrated, with higher CRs and an HHI value just over 1,000. However, it is reasonable to
assume that the marine engine manufacturing industry is perfectly competitive for the economic
analysis.
2-8
-------
Chapter 2: Industry Characterization
2.1.3.2 CI Marine Engine and Recreational Boat Manufacturers
We have determined that there are at least 16 companies that manufacture CI marine engines
for recreational vessels. Nearly 75 percent of diesel engines sales for recreational vessels in 2000
can be attributed to three large companies. Six of the identified companies are considered small
businesses as defined by the Small Business Administration SBA) size standard for NAICS code
333618 (less than 1000 employees). Based on sales estimates for 2000, these six companies
represent less than 5 percent of recreational marine diesel engine sales. Table 2.1-6 provides a
list of the diesel engine manufacturers identified to date by EPA.
Table 2.1-6
Annual Sales for Recreational Diesel Marine
Engine Manufacturers Identified by EPA, 2000/200125-26-27
Companies with greater than
1,000 employees
Annual
Sales8
($106)
Companies with less than
1,000 employees
Annual Sales"
($106)
Caterpillar, Inc. (Engines Div.)b $2,176.0
Cummins Engine Company, Inc. $6,600.0
Detroit Diesel Engines $2,358.7
Isotta Fraschini NA°
Deere & Company $13,137.0
Marine Corporation of America NA°
Mercruiser $68.6
MTU Aero Engine Components $7.9
Volvo Penta $275.0
Yanmar Diesel America Corporation $18.9
Alaska Diesel Electric/Lugger
American Diesel Corporation
Daytona Marine
Marine Power, Inc.
Peninsular Diesel Engines, Inc.
Westerbeke Corporation
$9.2
$5.0
$2.9
$7.0
NAC
$29.1
a Annual sales of listed companies include revenues received from the sale of all products sold by these companies, not
just revenues received from the sales of diesel marine engines.
b Companies in bold dominate the diesel engine market for recreational vehicles.
0 NA means Not Available.
Less precise information is available about recreational boat builders than is available about
engine manufacturers. Several sources were used, including trade associations, business
directories, and Internet sites when identifying entities that build and/or sell recreational boats.
We have also worked with an independent contractor to assist in the characterization of this
segment of the industry. Finally, we have also obtained a list of nearly 1,700 boat builders
known to the U.S. Coast Guard to produce boats using recreational gasoline and diesel engines.
At least 1,200 of these companies install gasoline-fueled engines and would therefore be subject
to the proposed evaporative emission standards. More that 90 percent of the companies
identified to date would be considered small businesses as defined by SBA size standards for
NAICS code 336612 (less than 500 employees). Table 2.1-7 provides a sample of recreational
2-9
-------
Draft Regulatory Support Document
boat manufacturers known to EPA.
2-10
-------
Chapter 2: Industry Characterization
Table 2.1-7
Annual Sales and Employment for a Sample of
Recreational Boat Manufacturers Identified by EPA, 2000/200128 293°
Company
Bayliner Marine Corporation
Beneteau USA Limited
Boston Whaler, Inc.
Brunswick Marine Group
Carver Boat Corporation
Catalina Yachts
Correct Craft, Inc.
Crestliner, Inc.
Fiberglass Unlimited
Fountain Powerboats, Inc.
Four Winns, Inc. LLC
Genmar Industries
Glastron Boats
Godfrey Marine
Grady- White Boats, Inc.
Hood Yacht Systems
Lowe Boats
Lund Boat Company
Magnum Marine Corporation
Mariah Boats, Inc.
MasterCraft Boat Company
Morgan Marine
Ocean Yachts, Inc.
Old Town Canoe Company
Palmer Johnson, Inc.
Porta-Bote International
Regal Marine Industries, Inc.
S2 Yachts, Inc.
Sabre Corporation
Sea Ark Boats, Inc.
Seaswirl Boats, Inc.
Skeeter Boats, Inc.
Smoker-Craft Boats, Inc.
Sport-Craft Boats, Inc.
Sunbird Boat Company, Inc.
Tracker Marine, LLP
Annual Salesa($106)
$450.0
$1.7
$6.0
$483.0
$149.8
$35.0
$35.0
$50.0
$1.0
$57.5
$46.6
$869.0
$58.0
$51.4
$55.0
NAb
$43.8
$60.4
$6.9
$31.7
$87.0
$37.1
$14.6
$11.5
$23.0
$3.6
$85.0
$78.0
$18.4
$6.0
$28.8
$45.0
$52.0
$23.0
$28.8
$57.0
Employment
2,500
10
600
2,900
1,300
250
250
350
16
390
500
6,500
650
550
500
NAb
380
525
60
275
500
400
150
100
200
32
700
600
160
100
250
200
400
200
250
2,400
2-11
-------
Draft Regulatory Support Document
a Annual sales of listed companies include revenues received from the sale of all products sold by these
companies, not just revenues received from the sales of recreational boats.
b NA means Not Available.
2.1.4 Markets
This section examines select historical market statistics for inboard and sterndrive boats and
engines. It presents domestic quantities, values, and unit prices for both boat types as well as
shipment data for inboard and sterndrive engines. Also presented are quantities and values of
exports and imports of both inboard and sterndrive boats and engines. The section concludes
with the current trends of the marine industry. EPA focuses on these two boat configurations
because they are available with diesel engines.
2.1.4.1 Quantity and Price Data
Quantities of shipments produced domestically, real values of shipments, and unit price data
are presented in Tables 2.1-8 through 2.1-10 for inboard runabouts, inboard cruisers, and
sterndrive boats equipped with SI and CI engines (disaggregated data were not available by
engine type). Real unit price data are calculated by simply dividing real value of shipments by
the quantity of shipments produced. Also provided are domestic shipment data for inboard and
sterndrive engines in Table 2.1-11 (price data were not available). While a fraction of inboard
boats are equipped with diesel engines (approximately 18 percent), recall that only 1 to 2 percent
of sterndrive boats contain diesel engines and that sterndrives with diesel engines are more
expensive than those operating with SI engines. Also note that virtually all diesel engines in
inboard boats are placed in cruisers. Only 1 to 2 percent of inboard runabouts contain CI
engines. Because these three boat categories may contain diesel engines, their market data are
discussed here.
An overall examination of the data for all three boat types shows that the quantity of
shipments, real value of shipments, and real unit values all increased over the 1980 to 2000 time
period. Comparing across these boat types shows that the average annual growth rates are
highest for quantities and shipment values for inboard runabouts (9.5 percent for the quantity of
shipments and close to 12 percent for the real value of shipments). The average growth rates for
these same variables are lowest for sterndrive boats (the quantity of shipments grew at an average
annual rate of under 4 percent and the average annual growth rate for the value of shipments was
5 percent). Also notable is that the unit price of inboard runabouts increased, on average, at a
lower rate than for inboard cruisers and sterndrives. Though the average annual growth rates are
positive across the variables presented, there is definite evidence of dips in the quantity of
shipments and real value of shipments for inboard cruisers, and in all three variables for
sterndrive boats. These trends are not existent for inboard runabouts. Before examining the
historical data presented for inboard cruisers and sterndrives, a closer examination at inboard
runabouts is warranted.
2-12
-------
Chapter 2: Industry Characterization
Table 2.1-8
Recreational Inboard Runabout Boats - Domestic Quantity of
Shipments, Value of Shipments, and Unit Values, 1980 - 2000 (1996$) 3132
Year
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Avg. Annual
Growth Rate
Quantity of Shipments
(units)
2,900
2,950
3,200
3,900
4,500
4,500
5,300
6,600
7,400
9,100
7,500
6,200
6,400
6,800
7,200
6,900
6,000
6,100
6,900
12,100
13,600
9.5%
Real Value of Shipments
($103)
$52,226
$55,860
$63,030
$71,217
$84,727
$92,238
$113,964
$137,669
$163,263
$215,846
$152,414
$129,380
$126,358
$141,809
$148,725
$150,673
$126,234
$133,733
$155,707
$293,742
$342,465
11.9%
Real Unit Value
($)
$18,009
$18,935
$19,697
$18,261
$18,828
$20,497
$21,503
$20,859
$22,063
$23,719
$20,322
$20,868
$19,743
$20,854
$20,656
$21,837
$21,039
$21,923
$22,566
$24,276
$25,181
1.9%
Of the three boat types presented here, domestic shipments and the real value of domestic
shipments grew at a higher annual rate, on average, for inboard runabouts. In 1980, just under
3,000 inboard runabouts were being manufactured and distributed in the U.S. The real value of
these boats (in 1996 dollars) was over $52 million, with the average inboard runabout equal to a
real value of $18,000. By 1990, both the quantity of shipments and the real value of shipments
more than doubled. Unit prices increased, but only by 12 percent. In 2000, quantity of
shipments, shipment values, and unit values hit their peak. U.S. shipments of inboard runabouts
were equal to 13,600, real value of shipments equaled over $342 million, and the real value was
just over $25,000.
2-13
-------
Draft Regulatory Support Document
Table 2.1-9
Recreational Inboard Cruiser Boats - Domestic Quantity of
Shipments, Value of Shipments, and Unit Values, 1980 - 2000 (1996$) 33 34
Year
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Avg. Annual
Growth Rate
Quantity of Shipments
(units)
5,300
5,450
5,125
7,485
10,780
12,200
12,700
13,100
13,500
12,300
7,500
3,600
3,550
3,375
4,200
5,460
5,350
6,300
6,600
7,000
8,000
5.0%
Real Value of Shipments
($103)
$802,253
$861,890
$854,167
$1,060,700
$1,604,094
$1,811,865
$1,894,840
$2,135,718
$2,355,750
$2,299,952
$1,589,672
$742,680
$675,032
$696,830
$927,793
$1,193,367
$1,215,268
$1,636,375
$1,631,720
$1,713,733
$2,123,768
7.9%
Real Unit Value
($)
$151,368
$158,145
$166,667
$141,710
$148,803
$148,514
$149,200
$163,032
$174,500
$186,988
$211,956
$206,300
$190,150
$206,468
$220,903
$218,565
$227,153
$259,742
$247,230
$244,819
$265,471
3.1%
Inboard cruisers are larger boats and hence have higher value of shipments and average unit
value measures. An examination of Table 2.1-9 shows that this market has grown over the 1980
to 2000 time period. Evidence of growth in this market can be seen by examining the average
annual growth rates. The real average price of an inboard cruiser was equal to slightly more than
$151,000 in 1980, but by the year 2000, prices reached a peak of $265,471 (a net price increase
of 75 percent). Real shipment values also showed a large increase starting at $802 million in
1980 and rising to over $2.1 billion in 2000. The reason for the large price increase is evident
because the rise in the quantity of shipments from 1980 to 2000 was not as dramatic as the rise in
the real value of shipments. The net increase in the quantity of shipments for the 1980 to 2000
time period was 50 percent.
During the mid to late 1980s, the quantity and real shipment values of inboard cruisers
2-14
-------
Chapter 2: Industry Characterization
steadily increased to reach their peak. In 1983, 7,485 inboard cruisers were manufactured with a
total real value of $1.6 billion. By 1988, shipments rose to 13,500 and the real value of
shipments exceeded $2.35 billion. The average value of this boat type in this same year was
$174,500. This surge in the market for inboard cruisers was followed by a large decline in the
quantities and values of shipments. By 1993, the domestic quantity of inboard cruisers fell to its
lowest level at 3,375 and real value of shipments was close to its lowest level at just under $700
million.
Table 2.1-10
Recreational Sterndrive Boats - Domestic Quantity of Shipments,
Value of Shipments, and Unit Values, 1980 - 2000 (1996$) 3S 36
Quantity of Shipments
Year (units)
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Avg.
Annual
Growth
Rate
56,000
51,000
55,000
79,000
108,000
115,000
120,000
144,000
148,000
133,000
97,000
73,000
75,000
75,000
90,000
93,000
64,500
92,000
91,000
79,600
78,400
3.7%
Real Value of Shipments
($103)
$1,080,702
$1,052,492
$1,039,167
$1,412,841
$2,031,008
$2,247,784
$2,481,280
$3,141,231
$3,230,840
$2,836,265
$2,062,421
$1,436,559
$1,347,147
$1,322,872
$1,738,313
$1,827,867
$1,925,248
$2,027,969
$2,046,755
$1,956,644
$2,106,395
5.0%
Real Unit Value
($)
$19,298
$20,637
$18,894
$17,884
$18,806
$19,546
$20,677
$21,814
$21,830
$21,325
$21,262
$20,553
$19,251
$17,580
$17,271
$18,920
$19,138
$29,264
$21,829
$22,063
$24,122
2.0%
The annual domestic quantities of sterndrive boat shipments far exceed the quantities of
inboard runabouts and inboard cruisers combined. They are mostly equipped with gasoline
2-15
-------
Draft Regulatory Support Document
engines and are in a similar price range as inboard runabouts. A closer examination of Table 2.1-
10 shows that this market peaked and dipped during the same years as the inboard cruiser market.
This general expansion of the market for recreational boats in the late 80s was due to higher
economic growth for the U.S. In 1988, shipments of sterndrives were equal to 148,000 (an 87
percent increase over the year 1983 quantity) and shipment values were equal to over $3.2 billion
(a 128 percent increase in the real shipment value in 1983). Also notable is that though unit
values of sterndrives are far less than those for inboard cruisers, the real value of shipments are
very close for these boat types (approximately $2.1 billion in the year 2000). The value of the
market for inboard runabouts is far smaller at a value of $342 million in 2000.
Table 2.1-11 below provides the quantity of shipments of inboard and sterndrive engines
combined. These data also combine gasoline and diesel engines. What is clear from this table is
that the shipment quantities tend to reflect the peaks and dips seen in the data for sterndrives and
inboard cruisers. Domestic engine shipments rose to their highest value in 1988 at a total of
211,900. They then fell over the remainder of the 1980s and early 1990s to quantities in the low
90 thousands. In the mid 1990s there was a slight rise in engine shipments to a total of 120,000
but in the year 2000, the quantity fell to just over 105,000.
Table 2.1-11
U.S. Shipments of Inboard and Sterndrive Engines, 1980 - 200137
Year
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
Quantity of Shipments
87,750
81,500
85,650
104,125
148,000
155,000
161,900
210,800
211,900
190,700
134,100
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Quantity of Shipments
92,400
94,600
94,700
114,000
120,000
120,000
116,100
104,500
108,500
110,400
105,800
2.1.4.2 Foreign Trade
Tables 2.1-12 and 2.1-13 present trade data for inboard and sterndrive boats. Over the 1992
to 2000 time frame, import values of these boat types grew. A large increase in the value of
inboard cruiser imports was evident from 1999 to 2000. Though they initially are larger, export
values for these boat types do not show the same rising trend. For both boat types, export values
dipped in the early 1990s and then steadily rose through the remainder of the decade. Inboard
export value never recovered to its 1992 level, but sterndrive exports did. In fact, the 2000 value
2-16
-------
Chapter 2: Industry Characterization
of sterndrive exports exceeded its value in 1992.
Further comparisons can be made between exports and imports of each boat type. As the
data in these tables show, inboard import values exceeded their export values during the latter
half of the 1990s. This was not always the case, as prior to 1996, export values were greater. In
1992, the value of inboard imports was only equal to 16 percent of the value of exports but by
1995, they caught up to exports and equaled 92 percent of inboard export values. In 2000,
inboard exports were equal to a fraction of their imports (37 percent).
Table 2.1-12
Import Values3 (S103) of Inboard and Sterndrive Boats, 1992 - 2000 3839
1992 1993 1994 1995 1996 1997 1998 1999 2000
Inboard 8,957 16,781 21,069 56,199 135,800 221,497 301,226 348,107 303,910
Runabouts
Inboard 32,859 87,997 113,858 143,620 142,007 90,184 113,173 151,170 220,214
Cruisersb
Inboards 41,816 104,778 134,927 199,819 277,807 311,681 414,399 499,277 524,124
Total
Sterndrive 10,900 7,965 9,479 15,224 12,090 11,637 22,494 27,894 30,139
Runabouts
Sterndrive 10,976 10,302 18,042 14,779 15,955 15,414 42,599 53,653 70,725
Cruisers0
Sterndrives 21,876 18,267 27,521 30,003 28,045 27,051 65,093 81,547 100,864
Total
a Import values are in nominal U.S. dollars.
b Data for inboard cruisers are for those over 24 feet in length.
0 Data for sterndrive cruisers are for those over 20 feet in length.
Table 2.1-13
U.S. Export Values* (S103) of Inboard and Sterndrive Boats, 1992 - 2000 4041
1992 1993 1994 1995 1996 1997 1998 1999 2000
Inboards 261,474 184,673 163,284 217,443 189,825 222,976 213,111 197,260 198,257
Sterndrives 189,463 127,382 135,229 186,230 191,327 199,364 198,675 236,326 198,349
* Export values are in nominal U.S. dollars.
In the case of Sterndrives, import values remained below the value of sterndrive exports over
the 1992 to 2000 time period. In 1992, imports were equal to approximately 12 percent of export
values. The value of imports did approach exports through the decade and by 2000, they were
equal to about 50 percent of the value of exports. What is notable is a large jump in the value of
sterndrive import values between the years 1997 and 1998. Imports rose from approximately $27
2-17
-------
Draft Regulatory Support Document
million to over $65 million in the span of this year. Sterndrive export values generally increased
through the year 1999 when they hit their peak at $236 million, however in the year 2000, they
fell to just below $200 million. Still, export values for sterndrives were twice the value of their
imports in this year.
Tables 2.1-14 and 2.1-15 present foreign trade data for inboard diesel and sterndrive
engines. Import data for inboard diesel engines were disaggregated by varying ranges of
horsepower (ranging from less than 150 to over 1000 horsepower) while inboard export data are
only available for diesel engines below 200 horsepower. Sterndrive engine data were not
available in disaggregated form. An examination of Table 2.1-14 shows that the total import
value of inboard diesel engines declined and rose over the 1990s. In the early part of the 1990s,
imports of inboard diesel engines steadily declined in value, but then rose dramatically in 1995.
This anomalous year was followed by a decline in import value which remained relatively
constant until it again rose in 2000. For sterndrive engines, import values grew dramatically in
the beginning of the 1990s as well. They then dipped during the mid 1990s only to rise again at
the end of the decade to its highest value.
Though Table 2.1-14 only provides inboard import data for diesels, it is clear that the value
of these engine imports exceed the value of sterndrive engine imports. We can infer that fewer
sterndrive engines were imported relative to inboard engines. Note however, that inboard
engines may also be used for boats with sterndrive engine configurations, which may partially
explain why the import values for inboard engines are much higher.
Export data for the various types of inboard diesel engines were not available, therefore we
are unable to make direct comparisons across the total import and export values of these engines.
Some comparison can be made between the import values of inboard diesel engines below or
equal to 150 horsepower and export values of inboard diesel engines under 200 horsepower since
these generally refer to the same set of engines. A comparison of the these values shows roughly
equal values of imports and exports of this engine type in the 1990s. Overall, export values are
slightly higher. Sterndrive engine import and export values can be directly compared as these
measures represent all foreign trade of this engine type to and from the U.S. From these tables,
we can see that export values of sterndrive engines far exceeded import values in the beginning
of the 1990s. However the value of imports for this engine type approached its export value by
1995. For the latter half of the 1990s, export values remained higher but the difference between
export and import values remained smaller.
2-18
-------
Chapter 2: Industry Characterization
Table 2
.1-14
U.S. Import Values* ($103) of Inboard Diesel
Engines and Sterndrive Engines, 1992 - 2000 42 43
1992
1993
1994
1995
1996
1997
1998
1999
2000
Inboard Diesels
< 150HP
150-199HP
200-3 12HP
313-499HP
500-999HP
> 1000HP
Inboard
Total
17,270
4,901
9,035
4,910
5,365
72,606
114,087
14,230
4,983
9,805
4,288
5,994
40,611
79,911
10,104
5,384
9,153
7,625
8,418
18,577
59,261
8,765
5,539
10,721
7,796
14,257
24,680
293,878
10,050
5,701
7,102
7,634
15,174
39,965
85,626
6,933
7,915
8,851
9,624
13,494
31,486
78,303
9,244
6,528
10,355
15,609
9,808
33, 777
85,321
13,992
6,114
13,032
21,332
10,836
29,002
94,308
15,084
6,916
8,756
38,506
12,725
43,698
125,685
Sterndrive Engines
Total
3,221
5,947
19,045
25,401
21,586
15,457
17,525
25,434
43,489
Import values are in nominal U.S. dollars.
Table 2.1-15
U.S. Export Values* ($103) of Diesel Inboard
Engines Under 200 HP and Sterndrive Engines, 1992 - 2000 444S
Inboard
Engines
Sterndrive
Engines
1992
11,174
25,186
1993
11,332
24,164
1994
8,962
25,024
1995
15,263
28,386
1996
13,976
26,980
1997
20,201
23,734
1998
18,665
17,089
1999
19,123
24,430
2000
23,543
30,427
: Export values are in nominal U.S. dollars.
2-19
-------
Draft Regulatory Support Document
2.2 Large SI Engines and Industrial Equipment
This section gives a general characterization of the Large SI industry. Large SI engines are
nonroad spark-ignition engines that have rated power higher than 25 horsepower (19 kW) and
that are not recreational engines or marine propulsion engines. They are typically derivatives of
automotive engines, but use less advanced technology and operate on LPG and CNG as well as
gasoline. Large SI engines are used in a wide variety of commercial uses. Because it is not
practical to present detailed information on all of these applications in this section, we focus
primarily on forklifts. This is reasonable because they are the dominant application for Large SI
engines. Also, as explained in greater detail in Section 9.7 of Chapter 9, the detailed economic
impact analysis performed for this sector focuses on forklifts. Other information presented in
this section describes some general characteristics of the Large SI sector.
2.2.1 The Supply Side
This section provides a description of the types of industrial equipment that may contain
Large SI engines, the major inputs used to manufacture this equipment, and the costs of
production.
2.2.1.1 Product Types and Populations
Large SI engines are used in a wide variety of applications, including forklifts, generators,
pumps, leaf blowers, sprayers, compressors, other material handling equipment, and agricultural
production. Table 6.2.2-1 in Chapter 6 presents our estimates of the 2000 U.S. population of the
various Large SI equipment applications. We estimated populations of engine and equipment
models using historical sales information adjusted according to survival and scrappage rates.
A 1996 study of the forklift market estimated that there were 491,321 engine-powered
forklifts in use in the United States in 1996 (Classes 4, 5, and 6; see below for an explanation of
these classes).46 That study estimated that 80 percent of this population used LPG (commonly
referred to as propane because propane is its primary constituent), with the rest running on either
gasoline or diesel fuel. If that 20 percent of that population are split evenly between gasoline and
diesel fuels, as we estimate, this means that the number of spark-ignition forklifts in 1996 was
about 442,000, or that about 90 percent of all forklifts were spark-ignition. As noted in Table
6.2.2.1, we estimate that about 95 percent of those spark-ignition forklifts are run LPG or CNG,
with the rest being run on gasoline. The high percentage of propane systems for forklifts can be
largely attributed to expenses related to maintaining fuel supplies. LPG cylinders can be readily
exchanged with minimal infrastructure cost. Installing and maintaining underground tanks for
storing gasoline has always been a significant expense, which has become increasingly costly due
to the new requirements for replacing underground tanks.
With regard to non-forklift applications, the split between LPG and gasoline is not as clear.
Large SI engines today are typically sold without fuel systems, which makes it difficult to assess
2-20
-------
Chapter 2: Industry Characterization
the distribution of engine sales by fuel type. Also, engines are often retrofitted for a different fuel
after the initial sale, making it still more difficult to estimate the prevalence of the different fuels.
Natural gas, a third option, is less common in Large SI engines even though natural gas and LPG
fuel systems are very similar. Natural gas supply systems typically offer the advantage of
pipeline service, but the cost of installing high-pressure refueling equipment is an obstacle to
increased use of natural gas. Table 6.6.2.1 contains our estimates of the use of LPG and CNG
for non-forklift applications; the rest are estimated to use gasoline. We estimate 100 percent
LPG/CNG use for oil field equipment, gas compressors, and refrigeration/AC. For construction,
general industrial, and other nonroad equipment, there may be a mix of central and noncentral
fueling; we therefore believe that estimating an even mix of LPG and gasoline for these engines
is most appropriate.
We estimate very low or no LPG/CNG use for agricultural and lawncare equipment. Lawn
and garden equipment is usually not centrally fueled and therefore operates almost exclusively on
gasoline, which is more readily available. Agriculture equipment is predominantly powered by
diesel engines. Most agriculture operators have storage tanks for diesel fuel. Those who use
spark-ignition engines in addition to, or instead of, the diesel models, would likely invest in
gasoline storage tanks as well, resulting in little or no use of LPG or natural gas for those
applications. An estimated distribution of fuel types for the individual applications are listed in
Table 6.2.2-1.
Large SI engines also vary considerably by size. Most of these engines are smaller than 100
horsepower, with the lower limit of the engine category at 25 horsepower. On an annual sales
basis, 34 percent of Large SI engines are less than 50 horsepower, and 80 percent are less than
100 horsepower. Only about 20 percent are larger than 100 horsepower, with the largest about
250 horsepower.
2.2.1.2 Engine Design and Operation
Most engines operate at a wide variety of speeds and loads, such that operation at rated
power (full-speed and full-load) is rare. To take into account the effect of operating at idle and
partial load conditions, a load factor indicates the degree to which average engine operation is
scaled back from full power. For example, at a 0.3 (or 30 percent) load factor, an engine rated at
100 hp would be producing an average of 30 hp over the course of normal operation. For many
nonroad applications, this can vary widely (and quickly) between 0 and 100 percent of full
power. Table 6.2.2-1 shows the load factors that apply to each nonroad equipment application.
Table 6.2.2-1 also shows annual operating hours that apply to the various applications.
These figures represent the operating levels that apply through the median lifetime of equipment.
2.2.1.3 Liquid-Cooled , Automotive-Derived Engines
The majority of Large SI engines are industrial versions of automotive engines and are
liquid-cooled. However, in the absence of emission standards there has been only limited
2-21
-------
Draft Regulatory Support Document
transfer of emission-control technology from automotive to industrial engines, and most of these
are equipped with only very basic emission control technology if any.
Producing an industrial version of an automotive engine typically involves fitting a common
engine block with less expensive systems and components appropriate for nonroad use.
Manufacturers remove most of the sophisticated systems in place for the high-performance, low-
emission automotive engines to be able to produce the industrial engine at a lower cost. For
example, while cars have used electronic fuel systems for many years, almost all industrial Large
SI engines still rely on mechanical fuel systems. Chapter 3 describes the baseline and projected
engine technologies in greater detail.
2.2.1.4 Air-Cooled Engines
Some manufacturers produce Large SI engines exclusively for industrial use. Most of these
are air-cooled. Air-cooled engines with less than one liter total displacement are typically very
similar to the engines used in lawn and garden applications. Total sales of air-cooled engines
over one liter have been about 9,000 per year, 85 percent of which are rated under 50 hp. While
these engines can use the same emission-control technologies as water-cooled engines, they have
unique constraints on how well they control emissions. Air-cooling doesn't cool the engine
block as uniformly as water-cooling. This uneven heating can lead to cylinder-to-cylinder
variations that make it difficult to optimize fuel and air intake variables consistently. Uneven
heating can also distort cylinders to the point that piston rings don't consistently seal the
combustion chamber. Finally, the limited cooling capacity requires that air-cooled engines stay
at fuel-rich conditions when operating near full power.
While air-cooled engines account for about 9 percent of Large SI engine sales, their use is
concentrated in a few specialized applications. Almost all of these are portable (non-motive)
applications with engine operation at constant speeds (the speed setting may be adjustable, but
operation at any given time is at a single speed). Many applications, such as concrete saws and
chippers, expose the engine to high concentrations of ambient particles that may reduce an
engine's lifetime. These particles could also form deposits on radiators, making water-cooling
less effective.
2.2.1.5 Forklift Truck Manufacturing
As noted above, forklifts are the most common application of Large SI engines. Forklifts
are self-propelled trucks equipped with platforms that can be raised and lowered. These trucks
are used for lifting, stacking, retrieving, and transporting materials and are typically powered by
either LPG, gasoline, diesel, or an electric motor. It is estimated that 80 percent of the forklift
trucks in these classes operate on LPG.47 The industry classifies forklifts in six categories, and
the types of forklifts with Large SI engines are those classified as Class 4, 5, and 6. They
represent those forklift truck classes that may be affected by the emissions control program.
Descriptions of Class 4, 5, and 6 forklifts are as follows48:
2-22
-------
Chapter 2: Industry Characterization
Class 4. Internal Combustion (1C) Engine Trucks - fork, counterbalanced, cushion tire, rider
trucks;
• Class 5. 1C Engine Trucks - fork, counterbalanced, pneumatic tire, rider trucks; and
• Class 6. Electric and 1C Engine Tractors - sit down rider, drawbar pull.
The major difference between Class 4 and Class 5 forklifts is the type of tire installed.
Pneumatic tires allow forklift trucks to be operated on varied terrain, while cushion tires are more
suitable for flat floor surfaces. All of these forklifts allow for the operator to sit down, thus
reducing operator fatigue or strain. Generally speaking, forklifts may differ in their design,
maximum lift capacity, location of the lift operator, type of tires installed, and by the type of fuel
used.
The costs of producing forklift trucks fall into three major categories: capital expenditures,
labor costs, and the costs of materials. Capital expenditures include the manufacturer's costs of
equipment and its installation; labor costs include the producer's costs associated with employees
wages and benefits; and the costs of materials are the costs of tangible and intangible inputs such
as internal combustion (1C) engines, steel for the truck frame, tires, rubber hosing and belting,
counterbalances, and energy. Table 2.2-1 shows the historical production costs for the industrial
truck, tractor, trailer, and stacker machinery manufacturing industry which includes forklift
manufacturers. This industry is identified by Standard Industrial Code (SIC) 3537 and the North
American Industrial Classification System (NAICS) Code 333924.
U.S. Department of Commerce statistics, set out in Table 2.2-1, show that the average value
of shipments (VOS) for this industry over the 1992 to 1999 time period is equal to approximately
$4.7 billion, with the highest value of shipments occurring in 1998. The cost of materials for this
industry is equal to an average of almost $3 billion (64 percent of VOS). The average cost of
labor is approximately $746 million (16 percent of VOS), while capital expenditures are equal to
an average value of $93 million (2 percent of VOS). Examination of this data clearly shows that
capital expenditures represent the smallest share of the value of shipments while the cost of
materials represents the largest share.
2-23
-------
Draft Regulatory Support Document
Table 2.2-1
Value of Shipments (VOS) and Production Costs for the SIC and
NAICS Codes that Include Forklift Manufacturers*, 1992 - 1999 ^,50,51,52,53,54,55
Year
1992
1993
1994
1995
1996
1997
1998
1999
Industry
Code
SIC 3537
SIC 3537
SIC 3537
SIC 3537
SIC 3537
NAICS 333924
NAICS 333924
NAICS 333924
Average
VOS
($106)
$2,754
$3,200
$4,054
$4,970
$4,866
$5,538
$6,248
$5,597
$4,653
Payroll
($106)
$499
$592
$628
$723
$742
$894
$944
$942
$746
% Gi-
VCS
18%
19%
15%
15%
15%
16%
15%
17%
16%
Cost of Materials
($106)
$1,701
$1,984
$2,700
$3,251
$3,076
$3,612
$4,112
$3,429
$2,983
%of
VOS
62%
62%
67%
65%
63%
65%
66%
61%
64%
Total Capital
Expenditures
($106)
$58
$43
$71
$94
$107
$140
$104
$127
$93
%of
VOS
2%
1%
2%
2%
2%
3%
2%
2%
2%
* Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars.
2.2.2 The Demand Side
This section provides information about the uses and consumers of Large SI engines and
forklift trucks. The various industrial sectors in which forklifts are used and the substitute
products for forklifts are also discussed.
Generally speaking, industrial SI equipment is considered a final good while Large SI
engines are referred to as intermediate goods. This is because the engines are manufactured to be
used as inputs to the production of industrial SI equipment. Consumers in the marketplace
demand industrial equipment which may contain Large SI engines, therefore their demand for
Large SI engines is derived from their demand for industrial equipment.
Manufacturers of industrial equipment have three options to obtain the SI engines they use
for equipment production. Their first options is to produce the SI engines used in their final
products. The second option is to purchase a partially finished engine and add on the fuel system
and perform the engine calibration in-house. The third options is to purchase a completed engine
and "drop" it in their equipment without modification. When equipment companies purchase
Large SI engines as an input to their production, they are considered the immediate consumers of
Large SI engines. However, if equipment manufacturers choose to produce Large SI engines as
inputs for their production of equipment, they have vertically integrated the production of a vital
input, SI engines, into their overall production process. Though they consume the engines in the
production of industrial equipment, they are, in this case, the suppliers of these engines via the
2-24
-------
Chapter 2: Industry Characterization
final product.
In the case of forklifts, engines are commonly purchased from outside companies. However,
the design and assembly of these engines may be completed in-house (i.e., adding the fuel system
and calibrating the engine). Sometimes the forklift manufacturer is the designer of the engines,
but in other cases, the forklift manufacturer may rely on its parent company to work on engine
design while it focuses exclusively on forklift production. This secondary arrangement is
common in large companies which may contain a subsidiary producer of forklift trucks. Because
engine designs may be specific, contractual arrangements may be made between engine
manufacturers and forklift producers so as to keep the supply of engines consistent.
2.2.2.1 Uses of Forklifts
The main function of forklift trucks is to lift and transport materials. Class 4, 5, and 6
forklifts are used in indoor settings, such as warehouses and stock rooms or in some outdoor
settings. Table 2.2-2 shows the population of forklift trucks by industry sector for the year 1995,
the most recent year for which industry data is available. The manufacturing sector uses the
largest share of forklifts followed next by wholesale trade. Together, these two industry sectors
accounted for over 60 percent of the U.S. total forklift population in 1995. This estimate is based
on industry shipments and allows for scrappage of older units.
Table 2.2-2
1995 Class 4, 5, and 6 Forklift Population by Industry Sector56
Industry Sector
Manufacturing
Wholesale Trade
Transportation, Communication, and
Utilities
Services
Retail Trade
Construction
Other
Total
Population
196,985
100,721
68,785
46,675
32,919
29,497
13,757
489,339
Percent
Share (%)
40.3%
20.6%
14.1%
9.5%
6.7%
6.0%
2.8%
100%
2.2.2.2 Substitution Possibilities for Forklifts
The most common substitute for Class 4, 5, and 6 1C engine forklifts are electric motor
forklifts, which fall into Classes 1, 2, and 3. Descriptions of these forklifts are as follows57:
Class 1. Electric Motor Rider Trucks;
• Class 2. Electric Motor Narrow Isle Trucks; and
2-25
-------
Draft Regulatory Support Document
Class 3. Electric Motor Hand Trucks.
Electric-powered forklifts are also used for lifting, transporting, and stacking of materials,
but they differ in design and lift capacity from Class 4, 5, and 6 lift trucks. Design differences
may lead a consumer to choose one type of forklift over another. For example, narrow isle trucks
are commonly found in warehouses that are designed to use less floor space and rely more on
vertical stacking. Rider-type forklift trucks are used when significant amounts of material must
be moved or where operator fatigue may be an issue. Hand trucks are used for lighter loads and
are operated using a handle.58 Generally, electric forklifts have lower material-handing capacity.
One advantage of Class 1, 2, and 3 forklifts is that they do not produce exhaust fumes while
in operation, thus making them well suited to indoor operations. However, electric forklifts rely
on batteries that must be recharged which may lead to times where forklifts are not available.
Changing out spent batteries to reduce recharge time is not generally practical because these
batteries are expensive (as much as $10,000 or more each) and can weigh 1,000 Ibs. While
electric forklifts can operate for about 8 hours on a charge, LPG forklifts can operate for about 12
hours before refueling. Consequently, electric forklifts may be a practical alternative only in
some applications.
Aside from electric powered forklifts, other modes of transporting materials may be
considered. For lighter loads, non-motorized hand pallet trucks and stacker machinery may be
acceptable substitutes. They are less expensive but have low load capacities. These types of
equipment also rely more heavily on manual labor.
2.2.2.3 Customer Concerns
As illustrated in Table 6.6.2.1, most Large SI engines are used in industrial applications.
These industrial customers have historically been most concerned about the cost of the engine
and equipment, and about reliability. In many cases, equipment users value uniform and familiar
technology because these characteristics simplify engine maintenance. As described in Chapter
5, equipment users have largely ignored the potential for improving fuel economy when they
make their purchase decisions. As a result most Large SI engines being sold today have
relatively simple carburetor technology that is similar to automotive technology of the early
1960s.
Another user concern relates to emissions. A large number of these engines are operated
indoors or in other areas with restricted airflow much of the time. For these applications,
customers generally want engines with lower CO emissions. Consequently, most engines used in
these applications are fueled with LPG or CNG. However, calibration or maintenance problems
in the field can cause dangerously high CO levels in these engines. Occasionally customers
purchase engines equipped with exhaust catalysts to protect operators from exposure to high
emission levels.
2-26
-------
Chapter 2: Industry Characterization
2.2.3 Industry Organization
It is important to gain an understanding of how the Large SI equipment and engine industries
may be affected by the emission control program. One way to determine how increase costs may
affect the market is to examine the organization of each industry. This section provides data to
measure the competitive nature of the forklift and Large SI engine industries and lists
manufacturers of these equipment and engines. It should be noted that while forklift
manufacturers will be affected by changing engine designs, only those companies that certify
their engines with EPA will be directly regulated.
This section does not contain detailed information on non-forklift application. While these
other sectors will be affected by the control program, it is not practical to report detailed
information for each.
2.2.3.1 Market Structure
Market structure is of interest because it determines the behavior of producers and
consumers in the industry. In perfectly competitive industries, no producer or consumer is able
to influence the price of the product sold. In addition, producers are unable to affect the price of
inputs purchased for use in production. This condition is most likely to hold if the industry has a
large number of buyers and sellers, the products sold and inputs used are homogeneous, and entry
and exit of firms is unrestricted. Entry and exit of firms are unrestricted for most industries,
except in cases where the government regulates who is able to produce output, where one firm
holds a patent on a product, where one firm owns the entire stock of a critical input, or where a
single firm is able to supply the entire market. In industries that are not perfectly competitive,
producer and/or consumer behavior can have an effect on price.
Concentration ratios (CRs) and the Herfindahl-Hirschman index (HHI) can provide some
insight into the competitiveness of an industry. The U.S. Department of Commerce reports these
ratios and indices for the six digit NAICS code level for the year 1997, the most recent year
available. Table 2.2-3 provides the four- and eight-firm concentration ratios (CR4 and CRS,
respectively), and the Herfindahl-Hirschman index for the industrial truck, tractor, trailer, and
stacker machinery manufacturing industry, the industry that includes producers of forklifts. This
industry is represented by NAICS code 333924. Concentration ratios are provided in percentage
terms while HHI are based on a scale formulated by the Department of Justice.
The criteria for evaluating the HHI are based on the 1992 Department of Justice Horizontal
Merger Guidelines. According to these criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are
considered moderately concentrated (i.e., moderately competitive), and those with HHIs above
1,800 are considered highly concentrated (i.e., less competitive). In general, firms in less
concentrated industries have more ability to influence market prices. Based on these criteria, the
industry that produces forklifts can be modeled as perfectly competitive for the purposes of the
economic impact analysis, since their HHI is 503.
2-27
-------
Draft Regulatory Support Document
Table 2.2-3
Measures of Market Concentration for the NAICS Code
that Includes Forklift Manufacturers, 1997 59
Description
NAICS
333924
CR4
38.5
CR8
52.3
HHI
503
vos
($106)
$5,538.33
Number of
Companies
434
2.2.3.2 Large SI Engine and Forklift Manufacturers
Using data from Power Systems Research for the period 1994-96, we have identified seven
principal manufacturers of Large SI engines. These are listed in Table 2.2-4, along with their
average annual sales volume. This table shows that sales volumes are relatively evenly
distributed among these seven manufacturers. The figures for "other" manufacturers presents
aggregated data from four additional companies: Volkswagen, Westerbeke, Hercules, and
Chrysler. While the market has changed over recent years, with some manufacturers dropping
out of the market, General Motors, Mitsubishi Motors, Ford Power Products, and Nissan
Industrial Engines continue to have roughly equal shares and represent between 60 and 70
percent of the annual sales of these engines in the United States.
Table 2.2-4
Engine Sales by Manufacturer (1994-1996)
Manufacturer
General Motors
Mitsubishi Motors
Ford Power Products
Nissan Industrial Engines
Wis-Con Total Power
Toyota
Mazda
Other
Total
Average Annual
Sales
19,500
15,600
14,000
13,800
12,100
11,800
8,200
7,200
102,300
Distribution
19%
15%
14%
13%
12%
12%
8%
6%
100%
Source: Power Systems Research Database
The degree to which engine manufacturers offer integrated engine and equipment models is
an important factor in determining how companies address the need to redesign their products.
Companies that use their own engine models to produce equipment can coordinate the engine
design changes with the appropriate changes in their equipment models. The principal integrated
2-28
-------
Chapter 2: Industry Characterization
manufacturers (Nissan, Mitsubishi, and Toyota) all produce forklifts. About 40 percent of Large
SI equipment sales are from integrated manufacturers.
Other forklift manufacturers have also been responsible for varying degrees of engine
design. Engine design expertise among these companies is so prevalent that some forklift
manufacturers may assume responsibility for certifying their engines, even though they buy the
engines mostly assembled from other manufacturers.
EPA has identified at least fourteen forklift manufacturers that use Large SI engines. The
majority of these companies produce Class 4 and 5 forklifts, though there are a handful that
manufacture Class 6 forklifts. Table 2.2-5 provides a listing of the forklift manufacturers and
their total annual sales (including sales abroad) for the most current year for which data were
available (2000 or 2001). The table shows that the companies range in size based on their annual
sales.
Table 2.2-5
Annual Sales for Forklift Manufacturing Companies, 2000/2001
60,61,62,63
Company
NACCO Materials Handling Group (owns Hyster and Yale)
Clark Material Handling Company
Mitsubishi Caterpillar Forklift America, Inc.
Nissan Forklift Corporation, North America
Toyota Industrial Equipment Manufacturing
Hyundai Construction Equipment - Material Handling Division
TCM Manufacturing USA
Komatsu Forklift USA, Inc.
Kalmar AC, Inc.
Linde Lift Truck Corporation
Drexel Industries, Inc.
TailiftUSA, Inc.
Blue Giant
Daewoo Heavy Industries America
Annual Sales
($106)
$1,292
$539
$172
$86
$83
$80
$50
$30
$27
$26
$26
$10
$9
$5
2.2.4 Markets
This section examines the historical market statistics for the forklift manufacturing industry.
Historical data on the quantity of domestic shipments and some price data of 1C engine forklifts
are provided. The quantity and values of exports and imports of non-electric forklift trucks are
presented as well.
2-29
-------
Draft Regulatory Support Document
2.2.4.1 Quantity and Price Data
Historical market data on the quantity of U.S. shipments of Class 4, 5, and 6 forklifts are
provided in Table 2.2-6 and were obtained from the Industrial Truck Association Membership
Handbook (2002). As this table shows, there has been an overall increasing trend in the quantity
of forklifts produced in the U.S. with an overall net increase of 118 percent from 1980 to 2000
and an average increase of just under 7 percent per year. During the 1990s, shipments increased
from almost 48,000 in 1990 to approximately 73,000 in 1995, but then dipped in 1996 to just
above 60,000. Since 1996, the general increasing trend in the quantity of SI engine forklifts
manufactured in the U.S. continued with a relatively small dip in 1999. For the purpose of this
economic impact analysis, we used 65,000 forklifts as our baseline quantity of forklifts produced
in 2000, based on production data for the past 10 years. For future year projections, we used the
growth rates contained in our NONROAD model.
Table 2.2-6
U.S. Shipments of Internal Combustion
Class 4, 5, and 6 Forklifts, 1980 - 2000 64
Year
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
Quantity of Shipments
39,448
31,885
18,553
26,245
45,338
47,844
46,195
47,945
48,535
55,104
47,702
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Averag
Quantity of Shipments
38,406
46,183
48,947
65,027
72,685
60,287
64,946
80,554
74,994
85,993
; Annual Growth Rate =
6.7%
Forklift truck prices can vary a great deal depending on their class, the manufacturer, the
model type, and selected options. Pricing data on various Class 4 and 5 forklift models were
obtained from the Handbook of New and Used Equipment Values - 1C Lift Trucks (Equipment
Watch, 2001). Current retail prices for various 1C forklifts with no options for the year 2001
varied from a low of $17,000 up to well over $100,000 for high end models. However, most
models were priced in the range of $25,000 to $50,000.
2-30
-------
Chapter 2: Industry Characterization
2.2.4.2 Foreign Trade
Export and import values and quantities for non-electric forklifts presented in Table 2.2-7
show increasing trends since 1989. Based on this information, the U.S. is a net importer of
forklifts as its value and quantity of imports exceeds it value and quantity of exports. Note,
however, that U.S. domestic production of forklifts far outweighs the quantity it imports. A
closer examination of the export value and quantity data show that while U.S. exports generally
increased over the 1989 to 2001 time period, there was a sharp decline in export quantity and
value in 1996. Exports of forklifts went from a total value of $194.3 million in 1995 to about
$91 million in 1996 (a similar decline is evident in the quantity of forklifts). Since 1996, both
the value and quantity of exports has increased with a slight dip occurring in 2001. U.S. imports
of forklifts has also shown a general increase in both value and quantity, however again, in 2001
a slight dip is evident.
The main importers of non-electric forklifts, related trucks, and parts of forklifts to the U.S.
are Japan, Canada, and the United Kingdom and the main countries the U.S. exports its forklifts
to are Canada, Mexico, and the United Kingdom.65
Table 2.2-7
Import and Export Quantities and Values* for Non-Electric
Self-Propelled Forklift and Other Trucks, 1989 - 200166
Year
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Average
Export Value
($106)
$113
$142
$148
$146
$144
$196
$194
$91
$146
$162
$150
$190
$168
$153
Export
Quantity
7,065
7,651
8,302
9,511
12,762
11,277
10,131
4,963
8,670
9,890
11,526
16,208
12,768
10,056
Import Value
($106)
NA
NA
NA
NA
NA
$301
$389
$375
$459
$611
$574
$612
$507
$294**
Import
Quantity
NA
NA
NA
NA
NA
19,496
22,824
19,214
21,820
29,251
26,741
30,751
23,381
14,883**
a Values are in nominal dollars.
b Average is computed for the years 1994 through 2001.
2-31
-------
Draft Regulatory Support Document
2.3 Snowmobile Market
Snowmobiles are normally one or two passenger vehicles that are used to transverse over
snow-covered terrain. They have a track in the rear similar to that of a bulldozer and runners
(similar to skis) in the front for steering. Snowmobiles are used primarily for recreational
purposes. However, a small number of them are produced and used for utility purposes, such as
search and rescue operations. Annual sales of snowmobiles in the U. S. have varied dramatically
over the years. Over 140.6 million units were sold in the U. S. in 2001.67
2.3.1 The Supply Side
This section provides a description of snowmobiles and their engines, the major inputs used
to manufacture this equipment, and the costs of production.
2.3.1.1 Product Types
There are several types of snowmobiles on the market. Snowmobiles types range from
children's models with very low horsepower to high-powered machines with engine sizes
approaching 1000 displacement cc. Snowmobiles are designed to appeal to a variety of
consumers including those who wish to cover rough mountainous terrain, those who seek speed,
those who wish to tour the countryside and the novice snowmobiler. Snowmobiles are offered in
one-seat and two-seat models and in luxury and low-cost varieties. Snowmobile manufacturers
seek to appeal to a wide range of potential snowmobile riders. This section will describe a few of
the components of the models on the market. There are a variety of engine options including
two-stroke or four-stroke, air or water cooled, and various engine displacements. Options
include electric start, reverse, specialized paints, and other items. For a more complete
description of typical snowmobile attributes see Section 9.4.
2.3.1.2 Engine Design and Populations
The vast majority of snowmobiles sold in the U.S. are powered by two-stroke engines
currently. Engine displacements range from 60 cc for an entry-level youth model to 998 cc for a
high-performance model. Based upon PSR snowmobile production data, snowmobiles produced
have been trending towards higher engine sizes with the average engine size increasing over 17
percent between the period 1990 and 2000. In 1996 over 44 percent of the snowmobiles
produced had engine sizes less than 500 cc displacement. In 2000, this percentage had dropped
to 23 percent. In general the larger the engine size, the more powerful for the 2-stroke engines
that dominate the snowmobile market today. The average engine size in 2002 was 570 cc
displacement.68
The number of models produced for a given engine size for the four major snowmobile
manufacturers is shown in Table 2.3-1.
2-32
-------
Chapter 2: Industry Characterization
Table 2.3-1
Engine Displacement for Major Snowmobile Manufacturers in the U.S. Market in 20006
Manufacturers
Arctic Cat, Inc.
Bombardier (Ski Doo)
Polaris Industries
Yamaha
Total
<300cc
852
2,638
2,533
0
6,023
<500cc
14,233
23,507
21,585
10,615
69,940
<700cc
41,253
20,017
34,067
16,483
111,820
700-lOOOcc
8,317
11,973
14,276
6,085
40,651
! Production data were taken from DELINK Database owned by Power Systems Research.
2.3.1.3 Two-Stroke vs Four-Stroke Cycle Engine Usage
The majority of snowmobiles are equipped with 2-stroke engines. For the 2003 models
currently available for sale, nine 4-stroke models are available. Each of the manufacturers offers
4-stroke models in their current sales inventory. For more details see Section 9.4.
2.3.1.4 Production Costs of Snowmobiles
Production costs for snowmobiles are not readily available. In lieu of cost of production
data for snowmobiles specifically, a discussion of the cost of production data for NAICS 366999
Other Transportation Equipment Manufacturing is presented. This category includes
snowmobiles, ATVs, golf carts, and other miscellaneous transportation equipment. As Table
2.3-2 shows, the average value of shipments (VOS) for these industries over the 1992 to 2000
time period is equal to approximately 4.5 billion dollars, with the highest value of shipments
occurring in 2000. The cost of materials for this industry is equal to an average of about 3 billion
dollars (65 percent of VOS). The average cost of labor is approximately 549 million (12 percent
of VOS), while capital expenditures are equal to an average value of 97 million (2 percent of
VOS). Examination of these data clearly shows that capital expenditures and payroll represent
the smallest shares of the value of shipments while the cost of materials represents the largest
share.
2-33
-------
Draft Regulatory Support Document
Table 2.3-2
Value of Shipments (VOS) and Production Costs for the SIC and
NAICS Codes that Includes Snowmobile Manufacturers, 1992 - 2000 70 7172 73 74
Year
1992
1993
1994
1995
1996
1997
1998
1999
2000
Industry
Code
SIC 3799
SIC 3799
SIC 3799
SIC 3799
SIC 3799
NAICS 336999
NAICS 336999
NAICS 336999
NAICS 336999
Average
VOS
($106)
3,087
3,807
3,947
4,539
5,179
4,437
5,033
5,645
6,245
4,568
Payroll
($106)
449
514
469
512
570
496
578
643
714
549
%of
VOS
15%
14%
12%
11%
11%
11%
11%
11%
11%
12%
Cost of Materials
($106)
1,969
2,422
2,611
3,056
3,368
2,803
3,236
3,766
4,195
3,047
%of
VOS
64%
64%
66%
67%
65%
63%
64%
67%
67%
65%
New Capital
Expenditures
($106)
62
86
98
86
103
97
122
106
117
97
%of
VOS
2%
2%
2%
2%
2%
2%
2%
2%
2%
2%
* Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars
2.3.2 The Demand Side
This section provides information on the uses of snowmobiles, various substitute products
on the market, and information concerning consumers who purchase snowmobiles.
2.3.2.1 Uses of Snowmobiles
There are a variety of snowmobile types currently produced and tailored to a variety of
riding styles. The majority of the overall snowmobile market is made up of high performance
machines. These snowmobiles have fairly high powered engines and are very light, giving them
good acceleration speed and handling. The performance sled come in several styles. Cross
country sleds are designed for aggressive trail and cross country riding. Mountain sleds have
longer tracks and wider runner stance for optimum performance in mountainous terrain. Finally,
muscle sleds are designed for top speeds (in excess of 120 miles per hour) over flat terrain such
as frozen lakes. Performance snowmobiles are generally designed for a single rider.
The second major style of snowmobile is designed for casual riding over groomed trails.
These touring sleds are designed for one or two riders and tend to have lower powered engines
than performance snowmobiles. The emphasis in this market segment is more on comfort and
convenience. As such, these sled feature more comfortable rides than performance machines and
tend to have features such as electric start, reverse, and electric warming hand grips.
2-34
-------
Chapter 2: Industry Characterization
The last and smallest segment of the snowmobile market is the utility sled segment. Utility
snowmobiles are designed for pulling loads and for use in heavy snow. Thus the engines are
designed more for producing torque at low engine speeds, which typically corresponds to a
reduced maximum speed of the snowmobiles. Utility snowmobiles are common in search and
rescue operations.
A typical snowmobile lasts thirteen years and travels approximately 17,000 miles over its
lifetime. The average snowmobile is used 57 hours per year.75
2.3.2.2 Substitution Possibilities
A number of substitute products to snowmobiles exist. Consumers can substitute across off-
road recreational vehicles. However, ATVs and off-highway vehicles may not be used safely in
the snow. Snow coaches are a substitute motorized product. Consumers may be interested in
engaging in outdoor activities, but may instead consider doing a non-motorized activity. For
example, consumers who are interested in being outside in the snow may engage in skiing or
sledding. Recreational indoor activity of many types are substitute possibilities for snowmobile
riding.
2.3.2.3 Customer Demographics and Customer Concerns
Based upon ISMA data, the average snowmobile owner is 42 years old, and had an average
annual income of $68,000 in 2001. The average snowmobile rider has 18 years experience in
riding. The majority of snowmobile owners are married. Approximately 63 percent of riders
trailer their snowmobiles to go riding.76
Good performance is very important to snowmobilers. This is especially true for the
performance segment of the market, where high power and low weight are crucial for the
enjoyment of the performance snowmobile enthusiast. The performance snowmobile segment is
driven by a constant demand for more power and lower weight. In the touring segment of the
market, performance in terms of power and weight is somewhat less important but still
significant. In all snowmobile market segments, durability and reliability are very important to
the customer.
The price of a snowmobile produced by the four major manufacturers currently ranges from
about $3,700 for entry level models to around $12,000 for some high performance models. The
average snowmobile price in 2001 was $6,360. Some of the high performance snowmobiles
produced by the small manufacturers can approach $20,000, but this is an extremely small niche
market. Since snowmobiles are a discretionary purchase, price is a factor in the consumers
decision to purchase.
2-35
-------
Draft Regulatory Support Document
2.3.3 Industry Organization
Because there are costs associated with the emission control program, it is important to
determine how the snowmobile industry may be affected. Industry organization is an important
factor which affects how a market may react to regulatory costs. This section provides a
description of the organization of the snowmobile industry.
2.3.3.1 Market Structure
Market structure is of interest because it determines the behavior of producers and
consumers in the industry. In perfectly competitive industries, no producer or consumer is able
to influence the price of the product sold. In addition, producers are unable to affect the price of
inputs purchased for use in production. This condition is most likely to hold if the industry has a
large number of buyers and sellers, the products sold and inputs used are homogeneous, and entry
and exit of firms is unrestricted. Entry and exit of firms are unrestricted for most industries,
except in cases where the government regulates who is able to produce output, where one firm
holds a patent on a product, where one firm owns the entire stock of a critical input, or where a
single firm is able to supply the entire market. In industries that are not perfectly competitive,
producer and/or consumer behavior can have an effect on price.
Concentration ratios (CRs) and the Herfindahl-Hirschman index (HHI) can provide some
insight into the competitiveness of an industry. The U.S. Department of Commerce reports these
ratios and indices for the six digit NAICS code level for the year 1997, the most recent year
available. Table 2.3-3 provides the four- and eight-firm concentration ratios (CR4 and CRS,
respectively), and the Herfindahl-Hirschman index for the NAICS code 336999, Other
Transportation Equipment Manufacturing, the industry category that includes producers of
snowmobiles. Note that the concentration ratio is reported in percentage terms while the HHI is
based on a scale developed by the Department of Justice. For this industry the CR4 was 50.7
percent and the CRS was 75.3 percent.
The criteria for evaluating the HHI are based on the 1992 Department of Justice Horizontal
Merger Guidelines. According to these criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are
considered moderately concentrated (i.e., moderately competitive), and those with HHIs above
1,800 are considered highly concentrated (i.e., less competitive). In general, firms in less
concentrated industries have more ability to influence market prices. Based on these criteria, the
NAICS category that includes firms that produce snowmobiles can be considered unconcentrated
or more competitive.
2-36
-------
Chapter 2: Industry Characterization
Table 2.3-3
Measures of Market Concentration for the NAICS Code
that Includes NAICS 336999 Manufacturers, 1997 77
Description
NAICS 336999
CR4
50.7
CR8
75.3
HHI
885.2
vos
($106)
$4,436,67
9
Number of
Companies
349
However, it is important to recognize that four producers dominate the snowmobile industry
or produce 99 percent of the worldwide snowmobiles produced and sold. This information
suggests that snowmobile manufacturing is highly concentrated with four manufacturers
dominating the market. However, when one considers firm behavior within the industry and the
availability of numerous product substitutes, the picture alters somewhat. While snowmobile
manufacturing is concentrated, snowmobiles represent a small fraction of total recreational
products available in the market place.
Market structure is important to assessing the potential impacts of a regulation on an
industry because it determines the behavior of producers and consumers within the industry.
Economists often estimate concentration ratios for the subject market or industry to assess the
competitiveness. More (less) concentrated markets are considered to be less (more) competitive.
The extremes are defined by perfect competition (many buyers/seller with no influence over
price) and monopoly (one seller with control over setting price). Between these two extremes are
varying degrees of imperfect competition, or oligopoly, that depend upon different assumptions
of strategic behavior among sellers within the market or industry. The competitiveness will
depend upon the definition of the subject market or industry with those being more (less) broadly
defined demonstrating more (less) competition. For example, the "snowmobile" market is
dominated by four major producers and may be considered less competitive. However, there are
likely to be many substitutes for snowmobiles when considering the broader "recreational
vehicles" or "recreational activities" markets. These substitutes increase the competitive nature
of the market or industry. In previous regulatory analysis, the Agency has modeled the
imperfectly competitive nature of pharmaceuticals (product differentiation) and cement (regional
barriers to entry) where there were commonly accepted and researched approaches. Rather than
add uncertainy to model outcomes by speculating on the strategic interactions of producers here,
we chose to model the markets as perfectly competitive. Generally speaking, this assumption
will tend to understate the price and output changes associated with regulation and may overstate
the profit loss of producers; however, the extent of the bias is unknown and direction may vary
by producer.78
2.3.3.2 Snowmobile Manufacturers
Manufacturers of snowmobile were formerly classified under the SIC code 3799 and are
now classified under NAICS code 336999, Other Transportation Equipment Manufacturing. The
2-37
-------
Draft Regulatory Support Document
Small Business Administration (SBA) uses SIC/NAICS categories to classify businesses as large
or small, depending on the number of employees or sales criteria. Snowmobile manufacturers
have the NAICS sub-classification 3369993414 and must have fewer than 500 employees to be
considered a small business by SBA. Snowmobile wholesale companies may also be impacted
by this regulation. Wholesale dealers of snowmobiles are categorized as NAICS classification
421110 - Automobile and Other Motor Vehicle Wholesales, and are considered small business if
they have fewer than 100 employees.
There are four major manufacturers of snowmobiles that account for almost the entire U.S.
market. These manufacturers are Arctic Cat, Bombardier (Ski Doo), Polaris and Yamaha.
Polaris is the largest snowmobile manufacturer by sales volume, followed by Arctic Cat,
Bombardier, and Yamaha. There are less than five snowmobile manufacturers that combined
make up significantly less than one percent of the U.S. snowmobile market. These snowmobile
manufacturers specialize in high performance snowmobiles and other unique designs (such as
stand-up snowmobiles).
Bombardier and Yamaha produce the engines used in the snowmobiles they sell. In
contrast, Polaris and Arctic Cat purchase engines for the snowmobiles they sell. Arctic Cat
typically purchases Suzuki engines, while Polaris purchases engines made by Fuji Corporation.
2.3.4 Snowmobile Retailers and Rental Firms
In contrast to the small number of manufacturers producing snowmobiles, there are over
1,500 registered snowmobile dealers in the United States according to ISMA data.
Approximately the same number operate in Canada and Scandinavia. These firms typically do
not sell snowmobiles exclusively, but also sell other recreational vehicle products such as ATVs
and motorcycles. Snowmobile retailers are included in NAICS category 441229 - All Other
Motor Vehicle Dealers, and are considered small business if annual sales revenues are less than
$6.0 million. In additional to retailers, rental firms exist that purchase snowmobiles to rent to the
occasional snowmobile rider. These firms are included in NAICS category 532292 -
Recreational Goods Rental, and are considered small business if the firm experiences sales less
than $6.0 million. Potentially, both retailers and rental firms may be impacted by the regulation
to the extent that the price of the snowmobiles the firms sell or rent increase.
2.3.5 Markets
This section examines the historical market data for the snowmobile industry. Historical
data on the quantity of domestic shipments and price data of snowmobiles are provided.
2.3.5.1 Quantity and Price Data
Historical market data on the quantity of snowmobiles sold in the U.S. are provided in Table
2.3-4. Data were obtained from ISMA.79 As this table shows, there has been an overall
increasing trend in the quantity of snowmobiles sold in the U.S. with an average annual increase
2-38
-------
Chapter 2: Industry Characterization
of 6 percent from 1990 to 2001. However, annual sales declined in 1991 and 1998 through 2000.
Sales of snowmobiles increased more than 76 percent between the years 1990 and 2001. Retail
dollars sales increased, on average, by 11 percent annually from 1990 to 2001. Snowmobile
retail dollars per unit have also increased, showing an annual average increase of 5 percent for
the same period.
Table 2.3-4
U.S. Units Sold, Retail Dollars and Retail Dollars Per Unit
Snowmobiles, 1990 - 2001 80
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
11 -year
Annual
Average
Change
1990 to
2001
Unit Sales
80,000
78,000
81,946
87,809
114,057
148,207
168,509
170,325
162826
147867
136,601
140,629
137,889
%
Change
Unit
Sales
—
(3%)
5%
7%
30%
30%
14%
1%
(4%)
(9%)
(8%)
3%
6%
76%
Retail Dollars
($106)
$300.0
$323.7
$356.0
$403.9
$558.9
$791.3
$905.2
$1,005.8
$975.1
$882.8
$821.0
$894.4
$747
% Change
Retail
Dollars
—
8%
10%
13%
38%
42%
14%
11%
3%
9%
7%
9%
11%
198%
Retail
Dollars/
Unit
$3,750
$4,150
$4,344
$4,600
$4,900
$5,339
$5,372
$5,905
$5,988
$5,970
$6,000
$6,360
$5,698
% Change Retail
Dollars/Unit
—
11%
5%
6%
7%
9%
1%
10%
1%
0%
1%
6%
5%
70%
*Dollar values and percent changes of dollar values presented are nominal values.
2.3.5.2 Foreign Trade
In general, export and import data are not available for the snowmobile market. Data for
SIC 3799 are available from the International Trade Commission. These data are presented on
Table 2.4-6, Import and Export Quantities and Values for ATVs, 1989-2001, in Section 2.4, All-
Terrain Vehicles, below. However, SIC 3799 includes snowmobiles, ATVs, golf carts and other
transportation equipment. Thus the trade data is not specific to snowmobiles. World wide sales
data for snowmobiles are presented in Table 2.3-5. During 2000 approximately 40 percent of
total worldwide production was produced by Bombardier and Yamaha, foreign companies with
the remainder of 60 percent produced by Arctic Cat and Polaris, domestic manufacturers.
2-39
-------
Draft Regulatory Support Document
Table 2.3-5
Worldwide Production, Sales, and Inventories of Snowmobiles 1990 - 200181
Year Worldwide Production
(103 units)
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
174.9
157.2
116.3
146.0
185.0
231.5
260.9
273.7
270.7
231.7
205.0
190.3
Worldwide Retail Sales
(103 units)
163.4
153.0
150.0
158.0
181.0
227.4
252.3
260.7
257.9
230.9
208.3
208.5
Worldwide
Inventory
(103 units)
55.5
59.7
27.9
16.0
18.6
22.6
31.1
44.2
56.9
57.7
54.4
36.1
2.4 All-Terrain Vehicles
All Terrain Vehicles (ATVs) are normally one-passenger open vehicles that are used for
recreational and other purposes requiring the ability to traverse over most types of terrain. Most
modern ATVs have four-wheels, and have evolved from three-wheeled designs that were first
introduced in the 1970s. According to data provided by the Motorcycle Industry Council (MIC),
production of ATVs sold in the U.S. has averaged about 390,000 units between 1996 and 2001.
However, ATV sales have increased during that time to more than 880,000 units in 2001. ATVs
therefore constitute the largest single category of non-highway recreational vehicles, though it is
difficult to calculate the total vehicle population at any given point in time since many states do
not require registration of ATVs.
2.4.1 The Supply Side
This section provides a description of ATVs and their engines, the major inputs used to
manufacture this equipment, and the costs of production.
2-40
-------
Chapter 2: Industry Characterization
2.4.1.1 Product Types
There are several types of ATVs on the market. This section will describe a few of the
components of the models on the market. There are a variety of engine options including two-
stroke or four-stroke, air or water cooled, and various engine displacements. Options also
include 5-speed manual or automatic transmissions.
2.4.1.2 Engine Design and Populations
The majority of ATVs sold in the U.S. are powered by single-cylinder, four-stroke cycle
engines of less than 40 horsepower, operating under a wide variety of operating conditions and
load factors. Engine displacements range from 50cc for an entry-level youth model to 660cc for
a high-performance adult model, but more than three-fourths of them fall in the 200-500cc range.
In the year 2000, ATV manufacturers used 225,246 engines between 200cc and 300cc
displacement (see Table 2.4-1). Of the engines produced, 64 percent were less than 400cc
displacement and 84 percent were less than SOOcc displacement. Over the past four years,
production of engines with greater than SOOcc displacement has increased from approximately 5
percent in 1996 to 16 percent in 2000.
Table 2.4-1
Engine Displacement for Major ATV Manufacturers in the U.S. Market in 200082
Manufacturers
Arctic Cat, Inc.
Honda
Kawasaki Motors
Polaris Industries
Suzuki
Yamaha
Total
<200cc
0
2,429
0
0
0
7,635
10,064
200 - 300cc
14,758
119,661
44,169
21579
9,346
15,733
225,246
300 - 400cc
4,896
7,561
6,780
54,834
0
26,977
101,048
400 - SOOcc
10,869
65,933
0
6,689
1,740
21,743
106,980
200 - 700cc
0
13,583
0
62,144
0
0
75,727
2.4.1.3 Two-Stroke vs Four-Stroke Cycle Engine Usage
Approximately 80 percent of all ATVs produced for U.S. consumption use four-stroke cycle
engines. Of the six major manufacturers, only Polaris, Suzuki and Yamaha used two-stroke
cycle engines at all. The remainder of the two-stroke engines in ATVs sold in U.S. are found in
entry-level or youth models, which are imported from the Far East or assembled in this country
from imported parts. In general, two-stroke engines are less expensive to produce than four-
stroke engines, thus providing a marketing advantage in the youth and entry-level categories. We
estimate that two-strokes make up roughly twenty percent of the market when the imported youth
models are included.
2-41
-------
Draft Regulatory Support Document
2.4.1.4 Production Costs of ATVs
As Table 2.4-2 shows, the average value of shipments (VOS) for this industry over the 1992
to 1999 time period is equal to approximately 4.6 billion dollars, with the highest value of
shipments occurring in 1999. The cost of materials for this industry is equal to an average of
about 3 billion dollars (65 percent of VOS). The average cost of labor is approximately 549
million (12 percent of VOS), while capital expenditures are equal to an average value of 97
million (2 percent of VOS). Examination of these data clearly shows that capital expenditures
and payroll represent the smallest shares of the value of shipments while the cost of materials
represents the largest share.
Table 2.4-2
Value of Shipments (VOS) and Production Costs for the SIC and
NAICS Codes that Includes ATV Manufacturers, 1992 - 2000 83 84 8S 86 87
Year
1992
1993
1994
1995
1996
1997
1998
1999
2000
Industry
Code
SIC 3799
SIC 3799
SIC 3799
SIC 3799
SIC 3799
NAICS 336999
NAICS 336999
NAICS 336999
NAICS 336999
Average
VOS
($106)
3,087
3,807
3,947
4,539
5,179
4,437
5,033
5,645
6,245
4,568
Payroll
($106)
449
514
469
512
570
496
578
643
714
549
%of
VOS
15%
14%
12%
11%
11%
11%
11%
11%
11%
12%
Cost of Materials
($106)
1,969
2,422
2,611
3,056
3,368
2,803
3,236
3,766
4,195
3,047
%of
VOS
64%
64%
66%
67%
65%
63%
64%
67%
67%
65%
New Capital
Expenditures
($106)
62
86
98
86
103
97
122
106
117
97
%of
VOS
2%
2%
2%
2%
2%
2%
2%
2%
2%
2%
* Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars.
2.4.2 The Demand Side
This section provides information on the uses of ATVs, various substitute products on the
market, and the consumers who purchase ATVs.
2.4.2.1 Uses of ATVs
As noted above, ATVs are used for recreational and other purposes. They are mainly used
for, riding on trails. Examples of non-recreational uses are for hauling and towing on farms,
ranches or in commercial applications. Some ATVs are sold with attachments that allow them to
take on some of the functions of a garden tractor or snow blower. ATVs are also used for
2-42
-------
Chapter 2: Industry Characterization
competitive purposes, although not to the same extent as off-highway motorcycles.
2.4.2.2 Alternate Uses of ATV Engines
Although a few ATV engine lines have been used in other applications, such as some
smaller on- and off-highway motorcycles, manufacturers have stated that ATV engines are
normally designed only for use in ATVs. ATV engines may share certain components with
motorcycles, snowmobiles and Personal Water Craft (PWC), but many major components such
as pistons, cylinders and crankcases differ within given engine displacement categories.
2.4.2.3 Substitution Possibilities
Consumers can substitute across off-road recreational vehicles. An off-highway motorcycle
as a substitute would allow the consumer to enjoy the same off-road recreation that they would
receive with an ATV. Consumers may be interested in engaging in outdoor activities, but may
instead consider doing a non-motorized activity. For example, consumers who are interested in
being outside may engage in hiking, running, or riding a bicycle. These non-motorized options
would allow the consumer to participate in outdoor activity, hence they may be considered
substitutes for less intensive off-highway pastime.
2.4.2.4 Customer Concerns
Except for the competitive segment of the market, performance seems to be somewhat less
important to ATV purchasers than it is to purchasers of snowmobiles or off-highway
motorcycles. Most youth models, which form a significant portion of the market, are normally
equipped with governors or other speed-limiting devices. Performance can be important for some
of the higher-end adult models, but handling is also an important consideration, particularly when
riding in dense wooded areas. Durability and reliability are also important to the customer, but
perhaps not as important as price.
The price of an ATV can range from about $1,200 for an entry-level youth model to around
$7,000 or more for a large, high performance machine. ATVs, like other recreational vehicles,
are basically discretionary purchases, although utility may enter into the equation more often than
in the case of off-highway motorcycles or snowmobiles. Cost is an important factor, particularly
in the youth or entry-level segments of the market, and significant cost increases could cause
people to spend their discretionary funds in other areas.
2.4.3 Industry Organization
Because there are costs associated with the emission control program, it is important to
determine how the ATV industry may be affected. Industry organization is an important factor
which affects how a market may react to regulatory costs. This section provides a description of
the organization of the motorcycle industry.
2-43
-------
Draft Regulatory Support Document
2.4.3.1 Market Structure
Market structure is of interest because it determines the behavior of producers and
consumers in the industry. In perfectly competitive industries, no producer or consumer is able
to influence the price of the product sold. In addition, producers are unable to affect the price of
inputs purchased for use in production. This condition is most likely to hold if the industry has a
large number of buyers and sellers, the products sold and inputs used are homogeneous, and entry
and exit of firms is unrestricted. Entry and exit of firms are unrestricted for most industries,
except in cases where the government regulates who is able to produce output, where one firm
holds a patent on a product, where one firm owns the entire stock of a critical input, or where a
single firm is able to supply the entire market. In industries that are not perfectly competitive,
producer and/or consumer behavior can have an effect on price.
Concentration ratios (CRs) and the Herfindahl-Hirschman index (HHI) can provide some
insight into the competitiveness of an industry. The U.S. Department of Commerce reports these
ratios and indices for the six digit NAICS code level for the year 1997, the most recent year
available. Table 2.4-3 provides the four- and eight-firm concentration ratios (CR4 and CRS,
respectively), and the Herfindahl-Hirschman index for the NAICS code 3369991, Other
Transportation Equipment Manufacturing, the industry category that includes producers of
ATVs. Note that the concentration ratio is reported in percentage terms while the HHI is based
on a scale developed by the Department of Justice. For this industry the CR4 was 50.7 percent
and the CRS was 75.3 percent.
The criteria for evaluating the HHI are based on the 1992 Department of Justice Horizontal
Merger Guidelines. According to these criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are
considered moderately concentrated (i.e., moderately competitive), and those with HHIs above
1,800 are considered highly concentrated (i.e., less competitive). In general, firms in less
concentrated industries have more ability to influence market prices. Based on these criteria, the
NAICS category that includes firms that produce ATVs can be considered unconcentrated or
more competitive.
Table 2.4-3
Measures of Market Concentration for the NAICS Code
that Includes ATV Manufacturers, 1997
Description
NAICS 336999
CR4
50.7
CRS
75.3
HHI
885.2
vos
($106)
$4,436,679
Number of
Companies
349
2-44
-------
2.4.3.2 ATV Manufacturers
Manufacturers of ATVs were formerly classified under the Standard Industrial Classification
(SlC)code 3799 and the North American Industrial Classification System (NAICS) code 336999,
Other Transportation Equipment Manufacturing. These codes are used by the Small Business
Administration (SBA) uses SIC/NAICS categories to classify businesses as large or small,
depending on the number of employees or sales criteria. ATV manufacturers have the NAICS
sub-classification 3369993101 and must have fewer than 500 employees to be considered a small
business by SBA. In addition to manufacturers, there are a number of importers of ATVs,
classified under NAICS code 42111, the code that also includes importers of automobiles, trucks,
motorcycles and motor homes. To be classified as a small business by SBA for this NAICS code,
an importer must have fewer than 100 employees.
Using data including the Power Systems Research (PSR) Database, Dun & Bradstreet
(D&B) Market Identifiers Online Database, and information from the MIC identified 16
manufacturers and 17 importers of ATVs. ATV producers and importers are listed in Table 2.4-
4. Six large manufacturers, Honda, Polaris, Kawasaki, Yamaha, Suzuki, and Arctic Cat
accounted for approximately 98 percent of all U.S. ATV production in calendar year 2000.
Four of the six major ATV manufacturers, Honda, Kawasaki, Yamaha and Suzuki, are
primarily automobile and/or on-highway motorcycle manufacturers who also produce ATVs, off-
highway motorcycles, snowmobiles, personal water craft (PWC) and other non-highway
vehicles. Polaris and Arctic Cat manufacture snowmobile, in addition to producing ATVs.
Polaris also produces on-highway motorcycles and Arctic Cat produces PWC.
The 10 other manufacturers account for the remaining two percent of U.S. production in
2000. Only three of these are non-U.S.-owned. Of these remaining producers, five are classified
as large businesses, and five as small businesses. Bombardier is a large Canadian snowmobile
manufacturer that has recently entered the ATV market. Cannondale is a large American bicycle
manufacturer that has recently begun production of ATVs as well. Hyosung and Tai Ling are
large Far Eastern manufacturers, who also manufacture motorcycles and motor scooters (in the
case of Hyosung). Roadmaster/Flexible Flyer is primarily a large bicycle and toy manufacturer
but it also produces youth ATVs that are sold in large discount stores.
There are also some 17 firms that import ATVs. Thirteen of these are U.S.-owned. Dun and
Bradstreet data on the numbers of employees are available for four of these companies, and
indicate that these are small businesses according to the SBA definition. Since none of these had
more than 40 employees and two had less than 20 employees, it seems safe to assume that the
others are also small businesses according to the SBA definition. The 17 importers and 5 small
manufacturers either import completed ATVs or assemble them in this country from imported
parts.
2-45
-------
Table 2.4-4
ATV Manufacturers/Importers
Firm Name
ATK
COSMOPOLITAN MOTORS
D.R.R. INC.
E-TON DISTRIBUTION LP
HOFFMAN GROUP INC.
J & J SALES
JEHM POWERSPORTS
KASEA MOTORSPORTS
MANCO PRODUCTS
MOTORRAD OF NORTH AMERICA
PANDA MOTORSPORTS
POWERGROUP INTERNATIONAL ALPHASPORTS
REINMECH MOTOR COMPANY, LTD
TRANSNATIONAL OUTDOOR POWER LLC
TWS-USA, INC
ULTIMAX LCC
UNITED MOTORS OF AMERICA, INC
AMERICAN SUNDIRO
ARCTIC CAT, INC.
BOMBARDIER
CANNONDALE CORP - BEDFORD
HONDA AMERICAN MANUFACTURING
HYOSUNG MOTORS AND MACHINERY
INTERNATIONAL POWERCRAFT
KAWASAKI MOTORS CORPORATION
KEEN PERCEPTION INDUSTRIES
MOSS
PANDA MOTORSPORTS
POLARIS INDUSTRIES
ROADMASTER /FLEXIBLE FLYER
SUZUKI
LAI LING MOTOR COMPANY
YAMAHA MOTOR MANUFACTURING CORP.
Type
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
2-46
-------
Chapter 2: Industry Characterization
2.4.3.3 Engine Manufacturers
Four of the major ATV producers, Honda, Kawasaki, Yamaha and Suzuki, manufacture
both engine and equipment. In addition to producing engines for itself, Suzuki manufactures
engines for Arctic Cat, and in fact owns a significant amount of Arctic Cat common stock.
Hyosung Motors and Machinery and the Tai Ling Motor Company also use Suzuki engines in
their ATVs. Although Polaris produces some of its own engines, a substantial number are
supplied by Fuji Heavy Industries, primarily an auto and truck manufacturer, and its U.S.
subsidiary, Robin Industries. Polaris owns a substantial amount of Robin common stock.
Other engine manufacturers include Rotax, a subsidiary of Bombardier Inc., a large
Canadian company. Bombardier/Rotax also produces engines for a wide variety of other
applications, including snowmobiles, motorcycles, ATVs, personal water craft (PWC), utility
vehicles and aircraft. A few small ATV manufacturers use Briggs or Kohler utility engines, but
these are covered by EPA's Small Spark Ignition (SI) Engine regulations and are not included in
this analysis.
2.4.4 Markets
This section examines the historical market data for the ATV industry. Historical data on
the quantity of domestic shipments and price data of ATVs are provided. The quantity and
values of imports and exports for ATVs are presented as well.
2.4.4.1 Quantity and Price Data
Historical market data on the quantity of ATVs sold in the U.S. are provided in Table 2.4-5.
Data were obtained from the Motorcycle Industry Council (MIC). As this table shows, there has
been an overall increasing trend in the quantity of ATVs sold in the U.S. with an average annual
increase of 17 percent from 1990 to 2001. Sales of ATVs increased more than 600% between
the years 1990 and 2001. Retail dollars increased, on average, by 22 percent from 1990 to 2001.
This is due to the huge increase in production. Retail dollars per unit has also increased, showing
an annual average increase of 5 percent for the same period. There was a steady rise of the retail
dollars/unit over this time period.
2-47
-------
Draft Regulatory Support Document
Table 2.4-5
U.S. Units Sold, Retail Dollars and Retail Dollars Per Unit ATVs, 1990 - 2001
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Annual
Average
Unit Sales
134,619
125,056
144,332
162,307
189,328
277,787
317,876
359,397
429,414
545,932
648,645
880,000
383,154
%
Change
Unit
Sales
(7%)
15%
12%
17%
48%
14%
13%
19%
27%
19%
12%
17%
Retail Dollars
($103)
$393.20
$371.32
$449.42
$563.18
$770.52
$1,282.47
$1,530.97
$1,759.77
$2,155.02
$2,805.70
$3,343.15
$3,734.91
$1,596.64
%
Change
Retail
Dollars
(5%)
21%
25%
37%
66%
19%
15%
22%
30%
19%
12%
22%
Retail
Dollars/
Unit
$2,921
$2,969
$3,114
$3,470
$4,070
$4,617
$4,816
$4,896
$5,019
$5,139
$5,154
$5,123
$4,276
% Change
Retail
Dollars/Unit
2%
5%
11%
17%
13%
4%
2%
3%
2%
0.3%
-0.6%
5%
2.4.4.2 Foreign Trade
Export and import values and quantities for ATVs are presented in Table 2.4-6. This table
shows that the export values started out on in an increasing trend for the first three years. Then
in 1992, export value dropped by 64 percent and fluctuated between $73 million and $95 million,
with the exception of the year 1997. Import quantity decreased until 1992 then remained
between 34 thousand and 45 thousand through 2001. The import value decreased each year from
1989 to 1993, it dropped again in 1995 and maintained an increasing trend from 1996 to 2001.
The import quantity generally decreased from 1989 to 1993 and started a general rebounding
trend. Note that the data presented relates to SIC 3799 and includes ATVs, snowmobiles, golf
carts and other transportation equipment.
2-48
-------
Chapter 2: Industry Characterization
Table 2.4-6*
Import and Export Quantities and Values for ATVs, 1989 - 200190
Export Value Export Quantity (103) Import Value
Year
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Average
($103)
$169,881
$196,344
$209,003
$134,356
$75,876
$72,787
$85,976
$92,806
$136,357
$85,742
$91,335
$94,783
$89,381
$118,048
161
95
75
35
40
45
43
42
45
34
42
40
42
56
($103)
$223,425
$156,239
$50,877
$31,786
$9,907
$13,549
$7,351
$9,272
$13,478
$19,174
$32,755
$48,433
$89,786
$54,310
Import
Quantity (103)
2,548
2,486
2,838
1,854
8
11
17
19
41
37
113
178
156
793
*Values shown relate to SIC 3799, which includes ATVs, snowmobiles, golf carts, and other transportation products.
2.5 Off-Highway Motorcycles
Off-highway motorcycles, commonly referred to as "dirt bikes," are recreational vehicles
designed specifically for use on unpaved surfaces. As such, they all have certain characteristics
in common, such as a large amount of clearance between the fenders and the wheels, tires with
aggressive knobby tread designs, and a lack of some of the equipment typically found on
highway motorcycles (e.g., lights, horns, turn signals, and often mufflers). Thus they normally
can not be licensed for on-highway use. There are a limited number of motorcycles, known as
dual-purpose motorcycles, that can be used for both on- and off-highway purposes. These can be
licensed for highway use, and so fall under the current highway motorcycle regulations, assuming
that they are powered by engines of 50cc or larger displacement. Off-highway motorcycles are
used for recreational riding, but substantial numbers are also used for competition purposes.
Some in fact can be used for little else, e.g., machines that are designed for observed trials
2-49
-------
Draft Regulatory Support Document
competition, which have no seats in the conventional sense of the term, and engine
characteristics that are totally unlike those of most other motorcycles. Only a few thousand
observed trials competition bikes are produced each year. Vehicles designed solely for
competition are exempt from this rule. EPA's noise regulations also exempt any off-highway
motorcycle that is designed and marketed solely for use in closed-course competition.
2.5.1 The Supply Side
This section provides a description of off-road motorcycles and engines, the major inputs
used to manufacture this equipment, and the costs of production.
The motorcycle manufacturing process generally begins with the delivery of motorcycle
engines and transmissions, from engine plants to the motorcycle assembly plant. At the plant, the
engines and transmissions are matched to designated vehicles on the assembly line. Motorcycle
engines are produced with 1 to 8 cylinders, with various configurations. Multi-cylinder engines
are manufactured in three basic configurations: in-line, opposed, and V-type. Each of these refer
to the position of one bank of cylinders in relation to the other. Motorcycles engines can be air or
water cooled; 2-stroke or 4-stroke; carbureted or fuel-injected. Engines may be manufactured
with variances in other design characteristics, including the number and placement of
carburetors, cams, and valves.
2.5.1.1 Product Types and Populations
The number of off-highway motorcycles produced for sale in the U.S. averaged about
71,415 units between 1990 and 2001. As is the case with ATVs, off-highway motorcycle
production increased considerably in later years, to more than 195,000 units in 2001 according to
the Motorcycle Industry Council (MIC). Since many states do not require registration of off-
highway motorcycles, it is difficult to estimate a total population of these vehicles operational at
any given time.
As noted above, off-highway motorcycles can be used for recreational purposes or for
competition. EPA defines vehicles that are "used solely for competition" as those with features
(not easily removable from the vehicle) that would make the vehicle's use in other recreational
activities unsafe, impractical, or highly unlikely.
Certain types of off-highway motorcycles are designed and marketed for closed-course
competition. These are commonly known as "motocross bikes." Some 12-14 percent of off-
highway motorcycles produced from 1996 to 2000 were motocross bikes. Other sources have
estimated motocross bikes to be closer to 30 percent of off-highway sales.91 Other types of
competition motorcycles are the observed trials machines mentioned above, which emphasize
handling ability rather than speed, and the so-called "enduro bikes." Enduro bikes are designed
for cross-country type racing, rather than closed-course competition. As such, they require some
of the equipment normally found on non-racing machines, such as spark arresters (required by
U.S. Forest Service regulations) and at least minimal lighting packages.
2-50
-------
Chapter 2: Industry Characterization
Whether for competition or recreational use, off-highway motorcycles are operated under
transient conditions that include a wide variety of speeds and load factors.
2.5.1.2 Engine Design and Operation
The off-road segment of the motorcycle market is dominated by vehicles with relatively
small engines. Off-highway motorcycle engines have traditionally been about two-thirds smaller
and less powerful than those used in on-highway cycles. In 1990 and 1998, approximately 88
percent of the off-highway motorcycles in use had an engine displacement less than 350cc. See
Table 2.5-1.
Table 2.5-1
Quantities of Off-road Motorcycles By Engine Displacement
1990 and 1998 92
Engine
Displacement
Under 125cc
125-349cc
350-499cc
450-749cc
Total
1990 Number of
Motorcycles
306,000
346,500
30,000
67,500
750,000
1990 %
of Total
40.8
46.2
4.0
9.0
100
1998 Number of
Motorcycles
367,200
680,500
34,700
113,600
1,196,000
1998 %
of
Total
30.7
56.9
2.9
9.5
100
In the year 2000, about 68 percent of the models produced were less than 300cc
displacement, and half of these were lOOcc or less. Percentages by engine displacement for the
top five producers are approximately the same as for the industry as a whole. The distribution of
engine sizes for these producers tends to be somewhat skewed, with a larger fraction of off-
highway motorcycles falling into the lower displacement ranges (see Figure 2.5-1). Unlike on-
highway motorcycles, our contractor found no off-highway engines larger than 700cc are
currently produced.
2-51
-------
S
Dfi
Q
Q
O
tt
Q.
UJ
O
o: Ł
o
>2
(D
O
O
I
o
co o
o
(O
I
o
o S O
O r \
m ^
i i
o
o
0
O
LJJ
O LJJ
CO
o
o
CM
i
O
o
o
o
o
CO
m o
CM CM
in
spuesnoi|j_
NOLLOnaOJdd
-------
Chapter 2: Industry Characterization
2.5.1.3 Two-Stroke vs Four-Stroke Cycle Engine Usage
Based on the PSR database, slightly more than half of the off-highway motorcycles
produced for sale in the United States are powered by four-stroke cycle engines. However,
estimates from the Motorcycle Industry Council (MIC) place the percentage of two-stroke sales
at more than 60 percent. The percentage of two-strokes varies considerably by manufacturer.
Honda, which accounts for more than 45 percent of this production, is predominantly a four-
stroke manufacturer. Four-strokes comprise about two-thirds of its production. For Yamaha, the
percentage is about 57 percent. The remainder of the foreign and domestic producers
manufacture more two-stroke engines than four-strokes. For the other top-five producers, KTM,
Kawasaki, and Suzuki, the percentage of two-stroke engines varies from 58 to 72 percent, and
can be even higher (up to 100 percent) for some of the remaining manufacturers.
Two-stroke engines are normally used in two primary applications: (1) racing machines,
because they tend to have a higher power-to-weight ratio than four-stroke engines (this is
important for competition, especially in the smaller displacement classes), and (2) youth model
or entry-level motorcycles, because two-strokes are cheaper to produce than four-strokes. Since
youth or entry-level motorcycles also tend to have smaller displacement engines, the higher
power-to-weight ratio of the two-stroke tends to provide slightly better performance. However,
there has been a growing tendency in recent years for manufacturers to bring out more new four-
stroke engines, particularly in the higher displacement ranges. This is also true in their
competition lines.
2.5.1.4 Use of Engines in Other Applications
Only a few engine lines, primarily among the top five producers, are used in both off-
highway and on-highway motorcycles. Part of the reason for this is because over half of the off-
highway bikes use two-stroke engines, whereas almost no two-stroke engines are found in on-
highway motorcycles. Also, as noted above, off-highway motorcycles generally have much
smaller displacement engines than their on-highway counterparts. Off-highway motorcycle
engines are closer in terms of engine size to ATV engines. However, ATVs also use
predominantly four-stroke engines and these are not as likely to be highly-tuned for performance
as are many off-highway motorcycle engines.
2.5.1.5 Off-Road Motorcycle Manufacturers
Motorcycle manufacturers are classified under the Standard Industrial Classification (SIC)
code 3751 and under the North American Industry Classification System (NAICS) code 336991,
Motorcycle, Bicycle and Parts Manufacturers. Motorcycle manufacturers have the subcode
3369913, which includes manufacturers of scooters, mopeds, and sidecars. To be classified as a
small business by the Small Business Administration (SBA) size standards, the manufacturer
must have fewer than 500 employees. Motorcycle importers are classified by subcode 4211101,
which also includes automobile importers, and has an SBA size cutoff of 100 employees to be
considered a small business.
2-53
-------
Draft Regulatory Support Document
Twenty five companies manufacture off-highway motorcycles. The five largest
manufacturers, Honda, Kawasaki, Yamaha, Suzuki, and KTM. accounted for approximately 85
percent of all production sold in the U.S. in calendar year 2000. These companies manufacture
automobiles and/or on-highway motorcycles, motorscooters, ATVs, Personal Water Crafts
(PWC), as well as off-highway motorcycles. Honda is by far the largest producer of off-highway
motorcycles, with over 45 percent of the total production for sale in the U.S. Figure 2.5-2 shows
the market shares for the top five and the other producers, and Table 2.5-2 presents a list of the
manufacturers of off-highway motorcycles.94
Figure 2.5-2
OFF-HIGHWAY MOTORCYCLES
PRODUCTION BY MFR
PERCENT OF TOTAL
15.2%
10.8%
9.8%
45.1%
HONDA
KAWASAKI
KTM
SUZUKI
YAMAHA
OTHERS
9.6%
9.5%
Source: ICF Consulting, Docket A-2000-01, Document II-A-84.
Of the 25 firms that manufacture off-highway motorcycles for the U.S. market, six are U.S.
manufacturers. With the exception of Cannondale, which is primarily a bicycle manufacturer, all
of these companies produce only motorcycles. Italy has five manufacturers. One of these,
Cagiva, is mainly a producer of on-highway motorcycles. Piaggio is primarily a motorscooter
manufacturer; Betamotor makes motorscooters and trials bikes. Lem and Polini manufacture
youth motorcycles. Spanish manufacturers of off-highway motorcycles that are imported to the
U.S. include Gas Gas Motos, primarily an observed trials bike manufacturer, and Montesa, which
is owned by Honda. Other manufacturing companies whose products are imported into the U.S.
market are also found in Austria, Belarus, Ireland, Korea, Sweden, Taiwan, and the United
Kingdom. KTM, an Austrian company with a U.S. branch, is one of the five major producers for
the U.S. market.
The 20 other manufacturers accounted for the remaining 15 percent of production for sale in
the U.S. Six of these firms, accounting for approximately 3 percent of total production for the
U.S. market, are located in this country. Dun and Bradstreet employee data are available for four
2-54
-------
Chapter 2: Industry Characterization
of the six U.S. manufacturers, indicating that these are small businesses according to the SB A
definition.
Our contractor has also identified 16 off-highway motorcycle importers. Eight of these are
U.S.-owned. Dun and Bradstreet data are available for five of the eight U.S. importers, indicating
that they are small businesses though it seems likely that all eight are small businesses.
Table 2.5-2
U.S. Off-Highway Motorcycle Manufacturers/Importers9S
Firm Name
Type
ACTION POLINI
BETA USA
ODY RACING PRODUCTS
OSMOPOLITAN MOTORS INC.
RE IMPORTS/E-LINE ACCESSORIES
GAS GAS NORTH AMERICA
HUSQVARNA USA
KASEA MOTORSPORTS
KTM SPORTMOTORCYCLE USA, INC.
MIDWEST MOTOR VEHICLES, INC.
TRANSNATIONAL OUTDOOR POWER, LLC
TRYALS SHOP
TWS-USA INC.
U.S. MONTESA
UNITED MOTORS OF AMERICA
VOR MOTORCYCLES USA
AMERICAN DIRT BIKE INC. (U.S.)
ATK MOTORCYCLES (U.S.)
BETAMOTOR SPA (ITALY)
:AGIVA MOTORCYCLE SPA (ITALY)
ANNONDALE CORP - BEDFORD (U.S.)
:CM MOTORCYCLES LTD (U.K.)
OBRA MOTORCYCLE MFG. (U.S.)
GAS GAS MOTOS SPA (SPAIN)
HM MOTORCYCLES (U.S.)
HONDA MOTORCYCLES (JAPAN)
HUSABERG MOTOR AB (SWEDEN)
HYOSUNG MOTORS AND MACHINERY (KOREA)
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
IMPORTER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
MANUFACTURER
2-55
-------
Draft Regulatory Support Document
KAWASAKI HEAVY INDUSTRIES (JAPAN) MANUFACTURER
KTM SPORT MOTORCYCLE AG (AUSTRIA) MANUFACTURER
LEM MOTOR SAS (ITALY) MANUFACTURER
MADFAST MOTORCYCLES (IRELAND) MANUFACTURER
MINSK MOTOVELOZAVOD (BELARUS) MANUFACTURER
MONTESA-HONDA ESPANA, SA (SPAIN) MANUFACTURER
PIAGGIO GROUP (ITALY) MANUFACTURER
POLINI (ITALY) MANUFACTURER
REV! MOTORCYCLES (U.S.) MANUFACTURER
SUZUKI (JAPAN) MANUFACTURER
TAI LING MOTOR COMPANY LTD. (TAIWAN) MANUFACTURER
VOR MOTORI (ITALY) MANUFACTURER
YAMAHA MOTOR COMPANY LTD. (JAPAN) MANUFACTURER
2.5.1.6 Engine Manufacturers
For the majority of off-highway motorcycles, the vehicle manufacturer is also the engine
manufacturer. However, a few motorcycle manufacturers use engines produced by other firms.
ATK Motorcycles and CCM Motorcycles Ltd. use Bombardier/Rotax engines, while the Tai Ling
Motor Company uses Suzuki engines. The Spanish manufacturer, Gas Gas Motos, noted
primarily for its observed trials machines, produces some of its own engines and buys others
from Cagiva, a large Italian manufacturer. One U.S. manufacturer, Rokon, markets a low-
production trail motorcycle resembling a large motorscooter. This vehicle type is intended for
hunters and fishermen. Rokon uses industrial-type engines made by Honda and other
manufacturers which are regulated under the EPA Small SI regulations. Therefore, Rokon is not
included here.
As Table 2.5-3 shows the average value of shipments (VOS) for this industry over the 1992
to 1999 time period is equal to approximately 2.8 billion dollars, with the highest value of
shipments occurring in 1998. The cost of materials for this industry is equal to an average of
almost 1.6 billion dollars (57 percent of VOS). The average cost of labor is approximately 347
million (19 percent of VOS), while capital expenditures are equal to an average value of 26.7
million (1 percent of VOS). Examination of this data clearly shows that capital expenditures
represent the smallest share of the value of shipments while the cost of materials represents the
largest share.
2-56
-------
Chapter 2: Industry Characterization
Table 2.5-3
Value of Shipments (VOS) and Production Costs for
the SIC and NAICS Codes that Include
Off-Highway Motorcycle Manufacturers, 1992 - 1999 969798"
Year
1992
1993
1994
1995
1996
1997
1998
1999
Industry
Code
SIC 3751
SIC 3751
SIC 3537
SIC 3537
SIC 3537
NAICS 336991
NAICS 336991
NAICS 336991
Average
VOS
($106)
1,878.9
1,878.3
2,632.1
2,832.9
3,094.0
3,382.6
3,343.8
3,066.1
2,776.8
Payroll
($106)
301.7
409.3
482.6
502.6
565.1
662.3
620.3
576.1
347.1
%of
VOS
16%
22%
18%
18%
18%
20%
19%
19%
19%
Cost of Materials
($106)
1,146.2
1,362.0
1,488.6
1,541.6
1,673.9
1,802.3
1,740.7
1,611.3
1,559.1
%of
VOS
61%
73%
57%
54%
54%
53%
52%
53%
57%
Total Capital
Expenditures
($106)
10.6
13.0
14.2
15.4
17.9
19.5
9.6
7.2
26.7
%of
VOS
1%
1%
1%
1%
1%
1%
0
0
1%
* Value of Shipments, Payroll, Cost of Materials, and Total Capital Expenditures are in nominal U.S. dollars
2.5.2 The Demand Side
This section provides information on the uses of off-highway motorcycles, the various
substitute products on the market, and the consumers who purchase off-highway motorcycles.
2.5.2.1 Uses of Off-Highway Motorcycles
Motorcycles are used for a for a variety of purposes, including recreation, touring,
commuting, and on- and off-road racing. There are generally three motorcycle model types, on-
highway, dual(both on highway and off-highway), and off-highway. On-highway motorcycles
are certified by the manufacture as being in compliance with the Federal Motor Vehicle Safety
Standards (FMVSS), and are designed for use on public roads. On-highway motorcycles include
scooters, but excludes mopeds (limited speed motor-driven cycles under 50cc, with or without
fully operative pedals). Dual motorcycles are certified by the manufacturer as being in
compliance with FMVSS, and are designed with the capability for use on public roads, as well as
off-highway recreational use. Off-highway motorcycles are not certified by the manufacturer to
be in compliance with FMVSS for on-highway use. This category includes competition
motorcycles. Table 2.5-4 show that off-highway motorcycles represents nearly 15% of the total
2-57
-------
Draft Regulatory Support Document
population in 1998 and nearly 18% in 1998.
Table 2.5-4
Estimated Population By Model Type
1990 and 1998 10°
MODEL TYPE
On-Highway
Dual
Off-Highway
Total
1990 NUMBER OF
MOTORCYCLES
3,650,000
(72.3%)
660,000
(13%)
750,000
(14.8%)
5,060,000
(100%)
1998 NUMBER OF
MOTORCYCLES
4,809,000
(73%)
565,000
(8.6%)
1,196,000
(18.2%)
6,570,000
(100%)
2.5.2.2 Substitution Possibilities
Consumers can substitute across off-road recreational vehicles. As a substitute, an ATV
would allow the consumer to enjoy the same off-road recreation that they would receive with an
off-highway motorcycle. Consumers may be interested in engaging in outdoor activities, but may
instead consider doing a non-motorized activity. For example, consumers who are interested in
being outside may engaging in hiking, running, or riding a bicycle. These non-motorized options
will also allow the consumer to participate in outdoor activity, but they may be considered
substitutes for less intensive off-highway past times. Indeed, any type of recreational activity
may be viewed as a substitute for off-highway motorcycle usage.
2.5.3 Industry Organization
Because there are costs associated with the emission control program, it is important to
determine how the off-highway motorcycle industry may be affected. Industry organization is an
important factor which affects how an industry may react to regulatory costs. This section
provides a description of the organization of the motorcycle industry.
2.5.3.1 Market Structure
Market structure is of interest because it determines the behavior of producers and
consumers in the industry. In perfectly competitive industries, no producer or consumer is able
to influence the price of the product sold. In addition, producers are unable to affect the price of
2-58
-------
Chapter 2: Industry Characterization
inputs purchased for use in production. This condition is most likely to hold if the industry has a
large number of buyers and sellers, the products sold and inputs used are homogeneous, and entry
and exit of firms is unrestricted. Entry and exit of firms are unrestricted for most industries,
except in cases where the government regulates who is able to produce output, where one firm
holds a patent on a product, where one firm owns the entire stock of a critical input, or where a
single firm is able to supply the entire market. In industries that are not perfectly competitive,
producer and/or consumer behavior can have an effect on price.
Concentration ratios (CRs) and the Herfindahl-Hirschman index (HHI) can provide some
insight into the competitiveness of an industry. The U.S. Department of Commerce reports these
ratios and indices for the six digit NAICS code level for the year 1997, the most recent year
available. Table 2.5-5 provides the four- and eight-firm concentration ratios (CR4 and CRS,
respectively), and the Herfindahl-Hirschman index for the Motorcycle, Bicycle, and Parts
Manufacturing industry, the industry that includes producers of off-highway motorcycles. This
industry is represented by NAICS code 336991. For this industry the CR4 was 67.5 percent and
the CRS was 76.7 percent.
The criteria for evaluating the HHI are based on the 1992 Department of Justice Horizontal
Merger Guidelines. According to these criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are
considered moderately concentrated (i.e., moderately competitive), and those with HHIs above
1,800 are considered highly concentrated (i.e., less competitive). In general, firms in less
concentrated industries have more ability to influence market prices. Though the HHI measure
for this industry is high, we have chosen to model is as a perfectly competitive market. We have
made this choice based on the number of recreational substitute available for off-highway
motorcycles.
Table 2.5-5
Measures of Market Concentration for the
NAICS Code that Includes Off-Highway Motorcycle Manufacturers, 1997 1Q1
Description
NAICS 336991
CR4
67.5
CRS
76.7
HHI
2,036.5
vos
($106)
$3,382,689
Number of
Companies
373
2.5.3.2 Motorcycle Manufacturers
As mentioned above, motorcycles are included under Standard Industrial Classification
(SIC) 3751. The U.S. motorcycle industry is relatively small compared to other industries such
as the automobile industry. There are over 40 U.S. firms (Table 2.5-2) engaged in the
manufacture and/or distribution of off-highway motorcycles. Six of these firms accounted for 90
percent of the new motorcycle units produced in the United States in 2000. Table 2.5-6 shows
the ranking and market share for the major producers in the industry for 1999 and 2000.
2-59
-------
Draft Regulatory Support Document
Table 2.5-6
Motorcycle Manufacturers by Market Share 1999-2000 102
BRAND
Honda
Harley-
Davidson
Yamaha
Suzuki
Kawasaki
BMW
All Others
1999 RANK
2
1
3
5
4
6
—
1999 MARKET
SHARE
24.1%
25.5%
17.8%
10.8%
11.8%
1.9%
8.1%
2000 RANK
1
2
3
4
5
6
—
2000 MARKET
SHARE
25.0%
23.0%
19.3%
11.2%
10.2%
1.7%
9.6%
In the off-highway segment, the top five manufacturers were Honda, Kawasaki, KTM,
Suzuki, and Yamaha. Table 2.5-7 shows the market share among the major producers. U.S. off-
highway motorcycle production by the top five firms steadily rose over the 1996 to 2000 time
period, with a slight dip in 1999.
Table 2.5-7
Off-Highway Motorcycle Units Manufactured by the Top Five Firms 1996-1999 103
Company
Honda
Suzuki
Yamaha
Kawasaki
KTM
Total
1996
45,694
17,022
23,862
12,687
2,778
102,043
1997
51,281
19,200
29,231
12,147
3,146
116,005
1998
56,678
18,694
25,230
13,249
3,783
117,634
1999
53,706
10,617
26,079
12,885
7,236
110,523
2000
68,924
11,039
20,406
14,560
14,747
129, 676
1996-
2000
TOTAL
276,283
76,572
124,808
66,528
31,690
575,881
1996-2000
MARKET
SHARE
48.0%
13.3%
21.7%
11.5%
5.5%
100%
2-60
-------
Chapter 2: Industry Characterization
2.5.3.3 Small Businesses
The motorcycle companies listed in Table 2.5-2 can be grouped into small and large
business categories using the Small Business Administration (SBA) general size standard
definitions for NAICS codes. The SBA defines a small business in terms of the employment or
annual sales of the owning entity and these thresholds vary by industry. Based on the size
standard for NAICS 336991, several of the motorcycle producers are considered small
businesses.
2.5.4 Markets
This section examines the historical market statistics for the off-highway motorcycle
manufacturing industry. Historical data on the quantity of domestic shipments and price data of
off-highway motorcycles are provided. The quantity and values of imports and exports for
motorcycles are presented as well.
2.5.4.1 Quantity and Price Data
Historical market data on the quantity of U.S. unit sales of off-highway motorcycles are
provided in Table 2.5-8. Data were obtained from the Motorcycle Industry Council (MIC). As
this table shows, there has been an overall increasing trend in the quantity of off-highway
motorcycles sold in the U.S. with an overall net increase of 290 percent and the retail value of
off-highway motorcycle increased by nearly 40 percent from 1990 to 2000.
2-61
-------
Draft Regulatory Support Document
Table 2.5-8
U.S. Units Sold, Retail Dollars and
Retail Dollars Per Unit Off-Highway Motorcycles, 1990 - 2001 104
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Unit Sales
39,221
37,363
39,345
39,863
40,991
40,791
45,266
49,168
59,930
77,875
120,501
195,250
Retail Dollars
$63,745,225
$63,670,177
$68,038,926
$75,033,960
$84,844,505
$94,125,405
$111,001,200
$119,041,853
$133,062,004
$170,303,959
$279,984,888
$334,983,201
Retail Dollars/Unit
$1,625
$1,704
$1,729
$1,882
$2,070
$2,308
$2,452
$2,421
$2,220
$2,187
$2,324
$2,253
* Values are in nominal dollars.
2.5.4.2 Foreign Trade
Export and import values and quantities for off -highway motorcycle are presented in Table
2.2-9. These data show increasing trends for export and import values since 1989. Note these
data reflect imports and exports for SIC 3751, motorcycles, bicycles, and parts.
2-62
-------
Chapter 2: Industry Characterization
Table 2.5-9
Import and Export Quantities and Values for Off-Highway Motorcycles, 1989 - 2001105
Year
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Average
Export Value
(1,000 Dollars)
$244,722
$419,911
$615,439
$671,331
$702,831
$711,053
$850,229
$906,040
$976,494
$918,277
$738,152
$798,357
$967,947
$732,368
Export
Quantity
( 1,000 Dollars)
$319
$480
$796
$846
$1,053
$739
$721
$626
$692
$662
$823
$673
$480
$685
Import Value
(1,000 Dollars)
$1,325,309
$1,216,239
$1,370,364
$1,574,380
$1,758,664
$1,800,564
$2,178,559
$2,046,358
$2,117,154
$2,445,434
$2,993,162
$3,898,859
$3,895,486
$2,201,579
Import
Quantity
(1,000 Units)
32,829
37,164
40,850
37,823
42,767
40,322
43,937
41,868
48,622
45,565
43,008
37,846
26,592
39,938
* Values are in nominal dollars and reflect values for SIC 3751 Motorcycles, Bicycles, and Parts.
2-63
-------
Draft Regulatory Support Document
Chapter 2 References
1. "Recreational Boating Industry Characterization," ICF Consulting, Contract No. 68-C-98-170,
WAN 0-5, August, 1999, Docket A-2000-01, Document U-A-9.
2. Boating Industry Magazine, "1997 Annual Industry Reviews: The Boating Business", January
1998. Docket A-2000-01, Document IV-A-116.
3. Dun & Bradstreet. 2001.
4. National Marine Manufacturers Association - http://www.nmma.org. Docket A-2000-01,
Document IV-A-51.
5. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the Diesel
Engine Rule on the U.S. Boat Market. Docket A-2000-01, Document IV-A-53.
6. U.S. EPA. 1999. Economic Impact Analysis of the Proposed Boat Manufacturing NESHAP.
Air Quality Strategies and Standards Division, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. EPA-452/R-99-005. November.
7. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-58.
8. U.S. Department of Commerce. 2000. 1998 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-59.
9. U.S. Department of Commerce. 1999. 1997 Economic Census-Boat Building Washington,
DC: Government Printing Office. Docket A-2000-01, Document IV-A-64.
10. U.S. Department of Commerce. 1998. 1996 Annual Survey of'Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-65.
11. U.S. Department of Commerce. 1997. 7995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-66.
12. U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-67.
13. U.S. Department of Commerce. 1995. 7993 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2-64
-------
Chapter 2: Industry Characterization
2000-01, Document IV-A-68.
14. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-58.
15. U.S. Department of Commerce. 2000. 1998 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-59.
16. U.S. Department of Commerce. 1999. 1997 Economic Census - Other Engine Equipment
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-62.
17. U.S. Department of Commerce. 1998. 1996 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-65.
18. U.S. Department of Commerce. 1997. 1995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-66.
19. U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-67.
20. U.S. Department of Commerce. 1995. 1993 Annual Survey oj"Manufactures -Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-68.
21. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/. Docket A-2000-01, Document IV-A-50.
22. National Marine Manufacturing Association. Boating 1998 - Facts and Figures at a Glance.
http://63.236.237.146/facts/boatingstats/boating98.html.. Docket A-2000-01, Document IV-A-
52.
23. U.S. Department of Commerce. 2001. 1997 Economic Census - Concentration Ratios in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
24. U.S. Department of Commerce. 2001. 1997 Economic Census - Concentration Ratios in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
2-65
-------
Draft Regulatory Support Document
25. Wards Business Directory. 2001.
26. Dun & Bradstreet. 2001.
27. American Business Directory. 2000.
28. Wards Business Directory. 2001.
29. Dun & Bradstreet. 2001.
30. American Business Directory. 2000.
31. National Marine Manufacturing Association. U.S. Recreational Boating Domestic Shipment
Statistics, 1970 - 1998. http://63.236.237.146/facts/historv.pdf.. Docket A-2000-01, Document
IV-A-49.
32. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/. Docket A-2000-01, Document IV-A-50.
33. National Marine Manufacturing Association. U.S. Recreational Boating Domestic Shipment
Statistics, 1970-1998. http://63.236.237.146/facts/historv.pdf Docket A-2000-01, Document
IV-A-49.
34. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/. Docket A-2000-01, Document IV-A-50.
35. National Marine Manufacturing Association. U.S. Recreational Boating Domestic Shipment
Statistics, 1970-1998. http://63.236.237.146/facts/historv.pdf Docket A-2000-01, Document
IV-A-49.
36. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/. Docket A-2000-01, Document IV-A-50.
37. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/. Docket A-2000-01, Document IV-A-50.
38. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Table 4 - Product Analysis: Inboard Boats,
Imports. Docket A-2000-01, Document IV-A-53.
39. National Marine Manufacturing Association. 2002. Supplemental Tables for Economic
Impact Analysis of the Diesel Engine Rule on the U.S. Boat Market. Sterndrive Boats, Imports.
Docket A-2000-01, Document IV-A-54.
40. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Table 4 - Product Analysis: Inboard Boats,
2-66
-------
Chapter 2: Industry Characterization
Exports. Docket A-2000-01, Document IV-A-53.
41. National Marine Manufacturing Association. 2002. Supplemental Tables for Economic
Impact Analysis of the Diesel Engine Rule on the U.S. Boat Market. Sterndrive Boats, Exports.
Docket A-2000-01, Document IV-A-54.
42. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Table 3 - Inboard Engines - Imports. Docket A-
2000-01, Document IV-A-53.
43. National Marine Manufacturing Association. 2002. Supplemental Tables for Economic
Impact Analysis of the Diesel Engine Rule on the U.S. Boat Market. Sterndrive Engines,
Imports. Docket A-2000-01, Document IV-A-54.
44. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Table 3 - Inboard Engines - Imports. Docket A-
2000-01, Document IV-A-53.
45. National Marine Manufacturing Association. 2002. Supplemental Tables for Economic
Impact Analysis of the Diesel Engine Rule on the U.S. Boat Market. Sterndrive Engines,
Exports. Docket A-2000-01, Document IV-A-54.
46."The Role of Propane in the Fork Lift/Industrial Truck Market: A Study of its Status, Threats,
and Opportunities," Robert E. Myers for the National Propane Gas Association, December 1996,
Docket A-98-01, Document II-D-2. Docket A-2000-01, Document H-A-86.
47. See footnote 46.
48. Industrial Truck Association. 2002. Membership Handbook. A copy of the relevant pages
can be found in Docket A-2000-01, Document No. IV-A-188.
49. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-58.
50. U.S. Department of Commerce. 2000. 1998 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-59.
51. U.S. Department of Commerce. 1999. 1997 Economic Census - Industrial Truck, Tractor,
Trailer, and Stacker Machinery Manufacturing. Washington, DC: Government Printing Office.
Docket A-2000-01, Document IV-A-63.
52. U.S. Department of Commerce. 1998. 1996 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
2-67
-------
Draft Regulatory Support Document
01, Document IV-A-65.
53. U.S. Department of Commerce. 1997. 1995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-66.
54. U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-67.
55. U.S. Department of Commerce. 1995. 1993 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-68.
56. See footnote 46.
57. See footnote 48.
58. U.S. International Trade Commission. 1996. Industry and Trade Summary: Forklift Trucks
and Related Vehicles. Washington, DC. Docket A-2000-01, Document IV-A-189
59. U.S Department of Commerce. 2001. 1997 Economic Census - Concentration Ratios in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
60. See footnote 48.
61. American Business Information. 2002. American Business Directory Electronic Database.
62. Dun & Bradstreet. 2002. Dun and Bradstreet Market Identifiers Electronic Database.
63. Ward's Business Directory, 53rd edition.
64. See footnote 48.
65. See footnote 13.
66. International Trade Commission. 2002. ITCDataWeb. Docket A-2000-01, Document IV-
A-187.
67. International Snowmobile Association . www.snowmobile.org. June 2002. Docket A-2000-
01, Document IV-A-128 and IV-A-136.
68. Production data were taken from OELINK Database owned by Power Systems Research.
69. Production data were taken from OELINK Database owned by Power Systems Research.
2-68
-------
Chapter 2: Industry Characterization
70. U.S. Department of Commerce. 2002. Statistics for Industry Groups and Industries: 2000.
Annual Survey of Manufacturers Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-117.
71. U.S. Department of Commerce. 1998. 1996 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-65.
72. U.S. Department of Commerce. 1997. 1995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-66.
73. U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-67.
74. U.S. Department of Commerce. 1995. 1993 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-68.
75. Memo to Docket A-98-01 from Line Wehrly, regarding Emission Modeling for Recreational
Vehicles. Docket A-2000-01, Document II-B-19.
76. International Snowmobile Manufacturers Association, www.snowmobile.org. June. Docket
A-2000-01, Documents IV-A-112 and Docket A-2000-01, Document IV-A-136.
77. US Department of Commerce. 1997 Economic Census - Concentration Ratio in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
78. USEPA. June 1998. "Regulatory Impact Analysis of Cement Kiln Dust Rulemaking." Final
report prepared by Research Triangle Institute for Office of Air Quality Planning and Standards,
RTF, North Carolina. See also USEPA. July 1996. "Economic Analysis of Air Pollution
Regulations: Portland Cement." Final report prepared by Research Triangle Institute for Office
of Air Quality Planning and Standards, RTF, North Carolina.
79. International Snowmobile Manufacturers Association, www.snowmobile.org. June 2002.
Docket A-2000-01, Document IV-A-128 and IV-A-136.
80. International Snowmobile Manufacturers Association, www.snowmobile.org. June 2002.
Docket A-2000-01, Document IV-A-128 and IV-A-136.
81. International Snowmobile Manufacturers Association, www.snowmobile.org. June 2002.
Docket A-2000-01, Document IV-A-128 and IV-A-136.
2-69
-------
Draft Regulatory Support Document
82. ICF Consulting. 2001. Industry Characterization of Non-Road Recreational Vehicles, Draft
Final Report. Prepared for U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, September 13, 2001. Docket A-2000-01, Document II-A-84.
83. U.S. Department of Commerce. 2002. Statistics for Industry Groups and Industries: 2000.
Annual Survey of Manufacturers Washington, DC: Government Printing Office. Docket A-
2000-01, Document IV-A-117.
84. U.S. Department of Commerce. 1998. 1996 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-65.
85. U.S. Department of Commerce. 1997. 1995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-66.
86. U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-67.
87. U.S. Department of Commerce. 1995. 1993 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-68.
88. US Department of Commerce. 1997 Economic Census - Concentration Ratio in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
89. Motorcycle Industry Council, Inc. 2002. Docket A-2000-01, Document No. IV-A-138.
90. International Trade Commission. 2002. ITC Data Web. Docket A-2000-01, Document IV-
A-193.
91. U.S. Environmental Protection Agency, Office of Transportation and Air Quality, 2001.
Regulatory Support Document: Control of Emissions from Unregulated Nonroad Engines. Ann
Arbor, MI. EPA 420-D-01-004. Docket A-2000-01, Document m-B-02 . This document is
available on our website, www.epa.gov/otaq/regs/nonroad/proposal
92. ICF Consulting. 2001. Industry Characterization of Non-Road Recreational Vehicles, Draft
Final Report. Prepared for U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, September 13, 2001. Docket A-2000-01, Document II-A-84.
93. ICF Consulting. 2001. Industry Characterization of Non-Road Recreational Vehicles, Draft
Final Report. Prepared for U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, September 13, 2001. Docket A-2000-01, Document II-A-84.
2-70
-------
Chapter 2: Industry Characterization
94. ICF Consulting. 2001. Industry Characterization of Non-Road Recreational Vehicles, Draft
Final Report. Prepared for U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, September 13, 2001. Docket A-2000-01, Document II-A-84.
95. ICF Consulting. 2001. Industry Characterization of Non-Road Recreational Vehicles, Draft
Final Report. Prepared for U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, September 13, 2001. Docket A-2000-01, Document II-A-84.
96.U.S. Department of Commerce. 1998. 1996 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-65.
97.U.S. Department of Commerce. 1997. 1995 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-66.
98.U.S. Department of Commerce. 1996. 1994 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-67.
99.U.S. Department of Commerce. 1995. 1993 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-68.
100. See footnote 90.
101.US Department of Commerce. 1997 Economic Census - Concentration Ratio in
Manufacturing. Washington, DC: Government Printing Office. Docket A-2000-01, Document
IV-A-57.
102. See footnote 90.
103.See footnote 90.
104. Motorcycle Industry Council, Inc. 2002. Docket A-2000-01, Document No. IV-A-138.
105.International Trade Commission. 2002. ITC Data Web. Docket A-2000-01, Document IV-
A-193.
2-71
-------
Draft Regulatory Support Document
2-72
-------
Chapter 3: Technology
Chapter 3: Technology
This chapter describes the current state of spark-ignition technology for engines, evaporative
emission technology, and compression-ignition technology for marine engines, as well as the
emission control technologies expected to be available for manufacturers. Chapter 4 presents the
technical analysis of the feasibility of the standards.
3.1 Introduction to 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 used other fuels. SI engines include engines fueled with liquefied
petroleum gas (LPG) and compressed natural gas (CNG).
3.1.1 Four-Stroke Engines
Four-stroke engines are used in many different applications. Virtually all automobiles and
many trucks are powered by four-stroke SI engines. Four-stroke engines are also very common
in motorcycles, all-terrain vehicles (ATVs), boats, airplanes, and numerous nonroad applications
such as lawn mowers, lawn and garden tractors, and generators, to name just a few.
A "four-stroke" engine gets it's 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, power,
and exhaust. Two of the strokes are downward (intake & power) and two of the strokes are
upward (compression & exhaust). Valves in the combustion chamber open and close to route
gases into and out of the combustion chamber or create compression.
The first step of the cycle is for an intake valve in the combustion chamber 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 creates a vacuum or suction in the cylinder, which draws air and fuel past the open
intake valve into the combustion chamber.
The intake valve then closes and the momentum of the crankshaft causes the piston to move
back up the cylinder from BDC to TDC, compressing the air and fuel mixture. This is the
"compression" stroke. As the piston nears TDC, at the very end of the compression stroke, the
air and fuel mixture is ignited by a spark from a spark plug and begins to burn. As the air and
3-1
-------
Draft Regulatory Support Document
fuel mixture burns, increasing temperature and pressure cause the piston to move back down the
cylinder, transmitting power to the crankshaft. This is referred to as the "power" stroke. The last
stroke in the four-stroke cycle is the "exhaust" stroke. At the bottom of the power stroke, an
exhaust valve opens in the combustion chamber and as the piston moves back up the cylinder,
the burnt gases are pushed out through the exhaust valve to the exhaust manifold, and the cycle is
complete.
3.1.2 Two-Stroke Engines
Two-stroke SI engines are widely used in nonroad applications, especially for recreational
vehicles, such as snowmobiles, off-highway motorcycles and ATVs. The basic operating
principle of the charge scavenged two-stroke engine (traditional two-stroke) is well understood;
in two-strokes the engine performs the operations of intake, compression, expansion and exhaust,
which the four-stroke engine requires four strokes to accomplish. Two-stroke engines have
several advantages over traditional four-stroke engines for use in recreational vehicles: high
power-to-weight ratios; simplicity; ease of starting; and lower manufacturing costs. However,
they also have much higher emission rates.
Another difference between two- and four-stroke engines is how the engines are lubricated.
Four-stroke engines use the crankcase as a sump for lubricating oil. Oil is distributed throughout
the engine by a pump through a series of small channels. Because the crankcase in a two-stroke
engine serves as the pump for the scavenging process, it is not possible to use it as an oil sump as
is the case for four-stroke engines. Otherwise, gasoline would mix with the oil and dilute it.
Instead, lubrication for two-stroke engines is provided by mixing specially-formulated two-stroke
oil with the incoming charge of air and fuel mixture. The oil is either mixed with the gasoline in
the fuel tank, or metered into the gasoline as it is consumed, using a small metering pump. As
the gasoline/oil mixture passes through the carburetor, it is atomized into fine droplets and mixed
with air. The gasoline quickly vaporizes, while the less volatile oil forms a fine mist of fine
droplets. Some of these droplets contact the crankshaft, piston pin, and cylinder walls, providing
lubrication. Most of the oil droplets, however, pass out of the crankcase and into the cylinder
with the rest of the incoming charge.
In a two-stroke engine, combustion occurs in every revolution of the crankshaft. Two-stroke
engines eliminate the intake and exhaust strokes, leaving only compression and power strokes.
This is due to the fact that two-stroke engines do not use intake and exhaust valves. Instead, they
have openings, referred to as "ports," in the sides of the cylinder walls. There are typically three
ports in the cylinder; an intake port that brings the air-fuel mixture into the crankcase; a transfer
port that channels the air and fuel mixture from the crankcase to the combustion chamber; and an
exhaust port that allows burned gases to leave the cylinder and flow into the exhaust manifold.
Two-stroke engines route incoming air and fuel mixture first into the crankcase, then into the
cylinder via the transfer port. This is fundamentally different from a four-stroke engine which
delivers the air and fuel mixture directly to the combustion chamber.
With a two-stroke engine, as the piston approaches the bottom of the power stroke, it
O O
3-2
-------
Chapter 3: Technology
uncovers exhaust ports in the wall of the cylinder. The high pressure burned combustion gases
blow into the exhaust manifold. At the same time, downward piston movement compresses the
fresh air and fuel mixture charge in the crankcase. As the piston gets closer to the bottom of the
power stroke, the transfer ports are uncovered, and fresh mixture of air and fuel are forced into
the cylinder while the exhaust ports are still open. Exhaust gas is "scavenged" or forced into the
exhaust by the pressure of the incoming charge of fresh air and fuel. In the process, however,
some mixing between the exhaust gas and the fresh charge of air and fuel takes place, so that
some of the fresh charge is also emitted in the exhaust. Losing part of the fuel out of the exhaust
during scavenging causes the very high hydrocarbon emission characteristics of two-stroke
engines.
At this point, the power, exhaust, and transfer events have been completed. When the piston
begins to move up, its bottom edge uncovers the intake port. Vacuum draws fresh air and fuel
into the crankcase. As the piston continues upward, the transfer port and exhaust ports are
closed. Compression begins as soon as the exhaust port is blocked. When the piston nears TDC,
the spark plug fires and the cycle begins again.
3.1.3 - Engine Calibration
For most current SI engines, the two primary variables that manufacturers can control to
reduce emissions are the air and fuel mixture (henceforth referred to as air-fuel ratio) and the
spark timing. For highway motorcycles, these two variables are the most common methods for
controlling exhaust emissions. However, for many nonroad engines and vehicles, the absence of
emission standards have resulted in air-fuel ratio and spark timing calibrations optimized for
engine performance and durability rather than for low emissions.
3.1.3.1 Air-fuel ratio
The calibration of the air-fuel mixture affects power, fuel consumption (referred to as Brake
Specific Fuel Consumption (BSFC)), and emissions for SI engines. The effects of changing the
air-fuel mixture are shown in Figure 3.1-1.1 Traditionally, in most nonroad SI applications,
manufacturers have calibrated their fuel systems for rich operation for two main advantages.
First, by running the engine rich, manufacturers can 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. Second, by making extra fuel available for combustion, it is possible to get more power
from the engine. At the same time, since a rich mixture lacks sufficient oxygen for full
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.
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
without increasing the risk of lean misfire. This reduces HC and CO emissions and fuel
consumption. Leaner air-fuel mixtures, however, increase NOx emissions due to the higher
3-3
-------
Draft Regulatory Support Document
temperatures and increased supply of oxygen.
Figure 3.1-1: Effects of Air-fuel Ratio on Power, Fuel Consumption, and Emissions
o
CL
O
LL
OT
CO
Lean
Stoichiometric
Rich
Power
BSFC
I
ro
o
ta
E
LU
"3
O
Lean
Stoichiometric
Rich
NOx
I
0.7 0.8 0.9 1.0 1.1
Fuel/Air Equivalence Ratio
3.1.3.2 Spark-timing:
1.2
1.3
0.7 0.8 0.9 1.0 1.1
Fuel/Air Equivalence Ratio
1.2
1.3
For each engine speed and air-fuel mixture, there is an optimum spark-timing that results in
peak torque. If the spark is advanced to an earlier point in the cycle, more combustion occurs
during the compression stroke. If the spark is retarded to a later point in the cycle, peak cylinder
pressure is decreased because too much combustion occurs later in the expansion stroke when it
generates little torque on the crankshaft. 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 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.
3.1.3.3 - Fuel Metering
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
5-4
-------
Chapter 3: Technology
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.
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, 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.
Multi-port gasoline injection: As the name suggests, multi-port fuel injection means that a
fuel injector is placed at each of the intake ports. A quantity of fuel is injected each time the
intake valve opens for each cylinder. This 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 over
15 years.
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
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.3
3.1.4 - Alternate Fuels
5-5
-------
Draft Regulatory Support Document
Gaseous-fuel engines
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 don't 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.
3.2 - Exhaust Emissions and Control Technologies
3.2.1 - Current Two-Stroke Engines
As discussed above, two-stroke engines are typically found in applications where light
weight, low cost, simplistic design, easy starting, and high power-to-weight ratio are desirable
attributes. Of the engines and vehicles and covered by this rulemaking, the engines found in
recreational vehicles tend to have a high percentage of two-stroke engines. For example, almost
all snowmobiles use two-stroke engines, while 40 percent of off-highway motorcycles are
equipped with two-strokes. Approximately 20 percent of all ATVs use two-stroke engines.
California ARB has had exhaust emission standards for off-highway motorcycles and ATVs
since 1996. However, the regulations allow the sales and use of non-certified vehicles within the
state. Thus, recreational vehicles equipped with two-stroke engines have essentially been
unregulated. As a result, two-stroke engines used in recreational vehicles are typically designed
for optimized performance and durability rather than low emissions. Current two-stroke engines
emit extremely high levels of HC and CO emissions. The scavenging of unburned fuel into the
exhaust contributes to the bulk of the HC emissions. Up to 30 percent" of the air and fuel mixture
(along with lubricating oil) can pass unburned from the combustion chamber to the exhaust,
resulting not only in high levels of HC, but also in high levels of particulate matter (PM). As
discussed above, two-stroke engines lubricate the engine by mixing specially-formulated two-
stroke oil with gasoline. As the gasoline/oil mixture passes through the carburetor, it is
atomized into fine droplets and mixed with air. The gasoline quickly vaporizes, while the less
volatile oil forms a fine mist of fine droplets. Some of these droplets contact the crankshaft,
piston pin, and cylinder walls, providing lubrication. Most of the oil droplets, however, pass out
of the crankcase and into the cylinder with the rest of the incoming charge. Much of this oil mist
will be trapped in the cylinder and burned along with the gasoline vapor. Since lubricating oil is
less combustible than gasoline, some of the oil will survive the combustion process in the
cylinder and be passed into the exhaust. In the hot exhaust, the oil may vaporize, however, as the
exhaust cools and through mixing with air after it is emitted, the oil vapor recondenses into very
fine droplets or particles and enter the atmosphere as PM.
Hare et al, 1974; Batoni, 1978; Nuti and Martorano, 1985
-------
Chapter 3: Technology
Another major source of unburned HC emissions from two-stroke engines is due to misfire
or partial combustion at light loads. Under light load conditions such as idle, the flow of fresh air
and fuel into the cylinder is reduced, and substantial amounts of exhaust gas are retained in the
cylinder. This high fraction of residual gas leads to incomplete combustion or misfire, which is
the source of the "popping" sound produced by two-stroke engines at idle and light loads. These
unstable combustion events are major sources of unburned HC at idle and light load conditions.15
High CO levels from two-stroke engines are a result of operating the engine at rich air and
fuel mixture levels to promote engine cooling and enhance performance. Two-stroke engines
typically have very low levels of NOx emissions due to relatively cool combustion temperatures.
Two-stroke engines have cooler combustion temperatures as a result of two phenomenon: rich air
and fuel mixture operation and internal exhaust gas recirculation. Two-stroke engines tend to
operate with a rich air and fuel mixture to increase power and to help cool the engine. Because
many two-stroke engines are air-cooled, the extra cooling provided by operating rich is a
desirable engine control strategy. Combustion with a rich air and fuel mixture results in some
incomplete combustion which means less efficient combustion and a lower combustion
temperature. High combustion temperature is the main variable in producing NOx emissions.
Two-stroke engines also tend to have a high levels of naturally occurring exhaust gas
recirculation due to the scavenging process where some of the burned gases are drawn back into
the cylinder rather than being emitted out into the exhaust. The addition of burned exhaust gas
into the fresh charge of air and fuel mixture in the combustion chamber also results in less
complete or efficient combustion, which lowers combustion temperatures and reduces NOx
emissions.
HC emissions for recreational vehicle two-stroke engines are approximately 25 times higher
than for recreational vehicle four-stroke emissions. CO levels are roughly the same for both
types of engines, while NOx levels are 1.5 times lower than four-stroke engine levels. Table 3.2-
1 shows two-stroke emission results for several off-highway motorcycles and ATVs tested by
and for EPA in grams per kilometer (g/km). Table 3.2-2 shows two-stroke emission results from
snowmobiles in grams per horsepower-hour (g/hp-hr).
Tsuchiya et al, 1983; Abraham and Prakash, 1992; Aoyama et al, 1977
3-7
-------
Draft Regulatory Support Document
Table 3.2-1
Baseline Two-Stroke Emissions From Off-Highway Motorcycles & ATVs (g/km)
MCor
ATV
ATV
ATV
ATV
MC
MC
MC
MC
MC
MC
MC
MC
MC
MC
MC
Manufacturer
Suzuki
Polaris
Polaris
KTM
KTM
KTM
Honda
Honda
Honda
Honda
Honda
KTM
KTM
KTM
Model
LT80
Scrambler 80
Trailblazer
125SX
125SX
200EXC
n/a
n/a
n/a
CR250R
n/a
250SX
250EXC
300EXC
Model
Year
1998
2001
2000
2001
2001
2001
1993
1993
1995
1997
1998
2001
2001
2001
Eng. Displ.
80 cc
90 cc
250 cc
125 cc
125 cc
200 cc
200 cc
200 cc
249 cc
249 cc
249 cc
249 cc
249 cc
298 cc
Average
HC
7.66
38.12
18.91
33.71
61.41
53.09
8.00
26.00
12.00
17.47
23.00
62.89
59.13
47.39
33.56
CO
24.23
25.08
44.71
31.01
32.43
39.89
16.00
28.00
21.00
36.62
36.00
49.29
40.54
45.29
33.51
NOx
0.047
0.057
0.040
0.008
0.011
0.025
0.010
1.010
0.010
0.004
0.010
0.011
0.016
0.0124
0.091
-------
Chapter 3: Technology
Table 3.2-2
Baseline Two-Stroke Emissions From Snowmobiles (g/hp-hr)
Source
Carroll 1999 (SwRI)
YNP
White etal. 1997
White etal. 1997
Hare & Springer 1974
Hare & Springer 1974
Hare & Springer 1974
Wright & White 1998
Wright & White 1998
ISMA#1
ISMA #2
ISMA #3
ISMA #4
ISMA #5
ISMA #6
ISMA #7
ISMA #8
ISMA #9
ISMA #10
ISMA #11
ISMA #12
ISMA #13
ISMA #14
ISMA #15
Eng. Displ.
480 cc
488 cc
440 cc
436 cc
335 cc
247 cc
440 cc
503 cc
600 cc
440 cc
600 cc
900 cc
698 cc
597 cc
695 cc
485 cc
340 cc
440 cc
600 cc
700 cc
593 cc
494 cc
699 cc
Average
HC
115
150
160
89
120
200
130
105
110
95
106
95
92
100
88
148
104
95
94
102
67
105
92
111
CO
375
420
370
142
235
63
380
400
218
312
196
215
298
328
345
385
297
294
262
355
288
400
276
298
NOx
0.69
0.42
0.50
1.40
1.80
3.40
0.42
0.73
0.86
1.62
1.30
0.84
0.34
0.30
0.24
0.56
0.84
0.56
0.81
0.69
0.57
0.43
0.50
0.86
PM
0.7
1.1
3.4
6.1
2.5
2.6
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
2.7
5-9
-------
Draft Regulatory Support Document
3.2.2 - Clean Two-Stroke Technologies
Technologies available for reducing two-stroke emissions can be grouped into several
categories: calibration improvements; combustion chamber modifications; improved scavenging
characteristics; advanced fuel metering systems; and exhaust aftertreatment technologies.
3.2.2.1 - Calibration Improvements
The vast majority of two-stroke engines used in recreational vehicles use a carburetor as the
means of metering the air and fuel that is supplied to the engine. The carburetion system
supplies a controlled mixture of air and fuel to the engine, taking into consideration engine
temperature and load and speed, while trying to optimize engine performance and fuel economy.
A carburetor is a mechanical fuel atomizing device. It uses the venturi or Bernoulli's principle,
which is based on pressure differences, to draw fuel into the air stream from a small reservoir
(known as the "bowl"). A venturi is a restriction formed in the carburetor throat. As air passes
through the venturi, it causes an increase in air velocity and creates a vacuum or low pressure.
The fuel in the bowl is under atmospheric pressure. The higher pressure fuel will flow to the
lower pressure (vacuum) created in the airstream by the venturi. The fuel is atomized (broken
into small droplets) as it enters the airstream.
As discussed above in section 3.1.3.1, the calibration of the air-fuel mixture affects power,
fuel consumption, and emissions. Traditionally, in most recreational vehicles using two-stroke
engines, manufacturers have calibrated their fuel systems for rich operation for two main
advantages. First, by running the engine rich, manufacturers can 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. Second, by making extra fuel available for combustion, it is possible to get more power
from the engine. At the same time, since a rich mixture lacks sufficient oxygen for full
combustion, it results in increased fuel consumption rates and higher HC and CO emissions.
One means of reducing HC and CO emissions from two-stroke engines is to calibrate the
air-fuel ratio for lower emissions. This means leaning the air-fuel mixture, so that there is more
oxygen available to oxidize HC and CO. This strategy appears simplistic, but the manufacturer
has to not only optimize the air-fuel ratio for emissions, but also allow acceptable performance
and engine cooling. This means that the air-fuel ratio must not be leaned to the point of causing
lean misfire or substantially reduced power. However, since it is common for manufacturers to
set-up their carburetors to operate overly rich, there is opportunity for better optimization of
carburetor air-fuel settings to account for performance, engine cooling and lower emissions.
3.2.2.2 - Combustion Chamber Modifications
For two-stroke engines, if modifications are made to air-fuel calibrations that result in leaner
operation, one of the main concerns is that the combustion temperature will increase and result in
engine damage. It is fairly common for two-stroke engines to seize the piston in the cylinder if
they operate at too high of combustion temperatures. Piston seizure results when combustion
3-10
-------
Chapter 3: Technology
chamber temperatures become excessive and the piston heats-up and expands until it becomes
lodged or seizes in the cylinder. Depending on the level of enleanment used to control HC and
CO emissions, it may be necessary to also incorporate modifications to the combustion chamber.
Combustion chamber and piston configuration can be improved to induce more swirl and squish
or turbulent motions during the compression stroke, as well as control the flow direction of the
air and fuel mixture as it enters the combustion chamber to minimize short-circuiting (unburned
fuel leaving thru the exhaust port). Increasing turbulence in the combustion chamber improves
thermal efficiency by increasing the rate of burning in the chamber, which results in lower
combustion temperatures. Improved combustion chamber and piston configurations can also
minimize the formation of pocket or dead zones in the cylinder volume where unburned gases
can become trapped. Many engine designs induce turbulence into the combustion chamber by
increasing the velocity of the incoming air-fuel mixture and having it enter the chamber in a
swirling motion (known as "swirl").
3.2.2.3 - Improved Scavenging Characteristics
As discussed above, the exhaust and intake events for two-stroke engines overlap
extensively, resulting in considerable amounts of unburned gasoline and lubricating oil passing
through the engine and out the exhaust into the atmosphere. As the piston moves downward
uncovering the exhaust port, a fresh charge of air and fuel enters the combustion chamber under
pressure from the transfer port and pushes the burned gases from the previous combustion event
out into the exhaust. Since the burned gases are pushed out of the chamber by the intake
mixture, some of the fresh air and fuel mixture being introduced into the chamber are also lost
through the exhaust port. The ideal situation would be to retain all of the fresh charge in the
cylinder while exhausting all of the burned gases from the last cycle. This is difficult in most
current two-stroke engine designs, since the cylinder ports and piston timing are generally
designed for high scavenging efficiency, in order to achieve maximum power and a smoother
idle, which results in higher scavenging losses and emissions. It is possible to reconfigure the
cylinder ports to fine tune the scavenging characteristics for lower emissions, but this involves
significant trade-offs with engine performance. There are, however, several techniques that can
be employed to improve scavenging losses.
Exhaust charge control technology modifies the exhaust flow by introducing one-way
control valves in the exhaust, or by making use of the exhaust pressure pulse wave. In order to
get increased power out of a two-stroke engine, it is imperative that the engine combust as much
air and fuel as possible. Scavenging losses from two-stroke engines (called "short-circuiting")
allow a large percentage of the air and fuel to leave the combustion chamber before they can be
combusted. Two-stroke engines used in recreational vehicles all tend to use an exhaust system
equipped with an "expansion chamber." An expansion chamber is typically made of two cones,
one diverging and the other converging, with a short straight section of pipe between the two
cones. As the exhaust pulse leaves the exhaust port and enters the exhaust pipe, it travels
through the diverging cone and expands. The expanded pulse travels through the straight section
of pipe and then meets the converging cone. Upon hitting the converging cone, the exhaust pulse
wave becomes a sonic wave and travels back into the combustion chamber, pushing some of the
3-11
-------
Draft Regulatory Support Document
burnt exhaust gases and fresh charge of air and fuel that escaped originally.
As part of the Society of Automotive Engineers (SAE) Clean Snowmobile Challenge 2001,
a college competition which encourages the development clean snowmobile technologies,
Colorado State University (CSU) developed a two-stroke snowmobile engine using a
supercharged "reverse uniflow" design. The reverse uniflow design incorporates an exhaust port
and a crankcase pressure activated intake valve. After the ignition of the charge occurs at TDC,
the high combustion pressures and expanding gases force the piston downward. As the bottom
of the piston covers the exhaust port, the pressure in the crankcase increases due to a decreasing
volume. The increasing pressure is transmitted to the check-valve diaphragm. As the piston
fully uncovers the exhaust port, the exhaust gases are expelled out of the port, and the cylinder
pressure goes to approximately atmospheric pressure. Due to the larger pressure in the crankcase
(and thus on the diaphragm) as compared to the cylinder, the check-valve opens and the
supercharged intake begins to runs into the cylinder. As the intake air is entering the cylinder,
expelling the exhaust gases out of the bottom ports, a fuel injector or carburetor provides fuel
into the intake air stream. After the piston reaches BDC, and begins to move back upwards, the
crankcase pressure decreases. Once the piston moves past the exhaust port, the crankcase
pressure returns to approximately atmospheric pressure, and the check-valve completely closes.
The piston continues up, compressing the air-fuel mixture until the point that ignition can once
again occur, completing the cycle.
3.2.2.4 - Advanced Fuel Metering Systems
The most promising technology for reducing emissions from two-stroke engines are
advanced fuel metering systems, otherwise known as fuel injection systems. For two-stroke
engines, there are two types of fuel injection systems available. The first system is electronic fuel
injection (EFI), similar to what exists on automobiles. This system consists of an electronic fuel
injector, an electronic fuel pump, pressurized fuel lines and an electronic control unit (ECU) or
computer. EFI also requires the use of various sensors to provide information to the ECU so
that precise fuel control can be delivered. These sensors typically monitor temperature, throttle
position and atmospheric pressure. The use of EFI can provide better atomization of the fuel and
more precise fuel delivery than found with carburetors, which can reduce emissions. EFI systems
also have the advantage of providing improved power and fuel economy, when compared to a
carburetor. However, EFI does not address the high emission resulting from short-circuiting or
scavenging losses.
The second type of fuel injection system, known as Direct Injection (DI), does address
scavenging losses. DI systems are very similar to EFI systems, since both are electronically
controlled systems. The main difference is that DI systems more fully atomize (i.e., break-down
into very small droplets) the fuel, which can greatly improve combustion efficiency resulting in
improved power and reduced emissions. DI engines pump only air into the cylinder, rather than
air and fuel. Finely atomized fuel is then injected into the combustion chamber once all of the
ports are closed. This eliminates the short-circuiting of fresh air and fuel into the exhaust port.
The biggest problem with DI is that there is very little time for air to be pumped into the cylinder
3-12
-------
Chapter 3: Technology
and fuel then injected after all of the ports have closed. This is overcome by the use of numerous
engines sensors, a high-speed electronic control module, and software which uses sophisticated
control algorithms.
DI systems have been in use for the past several years in some small motorcycle, scooter and
marine applications, primarily for personal watercraft (PWC) and outboard engines. There are
numerous variations of DI systems, but two primary approaches that are commercially available
today: high pressure injection and air-assisted injection. There are a number of companies who
have developed high pressure DI systems, but the most successful systems currently belong to
FICHT and Yamaha. The FICHT system uses a special fuel injector that is able to inject fuel at
very high pressure (e.g., over 250 psi). The fuel injector itself is essentially a piston that is
operated by an electromagnet. Fuel enters the injector at low pressure from an electric fuel pump
and is forced out of the injector nozzle at high pressure when the piston hammers down on the
fuel. The Yamaha system uses a high pressure fuel pump to generate the high fuel pressure. The
other DI approach that is most common in various engine applications is the air-assisted injection
system which has been developed by Orbital. The Orbital system uses pressurized air to help
inject the fuel into the combustion chamber. The system uses a small single cylinder
reciprocating air compressor to assist in the injection of the fuel. All three systems are currently
used in some marine applications by companies such as Kawasaki, Polaris, Sea-Doo, and
Yamaha. The Orbital system is also currently used on some small motorcycle and scooter
applications by Aprilla. Certification data from various engines certified with DI have shown
HC and CO emission reductions of 60 to 75 percent from baseline emission levels.
There is at least one other injection technology that has had success in small two-stroke SI
engines used in lawn and garden applications, such as trimmers and chainsaws. Compression
Wave technology, referred to as Low Emission (LE) technology, developed by John Deere, uses
a compressed air assisted fuel injection system, similar to the Orbital system, to reduce the
unburned fuel charge during the scavenging process of the exhaust portion of the two-stroke
cycle. The system has shown the ability to reduce HC and CO emissions by up to 75 percent
from baseline levels. Although this technology has not yet been applied to any recreational
vehicle engines, it appears to have significant potential, especially because of its simplistic
design and low cost. For a detailed description of the LE technology, refer to the Nonroad Small
SI regulatory support document.
3.2.2.5 - Exhaust Aftertreatment Technologies
There are two exhaust aftertreatment technologies that can provide additional emission
reductions from two-stroke engines: thermal oxidation (e.g., secondary air) and oxidation
catalyst. Thermal oxidation reduces HC and CO by promoting further oxidation of these species
in the exhaust. The oxidation usually takes place in the exhaust port or pipe, and may require the
injection of additional air to supply the needed oxygen. If the exhaust temperature can be
maintained at a high enough temperature (e.g., 600 to 700°C) for a long enough period,
substantial reductions in HC and CO can occur. Air injection at low rates into the exhaust
system has been shown to reduce emissions by as much as 77 percent for HC and 64 percent for
3-13
-------
Draft Regulatory Support Document
CO.1 However, this was effective only under high-power operating conditions, and the high
exhaust temperatures required to achieve this oxidation substantially increased the skin
temperature of the exhaust pipe, which can be a concern for off-highway motorcycle applications
where the operators legs could come in contact with the pipe.
Like thermal oxidation, the oxidation catalyst is used to promote further oxidation of HC
and CO emissions in the exhaust stream, and it also requires sufficient oxygen for the reaction to
take place. Some of the requirements for a catalytic converter to be used in two-stroke engines
include high HC conversion efficiency, resistance to thermal damage, resistance to poisoning
from sulfur and phosphorus compounds in lubricating oil, and low light-off temperature.
Additional requirements for catalysts to be used in recreational vehicle two-stroke engines
include extreme vibration resistance, compactness, and light weight.
Application of catalytic converters to two-stroke engines presents a problem, because of the
high concentrations of HC and CO in their exhaust. If combined with sufficient air, these high
pollutant concentrations result in catalyst temperatures that can easily exceed the temperature
limits of the catalyst. Therefore, the application of oxidation catalysts to two-stroke engines may
first require engine modifications to reduce HC and CO and may also require secondary air be
supplied to the exhaust in front of the catalyst.
Researchers of Graz University of Technology and the Industrial Technology Research
Institute (ITRI) in Taiwan have published data on the application of catalytic converters in small
two-stroke moped and motorcycle engines using catalytic converters. The Graz researchers
focused on reducing emissions using catalysts, as well as by improving the thermodynamic
characteristics of the engines, such as gas exchange and fuel handling systems, cylinder and
piston geometry and configurations, and exhaust cooling systems. For HC and CO emissions,
they found that an oxidation catalyst could reduce emissions by 88 to 96 percent. Researchers at
ITRI successfully retrofitted a catalytic converter to a 125 cc two-stroke motorcycle engine, and
demonstrated both effective emissions control and durability."1 The Manufacturers of Emission
Controls Association (MECA)in their publication titled "Emission Control of Two-and Three-
wheel Vehicles," published May 7, 1999, state that catalyst technology has clearly demonstrated
the ability to achieve significant emissions reductions from two-stroke engines. MECA points to
the success of two-stroke moped and motorcycle engines equipped with catalysts that have been
operating for several years in Taiwan, Thailand, Austria, and Switzerland.
1 White, J.J., Carroll,J.N., Hare, C.T., and Lourenco, J.G. (1991), "Emission Control
Strategies for Small Utility Engines," SAE Paper No. 911807, Society of Automotive Engineers,
Warrendale, PA, 1991.
m Hsien, P.H., Hwang, L.K., and Wang, H.W (1992), "Emission Reduction by
Retrofitting a 125 cc Two-Stroke Motorcycle with Catalytic Converter," SAE Paper No. 922175,
Society of Automotive Engineers, Warrendale, PA, 1992.
3-14
-------
Chapter 3: Technology
3.2.3 - Current Four-Stroke Engines
Four-stroke engines are the most common type of engine today. Large nonroad SI engines
are exclusively four-stroke. Recreational vehicles are also predominantly four-stroke. Four-
stroke engines have considerably lower HC emissions than two-stroke engines, due to the fact
that four-stroke engines do not experience short circuiting of raw fuel. CO emissions from four-
stroke engines is very similar to two-stroke engines, since CO emissions are the result of
inefficient combustion of the air-fuel mixture within the cylinder, typically resulting from rich
operation. Since the combustion of fuel within the cylinder of a four-stroke engine is more
efficient than that of a two-stroke engine, combustion temperatures are higher, which results in
higher NOx emission levels.
The four-stroke engines covered under this rulemaking are typically either automotive
engines (large nonroad SI) or motorcycle-like engines (including ATVs). Large nonroad SI
engines, off-highway motorcycles, ATVs, and snowmobiles have been unregulated federally.
Therefore, while they have relatively low HC emissions compared to two-stroke engines, they
can still have high levels of CO (due to rich air-fuel calibration) and NOx. Table 3.2-3 shows
baseline emission levels for four-stroke equipped off-highway motorcycles and ATVs.
3-15
-------
Draft Regulatory Support Document
Table 3.2-3
Baseline Four-Stroke Emissions From Off-Highway Motorcycles & ATVs (g/km)
MCor
ATV
MC
MC
MC
MC
ATV
ATV
ATV
ATV
ATV
ATV
ATV
ATV
ATV
ATV
ATV
Manufacturer
Yamaha
Yamaha
KTM
Husaberg
Kawasaki
Honda
Polaris
Yamaha
Polaris
Arctic Cat
Yamaha
Honda
Bombardier
Polaris
Yamaha
Model
WR250F
WR400
400EXC
FE501
Bayou
300EX
Trail Boss
Banshee
Sportsman H.O.
375 Automatic
Big Bear
Rancher
4X4 AWD
Sportsman
Raptor
Model
Year
20001
1999
2001
2001
1989
1997
1998
1998
2001
2001
2001
2001
2001
2001
2001
Eng. Displ.
249 cc
399 cc
398 cc
499 cc
280 cc
298 cc
325 cc
349 cc
499 cc
375 cc
400 cc
400 cc
500 cc
499 cc
660 cc
Average
HC
1.46
1.07
1.17
1.30
1.17
1.14
1.56
0.98
2.68
1.70
2.30
1.74
1.62
1.56
0.97
1.40
CO
26.74
20.95
28.61
25.81
14.09
34.60
43.41
19.44
56.50
49.70
41.41
33.98
20.70
19.21
16.56
28.33
NOx
0.110
0.112
0.050
0.163
0.640
0.155
0.195
0.190
0.295
0.190
0.170
0.150
0.740
0.420
0.210
0.245
3.2.4 - Clean Four-Stroke Technologies
The emission-control technologies for four-stroke engines are very similar to those used for
two-stroke engines. HC and CO emissions from four-stroke engines are primarily the result of
poor in-cylinder combustion. 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 3.1.3.1 and 3.1.3.2, can further reduce HC emissions and
lower CO emissions. The conversion from carburetor to EFI will also help reduce HC and CO
emissions. The use of exhaust gas recirculation on Large SI engines can reduce NOx emissions,
but is not necessarily needed for recreational vehicles, due to their relatively low NOx emission
levels. 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.
3-16
-------
Chapter 3: Technology
3.2.4.1. - 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 don't 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 excessive leakage of
lubricating oil into the combustion chamber.
3.2.4.2 - 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.5 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 since the addition of recirculated gas
increases intake pressure. Because the burned gas temperature is decreased, there is 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.6
For catalyst systems with high conversion efficiencies, the benefit of using EGR becomes
proportionally smaller. Also, including EGR as a design variable for optimizing the engine adds
significantly to the development time needed to fully calibrate engine models.
3.2.4.3. - Secondary air
Secondary injection of air into exhaust ports or pipes after cold start (e.g., the first 40 to 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, secondary air is injected into the exhaust system,
preferably in close proximity of the exhaust valve. Together with the oxygen of the secondary air
3-17
-------
Draft Regulatory Support Document
and the hot exhaust components of HC and CO, oxidation ahead of the catalyst can bring about
an efficient increase in the exhaust temperature which helps the catalyst to heat up quicker. The
exothermic reaction that occurs is dependent on several parameters (secondary air mass, location
of secondary air injection, engine A/F ratio, engine air mass, ignition timing, manifold and
headpipe construction, etc.), and ensuring reproducibility demands detailed individual
application for each vehicle or engine design.
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.
3.2.4.4 - Catalytic Aftertreatment
Over the last several years, there have been tremendous advances in exhaust aftertreatment
systems. Catalyst manufacturers are progressively moving to palladium (Pd) as the main
precious metal in automotive catalyst applications. Improvements to catalyst thermal stability
and washcoat technologies, the design of higher cell densities, and the use of two-layer washcoat
applications are just some of the advances made in catalyst technology. 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 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. For the NO reduction to occur efficiently, an overall rich or stoichiometric
air-fuel ratio is required. The NOx efficiency drops rapidly as the air-fuel ratio becomes leaner
than stoichiometric. If the air-fuel ratio can be maintained precisely at or just rich of
stoichiometry, 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 CO control even within this window. HC oxidation generally correlates with
CO conversion, though changing air-fuel ratios tend to affect CO emissions much more than HC
emissions.
There are several issues involved in designing catalytic control systems for the four-stroke
engines covered by this rulemaking. The primary issues are the cost of the system, packaging
constraints, and the durability of the catalyst. This section addresses these issues.
3-18
-------
Chapter 3: Technology
3.2.4.4.1. - System cost
Sales volumes of industrial and recreational equipment are small compared to automotive
sales. Manufacturers therefore have a limited ability to recoup large R&D expenditures for Large
SI and recreational engines. For this reason, we believe it is not appropriate to consider highly
refined catalyst systems that are tailored specifically to nonroad applications. For Large SI
engines, we have based the feasibility of the emission standards on the kind of catalysts that
manufacturers have already begun to offer for these engines. These systems are currently
produced in very low volumes, but the technology has been successfully adapted to Large SI
engines. The cost of these systems will decrease substantially when catalysts become
commonplace. Chapter 4 describes the estimated costs for a nonroad catalyst system.
3.2.4.4.2. - Packaging constraints
Large SI engines power a wide range of nonroad equipment. Some of these have no
significant space constraints for adding a catalyst. In contrast, equipment designs such as
forklifts have been fine-tuned over many years with a very compact fit. The same is even more
true for recreational vehicles, such as ATVs and motorcycles.
Automotive catalyst designs typically have one or two catalyst units upstream of the muffler.
This is a viable option for most nonroad equipment. However, if there is no available space to
add a separate catalyst, it is possible to build a full catalyst/muffler combination that fits in the
same space as the conventional muffler. With this packaging option, even compact applications
should have little or no trouble integrating a catalyst into the equipment design. The hundreds of
catalysts currently operating on forklifts and highway motorcycles clearly demonstrate this.
3.2.5 - Advanced Emission Controls
On February 10, 2000, EPA published new "Light-duty 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. (See Chapter 4 for a more detailed discussion of the
impact of the Light-duty Tier 2 final rule on VOC inventories.) 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. 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
3-19
-------
Draft Regulatory Support Document
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 Light-
duty Tier 2 final rule."
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.0 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 (or
traps). Each of these technologies, which are discussed below, offer 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
n http://www.epa.gov/otaq/tr2home.htmtfDocuments. EPA 420-R-99-023
0 McDonald, J., L. Jones, Demonstration of Tier 2 Emission Levels for Heavy Light-Duty Trucks, SAE
2000-01-1957.
3-20
-------
Chapter 3: Technology
demonstrated the ability to store heat for more than 12 hours.p 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.q 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.
3.2.5.1 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
p Burch, S.D., and IP. Biel, SULEV and "Off-Cycle" Emissions Benefits of a Vacuum-Insulated Catalytic
Convert, SAE 1999-01-0461.
q Laing, P.M., Development of an Alternator-Powered Electrically-Heated Catalyst System, SAE 941042.
3-21
-------
Draft Regulatory Support Document
intake and two exhaust valves to reduce pumping losses and improve their volumetric efficiency
and useful power output. Some highway motorcycles have used multiple valves for years,
especially the high-performance sport motorcycles.
In addition to gains in breathing, 4-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-out 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 4-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 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.
Variable-valve technology by itself may have somewhat limited effect on reducing
emissions, but combining it with optimized spark plug location and exhaust gas recirculation can
lead to substantial emission reductions.
3.3 - Evaporative Emissions
3.3.1 Sources of Evaporative Emissions
Evaporative emissions from nonroad SI equipment represents a small but significant part of
their NMHC emissions. The significance of the emissions varies widely depending on the engine
design and application. LPG-fueled equipment generally has very low evaporative emissions
because of the tightly sealed fuel system. At the other extreme, carbureted gasoline-fueled
equipment with open vented tanks can have very high evaporative emissions. 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.
RUNNING LOSSES: The hot engine and exhaust system can vaporize gasoline when the
engine is running.
3-22
-------
Chapter 3: Technology
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.
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.
Among the factors that affect emission rates are: (1) fuel metering (fuel injection or
carburetor); (2) the degree to which fuel permeates fuel lines and fuel tanks; (3) the proximity of
the fuel tank to the exhaust system or other heat sources; (4) whether the fuel system is sealed
and the pressure at which fuel vapors are vented; and (5) fuel tank volume.
3.3.1.1 - Diurnal and Running Loss 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 would be about 4.5 psi. In this example, the partial
pressure of the air would be 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 would be 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 would be 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.
Consider a typical fuel use cycle beginning with a full tank. 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
3-23
-------
Draft Regulatory Support Document
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 exhaust is very close to the fuel tank,
the fuel can actually begin to boil. When this happens, large amounts of gasoline vapor can be
vented directly to the atmosphere. Southwest Research Institute measured emissions from
several large nonroad gasoline engines and found them to vary from about 12 g/day up to almost
100 g/day. They also estimated that a typical large nonroad gasoline engine in the South Coast
Air Basin (the area involved in their study) would have an evaporative emission rate of about 0.4
g/kW-hr.
3.3.1.2 - Hot Soak Emissions
Hot soak emissions occur after the engine is turned off, especially during the resulting
temperature rise. For nonroad engines, 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.
3.3.1.3 - Refueling Emissions
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. At low
temperatures, the fuel vapor content of the vapor space that is replaced is lower than it is at
higher temperatures.
3.3.1.4 - Permeation
The polymeric material (plastic or rubber) of which many gasoline fuel tanks and fuel hoses
generally have 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. 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 rates are relatively low,
but emissions continue at a nearly constant rate, regardless of how much the vehicle or
equipment is used. Permeation-related emissions can therefore add up to a significant fraction of
the total emissions from gasoline powered vehicles.
3.3.2 Evaporative Emission Controls
Several emission-control technologies can be used to reduce evaporative emissions. The
advantages and disadvantages of the various possible emission-control strategies are discussed
3-24
-------
Chapter 3: Technology
below. Chapter 4 presents more detail on how we expect manufacturers to use these
technologies to meet the emission standards for the individual applications.
3.3.2.1 - 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.
For forklifts, the primary application of Large SI engines, Underwriters Laboratories
specifies that units operating in certain areas where fire risk is most significant must use
pressurized fuel tanks. Underwriters Laboratories requires that trucks use self-closing fuel caps
with tanks that stay sealed to prevent evaporative losses; venting is allowed for positive pressures
above 3.5 psi or for vacuum pressures of at least 1.5 psi.r These existing requirements are
designed to prevent evaporative losses for safety reasons. This same approach for other types of
engines would similarly reduce emissions for air-quality reasons.
An alternative to using a pressure relief valve to hold vapors in the fuel tank would be to use
a limited flow orifice. However, the orifice size may be so small that there would be a risk of
fouling. In addition, an orifice designed for a maximum of 2 psi under worst case conditions
may not be very effective at lower temperatures. One application where a limited flow orifice
may be useful is if it is combined with an insulated fuel tank as discussed below.
3.3.2.2 - Insulated Fuel Tank
Another option for reducing diurnal emissions is insulating the fuel tank. Rather than
capturing the vapors in the fuel tank, this strategy would minimize the fuel heating which
therefore minimizes the vapor generation. However, significant evaporative emissions would
still occur through the vent line due to diffusion even without temperature gradients. A limited-
flow orifice could be used to minimize the to loss of vapor through the vent line due to diffusion.
In this case, the orifice could be sized to prevent diffusion losses without causing pressure build-
up in the tank. Additional control could be achieved with the use of a pressure relief valve or a
smaller limited flow orifice. Note that an insulated tank could maintain the same emission
control with a lower pressure valve than a tank that was not insulated.
3.3.2.3 - Volume-Compensating Air Bag
Another concept for minimizing pressure in a sealed fuel tank is through the use of a
volume-compensating air bag. 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
TJL558, paragraphs 26.1 through 26.4
3-25
-------
Draft Regulatory Support Document
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. Once the fuel tank
cools as ambient temperature goes down, the resulting vacuum in the fuel tank will open the bag
back up. Depending on the size of the bag, pressure in the tank could be minimized; therefore,
the use of a volume-compenating air bag could allow a manufacturer to reduce the pressure limit
on its relief valve.
3.3.2.4 - Collapsible 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.
3.3.2.5 - Charcoal Canister
The primary evaporative emission control device used in automotive applications is a
charcoal canister. With this technology, vapor generated in the tank is vented through a charcoal
canister. The activated charcoal 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 charcoal canisters generally are about a liter in size and have the capacity to store three
days of vapor over the test procedure conditions.
For industrial applications, engines are typically used frequently which would limit the size
of canister needed; however, introducing an evaporative canister is a complex undertaking,
requiring extensive efforts to integrate evaporative and exhaust emission-control strategies.
Large SI engine manufacturers also often sell loose engines to equipment manufacturers, who
would also need to integrate the new technology into equipment designs.
3.3.2.6 - 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 industrial fuel tanks. Because of the motion of the
equipment, 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.
3.3.2.7 - Permeation Barriers
Another source of evaporative emissions is permeation through the walls of plastic fuel
tanks and rubber hoses.
3-26
-------
Chapter 3: Technology
3.3.2.7.1 Fuel Tanks
Blow molding is widely used for the manufacture of small fuel tanks of recreational
vehicles. 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
then pinched in a mold and inflated with an inert gas. In highway 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 two additional extruder screws which significantly increases the cost
of the blow molding process.
Multi-layer fuel tanks can also be formed using injection molding. 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. To add a barrier layer, a thin sheet of the barrier
material is 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. As an alternative, an additional extruder can be
added to inject the barrier layer prior to injecting the HDPE; however, this substantially increases
the cost of the process.
A less expensive alternative to coextrusion 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 EVOH or nylon, 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.
Another type of low permeation technology for fuel tanks would be to treat the surfaces of a
plastic fuel tanks with a barrier layer. Two ways of achieving this are known as fluorination and
sulfonation. The fluorination process causes a chemical reaction where exposed hydrogen atoms
are replaced by larger fluorine atoms which a barrier on 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 be 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. However, this method may not prove as effective as off-line
3-27
-------
Draft Regulatory Support Document
fluorination which treats the inside and outside surfaces.
Sulfonation is another surface treatment technology where 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, can also be
performed off-line. Either of these processes can be used to reduce gasoline permeation by more
than 95 percent.
3.3.2.7.2 Fuel Hoses
Fuel hoses produced for use in recreational vehicles are generally extruded nitrile rubber
with a cover for abrasion resistance. Lower permeability fuel hoses produced today for other
applications are generally constructed in one of two ways: either with a low permeability layer or
by using a low permeability rubber blend. By using hose with a low permeation thermoplastic
layer, permeation emissions can be reduced by more than 95 percent. Because the thermoplastic
layer is very thin, on the order of 0.1 to 0.2 mm, the rubber hose retains its flexibility. Two
thermoplastics which have excellent permeation resistance, even with an alcohol-blend fuel, are
ethylene-tetrafluoro-ethylene (ETFE) and tetra-fluoro-ethylene, hexa-fluoro-propylene, and
vinyledene fluoride (THV).
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. 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 recreational vehicle
design. Manufacturers have commented that they would need flexible hose to fit their many
designs, resist vibration, and to simplify the hose connections and fittings.
An alternative approach to reducing the permeability of marine hoses would be to apply a
surface treatment such as fluorination or sulfonation. This process would be performed in a
manner similar to discussed above for fuel tanks.
3.4 CI Recreational Marine Engines
In this section, we discuss how emissions can be reduced from compression-ignition (CI)
recreational marine engines. We believe recreational marine diesel engines can use the same
technology for reducing emissions that will be used to meet the standards for commercial marine
diesel engines.7 Because of the similarities between recreational and commercial diesel engines,
this chapter builds off the technological analysis in the Regulatory Impact Analysis for the
commercial diesel marine engine rule.8 This section discusses emissions formation, baseline
technology, control strategies for CI recreational marine engines.
3-28
-------
Chapter 3: Technology
3.4.1 Background on Emissions Formation from Diesel Engines
Most, if not all, of compression-ignition recreational marine engines use diesel fuel. For this
reason, we focus on recreational marine diesel engines in this section. In a diesel engine, the
liquid fuel is injected into the combustion chamber after the air has been heated by compression
(direct injection), or the fuel is injected into a prechamber, where combustion initiates before
spreading to the rest of the combustion chamber (indirect injection). The fuel is injected in the
form of a mist of fine droplets or vapor that mix with the air. Power output is controlled by
regulating the amount of fuel injected into the combustion chamber, without throttling (limiting)
the amount of air entering the engine. The compressed air heats the injected fuel droplets,
causing the fuel to evaporate and mix with the available oxygen. At several sites where the fuel
mixes with the oxygen, the fuel auto-ignites and the multiple flame fronts spread through the
combustion chamber.
NOx and PM are the emission components of most concern from diesel engines. Incomplete
evaporation and burning of the fine fuel droplets or vapor result in emissions of the very small
particles of PM. Small amounts of lubricating oil that escape into the combustion chamber can
also contribute to PM. Although the fuel-air ratio in a diesel cylinder is very lean, the air and
fuel are not a homogeneous charge as in a gasoline engine. As the fuel is injected, the
combustion takes place at the flame-front where the fuel-air ratio is near stoichiometry
(chemically correct for combustion). At localized areas, or in cases where light-ends have
vaporized and burned, molecules of carbon remain when temperatures and pressures in the
cylinder become too low to sustain combustion as the piston reaches bottom dead center.
Therefore, these heavy products of incomplete combustion are exhausted as PM.
NOx formation requires high temperatures and excess oxygen which are found in a diesel
engine. Therefore, the diesel combustion process can cause the nitrogen in the air to combine
with available oxygen to form NOx. High peak temperatures can be seen in typical unregulated
diesel engine designs. This is because the fuel is injected early to help lead to more complete
combustion, therefore, higher fuel efficiency. If fuel is injected too early, significantly more fuel
will mix with air prior to combustion. Once combustion begins, the premixed fuel will burn at
once leading to a very high temperature spike. This high temperature spike, in turn, leads to a
high rate of NOx formation. Once combustion begins, diffusion burning occurs while the fuel is
being injected which leads to a more constant, lower temperature, combustion process.
Because of the presence of excess oxygen, hydrocarbons evaporating in the combustion
chamber tend to be completely burned and HC and CO are not emitted at high levels.
Evaporative emissions from diesel engines are insignificant due to the low evaporation rate of
diesel fuel.
Controlling both NOx and PM emissions requires different, sometimes opposing strategies.
The key to controlling NOx emissions is reducing peak combustion temperatures since NOx
forms at high temperatures. In contrast, the key to controlling PM is higher temperatures in the
combustion chamber or faster burning. This reduces PM by decreasing the formation of
3-29
-------
Draft Regulatory Support Document
particulates and by oxidizing those participates that have formed. To control both NOx and PM,
manufacturers need to combine approaches using many different design variables to achieve
optimum performance. These design variables are discussed in more detail below.
3.4.2 Marinization Process
Like commercial marine engines, recreational marine engines are not generally built from
the ground up as marine engines. Instead, they are often marinized land-based engines. The
main difference between recreational and commercial marine engines is the application for which
they are designed. Commercial engines are designed for high hours of use. Recreational engines
are generally designed for higher power, but less hours of use. The following is a brief
discussion of the marinization process, as it is performed by either engine manufacturers or post-
manufacture marinizers (PMM).
3.4.2.1 Process common to all marine diesel engines
The most obvious changes made to a land-based engine as part of the marinization process
concern the engine's cooling system. Marine engines generally operate in closed compartments
without much air flow for cooling. This restriction can lead to engine performance and safety
problems. To address engine performance problems, these engines make use of the ambient
water to draw the heat out of the engine coolant. To address safety problems, marine engines are
designed to minimize hot surfaces. One method of ensuring this, used mostly on smaller marine
engines, is to run cooling water through a jacket around the exhaust system and the turbocharger.
Larger engines generally use a thick insulation around the exhaust pipes.
Hardware changes associated with these cooling system changes often include water
jacketed turbochargers, water cooled exhaust manifolds, heat exchangers, sea water pumps with
connections and filters, and marine gear oil coolers. In addition, because of the greater cooling
involved, it is often necessary to change to a single-chamber turbocharger, to avoid the cracking
that can result from a cool outer wall and a hot chamber divider.
Marinization may also involve replacing engine components with similar components that
are made of materials that are more carefully adapted to the marine environment. Material
changes include more use of chrome and brass including changes to electronic fittings to resist
water induced corrosion. Zinc anodes are often used to prevent engine components, such as raw-
water heat exchangers, from being damaged by electrolysis.
3.4.2.2 Process unique to recreational marine diesel engines
Other important design changes are related to engine performance. Especially for planing
hull vessels used in recreational and light duty commercial marine applications, manufacturers
strive to maximize the power-to-weight ratio of their marine engines, typically by increasing the
power from a given cylinder displacement. The most significant tool to accomplish this is the
fuel injection system: the most direct way to increase power is to inject more fuel. This can
3-30
-------
Chapter 3: Technology
require changes to the camshaft, cylinder head, and the injection timing and pressure.
Design limits for increased fuel to the cylinder are smoke and durability. Modifications
made to the cooling system also help enhance performance. By cooling the charge, more air can
be forced into the cylinder. As a result, more fuel can be injected and burned efficiently due to
the increase in available oxygen. In addition, changes are often made to the pistons, cylinder
head components, and the lubrication system. For instance, aluminum piston skirts may be used
to reduce the weight of the pistons. Cylinder head changes include changing valve timing to
optimize engine breathing characteristics. Increased oil quantity and flow may be used to
enhance the durability of the engine.
Depending on the stage of production and the types of changes made, the marinization
process can have an impact on the base engine's emission characteristics. In other words, a land-
based engine that meets a particular set of emission limits may no longer meet these limits after it
is marinized. This can be the case, for example, if the fuel system is changed to enhance engine
power or if the cooling system no longer achieves the same degree of engine cooling as that of
the base engine. Because marine diesel engines are currently unregulated, engine manufacturers
have been able to design their marine engines to maximize performance. Especially for
recreational marine engines, manufacturers often obtain power/weight ratios much higher than
for land-based applications.
Recreational engine manufacturers strive for higher power/weight ratios than are necessary
for commercial marine engines. Because of this, recreational marine engines use technology we
projected to be used by commercial marine engines to meet the Tier 2 emissions standards such
as raw-water aftercooling and electronic control. However, this technology is used to gain more
power rather than to reduce emissions. The challenge presented by the emission control program
will be to achieve the emission limits while maintaining favorable performance characteristics.
3.4.3 General Description of Technology for Recreational Marine Diesel Engines
We believe that the standards can be met using technology that has been developed for and
used on land-based nonroad and highway engines. The Regulatory Impact Analysis for the
commercial marine final rule includes a lengthy description of emission control technology for
diesel marine engines. Table 3.4-1 outlines this description. By combining the strategies shown
below, manufacturers can optimize the emissions and performance of their engines. We
anticipate that the same percent reductions achievable on commercial marine engines would be
achievable on recreational marine engines using the same technology. The same technology is
used in land-based applications to achieve even a higher magnitude of emission reduction. In
addition, this technology works consistently across the engine map encompassed by the NTE
zone. A more detailed analysis of the application of several of these technologies to recreational
marine engines is discussed in Chapter 4. The costs associated with applying these systems are
considered in Chapter 5.
3-31
-------
Table 3.4-1: Emission Control Strategies for Marine Diesel Engines
Technology
Combustion
optimization:
Advanced fuel
injection
controls
Improving
charge air
characteristics
Electronic
control
Exhaust gas
recirculation
Exhaust
aftertreatment
devices
(would require
"dry" exhaust)
Description HC CO NOx PM
timing retard-reduce peak cylinder temperatures by shortening 1 t 11 It
the premixed burning phase
reduced crevice volume-such as raising the top piston ring 1 1 « 1
geometry-match piston crown geometry to injector spray 1 1 1 1
increased compression ratio-raises cylinder pressures 1 i 1 1
increased swirl-control of air motion for better mixing 1 i t ,« 1
increased injection pressure-better atomization of fuel 1 1 t,« 1
nozzle geometry-optimize spray pattern 1 1 1 1
valve-closed orifice-minimize leakage after injection 1 « « 1
rate shaping-inject small amount of fuel early to begin 1
combustion to reduce premixed burning
common rail-high pressure rail to injectors, excellent control of 1 i 1 1
fuel rate, pressure, and timing
turbocharging-increases available oxygen in the cylinder but 1 i 1 1
heats intake air
jacket-water aftercooling-uses engine coolant to cool charged 1
air which increases available oxygen in cylinder
raw-water aftercooling-uses ambient water to cool charge air; « « 11
more effective than jacket-water aftercooling; may result in
additional maintenance such as changing anodes
better control of fuel system including rate, pressure, and timing 1 1 1 1
especially under transients; can use feedback loop
hot EGR-recirculated exhaust gas reduces combustion till
temperatures by absorbing heat and slowing reaction rates
cooled EGR-reduces volume of recirculated gases so to allow « ~ 11 t ,«
more oxygen in the cylinder
soot removal-soot in recirculated gases may cause durability « 1
problems at high EGR rates; gas filter or trap; oil filter
oxidation catalyst-oxidizes hydrocarbons and soluble organic 1 1 « 1
fraction of PM; will be poisoned by high levels of sulfur
paniculate trap-collect PM; use catalyst to regenerate at high 1 1 « 1
temperature
selective catalytic reduction-uses a catalyst and a reducing 1
-------
Chapter 3: Technology
Water water is mixed with fuel or injected into the cylinder; water has
emulsification a high heat capacity and will lower in-cylinder temperatures
Chapter 3 References
1. Heywood, J., "Internal Combustion Engine Fundamentals," McGraw-Hill, Inc., New York,
1988, pp.829-836, Docket A-2000-01, Document IV-A-110.
2. Heywood, pp.827-829, Docket A-2000-01, Document IV-A-110.
3. Saikalis, G., Byers, R., Nogi., T., "Study on Air Assist Fuel Injector Atomization and Effects
on Exhaust Emission Reduction," SAE Paper 930323, 1993, Docket A-2000-01, Document
II-A-55.
4. Energy and Environmental Analysis, "Benefits and Cost of Potential Tier 2 Emission
Reduction Technologies", Final Report, November 1997, Docket A-2000-01, Document H-A-01.
5. Southwest Research Institute, "Three-Way Catalyst Technology for Off-Road Equipment
Powered by Gasoline and LPG Engines," prepared for California ARB, California EPA, and
South Coast Air Quality Management District, (SwRI 8778), April 1999, Docket A-2000-01,
Document H-A-08.
6. Heywood, pp. 836-839, Docket A-2000-01, Document IV-A-110.
7. "Control of Emissions of Air Pollution from New Marine Compression-Ignition Engines at or
Above 37 kW; Final Rule," 64 FR 73318, December 29, 1999.
8. Final Regulatory Impact Analysis for "Control of Emissions of Air Pollution from New
Marine Compression-Ignition Engines at or Above 37 kW; Final Rule," November 1999, Docket
A-2000-01, Document H-A-78.
3-33
-------
Draft Regulatory Support Document
3-34
-------
Chapter 4: Feasibility of Proposed Standards
Chapter 4: Feasibility of Standards
Section 213(a)(3) of the Clean Air Act presents statutory criteria that EPA must evaluate in
determining standards for nonroad engines and vehicles. 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 emission standards are technically achievable
accounting for all the above factors.
It is important to note that the term "greatest degree of emission reduction achievable"
applies with respect to in-use emissions from each production engine at the end of engine's useful
life, rather than what is achievable under more ideal laboratory conditions. This means that the
standards that are being established in this rulemaking must account for production variability
and for deterioration in emission performance that will occur in use as the engines age and wear
over the applicable useful life periods. We have considered these factors in determining the
lowest emissions that will be feasible in the time frame required. Thus, in some cases, the
emission standards are somewhat higher than the lowest emissions observed during laboratory
testing. In general, we expect that manufacturers will design their engines and vehicles to be at
10- 20 percent below the applicable emission standard when produced to account for both
production variability and deterioration. Chapter 6 includes more information about our
expectations regarding compliance margins and deterioration rates.
4.1 CI Recreational Marine
The emission standards for CI recreational marine engines are summarized in the Executive
Summary. We believe that manufacturers will be able to meet these standards using technology
similar to that required for the commercial marine engine standards. This section discusses
technology currently used on CI recreational marine engines and anticipated technology to meet
the standards. In addition, this section discusses the emission test procedures and Not-to-Exceed
requirements.
4.1.1 Baseline Technology for CI Recreational Marine Engines
We developed estimates of the current mix of technology for CI recreational marine engines
based on data from the 1999 Power Systems Research (PSR) database and from conversations
with marine manufacturers. Based on this information, we estimate that 97 % of new marine
engines are turbocharged, and 80% of these turbocharged engines use aftercooling. The majority
of these engines are four-stroke, but about 14% of new engines are two-stroke. Electronic
4-1
-------
Draft Regulatory Support Document
controls have only recently been introduced into the marketplace; however, we anticipate that
their use will increase as customers realize the performance benefits associated with electronic
controls and as the natural migration of technology from on-highway to nonroad to marine
engine applications occurs.
Table 4.1-1 presents data1'2'3'4'5'6 from 25 recreational marine diesel engines based on the ISO
E5 duty cycle. This data shows to what extent emissions need to be reduced from today's CI
recreational marine engines to meet the standards.8 On average, we are requiring significant
reductions in HC+NOx and PM. However, this data seems to show that the diesel engine
designs will either have to be focused on NOx or PM due to the trade-off between calibrating to
minimize these pollutants. The CO standard will act more as a cap, but will require control to be
established.
s For most of the engines in Table 4.1-1, the standards are of 7.2 g/kW-hr HC+NOx, 5
g/kW-hr CO, and 0.2 g/kW-hr PM
4-2
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.1-1: Emissions Data from CI Recreational Marine Engines
Rated Power (kW)
120
132
142
162
164
170
186
209
230
235
265
276
287
321
324
336
336
447
447
474
537
820
1040
1080
1340
Control Management
electronic
mechanical
mechanical
mechanical
electronic
mechanical
mechanical
mechanical
electronic
mechanical
mechanical
mechanical
electronic
mechanical
mechanical
electronic
electronic
electronic
mechanical
electronic
electronic
electronic
electronic
electronic
electronic
Aftercooling
raw-water
raw-water
separate circuit
raw-water
raw-water
raw-water
raw-water
raw-water
raw-water
raw-water
jacket-water
raw-water
raw-water
raw-water
jacket-water
jacket-water
jacket-water
raw-water
jacket-water
raw-water
jacket-water
separate circuit
jacket-water
separate circuit
separate circuit
Emissions Data (g/kW-hr)
HC NOx CO PM
0.09 5.8 0.9
0.07 4.2 0.2
0.79 8.6 1.1
0.11 4.0 0.2
0.28 5.1 1.6
0.36 8.1 0.6 0.20
0.30 10.2 1.2 0.12
0.42 10.8 2.3 0.22
0.28 5.5 1.8 0.39
0.45 9.8 1.8 0.20
0.58 10.8 1.4
0.60 10.7 1.9 0.24
0.28 7.9 - 0.12
0.37 7.7 0.9 0.23
0.30 7.9 2.9 0.95
0.18 11.0 0.5 0.10
0.09 11.9 - 0.16
0.12 9.3 - 0.17
0.60 12.0 1.5 0.18
0.34 7.7 0.5 0.07
0.08 10.7 - 0.19
0.33 9.5 0.8 0.13
0.09 9.3 - 0.21
0.18 7.6 1.2 0.15
0.27 7.2 0.9 0.15
4-3
-------
Draft Regulatory Support Document
4.1.2 Anticipated Technology for CI Recreational Marine Engines
Marine engines are generally derived from land-based nonroad, locomotive, and to some
extent highway engines. In addition, recreational marine engines will be able to use technology
developed for commercial marine engines. This allows recreational marine engines, which
generally have lower sales volumes than other nonroad engines, to be produced more cost-
efficiently. Because the marine designs are derived from land-based engines, we believe that
many of the emission-control technologies which are likely to be applied to nonroad engines to
meet their Tier 2 and 3 emission standards will be applicable to marine engines. We also believe
that the technologies listed below will be sufficient for meeting both the new emission standards
and the Not to Exceed requirements discussed later in this chapter for the full useful life of these
engines.
We anticipate that timing retard will likely be used in most CI recreational marine
applications, especially at cruising speeds, to gain NOx reductions. The negative impacts of
timing retard on HC, PM, fuel consumption and power can be offset with improved fuel injection
systems with higher fuel injection pressures, optimized nozzle geometry, and potentially through
injection rate shaping. We do not expect marine engine manufacturers to convert from direct
injection to indirect injection due to these standards.
Regardless of environmental regulations, we believe that recreational marine engine
manufacturers will make more use of electronic engine management controls in the future to
satisfy customer demands of increased power and fuel economy. Through the use of electronic
controls, additional reductions in HC, CO, NOx, and PM can be achieved. Electronics may be
used to optimize engine calibrations under a wider range of operation. Most of the significant
research and development for the improved fuel injection and engine management systems
should be accomplished for land-based nonroad diesel engines which are being designed to meet
Tier 2 and Tier 3 standards. Common rail should prove to be a useful technology for meeting
even lower emission levels in the future, especially for smaller engines. Thus, the challenge for
this control program will be transferring land-based techniques to marine engines.
We project that all CI recreational marine engines will be turbocharged and most will be
aftercooled to meet emission standards. Aftercooling strategies will likely be mostly jacket-water
charge air cooling, and in some cases, we believe that separate cooling circuits for the
aftercooling will be used. We do not expect a significant increase in the use of raw-water charge
air cooling for marine engines as a result of this rule. We recognize that raw-water aftercooling
systems are currently in use in many applications. Chapter 5 presents one possible scenario of
how these technologies could be used on CI recreational marine engines to meet the standards.
By adopting standards that will not go into effect until 2006, we are providing engine
manufacturers with substantial lead time for developing, testing, and implementing emission
control technologies. This lead time and the coordination of standards with those for commercial
marine engines allows for a comprehensive program to integrate the most effective emission
control approaches into the manufacturers' overall design goals related to performance,
4-4
-------
Chapter 4: Feasibility of Proposed Standards
durability, reliability, and fuel consumption.
4.1.3 Emission Measurement Procedures for CI Recreational Marine Engines
In any program we design to achieve emissions reductions from internal combustion
engines, the test procedures we use to measure emissions are as important as the standards we put
into place. These test procedure issues include duty cycle for certification, in-use verification
testing, emission sampling methods, and test fuels.
4.1.3.1 Certification Duty Cycles
In choosing duty cycles for certification, we turned to the International Standards
Organization (ISO).7 For CI recreational marine engines, we based our standards on the ISO E5
duty cycle. This duty cycle is intended for "diesel engines for craft less than 24m length
(propeller law)."
We specify the E5 duty cycle for measuring emissions from CI recreational marine engines.
This cycle is similar to the E3 duty cycle which is used for commercial marine in that both cycles
have four steady-state test points on an assumed cubic propeller curve. However, the E5 includes
an extra mode at idle and has an average weighted power of 34% compared to the 69% for the
E3. This duty cycle is presented in Table 4.1-2.
Table 4.1-2: ISO E5 Marine Duty Cycle
Mode
1
2
3
4
5
% of Rated Speed
100
91
80
63
idle
% of Power at Rated Speed
100
75
50
25
0
Weighting Factor
0.08
0.13
0.17
0.32
0.30
4.1.3.2 Emission Control of Typical In-Use Operation
We are concerned that if a marine engine is designed for low emissions on average over a
small 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 boat which do not
necessarily lie on the test duty cycles. For instance, the test modes for the E5 duty cycle lie on
average propeller curves. However, a propulsion marine engine may never be fitted with an
"average propeller." In addition, a given engine on a boat may operate at higher torques than
average if the boat is heavily loaded. We are also aware that, before a boat comes to plane, the
engine operates closer to its full torque map than to the propeller curve.
4-5
-------
Draft Regulatory Support Document
We are applying the "Not-to-Exceed" (NTE) limit concept to recreational marine engines in
a way that is similar to commercial marine engines. This concept basically picks a zone of
operation under which a marine engine must not exceed the standard by a fixed percentage and is
discussed in more detail in the commercial marine FRM.8 Of course, the shape of the zone must
be adjusted to reflect recreational engine use.
Under this final rule, we have the authority to use test data from new or in-use engines to
confirm emissions compliance throughout an engine's useful life.
4.1.3.2.1 Engine operation included for NTE
The shape of the NTE zones are based on our understanding of how recreational marine
engines are used. Operation at low power is omitted from the NTE zone even though marine
engines operate here in use. This omission is because, by definition, brake-specific emissions
become very large at low power due to dividing by power values approaching zero.
We believe that the majority of marine engine operation is steady-state. We are therefore
including only steady-state operation in the NTE requirements. Also, these are technology-
forcing standards, so we expect engines to reduce emissions also under transient operation. If we
find that the effectiveness of this program is compromised due to high emissions under transient
operation, we will revisit this requirement in the future.
It should be noted that the emissions caps for operation in the NTE zone are based on the
weighted emissions over the E5 duty cycle. Because idle emissions are part of these weighted
values but not included in the NTE zone, it is likely that emissions in the NTE zone will be less
than the weighted average. This alone reduces the stringency of a "not-to-exceed" approach for
recreational when compared to commercial marine engines.
For compression-ignition engines, the NTE zone is defined by the maximum power curve,
actual propeller curves, and speed and load limits. The E5 duty cycle itself is based on a cubic
power curve through the peak power point. For the NTE zone, we define the upper boundary
using a speed squared propeller curve passing through the 115% load point at rated speed and the
lower boundary using on a speed to the fourth power curve passing through the 85% load point at
rated speed. We believe these propeller curves represent the range of propeller curves seen in
use.9 To prevent imposing an unrealistic cap on a brake-specific basis, we are limiting this
region to power at or above 25% of rated power and speeds at or above 63% of rated speed.
These limits are consistent with mode 4 of the E5 duty cycle. Figure 4.1-1 presents the NTE
zone for CI recreational marine engines.
4-6
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.1-1: NTE Zone for Recreational CI Marine Engines
(U
100%-
90%
80%
70%
60%
50%
40%
30%
20%
10%
1.50 x PEL
T3
CD
CD
Q.
CO
CO
CD
1.50 x PEL
CD
CD
Q.
CO
en
25% power
50% 60% 70% 80% 90%
engine speed
100%
We understand that an engine tested onboard a boat in use may not be operating as the
manufacturer intended because the owner may not be using a propeller that is properly matched
to the engine and boat. Also, the owner may have a boat that is overloaded and too heavy for the
engine. The boundaries in Figure 4.1-1 are intended to contain typical operation of recreational
diesel engines and exclude engines which are not used properly. Although the E5 uses a cubic
power curve engines generally see some variation in use. These boundaries are consistent with
operational data we collected.10
We are adopting emissions caps for the NTE zone that 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 1.00 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 steady-state tests included in this rule.
Consistent with the commercial requirements, we require that CI recreational marine engines
must meet a cap of 1.50 times the certified level for HC+NOx, PM, and CO for the speed and
power subzone below 45% of rated power and a cap of 1.20 times the certified levels at or above
4-7
-------
Draft Regulatory Support Document
45% of rated power. However, we are including an additional subzone, when compared with the
commercial NTE zone, at speeds greater than 95% of rated. We are adopting a cap of 1.50 times
the certified levels for this subzone. Our purpose for this additional subzone is to address the
typical recreational design for higher rated power. This power is needed to ensure that the engine
can bring the boat to plane.
We based the caps both on emissions data collected on the assumed propeller curve and on
data collected from a recreational marine diesel engine over a wide range of steady-state
operation. All of this data is cited earlier in this chapter. The data in Figures 4.1-2 through 4.1-4
show that, within the range of in-use testing points, HC+NOx and PM are generally well below
the E5 weighted averages. This is likely due to the effects of emissions at idle. For all of these
engines, modal CO results were below the standard. None of these engines are calibrated for
emissions control.
4-8
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.1-2: Mode/E5 Average HC+NOx
63% spd 80% spd 91% spd 100% spd
Figure 4.1-3: Mode/E5 Average PM
63% spd 80% spd 90% spd 100% spd
4-9
-------
0)
a
o
N
H
Z
"S
o
W)
s
o
H
•o
%
H
0)
a
a
o
r^
^
O
rt 1
LL
r~
S ®
°~ 1
O o>
81
•^ o
re is
01 S
to 8
UJ a:
UJ LZ
1- ^
Z/ »
O
1
M
•
T
T
CM -fr CD
OJ _ q co
CD CMV in
cp oq x CM
\ v"~
o\ \ — x in
CO » ^ \ ^
•^ X Ci X T-^
\ \
\ \
co x m CM vo co
CD CD 00 ^O CD
ci Xci ci V ci
V V
•^ \ CD T— \ o O
CD -CD CM . O) 00
ci x ci ^ \ci ci
\ \
\ \
\ \
S\ in co xr--
.CD ^ ^D
Ci X T-^ O O
\
\ ^
\ \
•<- \ o co i
CD ^JQ. _ _CD. 4
^~ CM O
"3- 1^ O CO
co CD oo in
ci ci •* -^
"" ^"" CD 00 Is* to IO ^f CO C^ T
oooooooo'c
j8/v\od pajBJ jo juaojad
^
0)
°' V
CD
CD
Q.
(/>
2 |
ii —
o
-i— •
5 1
CD
Q.
to
^J
IO
0
s
u_
^^
^^^ ^^
^Ł (D
Q. g
. 0)
0Q
M- C
1-
(0 c
Lt .2
+j
IA re
**' o«
^ §
[fJ S
Z CO
CO
LO OT OT CD h-
N-,. . _ O5 ^ CO T—
o ' o T-^ csi r^
Cp f^ »C7^
Q CD Ci^
t-^_ r , pg
in« r^ ' CM
ci' ci Y
1^ , 00 CD (7> CO
•* CD 00 CM, CD
ci ' , o ci T-: • T-:
° • P 0 S Ą
Ci • CD ^ ^ , T^
oo w . o oo "*"" "^- ^
ci ci . ci .-^ oo
ci o" " T-:
^0 ^0 ^0 ^0 ^0
O o O O O I
o oo to n- c>i
JSMOOd P8JBJ JO JU80J8d
o
- ^^
*J
- o -o
O> CD
CD
Q.
CO
ss 1
- O "CO
00 "-
M —
o
"c
CD
ss a
- O CD
I-- 0-
0
2_
to
^
5
LL
. |s«
x P:
0 fe
Z (A
51
I 0)
0 g
S»
_O ra
re o
(^ *"
U> LI
UJ LU
UJ 0
^~ _J
Z r^
T
T
T- 00 T^~
Cft _ 00 CM
^i~ T T
CKI O3X ^-
5 ° \ 2
\ X
r?x 8\ cS
T- X CD V CD
\ \
\ \
CO v ^ ^ \ T — CM
CM ^O ^5 CD O
^ xci ci fe ^
V V
CM \ oo co \ in co
co .in in . oo •*
T- x^ci Ci Y -^
\ \
\ \
O x o O V
\ \
4 \
to uru - J3. *
ci ci --
LO O 00 CD
-r^ -r^ O -^
~ ^~ 0) 00 ^^ CO IO ^f CO ^^ T
oooooooo'c
j8/v\od pajBJ jo juaojad
^
o>
° -o
CD
CD
Q.
S ®
^3 -i— •
CD
M—
0
0 1
CD
Q.
to
0
IO
o
3
X u_
^^^ ^^
^ Ł
OO
O
™|™ Q
> C
•81
^^
81
•^ o
ro tS
fV* 0)
10 8
LJJ Ł
UJ 5o
1— to
^.
in o o CM oo
•*. . . o i^ CD oo
•^ - T-: ci ci -^
5 8 '.85
T -^ Ci
^ 8 '.S
t- • ^ 'Ci
j- . t~~ en co oo
i
JSMOOd P8JBJ JO JU80J8d
o
- f—k
^_^
0
O> CD
Q.
CO
"O
vg CD
0^ -HJ
- O CO
*o
-1— 1
c
CD
-| 1
O
-------
Chapter 4: Feasibility of Proposed Standards
4.1.3.2.2 Ambient conditions during testing
Variations in ambient conditions can affect emissions from a marine engine. Such
conditions include air temperature, humidity, and (especially for diesels) water temperature. We
are applying the same ranges for these variables that apply to commercial marine engine. Within
the ranges, no corrections can be made for emissions. Outside of the ranges, emissions can be
corrected back to the nearest edge of the range. The ambient variable ranges are:
intake air temperature 13-35°C (55-95°F)
intake air humidity 7.1-10.7 g water/kg dry air (50-75 grains/lb.
dry air)
ambient water temperature 5-27°C (41 -80°F)
The air temperature and humidity ranges are consistent with those developed for NTE
testing of highway heavy-duty diesel engines. The air temperature ranges were based on
temperatures seen during ozone NAAQS exceedances.11 For NTE testing in which the air
temperature or humidity is outside of the range, emissions may be corrected back to the air
temperature or humidity range. These corrections must be consistent with the equations in Title
40 of the Code of Federal Regulations (CFR), except that these equations correct to 25°C and
10.7 grams per kilogram of dry air, while corrections associated with the NTE testing shall be to
the nearest outside edge of the specified ranges. For instance, if the temperature were higher than
35°C, a temperature correction factor may be applied to the emissions results to determine what
the emissions would be at 35°C.
For marine engines using aftercooling, we believe the charge air temperature is essentially
insensitive to ambient air temperature compared to the cooling effect of the aftercooler. SwRI
tested this theory and found that when the ambient air temperature was increased from 21.9 to
32.2°C, the cooling water to the aftercooler of a diesel marine engine only had to be reduced by
0.5°C to maintain a constant charge air temperature.12 According to the CFR correction factor,
there is only a ±3% variation in NOx in the NTE humidity range.
Naturally aspirated engines should be more sensitive to intake air temperature because the
temperature affects the density of the air into the engine. Therefore, high temperatures can limit
the amount of air drawn into the cylinder. Our understanding is that many engines operate in and
draw air from small engine compartments. This suggests that any naturally aspirated recreational
engines used today are already designed to operate with high intake air temperatures. In any
case, we do not believe that manufacturers will use naturally aspirated marine engines to meet the
new standards.
Ambient water temperature also may affect emissions due to its impact on engine and charge
air cooling. We based the water temperature range on temperatures that marine engines
experience in the U.S. in use. Although marine engines experience water temperatures near
freezing, we don't believe that additional emission control will be gained by lowering the
minimum water temperature below 5°C. At this time, we aren't aware of an established
4-11
-------
Draft Regulatory Support Document
correction factor for ambient water temperature. For this reason, NTE zone testing must be
within the specified ambient water temperature range.
We don't think that the range of ambient water temperatures discussed above will have a
significant effect on the stringency of the NTE requirements, even for aftercooled engines.
Following the normal engine test practice recommended by SAE for aftercooled engines, the
cooling water temperature would be set to 25±5°C.13 This upper portion of the NTE temperature
range is within the range suggested by SAE for engine testing. For lower temperatures,
manufacturers can use a thermostat or other temperature regulating device to ensure that the
charge air is not overcooled. In addition, the SAE practice presents data from four aftercooled
diesel engines on the effects of cooling medium temperature on emissions. For every 5°C
increase in temperature, HC decreases 1.8%, NOx increases 0.6%, and PM increases 0.1%.
We are aware that many marine engines are designed for operation in a given climate. For
instance, recreational vessels operated in Seattle don't need to be designed for 27°C water
temperatures. For situations such as this, manufacturers may petition for the appropriate
temperature ranges associated with the NTE zone for a specific engine design. In addition, we
understand there are times when emission control may need to be compromised for startability or
safety. Manufacturers are not responsible for the NTE requirements under start-up conditions.
In addition, manufacturers may petition to be exempt from emission control under specified
extreme conditions such as engine overheating where emissions may increase under the engine-
protection strategy.
4.1.3.3 Emissions Sampling
Aside from the duty cycle, the test procedures for marine engines are similar to those for
land-based nonroad engines. However, there are a few other aspects of marine engine testing
that need to be considered. Most recreational marine engines mix cooling water into the exhaust.
This exhaust cooling is generally done to keep surface temperatures low for safety reasons and to
tune the exhaust for performance and noise. Because the exhaust must be dry for dilute emission
sampling, the cooling water must be routed away from the exhaust in a test engine.
Even though many marine engines exhaust their emissions directly into the water, we base
our test procedures and associated standards on the emissions levels in the "dry" exhaust.
Relatively little is known about water scrubbing of emissions. We must therefore consider all
pollutants out of the engine to be a risk to public health. Additionally, we are not aware of a
repeatable laboratory test procedure for measuring "wet" emissions. This sort of testing is nearly
impossible from a vessel in-use. Finally, a large share of the emissions from this category come
from large engines which emit their exhaust directly to the atmosphere.
The established method for sampling emissions is through the use of full dilution sampling.
However, for larger engines the exhaust flows become so large that conventional dilute testing
requires a very large and costly dilution tunnel. One option for these engines is to use a partial
dilute sampling method in which only a portion of the exhaust is sampled. It is important that the
4-12
-------
Chapter 4: Feasibility of Proposed Standards
partial sample be representative of the total exhaust flow. The total flow of exhaust can be
determined by measuring fuel flow and balancing the carbon atoms in and out of the engine. For
guidance on shipboard testing, the MARPOL NOx Technical Code specifies analytical
instruments, test procedures, and data reduction techniques for performing test-bed and in-use
emission measurements.14 Partial dilution sampling methods can provide accurate steady-state
measurements and show great promise for measuring transient emissions in the near future. We
intend to pursue development of this method and put it in place prior to the date that the
standards in this final rule become enforceable.
Pulling a marine engine from a boat and bringing it to a laboratory for testing could be
burdensome. For this reason, we may perform in-use confirmatory testing onboard a boat. Our
goal would be to perform the same sort of testing as for the laboratory. However, engines tested
in a boat are not likely to operate exactly on the assumed propeller curve. For this reason,
emissions measured within the NTE zone must meet the subzone caps based on the certified
level during onboard testing. To facilitate onboard testing, manufacturers must provide a
location with a threaded tap where a sampling probe may be inserted. This location must be
upstream of where the water and exhaust mix at a location where the exhaust gases could be
expected to be the most homogeneous.
There are several portable sampling systems on the market that, if used carefully, can give
fairly accurate results for onboard testing. Engine speed can be monitored directly, but load may
have to be determined indirectly. For engines operating at a constant speed, it should be
relatively easy to set the engine to the points specified in the duty cycles.
4.1.3.4 Test Fuel Specifications
We are applying the recently finalized test fuel specifications for commercial marine engines
to recreational marine diesel engines. These fuel specifications are similar to land-based nonroad
fuel with a change in the sulfur content upper limit from 0.4 to 0.8 weight-percent (wt%). We
believe this will simplify development and certification burdens for marine engines that are
developed from land-based counterparts. This test fuel has a sulfur specification range of 0.03 to
0.80 wt%, which covers the range of sulfur levels observed for most in-use fuels. Manufacturers
will be able to test using any fuel within this range for the purposes of certification. Thus, they
will be able to harmonize their marine test fuel with U.S. highway (<0.05 wt%) and nonroad
(0.03 to 0.40 wt%), and European testing (0.1 to 0.2 wt%).
The intent of these test fuel specifications is to ensure that engine manufacturers design their
engines for the full range of typical fuels used by Category 1 marine engines in use. Because the
technological feasibility of the new emission standards is based on fuel with up to 0.4 wt%
sulfur, any testing done using fuel with a sulfur content above 0.4 wt% would be done with an
allowance to adjust the measured PM emissions to the level corresponding with a test using fuel
with 0.4 wt% sulfur. We do not expect the sulfur content to have a large impact on PM
emissions because only about 2 percent of the sulfur in the fuel is converted to direct sulfate
PM.15
4-13
-------
Draft Regulatory Support Document
The full range of test fuel specifications are presented in Table 4.1-3. Because testing
conducted by us is limited to the test fuel specifications, it is important that the test fuel be
representative of in-use fuels.
Table 4.1-3: Recreational Marine Diesel Test Fuel Specifications
Item
Initial Boiling Point, °C
10% point, °C
50% point, °C
90% point, °C
End Point, °C
Cetane
Gravity, API
Total Sulfur, % mass
Aromatics, % volume
Paraffins, Napthenes, Olefins
Flashpoint, °C
Viscosity @ 38 °C, centistokes
Procedure (ASTM)
D86-90
D86-90
D86-90
D86-90
D86-90
D613-86
D287-92
D 129-21 or D2622-92
D1319-89orD5186-91
D1319-89
D93-90
D445-88
Value (Type 2-D)
171-204
204-238
243-282
293-332
321-366
40-48
32-37
0.03-0.80
10 minimum
remainder
54 minimum
2.0-3.2
4.1.4 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 emission standards for CI recreational marine engines.
One important source of noise in diesel combustion is the sound associated with the
combustion event itself. When a premixed charge of fuel and air ignites, the very rapid
combustion leads to a sharp increase in pressure, which is easily heard and recognized as the
characteristic sound of a diesel engine. The conditions that lead to high noise levels also cause
high levels of NOx formation. Fuel injection changes and other NOx control strategies therefore
typically reduce engine noise, sometimes dramatically.
The impact of the new emission standards on energy is measured by the effect on fuel
consumption from complying engines. Many of the marine engine manufacturers are expected to
retard engine timing which increases fuel consumption somewhat. Most of the other technology
changes anticipated in response to the new standards, however, have the potential to reduce fuel
consumption as well as emissions. Redesigning combustion chambers, incorporating improved
4-14
-------
Chapter 4: Feasibility of Proposed Standards
fuel injection systems, and introducing electronic controls provide the engine designer with
powerful tools for improving fuel efficiency while simultaneously controlling emission
formation. To the extent that manufacturers add aftercooling to non aftercooled engines and shift
from jacket-water aftercooling to raw-water aftercooling, there will be a marked improvement in
fuel-efficiency. Manufacturers of highway diesel engines have been able to steadily improve fuel
efficiency even as new emission standards required significantly reduced emissions.
There are no known safety issues associated with the new emission standards. Marine
engine manufacturers will likely use only proven technology that is currently used in other
engines such as nonroad land-based diesel applications, locomotives, and diesel trucks.
4-15
-------
Draft Regulatory Support Document
4.2 Large Industrial SI Engines
This category of engines generally includes all nonrecreational land-based spark-ignition
engines rated above 19 kW that are not installed in motor vehicles or stationary applications. In
an earlier memorandum, we described the rationale for developing emission measurement
procedures for transient and off-cycle engine operation.16 Information from that memorandum is
not repeated here, except to the extent that it supports decisions about the selecting the numerical
emission standards.
The emission standards for Large SI engines are listed in the Executive Summary. The
following paragraphs summarize the data and rationale supporting the standards.
4.2.1 2004 Standards
Engine manufacturers are currently developing technologies and calibrations to meet the
2004 standards that apply in California. We expect manufacturers to rely on electronically
controlled, closed-loop fuel systems and three-way catalysts to meet those emission standards.
As described below, emission data show that water-cooled engines can readily meet the
California ARE standards (3 g/hp-hr NMHC+NOx; 37 g/hp-hr CO).
Manufacturers will have just over one year to prepare engines for nationwide sales starting
in 2004. Implementing new standards with such a short lead time is only possible because
manufacturers have been aware of their need to comply with the California ARB standards as
well as our proposal to implement those standards nationwide. With no need to further modify
engine designs, manufacturers should have time before 2004 to plan for increasing production
volume for nationwide sale of engines that can meet the 2004 California ARB standards.
Adopting standards starting in 2004 allows us to align near-term requirements with those
adopted by California ARB. This also provides early emission reductions and gives
manufacturers the opportunity to amortize their costs over a broader sales volume before
investing in the changes needed to address the long-term standards described below.
4.2.2 2007 Standards
The 2004 standards described above will reduce emissions from Large SI engines, but we
believe these levels don't fulfill the requirement to adopt standards achieving the "greatest degree
of reduction achievable" from these engines in the long term. With additional time to optimize
designs to better control emissions, manufacturers can optimize their designs to reduce emissions
below the levels required by the 2004 standards. We are also adopting new procedures for
measuring emissions starting in 2007, which will require further efforts to more carefully design
and calibrate emission-control systems to achieve in-use emission reductions. The following
discussion explains why we believe the 2007 emission standards are feasible.
4-16
-------
Chapter 4: Feasibility of Proposed Standards
The biggest uncertainty in adopting emission standards for Large SI engines was the degree
to which emission-control systems deteriorate with age. While three-way catalysts and closed-
loop fueling systems have been in place in highway applications for almost 20 years, we needed
to collect information showing how these systems hold up under nonroad use. To address this,
we participated in an investigative effort with Southwest Research Institute (SwRI), California
ARE, and South Coast Air Quality Management District, as described in the memorandum
referenced above.17 The engines selected for testing had been retrofitted with emission-control
systems in Spring 1997 after having already run for 5,000 and 12,000 hours. Both engines are in-
line four-cylinder models operating on liquefied petroleum gas (LPG)—a 2-liter Mazda engine
rated at 32 hp and a 3-liter GM engine rated at 45 hp. The retrofit consisted of a new,
conventional three-way catalyst, electronic controls to work with the existing fuel system, and the
associated sensors, wiring, and other hardware. The electronic controller allowed only a single
adjustment for controlling air-fuel ratios across the range of speed-load combinations.
Laboratory testing consisted of measuring steady-state and transient emission levels, both
before and after taking steps to optimize the system for low emissions. While the engines'
emission-control systems originally focused on controlling CO emissions, the testing effort
focused on simultaneously reducing HC, NOx, and CO emissions. This testing provides a good
indication of the capability of these systems to control emissions over an engine's full useful life.
The testing also shows the degree to which transient emissions are higher than steady-state
emission levels for Large SI engine operation. Finally, the testing shows how emission levels
vary for different engine operating modes. Emission testing included engine operation at a wide
range of steady-state operating points and further engine operation over several different transient
duty cycles. Much of the emissions variability at different speeds and loads can be attributed to
the basic design of the controller, which has a single, global calibration setting. This data
showing the variability of emissions is necessary to support the field-testing emission standards,
as described further below.
4.2.2.1. Steady-state testing results
Testing results from the aged engines at SwRI showed very good emission control capability
over the full useful life. Test results with emission control hardware on the aged engines lead to
the conclusion that the systems operated with relatively stable emission levels over the several
thousand hours. As shown in Table 4.2-1, the emission levels measured by SwRI are consistent
with results from a wide variety of measurements on other engines. The data listed in the table
includes only LPG-fueled engines. See Section 4.2.2.6 for a discussion of gasoline-fueled
engines.
4-17
-------
Draft Regulatory Support Document
Table 4.2-1
Steady-State Emission Results from LPG-fueled Engines
Test engine
Mazda 2L18
GM3L
Engine B
GFI19
Toyota/ECS 2L20
GM/Impco 3L21
HC+NOx*
g/hp-hr
0.51
0.87
0.22
0.52
NMHC+NOx
1.14
0.26
CO
g/hp-hr
3.25
1.84
2.79
2.23
0.78
0.21
Notes**
4,000 hours, add-on retrofit
5,600 hours, add-on retrofit
250 hours
5,000 hours
zero-hour; ISO Cl duty cycle
for nonroad diesel engines
zero-hour
*Measurements are THC+NOx, unless otherwise noted.
**Emissions were measured on the ISO C2 duty cycle, unless otherwise noted.
This data set supports emission standards significantly more stringent than the 2004
standards. However, considering the need to focus on transient emission measurements, we
believe it is not appropriate to adopt more stringent emission standards based on the steady-state
duty cycles. Stringent emission standards based on certain discrete modes of operation may
inappropriately constrain manufacturers from controlling emissions across the whole range of
engine speeds and loads. We therefore intend to rely more heavily on the transient testing to
determine the stringency of the emission-control program.
4.2.2.2 Transient testing results
The SwRI testing is the only known source of information for evaluating the transient
emission levels from Large SI engines equipped with emission-control systems. Table 4.2-2
shows the results of this testing. The transient emission levels, though considerably lower than
the 2004 standards, are higher than those measured on the steady-state duty cycles. A
combination of factors contribute to this. First, these engines are unlikely to maintain precise
control of air-fuel ratios during rapid changes in speed or load, resulting in decreased catalyst-
conversion efficiency. Also, the transient duty cycle includes operation at engine speeds and
loads that have higher steady-state emission levels than the seven modes constituting the C2 duty
cycle. Both of these factors also cause uncontrolled emission levels to be higher, so the
measured emission levels with the catalyst system still show a substantial reduction in emissions.
Additional emission data measured during transient operation is shown in Section 4.2.2.7 for
selecting the numerical values for the standards.
4-18
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.2-2
Transient Test Results from SwRI Testing
Engine*
Mazda
GM
Duty Cycle
Variable-speed, variable-load
Constant-speed, variable-load
Variable-speed, variable-load
THC+NOx
g/hp-hr
1.1
1.5
1.2
CO
g/hp-hr
9.9
8.4
7.0
*Based on the best calibration on the engine operating with an aged catalyst.
4.2.2.3 Off-cycle testing results
Engines operate in the field under both steady-state and transient operation. Although these
emission levels are related to some degree, they are measured separately. This section therefore
first considers steady-state operation.
Figures 4.2-1 through 4.2-6 show plots of emission levels from the test engines at several
different steady-state operating modes. This includes the seven speed-load points in the ISO C2
duty cycle, with many additional test points spread across the engine map to show how emissions
vary with engine operation. The plotted emission level shows the emissions at each normalized
speed and normalized load point. The 100-percent load points at varying engine speeds form the
engine's lug curve, which appears as a straight line because of the normalizing step.
Figure 4.2-1 shows the THC+NOx emissions from the Mazda engine when tested with an
aged catalyst. While several points are higher than the 0.51 g/hp-hr level measured on the C2
duty cycle, the highest levels observed from the Mazda engine are around 2.3 g/hp-hr. The
highest emissions are generally found at low engine speeds. Emission testing on the Mazda
engine with a new catalyst showed very similar results on the C2 duty cycle, so testing was not
done over the whole range of steady-state operating points shown in Figure 4.2-1.
CO emissions from the same engine had a similar mix of very low emission points and
several higher measurements. The CO levels along the engine's lug curve (100 percent load)
range 12 to 22 g/hp-hr, well above the other points, most of which are under 4 g/hp-hr. The
corner of the map with high-speed and low-load operation also has a high level of 9 g/hp-hr.
These high-emission modes point to the need to address control of air-fuel ratios at these
extremes of engine operation.
If CO emissions at these points were an inherent problem associated with these engines, we
could take that into account in setting the standard. Figure 4.2-4 shows, however, that the GM
engine with the same kind of aged emission-control system had emission levels at most of these
points ranging from 0.7 to 4.7 g/hp-hr. The one remaining high point on the GM engine was
4-19
-------
Draft Regulatory Support Document
11.6 g/hp-hr at full load and low speed. A new high-emission point was 28 g/hp-hr at the lowest
measured speed and load. Both of these points are much lower on the same engine with the new
catalyst installed (see Figure 4.2-6). These data reinforce the conclusion that adequate
development effort will enable manufacturers to achieve broad control of emissions across the
engine map.
Figure 4.2-3 shows the THC+NOx emissions from the GM engine when tested with the
aged catalyst. Emission trends across the engine map are similar to those from the Mazda
engine, with somewhat higher low-speed emission levels between 2.3 and 4.4 g/hp-hr at various
points. Operation on the new catalyst shows a significant shifting of high and low emission
levels at low-speed operation, but the general observation is that the highest emission levels
disappear, with 2.3 g/hp-hr being again the highest observed emission level over the engine map
(see Figure 4.2-5).
4-20
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.2-1
Mazda/old cat.--NOx+HC
g/hp-hr
100
80
as 60
40
20
0
1
1.89
1.61
1.92
2.08
2.28
1.43
0 20
C2 = 0.51 g/hp-hr
0.6
0.46
0.95
0.87
1.11
1.67
2.26
30 40
0.77
0.53
0.41
0.35
0.33
0.14
1.24
50 60 70
Speed
0.57
0.31
0.31
0.5
0.62
0.72
0.28
80 90
0.25
0.27
0.25
0.43
0.63
0.81
0.54
100 110
Figure 4.2-2
Mazda/old cat--CO
100 2224
80 1'07
0.23
§ 6° 0.33
_i
40 0.64
20 °'51
1.3
10 20
C2=3.25 g/hp-hr
11.52
2.28
1.27
0.88
0.56
0.04
0.19
30 40
g/hp-hr
15.24
8.07
4.06
2.44
0.91
0.79
0
50 60 70
Speed
18.98
4.17
3.01
3.87
3.61
2.89
1.61
80
2.49
3.87
3.88
3.9
4.47
7.6
9.08
90 100 110
-------
Draft Regulatory Support Document
Figure 4.2-3
GM/old cat.--NOx+HC
100 3.5
80 16
1.6
as 60 1R
o 1.6
_i
40 2.3
20 23
4.4
n
10 20
C2=0.87 ghp-hr
1.3
1.2
1.1
1.2
0.9
0.6
1.6
30 40
g/hp-hr
0.8
0.7
1.2
0.9
0.7
0.5
1.1
50 60 70
Speed
0.8
0.7
0.6
0.6
0.7
0.5
0.7
80
0.9
0.9
0.6
0.5
0.4
0.7
0.4
90 100 110
Figure 4.2-4
GM/old cat.--CO
100 11.6 0.7
•3 Q 91
80 ^y z''
4.3 2.4
as 60 41 _,
o 4.1 3.5
40 6.0 3.6
20 3.5 3.9
28.0 5.1
10 20 30 40
C2=1. 84 ghp-hr
g/hp-hr
4.7
0.6
1.7
1.6
2.1
1.1
1.4
50 60 70
Speed
4.5
0.7
0.6
0.8
0.3
2.8
10.3
80
0.7
0.7
1.3
1.8
1.4
6.2
4.3
90 100 110
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.2-5
GM/new cat.-NOx+HC
100 0.57 0.92
80 2'25 °'75
2.25 0.82
1 6° 1.93 0.79
_l
40 1.61 0.83
__ 1.33 0.66
1.47 1.17
10 20 30 40
C2=0.35 ghp-hr
g/hp-hr
0.32
0.28
0.19
0.25
0.30
0.13
0.25
50 60 70
Speed
0.26 0.14
0.18 0.11
0.19 0.08
0.20 0.05
0.06 0.04
0.13 0.08
0.65 0.16
80 90 100 110
Figure 4.2-6
GM/new cat.
100 4.08
80 °'55
0.33
03 60 „ QQ
O 0.33
_i
40 0.24
20 °'11
0.45
10 20
C2=0.28 ghp-hr
0.16
1.03
0.92
0.70
0.72
1.04
0.44
30 40
g/hp-hr
2.65
0.81
0.37
0.00
0.93
0.29
0.73
--CO
1.78
0.23
0.47
0.65
0.10
0.23
6.70
50 60 70 80
Speed
0.06
1.15
0.44
0.21
0.12
0.28
0.26
90 100 110
-------
Draft Regulatory Support Document
Field testing will typically also include transient emission measurement. Field-testing
measurement may include any segment of normal operation with a two-minute minimum
sampling period. This does not include engine starting, extended idling, or other cold-engine
operation. Table 4.2-3 shows a wide variety of transient emission levels from the two test
engines. While the engines were tested in the laboratory, the results show how emissions vary
under normal operation when installed in nonroad equipment. These segments could be
considered as valid field-testing measurements to show that an engine meets emission standards
in the field when tested in nonroad equipment in which the engines are installed. Several
segments included in the table were run with a hot start, which could significantly increase
emission levels, depending on how long the engine runs in open loop after starting. This is
especially important for CO emissions. Even with varied strategies for soaking and warming up
engines, emission levels are generally between 1 and 2 g/hp-hr THC+NOx and between 4 and 13
g/hp-hr CO. Emission levels don't seem to vary dramatically between cycle segments, even
where engine operation is significantly different.
Table 4.2-3
Transient Emission Measurements from SwRI Testing
Engine
Mazda
GM
Test Segment
"typical" forklift (5 min.)
"high-transient" forklift (5 min.)
highway certification test
backhoe/loader cycle
"typical" forklift (5 min.)
"high-transient" forklift (5 min.)
highway certification test
backhoe/loader cycle
THC+NOx
g/hp-hr
2.0
1.3
1.2
1.3
1.3
2.0
1.0
1.0
CO,
g/hp-hr
5.7
4.3
4.6
9.1
9.5
12.6
4.4
3.8
Notes
hot start
hot start
hot start
20-minute soak before test
hot start
hot start
3-minute warm-up; 2-minute soak
3-minute warm-up; 2-minute soak
4.2.2.4 Ambient conditions
While certification testing involves engine operation in a controlled environment, engines
operate in conditions of widely varying temperature, pressure, and humidity. To take this into
account, we are broadening the range of acceptable ambient conditions for field-testing
measurements. Field-testing emission measurements must occur with ambient temperatures
between 13° and 35° C (55° and 95° F), and with ambient pressures between 600 and 775
millimeters of mercury (which should cover almost all normal pressures from sea level to 7,000
feet above sea level). Tests will be considered valid regardless of humidity levels. This allows
4-24
-------
Chapter 4: Feasibility of Proposed Standards
testing under a wider range of conditions in addition to helping ensure that engines are able to
control emissions under the whole range of conditions under which they operate.
The SwRI test data published here are based on testing under laboratory conditions typical
for the test location. Ambient temperatures ranged from 70 to 86° F. Barometric pressures were
in a narrow range around 730 mm Hg. Humidity levels ranged from about 4 to 14 g of water per
kg dry air, but all emission levels were corrected to a reference condition of 10.7 g/kg. Most
testing occurred at humidity levels above 10.7, in which case actual NOx emission levels were up
to 7 percent lower than reported by SwRI after correction.. In the driest conditions, measured
NOx emission levels were up to 10 percent higher than reported. The field-testing standards take
into account the possibility of a humidity effect of increasing NOx emissions. We are not aware
of any reasons that varying ambient temperatures or pressures will have a significant effect on
emission levels from spark-ignition engines.
4.2.2.5 Durability of Emission-Control Systems
SwRI tested engines that had already operated for the full useful life period with functioning
emission-control systems. Before being retrofitted with catalysts and electronic fuel systems,
these engines had already operated for 5,000 and 12,000 hours, respectively. The tested systems
therefore provide very helpful information to show the capability of the anticipated emission-
control technologies to function over a lifetime of normal in-use operation.
The testing effort required selection, testing, and re-calibration of installed emission-control
systems that were not designed specifically to meet emission standards. These systems were
therefore not necessarily designed for simultaneously controlling NOx, HC, and CO emissions,
for lasting 5,000 hours or longer, or for performing effectively under all conditions and all types
of operation that may occur. The testing effort therefore included a variety of judgments, and
adjustments to evaluate the emission-control capability of the installed hardware. This effort
highlighted several lessons that should help manufacturers design and produce durable systems.
Selecting engines from the field provided the first insights into the functionality of these
systems. Tailpipe ppm measurements showed that several engines had catalysts that were
inactive (or nearly inactive). These units were found to have loose catalyst material inside the
housing, which led to a significant loss of the working volume of the catalyst and exhaust flow
bypassing the catalyst material. Dimensional measurements showed that this resulted from a
straightforward production error of improperly assembling the catalyst inside the shell.22 This is
not an inherent problem with catalyst production and is easily addressed with automated or more
careful manual production processes. The catalyst from the GM engine selected for testing had
also lost some of its structural integrity. Almost 20 percent of the working volume of the catalyst
had disappeared. This catalyst was properly re-assembled with its reduced volume for further
testing. This experience underscores the need for effective quality-control procedures in
assembling catalysts.
Substituting a new catalyst on the aged system allowed emission measurements that help us
4-25
-------
Draft Regulatory Support Document
estimate how much the catalysts degraded over time. This assessment is rather approximate,
since we have no information about the zero-hour emissions performance of that exact catalyst.
The new catalysts, which were produced about three years later under the same part numbers and
nominal characteristics, generally performed in a way that was consistent with the aged catalysts.
Not surprisingly, the catalyst with the reduced working volume showed a higher rate of
deterioration than the intact catalyst. Both units, however, showed very stable control of NOx
and HC emissions. CO deterioration rates were generally higher, but the degree of observed
deterioration was very dependent on the particular duty cycle and calibration for a given set of
emission measurements.
Measured emission levels from the aged catalysts shows what degree of conversion
efficiency is possible for each pollutant after several thousand hours of operation. The emission
data from the new catalysts suggest that manufacturers probably need to target low enough zero-
hour CO emission levels to account for significant deterioration. The data also show that catalyst
size is an important factor in addressing full-life emission control. The nominal sizes of the
catalysts on the test engines were between 50 and 55 percent of total engine displacement. The
cost analysis in Chapter 5 is based on initial compliance with a catalyst sized at 60 percent of
total engine displacement. We expect manufacturers to reduce catalyst size as much as possible
to reduce costs without risking the possibility of high in-use emissions.
Another important issue relates to degradation associated with fuel impurities, potential lack
of maintenance, and wear of oxygen sensors. Fuel system components in LPG systems are prone
to fuel deposits, primarily from condensation of heavy hydrocarbon constituents in the fuel. The
vaporizer and mixer on the test engines showed a typical degree of fuel deposits from LPG
operation. The vaporizer remained in the as-received condition for all emission measurements
throughout the test program. Emission tests before and after cleaning the mixer give an
indication of how much the deposits affect the ability of the closed-loop fueling system to keep
the engine at stoichiometry. For the GM engine operating with the aged catalyst, the combined
steps of cleaning the mixer and replacing the oxygen sensor improved overall catalyst efficiency
on the C2 duty cycle from 55 to 61 percent for NOx. CO conversion efficiency improved only
slightly. For the Mazda engine, the single step of cleaning the mixer slightly decreased average
catalyst efficiency on the C2 duty cycle for NOx emissions; HC and CO conversion efficiency
improved a small amount (see Table 4.2-4). Engines operating with new catalysts showed the
same general patterns. These data show that closed-loop fueling systems can be relatively
tolerant of problems related to fuel impurities.
4-26
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.2-4
Average C2 Catalyst Conversion Efficiencies Before and After Maintenance
Engine
GM
Mazda
Pollutant
NOx
CO
HC
NOx
CO
HC
OLD CATALYST
before
maintenance
54.7%
96.3%
93.8%
62.3%
96.9%
86.9%
after
maintenance
61.1%
98.1%
93.6%
61.5%
98.9%
93.2%
NEW CATALYST
before
maintenance
45.6%
99.3%
93.6%
60.3%
99.6%
86.2%
after
maintenance
56.1%
99.5%
93.7%
60.1%
99.6%
94.3%
Manufacturers may nevertheless be concerned that some in-use operation can cause fuel
deposits that exceed the fuel system's compensating ability to maintain correct air-fuel ratios.
Two technologies are available to address this concern. First, the required diagnostic systems
inform the operator if fuel-quality problems are severe enough to prevent the engine from
operating at stoichiometry. A straightforward cleaning step would restore the fuel system to
normal operation. Manufacturers may also be able to monitor mixer performance directly to
detect problems with fuel deposits, rather than depending on air-fuel ratios as a secondary
indicator. In any case, by informing the operator of the need for maintenance, the diagnostic
system reduces the chance that the manufacturer will find high in-use emissions that result from
fuel deposits.
The second technology to consider is designed to prevent fuel deposits from forming. A
commercially available thermostat regulates fuel temperatures to avoid any high-temperature or
low-temperature effects. In addition, some industry participants have made the general
observation that some engine models are more susceptible to fuel deposits than others,
suggesting that there may be other engine-design parameters that may help prevent these
problems.
Maintaining the integrity of the exhaust system another basic but essential element of
keeping control of air-fuel ratios. Any leaks in the exhaust pipe between the exhaust valves and
the oxygen sensor would allow dilution air into the exhaust stream. The extra oxygen from the
dilution air would cause the oxygen sensor to signal a need to run at a air-fuel ratio that is richer
than optimal. If an exhaust leak occurs between the oxygen sensor and the catalyst, the engine
will run at the correct air-fuel ratio, but the extra oxygen would affect catalyst conversion
efficiencies. As evidenced by the test engines, manufacturers can select materials with sufficient
quality to prevent exhaust leaks over the useful life of the engine.
4-27
-------
Draft Regulatory Support Document
4.2.2.7 Emission standards
4.2.2.7.1 Technology Basis
Three-way catalyst systems with electronic, closed-loop fuel systems have a great potential
to reduce emissions from Large SI engines. We believe these technologies are capable of the
greatest degree of emission reduction achievable from these engines in the projected time frame,
considering the various statutory factors. In particular, we are not basing the emission standards
on the emission-control capability from any of the following technologies.
- Spark timing
- Combustion-chamber redesign
- Gaseous fuel injection
- Exhaust gas recirculation
Incorporating these technologies with new engines could further reduce emissions; however,
Large SI engine manufacturers typically produce 10,000 to 15,000 units annually, which limits
the resources available for an extensive development program. Considering the limited
development budgets for improving these engines, we believe it is more important to make a
robust design with basic emission-control hardware than to achieve very low emission levels
with complex hardware at a small number of steady-state test modes. Even without these
additional technologies, we anticipate that manufacturers will be able to reduce emissions by
about 90 percent from uncontrolled levels. Further optimizing an engine with a full set of
emission-control hardware while meeting transient and field-testing emission standards is more
of a burden than Large SI manufacturers can bear in the projected time frame. Manufacturers
producing new engines may find it best to use some of these supplemental technologies to
achieve the desired level of emission control and performance at an acceptable cost.
4.2.2.7.2 Duty-cycle emission standards
Given the control technology, as described above, there is a need to select emission
standards that balance the tradeoff between NOx and CO emissions. Both NOx and CO vary
with changing air-fuel ratios, but in an inverse relationship. This is especially important
considering the degree to which these engines are used in enclosed areas.
Commenters representing states and environmental groups stressed the need to control
HC+NOx emissions to address concerns for meeting ambient air quality standards for ozone.
We are accordingly setting an HC+NOx emission standard of 2.0 g/hp-hr (2.7 g/kW-hr), which is
somewhat more stringent than the proposed standard. We are adopting a slightly higher CO
emission standard than proposed, which reflects the tradeoff between NOx and CO emissions.
Further, we are adopting provisions that will encourage manufacturers to reduce HC+NOx even
further by allowing higher CO levels where a manufacturer certifies to lower HC+NOx levels.
Under this approach, customers desiring to protect workers or others in close proximity to the
engines can choose engine models that offer the maximum control of CO emissions. Conversely,
4-28
-------
Chapter 4: Feasibility of Proposed Standards
if individual exposure to CO emissions is less of a concern, manufacturers have a strong
incentive to maximize control of HC+NOx emissions.
Table 4.2-5 shows the range of measured emission values from the engines tested with
optimized emission controls. In general, the engines with higher CO values and lower HC+NOx
values were calibrated with slightly richer air-fuel ratios, with all other engine parameters
unchanged. The measured emission levels include a variety of duty cycles, but this doesn't seem
to affect the observed trends. Also, Table 4.2-5 notes the length of time the engine was turned
off before starting the transient duty cycle. All the data points shown are from measurements
with the aged catalysts. Several measurements with the new catalyst showed that engines were
able to achieve very low levels of both NOx and CO emissions.
Table 4.2-5
Range of Measured Emission Levels (g/hp-hr)
Engine*
GM
GM
GM
GM
GM
GM
Mazda
Mazda
Mazda
Mazda
Mazda
Ma 7 Aft
HC
0.30
0.27
0.41
0.29
0.27
0.28
0.34
0.58
0.61
0.66
0.6
051
NOx
3.82
4.14
5.91
5.89
4.42
5.33
0.88
0.15
0.19
0.14
0.35
0.7
HC+NOx
4.12
4.41
6.32
6.18
4.69
5.61
1.22
0.73
0.8
0.8
0.95
1.21
CO
0.66
0.68
0.83
0.86
0.87
0.89
4.61
6.66
6.97
7.5
7.61
776
Cycle
Backhoe-loader
Backhoe-loader
Backhoe-loader
Large SI Composite
Highway FTP
Highway FTP
Highway FTP
Large SI Composite
Large SI Composite
Large SI Composite
Large SI Composite
Welder
soak, min.
4
2
20
6
3
3
5
5
5
5
7
4
*Both engines operated on LPG for all tests.
Figure 4.2-7 shows an attempt to apply a curve-fit to the data points. Using a log-log
relationship as shown yielded an R-square value of 0.93, indicating a relatively good fit to the
data. Table 4.2-6 and Figure 4.2-8 show the curve relating CO and HC+NOx emission levels
using the mathematical relationship. This involves starting with a set of HC+NOx emission
levels, then calculating the corresponding CO emission levels.1 Finally, both CO and HC+NOx
emission levels are increased by 10 percent to account for a compliance margin around the
measured data points. These standards apply to all steady-state and transient duty-cycle testing
for certification, production-line, and in-use testing.
'While somewhat roundabout mathematically, solving for CO values from the logarithmic
equation is most easily done by converting the curve-fit to an equation based on the natural log
function. Using logarithm relationships yields the equivalent relationship (in metric units):
(HC+NOx) x CO0784 = 8.57 or CO = (8.57 - (HC+NOx))1276.
4-29
-------
Draft Regulatory Support Document
Figure 4.2-7
1,
0 8
n 06
~z.
C ) 04-
o 02
0_
.n 9
Log HC+NOx vs Log CO
~vo
* °
Log (HC+NOx) = -0.78 Log CO + 0.63
R-square = 0.93
o o
v
0°°
-\>'f~ i 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-0.4 -0.2 0 0.2 0.4 0.6 0.8
Log CO
1
Table 4.2-6
Sample Standards Using the
Optional Duty-cycle Standards(g/kW-hr)
HC+NOx
2.70
2.20
1.70
1.30
1.00
0.80
CO
4.4
5.6
7.9
11.1
15.5
20.6
4-30
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.2-8
35
30
JE 25
| 20
°? 15
8 10
5
0
CO vs. NOx+HC
Duty-cycle standards
0.5
1.5 2 2.5
HC+NOx, g/kW-hr
3.5
We generally set standards by focusing on attaining ambient air quality in broad outdoor
areas. The HC+NOx standard of 2.7 g/kW-hr is consistent with this focus and achieves
significant reductions in ozone precursor emissions. Moreover, any of the emission levels shown
in Table 4.2-6 provide large reductions in CO, NO, and NO2 to address any concerns for
individual exposures.
4.2.2.7.3 Engine protection
The table of standards above does not take into account the fact that some engines are
unable to maintain sustained stoichiometric operation at high engine loads. Engines running rich
at high load typically continue to have low HC+NOx emissions, but CO emissions increase
substantially. However, operation over the transient duty cycle involves very little sustained
high-load operation. Table 4.2-7 shows the total time during the 20-minute cycle with engine
loads exceeding various thresholds. This alone shows that the standard for testing over the
transient duty cycle needs little or no adjustment to account for rich operation under high-load
conditions. Delaying rich operation would further ensure that emission-controls continue to
function properly while still protecting against overheating. As a result, we don't believe that
4-31
-------
Draft Regulatory Support Document
emission standards for the transient emission test should be adjusted to account for engine-
protection strategies.
Table 4.2-7
Evaluation of High-Load Operation Over the Transient Duty Cycle
Torque threshold
(percent of maximum
at a given speed)
90%
85%
80%
75%
Total time over torque
threshold (seconds)
16
23
41
67
Percent of
20-minute cycle
1.3
1.9
3.4
5.6
Average number of
seconds during each minute
0.8
1.2
2.0
3.4
The steady-state duty cycles, however, have a fixed weighting to account for emission
levels at high load operation. Also, delaying enrichment does not help with steady-state
emissions, because emissions are measured only after engine operation and emission levels have
stabilized. We are therefore setting a maximum CO level of 31 g/kW-hr during steady-state
testing for engines needing protection strategies. This corresponds to the highest CO emission
level we are allowing under field-testing standards, as noted in Table 1 and described further
below. This less stringent standard would apply to all steady-state testing with the C2 or D2 duty
cycles for certification, production-line, or in-use testing. The emission standards described in
Table 1 would still apply to these engines when tested over the transient duty-cycle. We are also
applying the field-testing standards equally to different engines, regardless of whether or not they
are certifying to a less stringent CO emission standard for steady-state testing. This reflects our
expectation that engines undergoing normal operation in the field will continue to meet emission
standards.
Ford submitted test data with their gasoline engine showing that their emission levels
comply with this less stringent CO standard for steady-state testing. For example, with a
measured emission level of 23.9 g/kW-hr, they would have roughly a 20-percent compliance
margin relative to a standard of 31 g/kW-hr. The proposed curve of candidate emission
standards incorporated a 10-percent compliance margin, even though the measured emissions
were from aged engines not designed to meet emission standards. Our emission modeling
typically incorporates an assumed 20-percent compliance margin for spark-ignition engine
emissions.
In addition, as described in the preamble to the final rule, we are adopting a combination
of provisions to ensure that manufacturers will take steps to allow enrichment only under
exceptional circumstances. This is necessary to ensure that engines in nonroad equipment don't
operate substantially under engine-protection regimes leading to compromised control of
4-32
-------
Chapter 4: Feasibility of Proposed Standards
emissions.
4.2.2.7.4 Field-testing emission standards
Manufacturers may do testing under the in-use testing program using field-testing
procedures. This has the potential to substantially reduce the cost of testing. Setting an emission
standard for testing engines in the field requires that we take into account all the variability
inherent in testing outside the laboratory. As discussed further below, this includes varying
engine operation, and a wider range of ambient conditions, and the potential for less accurate or
less precise emission measurements and calculations. Also, while the field-testing standards and
procedures are designed for testing engines installed in equipment, engines can also be tested on
a dynamometer to simulate what would happen in the field. In this case, extra precautionary
steps would be necessary to ensure that the dynamometer testing could be characterized as
"normal operation." Also, the less stringent field-testing standards would apply to any simulated
field-testing on a dynamometer to take emission-measurement variability into account, as
described below.
The SwRI test engines also show that Large SI engines are capable of controlling
emissions under the wide range of operation covered by the field-testing provisions. A modest
amount of additional development will be necessary to address isolated high-emission points
uncovered by the testing. We believe that manufacturers will be able to reduce emissions as
needed to meet the 2007 emission standards by spending time improving the precision of their
engine calibrations, perhaps upgrading to more sophisticated control software to achieve this.
Field testing may also include operation at a wider range of ambient conditions than for
certification testing. Selecting emission standards for field testing that correspond with the duty-
cycle standards requires consideration of the following factors:
The data presented above show that emissions vary for different modes of engine
operation. Manufacturers will need to spend time addressing high-emission
points to ensure that engines are not overly sensitive to operation at certain speeds
or loads. The data suggest that spark-ignition engines can be calibrated to
improve control at the points with the highest emission rates.
Established correction factors allow for adjustment to account for varying ambient
conditions. Allowing adjustment of up to 10 percent adequately covers any
potential increase in emissions resulting from extreme conditions.
While emission measurements with field-testing equipment allow more flexibility
in testing, they are not as precise or as accurate as in the laboratory; the
regulations define specifications to limit the error in emission measurements. For
most mass-flow and gas analyzer hardware, these tolerance remain quite small.
Measurements and calculations for torque values introduce a greater potential for
error in determining brake-specific emission levels. The tolerance for onboard
torque readings allows for a 15-percent error in understating torque values, which
would translate into a 15-percent error in overstating brake-specific emissions.
4-33
-------
Draft Regulatory Support Document
Taking all these factors into account, we believe it is appropriate to allow for a 40-percent
increase in HC+NOx emissions relative to the SwRI measured values to account for the factors
listed above. CO emissions are generally somewhat more sensitive to varying engine operation,
so a 50-percent adjustment is appropriate for CO. The approach for field-testing standard
follows the format described for duty-cycle testing. This results in an HC+NOx standard of 3.8
g/kW-hr (2.8 g/hp-hr), with scaled values for the CO standard, as shown in Table 4.2-8 and
Figure 4.2-9.
These same numerical field-testing standards apply to natural gas engines. Much like for
certification, we are excluding methane measurements from natural gas engines. Since there are
currently no portable devices to measure methane (and therefore nonmethane hydrocarbons), the
3.8 g/kW-hr field-testing standard and the values in Table 4.2-8 apply only to NOx emissions for
natural gas engines.
Table 4.2-8
Sample Standards Using the
Optional Field-testing Standards(g/kW-hr)
HC+NOx
3.80
3.10
2.40
1.80
1.40
1.10
CO
6.5
8.5
11.7
16.8
23.1
31.0
4-34
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.2-9
35
30
j= 25
| 20
°! 15
8 10
5
0
CO vs. NOx+HC
Field-testing standards
0.5
1.5 2 2.5
HC+NOx, g/kW-hr
3.5
4.2.2.7.5 Evaporative emissions
Several manufacturers are currently producing products with pressurized fuel tanks to
comply with Underwriters Laboratories specifications. Most fuel tanks in industrial applications
are made of a thick-gauge sheet metal or structural steel, so increasing fuel pressures within the
anticipated limits poses no risk of bursting or collapsing tanks. For those few applications that
use plastic fuel tanks, equipment manufacturers already use or could easily use blow-molded
tanks that are also able to withstand substantial pressure buildup. If an exceptional application
relies on a fuel tank that must keep internal pressures near ambient levels, a volume-
compensating bag would allow for adequate suppression of fuel vapors with minimal pressure
buildup."
Testing with pressurized fuel tanks shows emission data related to sealing fuel tanks. The
tests included 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
""New Evaporative Control System for Gasoline Tanks," EPA Memorandum from
Charles Moulis to Glenn Passavant, March 1, 2001, Docket A-2000-01, document II-B-16.
4-35
-------
Draft Regulatory Support Document
adjustments to the spring tension in the pressure relief valve. We performed these tests on an
aluminum fuel tank to remove the variable of permeation. As shown in Figure 4.2-10, there was
a fairly linear relationship between the pressure setting of the valve and the emissions measured
over the proposed test procedure, which we would expect based on the theoretical relationships.
At 3.5 psi, this relationship extrapolates to a value of 0.2 g/gallon/day.
Figure 4.2-10: Effect of Pressure Cap on Diurnal Emissions
1 KC\ -,
\ .OU
<
1 40
•^ 1 on
^» I .Ł.\J
re
"n -i nn
•i; 1 .UU
S» n on
_O) U.oU
S DfiO
^^ U.OU
T n AC\
^t. U.'fU
Don
.ZU
Onn
^^^*^^*
4^^^^^^
^"^^*^^^^
""^^^^^^
y = -0.41 53x+ 1.5 4 ' —
R2 = 0.924
.UU i i i i
0 0.5 1 1.5 2 2.5
pressure relief setting [psi]
Ą.2.2.7.6 Conclusion
Manufacturers have been developing emission-control technologies to meet the 2004
emission standards since October 1998, when California ARB adopted the same standards. We
expect that manufacturers will add three-way catalysts to their engines and use electronic closed-
loop fueling systems. These technologies have been available for industrial engines for many
years.
The SwRI testing program was based on aged engines and involved no effort to fine-tune
air-fuel ratios or emission levels across the engine map. We expect that manufacturers will be
able to control emission levels more broadly across the range of engine speeds and loads by
improving control of air-fuel ratios at different operating modes. These improvements will
reduce both steady-state and transient emission levels. The 2007 emission standards are based
directly on the data presented above. The test results therefore show that these Large SI engines
are capable of meeting the 2007 emission standards for both steady-state and transient duty
cycles. Similarly, the data presented above show how off-cycle emissions vary for engines that
have been designed for effective control of air-fuel ratios across the range of normal operation.
Here too, the test engines generally had emission levels consistent with the 2007 field-testing
4-36
-------
Chapter 4: Feasibility of Proposed Standards
standards, with certain limited exceptions as noted above.
The SwRI testing program involved about eight weeks of development effort to
characterize and modify two engines to for optimized emissions on the steady-state and transient
duty cycles, and for all kinds of off-cycle operation. Both of the test engines had logged several
thousand hours of operation using off-the-shelf technologies that have been available for nonroad
engines for many years. Several hardware and software adjustments were made to maintain
optimal air-fuel ratios for effective control of all pollutants under all operating modes. Some
further development effort will be necessary to address the few isolated modes with high
emission levels, as described earlier in this section. Manufacturers may save development time
by upgrading to the modestly more expensive controller with independent air-fuel control
capability in different speed-load zones. This would achieve the same result, but would
potentially reduce the cost of meeting the standards by reducing engineering time. We believe
that the several years until 2007 allow enough lead time for manufacturers to carry out this
development effort for all their engines.
We expect the SwRI testing program to provide extensive, basic information on
optimizing the subject engines for low emissions, so manufacturers will need significantly less
time and testing resources to modify additional engine models. For example, the SwRI testing
shows how emissions change over varying speeds and loads; as a result, future testing can focus
on far fewer test points to characterize a calibration. The test results also show how
manufacturers will need to balance calibrations for controlling emissions of different pollutants
across the range of engine speeds and loads.
The emission standards for Large SI engines are significantly more stringent than those
we are adopting for recreational vehicles and those we have already adopted for lawn and garden
engines. We believe this is appropriate, for several reasons. First, the similarity to automotive
engines makes it possible to use basic automotive technology that has already been adapted to
industrial use. Second the cost of Large SI equipment is typically much higher than the
recreational or other light-duty products, so there is more capability for manufacturers to pass
along cost increases in the marketplace. Third, the Large SI emission standards correspond with
a substantial fuel savings, which offset the cost of regulation and provide a great value to the
many commercial customers.
4.2.3 Impacts on Noise, Energy, and Safety
The Clean Air Act directs us to consider potential impacts on noise, energy, and safety
when establishing the feasibility of emission standards for nonroad engines.
As automotive technology demonstrates, achieving low emissions from spark-ignition
engines can correspond with greatly reduced noise levels. Electronically controlled fuel systems
are able to improve management the combustion event, and catalysts can be incorporated into
existing equipment designs without compromising the muffling capabilities in the exhaust.
4-37
-------
Draft Regulatory Support Document
Adopting new technologies for controlling fuel metering and air-fuel mixing will lead to
substantial improvements in fuel consumption rates. We project fuel consumption improvements
that will reduce total nationwide fuel consumption by about 300 million gallons annually once
the program is fully phased in. While a small number of engines already have these
technologies, it seems that the industrial engine marketplace has generally not valued fuel
economy highly enough to create sufficient demand for these technologies.
We believe the technology discussed here will have no negative impacts on safety.
Electronic fuel injection is almost universally used in cars and trucks in the United States with
very reliable performance. In addition, we expect cases of CO poisoning from these engines to
decrease as a result of the reduced emission levels.
4-38
-------
Chapter 4: Feasibility of Proposed Standards
4.3 Snowmobile Engines
The following paragraphs summarize the data and rationale supporting the emission
standards for snowmobiles, which are listed in the Executive Summary.
4.3.1 Baseline Technology and Emissions
Snowmobiles are equipped with relatively small high-performance two-stroke two and
three cylinder engines that are either air- or liquid-cooled. The main emphasis of engine design
is on performance, durability, and cost. Because these engines are currently unregulated, they
have no emission controls. The fuel system used on these engines are almost exclusively
carburetors, although a small number have electronic fuel injection. Two-stroke engines
lubricate the piston and crankshaft by mixing oil with the air and fuel mixture. This is
accomplished by most contemporary 2-stroke engines with a pump that sends two-cycle oil from
a separate oil reserve to the carburetor where it is mixed with the air and fuel mixture. Some less
expensive two-stroke engines require that the oil be mixed with the gasoline in the fuel tank. In
fact, because performance and durability are such important qualities for snowmobile engines,
they all operate with a "rich" air and fuel mixture. That is, they operate with excess fuel, which
enhances performance and allows engine cooling which promotes longer lasting engine life.
However, rich operation results in high levels of HC, CO, and PM emissions. Also, two-stroke
engines tend to have high scavenging losses, where up to a third of the unburned air and fuel
mixture goes out of the exhaust resulting in high levels of raw HC.
We developed average baseline emission rates for snowmobiles based on the results of
emissions testing of 23 snowmobiles.23 Current average snowmobile emissions rates are 397
g/kW-hr (296 g/hp-hr) CO and 149 g/kW-hr (111 g/hp-hr) HC.
4.3.2 Potentially Available Snowmobile Technologies
A variety of technologies are currently available or in stages of development to be
available for use on 2-stroke snowmobiles. These include engine modifications, improvements
to carburetion (improved fuel control and atomization, as well as improved production
tolerances), enleanment strategies for both carbureted and fuel injected engines, pulse air, and
semi-direct and direct fuel injection. In addition to these 2-stroke technologies, it is also feasible
to convert from using 2-stroke engines to 4-stroke engines. Each of these is discussed in the
following sections.
4.3.2.1 Engine Modifications
There are a variety of engine modifications that could reduce emissions from two-stroke
engines. The modifications generally either increase trapping efficiency (i.e., reduce fuel short-
circuiting) or improve combustion efficiency. Those modifications that increase trapping
efficiency include optimizing the intake, scavenge and exhaust port shape and size, and port
4-39
-------
Draft Regulatory Support Document
placement, as well as optimizing port exhaust tuning and bore/stroke ratios. Optimized
combustion charge swirl, squish, and tumble improve the combustion of the intake charge.
Various snowmobile manufacturers have told us that they believe these modifications have the
potential to reduce emissions by up to 40 percent, depending on how well the unmodified engine
is optimized for these changesv.
4.3.2.2 Carburetion Improvements
There are several things that can be done to improve carburetion in snowmobile engines.
First, strategies to improve fuel atomization promote more complete combustion of the fuel/air
mixture. Additionally, production tolerances can be improved for more consistent fuel metering.
Both of these allow for more accurate control of the air/fuel ratio. In conjunction with these
improvements in carburetion, the air/fuel ratio can be leaned out somewhat. Snowmobile
engines are currently calibrated with rich air/fuel ratios for durability reasons. Manufacturers
have stated that based on their experience, leaner calibrations can reduce CO and HC emissions
by up to 20 percent, depending on how lean the unmodified engine is prior to recalibrationw.
Small improvements in fuel economy can also be expected with recalibration.
The calibration changes just discussed (as well as some of the engine modifications
previously discussed) also reduce snowmobile engine durability, though many possible engine
improvements could regain any lost durability that occurs with leaner calibrations. These include
changes to the cylinder head, pistons, ports and pipes to reduce knock. In addition, critical
engine components can be made more robust to improve durability.
The same calibration changes to the air/fuel ratio just discussed for carbureted engines
can also be employed, possibly with more accuracy, with fuel injection. At least one major
snowmobile manufacturer currently employs electronic fuel injection on several of its
snowmobile models.
4.3.2.3 Pulse Air
Pulse air injection into the exhaust stream mixes oxygen with the high temperature HC
and CO in the exhaust. The added oxygen allows the further combustion of these exhaust
constituents between the combustion chamber and tailpipe exhaust. Our testing of pulse air on
four-stroke ATV engines indicated that reductions of 30-70% for HC and 30-80% for CO are
possible. We believe similar reductions could be expected for engines used in snowmobile
applications. We expect some modest reductions in two-stroke applications as well.
v See "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
wSee "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
4-40
-------
Chapter 4: Feasibility of Proposed Standards
4.3.2.4 Direct and Semi-direct Fuel Injection
In addition to rich air/fuel ratios, one of the main reasons that emissions from two-stroke
engines are high is scavenging losses, as described above. One way to reduce or eliminate such
losses is to inject the fuel into the cylinder after the exhaust port has closed. This can be done by
injecting the fuel into the cylinder through the transfer port (semi-direct injection) or directly into
the cylinder (direct injection). Both of these approaches are currently being used successfully in
two-stroke personal watercraft (PWC) engines. Bombardier has developed a semi-direct
injection engine for snowmobiles that will be available in several different models for the 2003
model year. Manufacturers have indicated to us that two-stroke engines equipped with direct
fuel injection systems could reduce HC emissions by 70 to 75 percent and reduce CO emissions
by 50 to 70 percent. Certification results for 2002 model year PWC support the manufacturers
projections, as shown in Table 4.3-1. This table shows the paired certification data from some
PWC engines in both uncontrolled and direct injection configurations. The percent difference in
FEL column refers to the HC + NOx FEL. This is a pretty good surrogate for HC since most of
the HC + NOx level is made up of HC, as can be seen from the table.
Table 4.3-1
Certification Levels of Direct Injection vs. Uncontrolled Engines
Mfr
Kawasaki
Polaris
Bombardier
Polaris
%
difference
in FEL
67%
72%
73%
65%
size
(liter)
1.071
1.071
0.78
0.70
0.9514
0.9513
1.16
1.16
power
(kW)
95.6
88.3
Not
Reported
Not
Reported
88.9
89.5
85.26
93.25
FEL
(HC +
NOx)
46.0
140.0
47.1
165
36.8
137.8
46.3
134.0
HCcert
level
38.4
136.76
33.2
158.8
24.5
136.7
37.46
130.8
CO cert
level
103.1
241.8
135.2
217.0
100.1
330.6
100.4
359.3
Technology
Direct injection,
electronic control
Carburetor
Direct injection
Carburetor
Direct injection,
electronic control
Carburetor
Direct injection
Carburetor
Substantial improvements in fuel economy could also be expected with these
technologies. We believe these technologies hold promise for application to snowmobiles.
four of the major snowmobile manufacturers have indicated that they consider direct fuel
injection as a viable technology for controlling emissions and are currently either analyzing
various direct injection systems or are in the process of developing their own system.
All
4-41
-------
Draft Regulatory Support Document
Manufacturers must address a variety of technical design issues for adapting the technology to
snowmobile operation, such as operating in colder ambient temperatures and at variable altitude.
Manufacturers have also stated that the direct injection systems used in many of their PWC
cannot simply be placed into their snowmobiles because of inherent differences in snowmobile
and PWC engines. Primarily the fact that PWC engines operate at considerably lower engine
speeds than snowmobile engines. PWC engines typically operate at maximum engine speeds of
6,000 rpm, compared to engine speeds of almost double that for snowmobiles. This poses a
problem because some of the current direct injection designs can't properly operate at such high
engine speeds. While these are all legitimate concerns, we believe that this technology can be
adapted without significant problems. Bombardier's use of direct fuel injection in several
snowmobile models in the 2003 model year demonstrates that these issues have been resolved
enough for Bombardier to be comfortable selling snowmobiles with such engines. However,
direct fuel injection is a complex technology and there are several different types of approaches
to designing these systems and not all manufacturers have the same access to the various
systems. Therefore, it appears important to provide manufacturers with sufficient lead time to
resolve all of the potential issues with direct injection so that it can be widely available for all
snowmobile models, instead of a few niches models for a select manufacturer or two. That is
why we believe it is appropriate to give manufacturers until 2012. This will give manufacturers
sufficient time to incorporate these development efforts into their overall research plan and apply
these technologies to a substantial percentage of their snowmobiles.
4.3.2.5 Four-Stroke Engines
In addition to the two-stroke technologies just discussed, the use of four-stroke engines in
snowmobiles is feasible. Four-stroke engines have been used in numerous recreational vehicle
applications for years. Four-stroke engines have also been used in limited numbers over the
years in snowmobiles. In 1999, Arctic Cat released a four-stroke touring sled. Polaris followed
two years later with their four-stroke touring sled in 2001. Table 4.3-2 provides emission results
from a 2001 Arctic Cat four-stroke touring sled and a 2001 Polaris Frontier (four-stroke), both
owned and tested by the National Park Service (NFS) at Southwest Research Institute. Table
4.3-3 presents certification data from four 2002 PWC's equipped with four-stroke engines. The
engines in these PWC are higher output engines than the Arctic Cat and Polaris snowmobile
four-stroke engines and have emission results very similar to that which a high-output four-stroke
snowmobile engine could expect to emit.
Table 4.3-2
Four-Stroke Snowmobile Emissions
Manufacturer
Arctic Cat
Polaris
Model
4-Stroke
Touring
Frontier
Engine
Displacement
660 cc
784 cc
HC
(g/kW-hr)
6.2
3.2
CO
(g/kW-hr)
79.9
79.1
NOx
(g/kW-hr)
15.0
7.0
4-42
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.3-3
Four-Stroke PWC Certification Emission results
Manufacturer
Honda
Honda
Bombardier
Yamaha
Model
AquaTraxF-12
Aqua Trax F-
12X
GTX 4-TEC
FX140
Engine
Displacement
1,244 cc
1,244 cc
1,504 cc
998 cc
HC
(g/kW-hr)
11.2
10.7
9.6
16.6
CO
(g/kW-hr)
266.0
235.3
161.7
255.1
NOx
(g/kW-hr)
3.8
4.6
5.0
5.9
Much has changed in the time since we published our proposed standards. In October
2001, when we published our proposed standards for snowmobiles, there was only one
manufacturer that had introduced a four-stroke snowmobile (the Polaris Frontier was released
soon after). Today, all four of the major snowmobile manufacturers have developed a four-
stroke engine for snowmobiles. In fact, the 2003 model year will see four-stroke engines in
several models from all four manufacturers. The models will range from touring sleds to sport,
mountain, and high-performance models. Since four-stroke engines do not rely on scavenging of
the exhaust gases with the incoming air/fuel mixture, they have inherently lower HC emissions
compared to two-strokes (up to 90 percent lower). Four-stroke engines can also have reductions
in CO emissions, depending on the power output of the engines and the engine calibration. A
smaller four-stroke engine calibrated to operate at or near stoichiometry could reduce CO
emissions significantly. This is demonstrated above in Table 4.3-2, since both of these
snowmobiles use four-stroke engines equipped with closed-loop control EFI systems which try to
maintain the air and fuel mixture at or near stoichiometry. A larger four-stroke engine calibrated
for maximum power could generate CO emission levels closer to a comparably powered two-
stroke engine. Table 4.3-3 above, demonstrates this. Although the engines in this table are from
PWCs, they are high-output four-stroke engines producing horsepower in excess of 100 hp, that
are very similar to what could be expected to be used in a high-performance snowmobile. The
CO emissions from the four PWC engines are considerably higher than the CO levels from the
two lower powered four-stroke snowmobiles. Four-stroke engines have a lower power density
compared to two-stroke engines. Two-stroke engines have a power stroke every other stroke
compared to a power stroke every fourth stroke for a four-stroke engine. Thus, a comparably
powered four-stroke engine requires almost a third more engine displacement, to equal the power
of a two-stroke engine. The impact this has on snowmobile applications is that a four-stroke
engine is already heavier than a two-stroke engine because of the valve-train system. In order to
have comparable power output with a two-stroke, a four-stroke engine needs to have a larger
4-43
-------
Draft Regulatory Support Document
displacement. This is achieved through an increase in the cylinder bore and/or stroke or by
adding more cylinders, which all have the potential effect of adding even more weight. Thus, for
a four-stroke to be competitive with a two-stroke engine, manufacturers need to find a way to
reduce weight in the engine and elsewhere in the snowmobile. This could entail the use of
lighter materials in the engine and chassis or reducing the size of the fuel tank to take advantage
of the superior fuel efficiency of the four-stroke engine while maintaining the same cruising
time/range.
Another way to increase the output from a four-stroke engine is to use a turbocharger or
supercharger. Both of these devices act as air compressors, providing increased air density in the
engines's combustion chambers, which allows more efficient burning of air and fuel and results
in higher horsepower output. A turbocharger uses exhaust gases to compress air, while a
supercharger is mechanically driven using a belt between the supercharger and typically the
camshaft. Honda is currently selling a turbocharged version of their four-stroke personal
watercraft. A turbocharger or supercharger could provide an increase in power without having to
increase the engine displacement. Regardless of the strategy used, it is apparent that four-stroke
engines will have a larger role in snowmobile applications than originally thought.
However, it is important to provide sufficient lead time for the development and
implementation of some four-stroke engines in snowmobiles, similar to the concern with direct
fuel injection. For example, in the case of the Yamaha four-stroke snowmobile, a considerable
amount of effort and resources went into designing a new snowmobile from the ground up
specifically to accommodate the size, weight and power characteristics of a four-stroke engine.
A completely new chassis was designed which allowed the somewhat heavier engine to be placed
lower and further back than is typical for two-stroke snowmobiles. This was necessary to
maintain the kind of handling characteristics required of a high performance snowmobile. While
a stock four-stroke engine can be placed into an existing snowmobile model and made to work
acceptably, as can be seen in the Polaris and Arctic Cat four-stroke offerings, such designs are
only practical for lower powered touring snowmobiles. Since the vast majority of the
snowmobile market is in higher performance sleds, we believe that the conversion of all
snowmobiles to four-strokes would require that many current snowmobile chassis be replaced
with new models designed from the ground up. This could be a substantial undertaking for the
snowmobile industry given the number of models it offers and niche markets it currently serves.
That is why we believe the delay of our proposed Phase 2 standards by two years will give
manufacturers time to incorporate these development efforts into their overall research plan as
they apply these technologies to their snowmobiles.
4.3.3 Test and Measurement Issues
4.3.3.1 Test procedure
We are generally adopting the snowmobile test procedure developed by Southwest
Research Institute in cooperation with the International Snowmobile Manufacturers Association
for all snowmobile emissions testing.24 This test procedure consists of two main parts; the duty
4-44
-------
Chapter 4: Feasibility of Proposed Standards
cycle that the snowmobile engine operates over during testing and other testing protocols
involving the measurement of emissions (sampling and analytical equipment, specification of test
fuel, atmospheric conditions for testing, etc.). While the snowmobile duty cycle was developed
specifically to reflect snowmobile operation, many of the testing protocols are well established in
other EPA emissions programs and have been simply adapted where appropriate for
snowmobiles.
The snowmobile duty cycle was developed by instrumenting several snowmobiles and
operating them in the field in a variety of typical riding styles, including aggressive (trail),
moderate (trail), double (trail with operator and one passenger), freestyle (off-trail), and lake
driving. A statistical analysis of the collected data produced the five mode steady-state test cycle
shown in Table 4.3-4. The snowmobiles used to generate this data were not derived from
members of the general public found openly operating in these riding styles, but were
snowmobiles operated by contractor personnel in staged set-ups of these riding styles. This duty
cycle was used to generate the baseline emissions levels for snowmobiles, and we believe it is the
most appropriate cycle for demonstrating reductions in snowmobile emissions at this time.
Table 4.3-4
Snowmobile Engine Test Cycle
Mode
Normalized
Speed
Normalized
Torque
Relative
Weighting
1
1
1
12
2
0.85
0.51
27
3
0.75
0.33
25
4
0.65
0.19
31
5
Idle
0
5
The other testing protocols are largely derived from our regulations for marine outboard
and personal watercraft engines.25 The testing equipment and procedures from that regulation are
largely appropriate for snowmobiles. However, unlike snowmobiles, outboard and personal
watercraft engines tend to operate in fairly warm ambient temperatures. Thus, some provision
needs to be made in the snowmobile test procedure to account for the colder ambient
temperatures typical of snowmobile operation. Since snowmobile carburetors are jetted for
specific ambient temperatures and pressures, we could take one of two general approaches. The
first is to require testing at ambient temperatures typical of snowmobile operation, with
appropriate jetting. A variation of this option is to simply require that the engine inlet air
temperature be representative of typical snowmobile operation, without requiring that the entire
test cell be at that temperature. The second is to allow testing at higher temperatures than
typically experienced during snowmobile operation, with jetting appropriate to the warmer
ambient temperatures.
4-45
-------
Draft Regulatory Support Document
Manufacturers shared confidential emission data with us that indicated that there was no
difference between testing snowmobiles with cold inlet air and testing at higher temperatures
with carburetor jetting adjusted for the warmer temperature. We also did some limited testing
which substantiates the manufacturer's claim. Some manufacturers argued that even though
there was no difference between the test methods, we should still require testing with cold inlet
air because it would be more representative. Other manufacturers felt that the increased cost of
cold inlet air testing made this approach undesirable. We decided that since there was ample
evidence that two approaches would produce similar results with the technologies we expect to
be used and that it did not make sense to require manufacturers to incur the cost of cold inlet air
testing if it wouldn't provide any additional benefit. Therefore, we are allowing manufacturers to
test at warmer (i.e., typical test cell temperature 68°F-86°F) with carburetor jetting set to the
appropriate temperature.
4.3.3.2 HC is a Good Proxy for Fine PM Emissions
We believe the best way to regulate fine PM emissions from current snowmobile engines
is to set standards based on HC emissions. Unlike other recreational vehicles, the current fleet of
snowmobiles consists almost exclusively of two-stroke engines. Two-stroke engines inject
lubricating oil into the air intake system where it is combusted with the air and fuel mixture in
the combustion chamber. This is done to provide lubrication to the piston and crankshaft, since
the crankcase is used as part of the fuel delivery system and cannot be used as a sump for oil
storage as in four-stroke engines. As a result, in addition to products of incomplete combustion,
two-stroke engines also emit a mixture of uncombusted fuel and lubricant oil. HC-related
emissions from snowmobiles increase PM concentrations in two ways. Snowmobile engines
emit HCs directly as particles (e.g., droplets of lubricant oil). Snowmobile engines also emit HC
gases, as well as raw unburned HCs from the fuel which either condense in cold temperatures to
particles or react chemically to transform into particles as they move in the atmosphere. As
discussed above, fine particles can cause a variety of adverse health and welfare effects,
including visibility impairment.
We believe HC measurements will serve as a reasonable surrogate for fine PM
measurement for snowmobiles for several reasons. First, emissions of PM and HC from these
engines are related. Test data show that over 70 percent of the average volatile organic fraction
of PM from a typical 2-stroke snowmobile engine is organic hydrocarbons, largely from
lubricating oil components." The HC measurements (which use a 191 Celsius/375.8 degree
xMemo to Docket, Mike Samulski. "Hydrocarbon Measurements as an Indicator for
Particulate Matter Emissions in Snowmobiles," September 6, 2002, Docket A-2000-01;
Document IV-B.
Carroll, IN, JJ White, IA Khalek, NY Kado. Characterization of Snowmobile Particulate
Emissions. Society of Automotive Engineers Technical Paper Series. Particle Size Distribution
in the Exhaust of Diesel and Gasoline. SP-1552, 2000-01-2003. June 19-22, 2000.
4-46
-------
Chapter 4: Feasibility of Proposed Standards
Fahrenheit heated FID) would capture the volatile component which in ambient temperatures
would be particles (as droplets).
Second, many of the technologies that will be employed to reduce HC emissions are
expected to reduce PM (e.g., 4-stroke engines, pulse air, and direct fuel injection techniques).
The organic emissions are a mixture of fuel and oil, and reductions in the organic emissions will
likely yield both HC and PM reductions. For example, the HC emission factor for a typical 2-
stroke snowmobile is 111 g/hp-hr. The HC emission factor for a direct fuel injection engine is
21.8, and for a 4-stroke is 7.8 g/hp-hr, representing a 80 percent and 99 percent reduction,
respectively. Similarly, the PM emission factor for a typical 2-stroke snowmobile is 2.7 g/hp-hr.
The corresponding PM emission factor for a direct fuel injection engine is 0.57, and for a 4-
stroke is 0.15 g/hp-hr, representing a 75 percent and 93 percent reduction, respectively. HC
measurements would capture the reduction from both the gas and particle (at ambient
temperature) phases.
Thus, manufacturers will generally reduce PM emissions as a result of reducing HC
emissions, making separate PM standards less necessary. Moreover, PM standards would only
cover the PM directly emitted at the tailpipe. It would not measure the gaseous or semi-volatile
organic emissions which would condense or be converted into PM in the atmosphere. By
contrast HC measurements would include the gaseous HC which could condense or be converted
into PM in the atmosphere. Thus, the HC measurement would be a more comprehensive
measurement. HC standards actually will reduce secondary PM emissions that would not
necessarily be reduced by PM standards.
Finally, from an implementation point of view, PM is not routinely measured in
snowmobiles, and there is no currently established protocol for measuring PM and substantial
technical issues to overcome to create a new method. Establishing additional PM test procedures
would entail additional costs for manufacturers. HC measurements are more routinely performed
on these types of engines, and these measurements serve as a more reliable basis for setting a
numeric standard. Thus, we believe that regulation of HC is the best way to reduce PM
emissions from current snowmobile engines.
We included a NOx standard for snowmobiles as part of the long-term program. NOx
emissions from current snowmobiles are very small, especially compared to HC. This standard
will essentially cap NOx emissions from these engines to prevent backsliding in advanced
technology engines. We are not promulgating standards that would require substantial reductions
in NOx because we believe that non-aftertreatment based standards which force substantial NOx
reductions could put upward pressure on HC emissions and would not necessarily lead to
reductions in ambient PM. Given the overwhelming level of HC, CO and PM compared to NOx,
and the secondary PM expected to result from high HC levels, it would be premature and
possibly counterproductive to promulgate NOx standards that require significant NOx reductions
from snowmobiles at this time. We have therefore decided to structure our long term HC+NOx
standard for 2012 and later model year snowmobiles to require only a cap on NOx emissions
from the advanced technology engines which will be the dominant technology in the new
4-47
-------
Draft Regulatory Support Document
snowmobiles certified at that time.
4.3.4 Impacts on Noise, Energy, and Safety
The Clean Air Act directs us to consider potential impacts on noise, energy, and safety
when establishing the feasibility of emission standards for nonroad engines.
As automotive technology demonstrates, achieving low emissions from spark-ignition
engines can correspond with greatly reduced noise levels. Four-stroke engines can have
considerably lower sound levels than two-stroke engines. Electronically controlled fuel systems
are able to improve management of the combustion event which can help lower noise levels.
Adopting new technologies for controlling fuel metering and air-fuel mixing will lead to
substantial improvements in fuel consumption rates for two-stroke engines as well as for four-
stroke engines. Four-stroke engines have far less fuel consumption than two-stroke engines.
Average mileage for a baseline two-stroke snowmobile is 12 miles per gallon (mpg). Average
mileage for a four-stroke snowmobile is 18 mpg and up to 20 mpg for a two-stroke with direct
injection. We project that these fuel consumption benefits will reduce total nationwide fuel
consumption by more than 50 million gallons annually once the program is fully phased in.
We believe the technology discussed here will have no negative impacts on safety.
Electronic fuel injection is almost universally used in cars, trucks and highway motorcycles in the
United States with very reliable performance. While the manufacturers have expressed some
concern about heavier weight and cold-starting for four stroke engines we believe these are not
significant concerns. There are already four-stroke models in production today and obviously
they are not being introduced into commerce with known safety concerns. A two-stroke
snowmobile has a fuel tank of about 12 gallons. A four-stroke could have a fuel tank of 8
gallons and maintain the same driving time/range. This would lead to a weight reaction of 25
pounds to help offset concerns about increased weight of four-stroke snowmobiles. If cold
starting of four strokes is an issue, it can be resolved with the assistance of an electronic starter or
a dry sump oil system that stores oil in a separate tank rather than in the crankcase, thus
eliminating the concern over high viscous oil adding excessive resistance to the starting process.
4.3.5 Conclusions
4.3.5.1 Phase 1 Standards
For the Phase 1 standards which start in the 2006 model year, we are allowing a phase-in
schedule that requires 50 percent of a manufacturers snowmobile fleet to meet the standards in
the 2006 model year and 100 percent to meet the standards in the 2007 model year. Snowmobile
manufacturers will have three main emission control technologies for meeting these standards:
modified two-stroke technologies (combination of engine modifications and fuel system
4-48
-------
Chapter 4: Feasibility of Proposed Standards
improvements), direct fuel injection, and four-stroke engine technology. We expect that the
Phase 1 emission standards will be met through a combination or mixture of these three emission
control strategies. All three of these strategies have been proven to be feasible and are already
available on some sleds today. Four-stroke engines and direct fuel injection technology have
already been demonstrated to be capable of achieving emission reductions well in excess of our
standards. Significant reductions are also achievable using modified two-stroke technologies.
For the 2006 model year, we expect manufacturers to rely most heavily on modifications
to existing two-stroke engines with a small amount (e.g., 10 percent) of direct injection two-
stroke engines and four-stroke engines (e.g., another 10 percent). In the context of an averaging
program, the use of direct injection technology and four-stroke engines will not only be necessary
to meet the standards, but may also allow some manufacturers to leave a small percentage of
engines unchanged, most specifically, inexpensive entry-level sleds that manufacturers have
argued are very cost sensitive. Such an approach may be necessary given the lead time and the
fairly large number of engine models to be modified and certified. Table 4.3-5 provided below
presents a potential technology mix scenario for the Phase 1 standards. The average reduction
level at the bottom of the table represents average reductions for a manufacturer's entire fleet
which already incorporates compliance margin and useful life consideration, since each engine
family FEL will have a unique compliance margin. The percent reduction presented in the table
is based on HC and CO. Obviously, a manufacturer could change the technology mix based on
cost and performance considerations.
Table 4.3-5
Potential Snowmobile Technology Mix for Phase 1 Standards
Technology
Minimal Control
Engines*
Carburetor/EFI
Recalibration + Engine
Modifications
Direct Injection
Four-Stroke
Percent Usage
20%
60%
10%
10%
Percent
Reduction
HC
0%
30%
75%
90%
Percent
Reduction
CO
0%
30%
70%
50%
Average Reduction
Fleet %
Reduction
HC
0%
18%
7.5%
9%
35%
Fleet %
Reduction
CO
0%
18%
7%
5%
30%
* Some minimal control may be required to account for deterioration and to ensure certification
FELs are met in production.
4-49
-------
Draft Regulatory Support Document
4.3.5.2 Phase 2 Standards
We are also finalizing Phase 2 standards in the 2010 model year that will serve as
transitional standards to our more stringent Phase 3 standards. As for the Phase 1 standards, we
believe manufactures will rely on a mixture of technologies, with the focus on modified two-
stroke technologies, perhaps including pulse air injection, direct fuel injection, and four-stroke
engines. We expect that to meet the 2010 standards, manufacturers will employ more of the
advanced technologies such as direct injection and four-stroke engines and less of the modified
two-stroke technologies. We anticipate manufacturers will have numerous technology mix
scenarios that they will consider. Table 4.3-6 provided below presents a potential technology
mix scenario for the Phase 2 standards. Obviously, a manufacturer could change the technology
mix based on cost and performance considerations. As for the Phase 1 standards, the use of
advanced technologies such as direct injection and four-stroke engines, in the context of our
averaging program, may allow some manufacturers to have a small percentage of engines with
minimal change. As discussed above in sections 4.3.2.4 and 4.3.2.5, we believe the biggest task
manufacturers will face in meeting our standards will be the converting of their large current fleet
of snowmobiles equipped with unregulated two-stroke engines to snowmobiles equipped with
advanced clean technologies, such as direct injection and four-stroke engines.
Table 4.3-6
Potential Snowmobile Technology Mix for 2010 Standards
Technology
Minimal Control
Engines*
Carburetor/EFI
Recalibration + Engine
Modifications + Pulse Air
Injection
Direct Injection
Four-Stroke
Percent Usage
20%
30%
35%
15%
Percent
Reduction
HC
0%
35%
75%
90%
Percent
Reduction
CO
0%
35%
70%
50%
Average Reduction
Fleet %
Reduction
HC
0%
10.5%
26%
13.5%
50%
Fleet %
Reduction
CO
0%
10.5%
24.5%
7.5%
43%
* Some minimal control may be required to account for deterioration and to ensure certification
FELs are met in production.
4-50
-------
Chapter 4: Feasibility of Proposed Standards
4.3.5.3 Phase 3 Standards
We are finalizing Phase 3 standards in the 2012 model year that we believe will require a
significant percentage of snowmobile models to be equipped with advanced technologies. As
with our Phase 1 and Phase 2 standards, we believe manufactures will rely on a mixture of
technologies, with the focus on direct fuel injection and four-stroke engines. While we expect
that to meet the 2012 standards manufacturers will employ considerably more of the advanced
technologies such as direct injection and four-stroke engines, they may still use a relatively small
amount of the modified two-stroke technologies. To provide manufacturers with additional
flexibility, we are allowing the Phase 3 standards to be met by using the following equation:
( (HC+ NOx)STD-\5'] ( COSTD\
100= 1-- — X100+ 1- — xlOO
1. 150 ) \ 400 )
Under this equation, the sum of reductions in HC+NOx and CO must equal or exceed 100
percent on a corporate average basis. Corporate average HC levels cannot exceed 75 g/kW-hr as
in the Phase 2 requirement. We believe this will allow manufacturers to use a broader variety of
technology mixes than our proposed Phase 2 standards. Tables 4.3-7 and 4.3-8 provided below
present a couple of potential technology mix scenarios for the Phase 3 standards. For the Phase
3 standards, we are including a HC+NOx requirement. This was done because, as the tables
below will show, the number of four-stroke snowmobiles is anticipated to significantly increase
compared to the number used to meet our Phase 1 and Phase 2 standards. Four-stroke engines
emit significantly higher levels of NOx emissions than two-stroke engines. In order to make sure
that NOx emissions do not become a problem as a result of the increase in the number of four-
stroke snowmobiles, we decided to establish a NOx standard as well. The NOx standard is set at
a level that makes it more of a cap, 15 g/kW-hr. This level should be inherently achievable for
the majority of four-stroke engines. However, should a manufacturer attempt to design a four-
stroke snowmobile that operates with a very lean air and fuel mixture to get even further HC
reductions, this standard will prevent backsliding. NOx emissions from two-stroke engines are
inherently well below the 15 g/kW-hr level.
We do not believe that incorporating the 15 g/kW-hr NOx standard as part of the
HC+NOx standard will provide any incentive to increase HC significantly. NOx emissions from
four-stroke engines are sufficiently close to 15 g/kW-hr that there will be little ability to increase
HC even marginally. For two-stroke engines, while the 15 g/kW-hr level for NOx is well above
typical two-stroke NOx emissions, it is still well below two-stroke HC emissions and does not
provide enough of a margin to avoid use of advanced technologies on most engines. At most, it
may provide a slight compliance cushion for these engines.
4-51
-------
Draft Regulatory Support Document
Table 4.3-7
Potential Snowmobile Technology Mix for Phase 3 Standards
Technology
Carburetor/EFI
Recalibration + Engine
Modifications + Pulse Air
Injection
Direct Injection
Four-Stroke
Percent Usage
20-30%
50%
20%
Percent
Reduction
HC
23-35%
75%
90%
Percent
Reduction
CO
23-35%
70%
50%
Average Reduction
Fleet %
Reduction
HC
7-11.5%
37.5%
18%
63%
Fleet %
Reduction
CO
7-11.5%
35%
10%
52%
Table 4.3-8
Potential Snowmobile Technology Mix for Phase 3 Standards
Technology
Carburetor/EFI
Recalibration + Engine
Modifications + Pulse Air
Injection
Direct Injection
Four-Stroke
Percent Usage
0-20%
10%
70%
Percent
Reduction
HC
0-35%
75%
90%
Percent
Reduction
CO
0-35%
70%
50%
Average Reduction
Fleet %
Reduction
HC
0-7%
7.5%
63%
71%
Fleet %
Reduction
CO
0-7%
7%
35%
42%
Clearly the technologies necessary to meet our 2012 standards are feasible, and in many
cases the technologies are already being used on various snowmobile applications. As these
technologies have been shown to provide emission reductions at or beyond the reductions needed
to meet the standards, the standards are clearly feasible given the appropriate lead time even
when considering production variability and emissions deterioration. The challenge
manufacturers will face will be deciding which technologies to use for different applications and
how consumers will respond to those technologies. In our testing efforts we attempted to order
one of the new 2003 Yamaha RX-1 high performance four-stroke snowmobiles, but were
surprised to find out that local dealers said there would be a six month wait to get one due to the
high demand. We verified with Yamaha that they indeed have commitments for virtually every
one of the new RX-1 models they are making and it's not a limited run, but rather a full scale
production build. Therefore, if the Yamaha case is any indication, we believe there are a number
of viable technologies available to meet our 2010 standards and the public is not only going to
4-52
-------
Chapter 4: Feasibility of Proposed Standards
accept them, but embrace them.
Tables 4.3-7 and 4.3-8 are meant to show some possible technology mix scenarios that
manufacturers may choose to comply with the Phase 3 standards in 2012. Implicit in these tables
is the possibility that, under the averaging program, there may still be some largely unmodified
two-stroke engines sold under the Phase 3 program. There are several reasons why a
manufacturer might choose to continue to sell a small number of baseline technology
snowmobiles under the Phase 3 program. First, it may prove significantly more expensive to
reduce the emissions of a particular engine family relative to a manufacturer's other product
offerings, and the manufacturer may simply choose to apply additional technology to some of its
other models rather than put the extra effort and expense into reducing emissions from every one
of its models. Second, a particular engine family may not respond as well to technology changes
as other engine families, and the manufacturer may choose to apply additional technology to
some of its other offerings rather than spending the resources to overcome the technological
hurdles associated with a particular engine family. This could be because the technologies may
affect the performance of the particular snowmobile model, including increased weight and
startability concerns, and thus need further refinement for implementation. Finally, a
manufacturer may intend to discontinue a particular engine family in the near future and may
choose to focus its efforts on its other product offerings rather than spend the resources to reduce
emissions from an engine family that is scheduled to be discontinued.
While it is possible that there may be some baseline technology snowmobiles in the
product mix under the Phase 3 program, we expect that sales of such snowmobiles will be
minimal for the following reasons. First, as Tables 4.3-7 and 4.3-8 show, we expect that
compliance with the Phase 3 standards will require that at least 70 percent of snowmobile
production employ some form of advanced technology such as direct injection two-stroke
technology, or four-stroke engines. There may be some uncertainty amongst manufacturers as to
whether they will be able to sell enough snowmobiles with advanced technology to allow for
including baseline technology snowmobiles in their product mix. Manufacturers will likely
choose to apply some level of emissions control to every snowmobile they sell in order to assure
compliance with the Phase 3 standards on average. Similarly, there is no assurance that the
advanced technologies will reduce emissions as well as expected on all engine families in the
time frame provided, and we expect that manufacturers will also choose to apply some level of
technology to every snowmobile in order to provide a compliance margin in case some
technologies or particular applications of technologies do not perform as expected.
4.4 All-Terrain Vehicles/Engines
The following paragraphs summarize the data and rationale supporting the emission
standards for ATVs, which are listed in the Executive Summary.
4-53
-------
Draft Regulatory Support Document
4.4.1 Baseline Technology and Emissions
ATVs have been in popular use for over 25 years. Some of the earliest and most popular
ATVs were three-wheeled off-highway motorcycles with large balloon tires. Due to safety
concerns, the three-wheeled ATVs were phased-out in the mid-1980s and replaced by the current
and more popular vehicle known as "quad runners" or simply "quads." Quads resemble the
earlier three-wheeled ATVs except the single front wheel was replaced with two wheels that are
controlled by a steering system. The ATV steering system uses motorcycle handlebars, but
otherwise looks and operates like an automotive design. The operator sits on and rides the quad
much like a motorcycle. The engines used in quads tend to be very similar to those used in off-
highway motorcycles - relatively small single cylinder two- or four-stroke engines that are either
air- or liquid-cooled. Recently, some manufacturers have introduced ATVs equipped with
larger four-stroke two-cylinder V-twin engines. Quads are typically divided into two types: utility
and sport. The utility quads are designed for recreational use but have the ability to perform
many utility functions such as plowing snow, tilling gardens, and mowing lawns to name a few.
They are typically heavier and equipped with relatively large four-stroke engines and automatic
transmissions with reverse gear. Sport quads are smaller and designed primarily for recreational
purposes. They are equipped with two- or four-stroke engines and manual transmissions.
Although ATVs are not currently regulated federally, they are regulated in California.
The California ATV standards are based on the FTP cycle just like highway motorcycles,
however, California allows manufacturers to optionally certify to a steady-state engine cycle
(SAE J1088) and meet the California non-handheld small SI utility engine standards.
Manufacturers have felt that these standards are unattainable with two-stroke engine technology.
Therefore, all of the ATVs certified in California are equipped with four-stroke engines.
California ultimately allowed manufacturers to sell uncertified engines as long as those ATVs
and motorcycles equipped with uncertified engines were operated exclusively on restricted public
lands and at specified times of the year. This allowed manufacturers to continue to produce and
sell two-stroke ATVs in California. Thus, the main emphasis of ATV engine design federally,
and for two-stroke powered ATVs in California, is on performance, durability, and cost.
Although some manufacturers offer some of their California models nationwide, most ATVs sold
federally have no emission controls.
ATVs predominantly use four-stroke engines (e.g., 80 percent of all sales are four-stroke).
The smaller percentage of two-stroke engines are found primarily in the small engine
displacement "youth" models. Of the seven major ATV manufacturers, only two make two-
stroke ATVs for adults. These models are either inexpensive entry models or high-performance
sport models. The fuel system used on ATVs, whether two- or four-stroke, are almost
exclusively carburetors, although at least one manufacturer has introduced a four-stroke ATV
with electronic fuel injection. Although ATVs are mostly four-stroke equipped, they still can
have relatively high levels of HC and extremely high levels of CO, because many of them
operate with a "rich" air and fuel mixture, which enhances performance and allows engine
cooling, which promotes longer lasting engine life. This is also true for two-stroke equipped
ATVs. Rich operation results in high levels of HC, CO, and PM emissions. In addition, two-
4-54
-------
Chapter 4: Feasibility of Proposed Standards
stroke engines lubricate the piston and crankshaft by mixing oil with the air and fuel mixture.
This is accomplished by most contemporary 2-stroke engines with a pump that sends two-cycle
oil from a separate oil reservoir to the carburetor where it is mixed with the air and fuel mixture.
Some less expensive two-stroke engines require that the oil be mixed with the gasoline in the
fuel tank. Because two-stroke engines tend to have high scavenging losses, where up to a third
of the unburned air and fuel mixture goes out of the exhaust, lubricating oil particles are also
released into the atmosphere, becoming HC particles or particulate matter (PM). The scavenging
losses also result in high levels of raw HC. This is in contrast to four-stroke engines that use the
crankcase as an oil sump and a pump to distribute oil throughout the engine, resulting in virtually
noPM..
We tested 11 four-stroke and three two-stroke ATVs over the FTP. Tables 4.4-1 and 4.4-
2 shows that the HC emission rate for the four-stroke ATVs is significantly lower than for the
two-stroke ATVs, whereas the NOx emissions from the two-strokes were considerably lower
than from the four-strokes. The CO emissions were also lower for the two-stroke ATVs. The
four-stroke ATVs that we tested that had high levels of CO also happened to be 50-state certified
vehicles, meaning they are California vehicles sold nationwide. Because there are California
standards for HC+NOx, manufacturers have tended to calibrate the ATVs fuel system to run even
richer than normal to meet the NOx standard. Since the CO standard in California is relatively
high, these ATVs can run rich and still meet the CO standards. Another observation that can be
made from the test results is that of the 11 four-stroke models tested, the four ATVs with the
lowest emissions were sport models. The other seven models were all utility models. The four
sport models, the Yamaha Warrior and Raptor, the Honda 300EX, and Polaris Trail Boss had an
average HC+NOx level of 1.35 g/km, below our 1.5 g/km standard, and an average CO level of
28.5 g/km, only slightly above our standard of 25 g/km. In fact, the Warrior and Raptor already
meet our standards with considerable headroom. The average HC+NOx and CO emissions levels
for the seven utility models were 2.20 g/km and 33.7 g/km, respectively. This may indicate that
when testing over the highway motorcycle test procedure, utility ATVs may be at a disadvantage
compared to the sport models because of their lower power-to-weight ratio and use of
continuously variable transmissions. Even when tested over the less strenuous Class I highway
motorcycle test cycle, the utility ATVs appeared to be operating at higher loads than the sport
models. Although we didn't examine all of the ATVs, the Warrior operated at a slightly leaner
air and fuel mixture than the Polaris Sportsman. This could be model or manufacturer specific,
but if this is at all indicative of how sport and utility ATVs fuel systems are calibrated, the fact
that utility ATVs already operate very rich could be exacerbated when operated over the FTP,
resulting in the higher HC and CO levels that we observed.
4-55
-------
Draft Regulatory Support Document
Table 4.4-1
Four-Stroke ATV Emissions (g/km)
Make
Kawasaki
Honda
Polaris
Yamaha
Polaris
Arctic Cat
Yamaha
Honda
Bombardier
Polaris
Yamaha
Model
Bayou
300EX
Trail Boss
Warrior
Sportsman
375 Automatic
Big Bear
Rancher
4X4 AWD
Sportsman
Raptor
Model
Year
1989
1997
1998
1998
2001
2001
2001
2001
2001
2001
2001
Eng. Displ.
280 cc
298 cc
324 cc
349 cc
499 cc
375 cc
400 cc
400 cc
500 cc
499 cc
660 cc
Average
HC
1.17
1.14
1.56
0.98
2.68
1.70
2.30
1.74
1.62
1.56
0.97
1.58
CO
14.09
34.60
43.41
19.44
56.50
49.70
41.41
33.98
20.70
19.21
16.56
31.78
NOx
0.640
0.155
0.195
0.190
0.295
0.190
0.170
0.150
0.740
0.420
0.210
0.305
Table 4.4-2
Two-Stroke ATV Emissions (g/km)
Make
Suzuki
Polaris
Polaris
Model
LT80
Scrambler
Trailblazer
Model Year
1998
2001
2000
Eng. Displ.
79 cc
89 cc
250 cc
Average
HC
7.66
38.12
18.91
21.56
CO
24.23
25.08
44.71
31.34
NOx
0.047
0.057
0.040
0.048
4.4.2 Potentially Available ATV Technologies
A variety of technologies are currently available or in stages of development to be
available for use on two-stroke ATVs, such as engine modifications, improvements to
carburetion (improved fuel control and atomization, as well as improved production tolerances),
enleanment strategies for both carbureted and fuel injected engines, and semi-direct and direct
fuel injection. However, it is our belief that manufacturers will choose to convert their two-
stroke engines to four-stroke applications, because of the cost and complexity of the above
4-56
-------
Chapter 4: Feasibility of Proposed Standards
mentioned technologies necessary to make a two-stroke engine meet our standards. We believe
that to meet our ATV standards, manufacturers will use four-stroke engines. Depending on the
size, performance and calibration of the engine, they will also need to make improvements to the
fuel system, consisting of improved carburetor tolerances and a leaner air and fuel mixture, and
in some cases the use of pulse air injection.
4.4.2.1 Engine Modifications
There are a variety of engine modifications that could reduce emissions from two-stroke
and four-stroke engines. The modifications generally either increase trapping efficiency (i.e.,
reduce fuel short-circuiting) or improve combustion efficiency. Those modifications for two-
stroke engines that increase trapping efficiency include optimizing the intake, scavenge and
exhaust port shape and size, and port placement, as well as optimizing port exhaust tuning and
bore/stroke ratios. Optimized combustion charge swirl, squish and tumble would serve to
improve the combustion of the intake charge for both two- and four-stroke engines.
Manufacturers have indicated that they believe these modifications for two-stroke engines have
the potential to reduce emissions by up to 40 percent, depending on how well the unmodified
engine is optimized for these things, but would be insufficient alone to meet our standardsy.
4.4.2.2 Carburetion Improvements
There are several things that can be done to improve carburetion in ATV engines. First,
strategies to improve fuel atomization would promote more complete combustion of the fuel/air
mixture. Additionally, production tolerances could be improved for more consistent fuel
metering. Both of these things would allow for more accurate control of the air/fuel ratio. In
conjunction with these improvements in carburetion, the air/fuel ratio could be leaned out some.
ATV engines are currently calibrated with rich air/fuel ratios for durability and performance
reasons. According to manufacturers, based on their experience, leaner calibrations could serve
to reduce CO and HC emissions by up to 20 percent, depending on how lean the unmodified
engine is prior to recalibration2. Small improvements in fuel economy could also be expected
with recalibration.
The calibration changes just discussed (as well as some of the engine modifications
previously discussed) could create concerns about ATV engine durability. There are many
engine improvements that could be made to regain any lost durability that occurs with leaner
calibration. These include changes to the cylinder head, pistons, pipes and ports for two-stroke
and valves for four-stroke, to reduce knock. In addition, critical engine components could be
made more robust with improvements such as better metallurgy to improve durability.
y See "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
z See "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
4-57
-------
Draft Regulatory Support Document
The same calibration changes to the air/fuel ratio just discussed for carbureted engines
could also be employed, possibly with more accuracy, with the use of fuel injection. At least one
ATV manufacturer currently employs electronic fuel injection on one of its ATV models.
4.4.2.3 Direct and Semi-Direct Fuel Injection
In addition to rich air/fuel ratios, one of the main reasons that two-stroke engines have
such high levels of HC emissions is scavenging losses, as described above. One way to reduce or
eliminate such losses is to inject the fuel into the cylinder after the exhaust port has closed. This
can be done by injecting the fuel into the cylinder through the transfer port (semi-direct injection)
or directly into the cylinder (direct injection). Both of these approaches are currently being used
successfully in two-stroke personal watercraft engines and some are showing upwards of 70
percent reductions in emissions. Direct injection is also being used by some motorcycle
manufacturers (e.g., Aprilla) on small mopeds, scooters, and motorcycles to meet emission
standards for two-strokes in Europe and Asia. A new start-up company called Rev! Motorcycles
plans to manufacturer high-performance recreational and competition off-highway motorcycles
with direct fuel injection two-stroke engines in the next year or so (for more, see Section 4.7.2.3).
They have not indicated whether they will manufacturer any ATVs. Substantial improvements in
fuel economy could also be expected with these technologies. However, there are some issues
with ATV operation (larger displacement engines that experience more transient operation than
watercraft and small mopeds) that make the application of the direct injection technologies
somewhat more challenging for ATVs than for personal watercraft and small displacement
scooters. The biggest obstacle for this technology is that the many of the two-stroke equipped
ATVs are youth models which emphasize low price. Direct injection is relatively expensive and
may not be considered to be cost effective for these engines.
4.4.2.4 Four-Stroke Engines
Four-stroke engines produce significantly lower levels of HC emissions than two-stroke
engines. This is primarily due to the fact that two-stroke engines experience high scavenging
losses that allow up to a third of the unburned air and fuel mixture to escape into the atmosphere
during the combustion process. Since four-stroke engines have a valve-train system and
introduce the air and fuel mixture into the combustion chamber when the exhaust valve is closed
or almost closed, there is very little scavenging of unburned fuel. Thus, four-stroke engines have
superior HC control to conventional two-stroke engines. Four-stroke engines have comparable
CO performance to two-stroke engines. CO emissions result from incomplete combustion due to
an excess of fuel in the air and fuel mixture. Thus, CO emissions are a function of air and fuel
mixture. Current unregulated four-stroke and two-stroke engines both operate with a rich air and
fuel mixture, resulting in high levels of CO emissions. Therefore, four-stroke engines do not
have inherently low CO emission levels. Four-stroke engines also generate higher NOx emission
levels than two-stroke engines. This is because NOx emissions are a function of temperature.
Higher combustion temperatures generate higher NOx emission levels. Four-stroke engines have
more complete combustion than conventional two-stroke engines, which results in higher
combustion temperatures and higher NOx emission levels. Thus, four-stroke engines are an
4-58
-------
Chapter 4: Feasibility of Proposed Standards
excellent choice for significantly reducing HC emissions. However, to reduce CO emissions, a
four-stroke engine may need some fuel system calibration changes, engine modifications, or the
use of secondary air or a catalyst. To reduce NOx emissions from a four-stroke engine would
require fuel system calibration changes, engine modifications, exhaust gas recirculation (EGR),
or a catalyst.
Since 80 percent of all ATVs sold each year are four-stroke, there is no question about the
feasibility of using four-stroke engine technology for ATVs. Conversion from two-stroke to
four-stroke engine technology also results in improvements to fuel consumption and engine
durability. These benefits could be especially valuable to consumers who purchase utility ATVs.
The ATV models that are currently equipped with two-stroke engines tend to be small-
displacement youth models, entry-level adult ATVs and high-performance adult sport ATVs.
While most youth ATVs are equipped with two-stroke engines, there are several manufactures
who offer four-stroke models. Youth ATVs are regulated by the Consumer Product Safety
Commission (CPSC). Although the regulations are voluntary, manufacturers take them very
seriously, and one of the their requirements is that youth ATV speeds be governed. For "Y6"
ATVs (i.e., age 6 and up) the maximum speed is 15 miles per hour (mph) and for "Y12" ATVs
(i.e., age 12 and up), the maximum speed is 30 mph. By Consent Decree these are limited to 50
cc and 90 cc, respectively. Some manufacturers have argued that because of these constraints,
they need to use light-weight two-stroke engines, which have higher power-to-weight ratios than
four-stroke engines, in order to have sufficient power to operate the ATV. However, as
mentioned earlier, some manufacturers already use four-stroke engines in these applications
without any problem. The power required to meet the maximum speed limits for these little
ATVs is low enough that a four-stroke engine is more than adequate. The real issue appears to
be cost. Manufacturers argue that youth ATVs are price sensitive and that minor increases in
cost would be undesirable. Four-stroke engines are more expensive than similarly powered two-
stroke engines. This appears to be the issue with entry-level adult ATVs as well. Those
manufacturers that offer two-stroke entry-level ATVs also offer similar entry-level machines
with four-stroke engines. The argument is that consumers of their product like having the ability
to choose between engine types. In addition, manufacturers have expressed concern that these
smaller engines have lower cylinder surface to volume area ratios than larger displacement
engines, thus increasing the difficulty of in-cylinder control of HC emissions. That is one of the
reasons that we 1) are allowing engines under 99 cc to stay in the relatively less stringent utility
engine program and 2) that we permit averaging across the entire spectrum of ATV
vehicles/engines if they certify to the FTP-based standards.
Adult sport ATVs equipped with two-stroke engines were at one time considered the only
ATVs that were capable of providing true high-performance. However, advancements in four-
stroke engine technology for ATVs and off-highway motorcycles have now made it possible for
larger displacement high-powered four-stroke engines to equal, and in some cases surpass, the
performance of the high-powered two-stroke engines. Again, the argument for two-stroke
engines appears to be a matter of choice for consumers. However, since only two manufacturers
produce two-stroke adult ATVs, we believe that the relatively low sales volumes for these
4-59
-------
Draft Regulatory Support Document
models will make it cost prohibitive to reduce two-stroke emissions to the levels necessary to
meet our standards. Nonetheless, the credit exchange program (ABT) we are including for ATV
s creates the possibility for manufacturers to retain some lower emission two-stroke ATVs and
offset their higher emissions with reductions from 4-stroke models.
4.4.2.5 Air Injection
Secondary pulse air injection involves the introduction of fresh air into the exhaust pipe
immediately after the exhaust gases exit the engine. The extra air causes further combustion to
occur as the gases pass through the exhaust pipe, thereby oxidizing more of the HC and CO that
escape the combustion chamber. This type of system is relatively inexpensive and
uncomplicated because it does not require an air pump; air is drawn into the exhaust through a
one-way reed valve due to the pulses of negative pressure inside the exhaust pipe. Secondary
pulse-air injection is one of the most effective non-catalytic, emissions control technologies;
compared to engines without the system, reductions of 30-70% for HC and 30-80% for CO are
possible with pulse-air injection.
This technology is fairly common on highway motorcycles and is used on some off-
highway motorcycle models in California to meet the California off-highway motorcycle and
ATV emission standards. We believe that secondary air injection will not be necessary to meet
our standards for all models, but will be a viable control technology for some machines. We
tested three different four-stroke ATVs with secondary air. A 1998 Yamaha Warrior sport
model, a 2001 Polaris Sportsman High Output (H.O.) utility model, and a 2001 Polaris
Sportsman utility model. Initially we didn't have access to a pulse air system so we used shop air
introduced into the exhaust manifold at various flow rates to simulate air injection. To save time
and money, we performed our tests over the hot 505 section of the Class I Motorcycle cycle.
This is a warmed-up version of the first bag or 505 seconds of the FTP test cycle. The initial
tests with shop air indicated that air injected into the exhaust stream could reduce HC emissions
from 5-percent to 60-percent depending on the vehicle and the amount of air injected. For
example, the Warrior was very responsive to air injection. We tested at flow rates of 10, 20, 30,
and 40 cubic feet per minute (cfm). HC emissions were reduced from 25-percent to 60-percent
depending on the flow rate. Figure 4.4-1 illustrates these reductions. We also experimented with
the air and fuel mixture and found that if we leaned the mixture slightly, the air injection had an
even greater effect, reducing HC emissions by 83-percent from the uncontrolled baseline level
with 40 cfm of air. Our next task was to determine how the various flow rates we tested
compared to the capabilities of a pulse air system. A pulse air system uses a system of check
valves which uses the normal pressure pulsations in the intake manifold to draw in air from
outside and inject into the exhaust manifold. A reed valve is used in the exhaust manifold to
prevent reverse airflow of exhaust gases through the system. A valve called the "air injection"
valve reacts to high intake manifold vacuum and will cut-off the supply of air during engine
decelerations, thereby preventing after burn in the exhaust system.
4-60
-------
Chapter 4: Feasibility of Proposed Standards
Figure 4.4-1
Emission Impact of Air Injection
Yamaha Warrior
g 0.30
20cfm
Air Injection Rates
[•HC
Since generic pulse air systems can't be simply purchased from the store or dealership,
we had to modify an existing pulse air system to work on our test ATVs. We purchased a pulse
air system for a 1995 BMW 100R. Because this is a multi-cylindered engine, we had to make
some modifications to get it to work with a single-cylinder ATV engine. We were able to
successfully install the pulse air system onto the Warrior and performed several hot 505 test runs
to see how the pulse air system compared with the various flow rates of shop air. For our shop
air tests, we injected a constant flow rate over the entire 505 seconds of the test. Because a pulse
air system relies on drawing air into the exhaust system during negative pressure pulses in the
cylinder, increasing the engine speed increases the magnitude of the positive pressure pulses
resulting in increased back-pressure which can make a pulse air system ineffective. Our biggest
concern was that a pulse air system might not have the same overall flow capacity as our shop air
experiments since the pulse air system is only capable of drawing air into the exhaust manifold
during lower speeds where increased exhaust back-pressure is decreased. Due to timing
constraints, we only tested the Warrior with the pulse air system in conjunction with the enleaned
carburetor setting. The carburetor was enleaned by raising the jet needle one clip notch. When
we raised the clip two notches, the engine ran too lean and performance and driveability were
affected. With pulse air and the slightly lean calibration, the Warrior had emissions comparable
to the 20-30 cfm shop air results. Figure 4.4-2 shows the results between shop air and the pulse
air results. When the Warrior was tested over the full FTP with pulse air and the slightly lean
calibration, HC and CO emissions were reduced from baseline levels, while NOx increased. HC
was reduced by 73-percent, CO was reduced by 83-percent and NOx was increased by 47-
percent. The NOx emission increase is most likely due to the leaner air and fuel mixture. The
HC+NOx level was reduced by 54-percent from the baseline level as shown in Table 4.4-3.
4-61
-------
Draft Regulatory Support Document
Figure 4.4-2
Comparison Between Injected Air and Pulse Air
Warrior
0.70
505 Test Cycle
0.36
Pulse Air w/enleanment
0.290
0.00
Baseline
10cfm
20 cfm 30 cfm
Air Injection Rates
40 cfm
Pulse Air
Table 4.4-3
Yamaha Warrior Emissions with and without Pulse Air Injection
Test Configuration
Baseline
Pulse Air
w/enleanment
HC
0.98
0.26
CO
19.44
3.33
NOx
0.19
0.28
HC+NOx
1.17
0.54
The two Polaris Sportsman models proved to be more problematic than the Warrior. As
discussed above, the utility ATVs all had higher baseline emissions levels than the sport models.
The Polaris Sportsman High Output (H.O.) had the highest baseline emissions of any of the
ATVs we tested. HC+NOx emissions were 3.0 g/km, almost 100-percent higher than our
standard of 1.5 g/km, while CO was 56.5 g/km, 125-percent higher than the standard of 25 g/km.
The regular Sportsman was cleaner than the H.O. model with a HC+NOx level of 1.98 g/km and
a CO level of 19.2 g/km. As a result of these higher baseline emissions, the two Sportsman
models were at a disadvantage compared to the relatively clean Warrior. When supplying shop
4-62
-------
Chapter 4: Feasibility of Proposed Standards
air to the two Sportsman models we saw varied results. The higher emitting H.O. model
responded to air injection. However, the emissions were still so high that we stopped any further
testing and focused on catalyst use for this model. The regular Sportsman model was less
receptive to air injection. In fact the same levels of flow that resulted in sharp reductions for the
Warrior had only minimal effects for this vehicle. Further investigation indicated that the air and
fuel mixture was too rich for the injected air to have any significant effect. We tried to lean-out
the air and fuel mixture by raising the jet needle clip to the top of the needle, similar to what we
did for the Warrior, but there was no response. We had to use a different, leaner main jet, in
order to successfully lean-out the air and fuel mixture. With the air and fuel mixture leaner, we
ran several tests with shop air and found that the Sportsman was more receptive to air injection,
so we decided to install the BMW pulse air system that we modified for the Yamaha Warrior to
the Sportsman. We ran a full FTP with the pulse air system and the leaner main jet installed and
found that emissions were reduced considerably. HC and CO were reduced by 71-percent and
68-percent, respectively. NOx emissions increased by 45-percent. Limited time prevented us
from further investigating ways to reduce the air and fuel mixture. However, as Table 4.4-4
shows, the Sportsman was able to meet the standard using this approach.
Table 4.4-4
Polaris Sportsman Emissions with and without Pulse Air Injection
Test Configuration
Baseline
Pulse Air
w/enleanment
HC
1.56
0.49
CO
19.21
6.12
NOx
0.42
0.60
HC+NOx
1.98
1.09
4.4.2.6 Catalyst Technology
For our proposal, we proposed Phase 2 standards of 1.0 g/km HC+NOx. To achieve a
standard of 1.0 g/km, manufacturers will actually have to design their emission control system to
meet an emission level lower than the standard to account for deterioration and provide an
acceptable certification emission margin. Manufactures typically aim for a certification
emissions margin of 20 percent. Our NONROAD emission model uses a deterioration factor of
1.17 for four-stroke ATV engines. Taking these factors into consideration would result in a
potential emission level design goal of approximately 0.7 g/km. To meet this level of HC+NOx
control, we projected in our proposal that it might be necessary for some ATV models to use a
catalyst. To establish the feasibility of using a catalyst on an ATV, we tested the Polaris
Sportsman High Output (HO) ATV equipped with several different catalysts. The Sportsman is a
large utility ATV equipped with a 500 cc (HO) four-stroke engine and is one of the larger ATV
models currently offered in the market. We chose this model to demonstrate catalyst viability
because, as mentioned above, it had the highest baseline emissions of any of the ATVs we tested,
and it is a California certified vehicle that is sold nationwide. We tested the Polaris with three
different catalysts. Two of the catalysts were three-way catalysts with metal substrates and cell
4-63
-------
Draft Regulatory Support Document
densities of 200 cells/in2. One of the catalyst's had a Pt/Rh washcoat, while the other used a Pd-
only washcoat. The third catalyst was an oxidation catalyst with a ceramic substrate and a cell
density of 400 cells/in2. Table 4.4-5 shows that emissions were significantly reduced when the
various catalysts were installed on the Sportsman. However, even though there was a significant
reduction in emissions, the ATV was still unable to meet the proposed 1.0 g/km HC+NOx
standard, let alone the design target of approximately 0.7 g/km.
Table 4.4-5
Polaris Sportsman 500 Emissions with Various Catalysts
Catalyst
Baseline
TWC (Pd-only)
TWC (Pt/Rh)
Oxidation
HC
2.68
1.27
1.29
1.38
CO
56.5
35.27
32.6
28.87
NOx
0.3
0.05
0.04
0.02
HC+NOx
2.98
1.32
1.33
1.4
The three catalysts that we used had volumes ranging from 400 to 500 cc. Most highway
motorcycles typically use catalysts with a catalyst-to-engine volume ratio of one half. In other
words, they typically use a catalyst that has a volume approximately half of the engine's
displacement. For our catalyst cost estimation in the proposal, we argued that this would be a
good assumption for ATVs as well. We estimated that for ATVs, the catalyst size necessary to
meet our proposed HC+NOx standard of 1.0 g/km would be equal to half of the engine
displacement. We projected an average catalyst volume of 200 cc. The catalysts that we tested
were roughly double the size of catalysts we projected would be necessary to meet our standards.
We chose to use these catalysts not because of their size, but because of their availability. All
three catalysts are used in production highway motorcycle applications and were provided to us
by catalyst manufacturers. The highway motorcycles that these catalyst are from have an engine
displacement of approximately 900 cc. The implication of this is that even with catalysts twice
as large as we projected would be necessary to meet our 1.0 g/km standard, the emission
reductions for this ATV were still about 33-percent short of the standard.
Due to rulemaking schedule constraints, we had limited time to perform the testing and
analyses that we felt were necessary to support the proposed standards. One of the consequences
of this timing was that were unable to test the Sportsman with the various catalysts with pulse air
injection and a leaner air and fuel mixture. It is quite possible, that had we been able to perform
those tests we would have found that the emissions from the Sportsman could be brought down
to levels below the proposed Phase 2 standards. However, with our limited success with air
injection and enleaning of the air and fuel mixture with the two Sportsman models, it is also
possible that these additional strategies would not have helped quickly. We are confident that
the use of a catalyst has the potential to significantly reduce emissions for many ATV
applications, but at this time we can not confidently claim they will work for all applications
without further investigation.
4-64
-------
Chapter 4: Feasibility of Proposed Standards
4.4.3 Test Cycle/Procedure
For ATVs, we specify the current highway motorcycle test procedure for measuring
emissions. The highway motorcycle test procedure is the same test procedure as used for light-
duty vehicles (i.e., passenger cars and trucks) and is referred to as the Federal Test Procedure
(FTP). The FTP for a particular class of engine or equipment is actually the aggregate of all of
the emissions tests that the engine or equipment must meet to be certified. However, the term
FTP has also been used traditionally to refer to the exhaust emission test based on the Urban
Dynamometer Driving Schedule (UDDS), also referred to as the LA4 (Los Angeles Driving
Cycle #4). The UDDS is a chassis dynamometer driving cycle that consists of numerous "hills"
which represent a driving event. Each hill includes accelerations, steady-state operation, and
decelerations. There is an idle between each hill. The FTP consists of a cold start UDDS, a 10
minute soak, and a hot start. The emissions from these three separate events are collected into
three unique bags. Each bag represents one of the events. Bag 1 represents cold transient
operation, bag 2 represents cold stabilized operation, and bag 3 represents hot transient operation.
Highway motorcycles are divided into three classes based on engine displacement, with
Class I (50 to 169 cc) being the smallest and Class HI (280 cc and over) being the largest. The
highway motorcycle regulations allow Class I motorcycles to be tested on a less severe UDDS
cycle than the Class II and in motorcycles. This is accomplished by reducing the acceleration
and deceleration rates on some the more aggressive "hills" and by reducing the top speed from 56
miles per hour to 35 mile per hour. California requires ATVs to be tested over the Class I
motorcycle cycle. Our testing has shown that some utility ATVs are at a disadvantage when
tested over the Class n and HI cycles because utility ATVs use continuously variable
transmissions (CVT), similar to snowmobiles. These transmissions tend to be geared towards
lower speed operation for ATVs with high torque generation at lower engine speeds. This is so
they can perform a broad variety of utilitarian tasks, such as plowing snow, hauling loads, cutting
grass and other high load activities. As a result, when operated over the Class n or HI motorcycle
test cycle, these vehicles operate under a much higher load than would be typically expected in
real-world operating conditions. Operating under higher loads means the engine runs at a richer
air and fuel mixture and generates higher levels of emissions. We received comments from
manufacturers stating that if keep the FTP as the main ATV test cycle, that we should only
require the Class I cycle, similar to California. As a result of these comments and our own
experience testing various ATVs over the FTP, we have decided to require Class I motorcycle
test cycle rather than using all three cycles depending on the engine displacement as proposed.
Some manufacturers have noted that they do not currently have chassis-based test
facilities capable of testing ATVs. Manufacturers have noted that requiring chassis-based testing
for ATVs would require them to invest in additional testing facilities which can handle ATVs,
since ATVs do not fit on the same chassis dynamometer roller(s) as motorcycles used in chassis
testing. Some manufacturers also have stated that low pressure tires on ATVs would not stand
up to the rigors of a chassis dynamometer test. California provides manufacturers with the
option of certifying ATVs using the engine-based, utility engine test procedure (SAE J1088), and
4-65
-------
Draft Regulatory Support Document
most manufacturers use this option for certifying their ATVs. Manufacturers have facilities to
chassis test motorcycles and therefore California does not provide an engine testing certification
option for off-highway motorcycles.
We have tested numerous ATVs over the FTP and have found that several methods can
be used to test ATVs on chassis dynamometers. The most practical method for testing an ATV
on a motorcycle dynamometer is to disconnect one of the drive wheels and test with only one
drive wheel in contact with the dynamometer. For chassis dynamometers set-up to test light-duty
vehicles, wheel spacers or a wide axle can be utilized to make sure the drive wheels fit the width
of the dynamometer. We have found that the low pressure tires have withstood dynamometer
testing without any problems.
We acknowledge that a chassis dynamometer could be costly to purchase and difficult to
put in place in the short run, especially for some smaller manufacturers. ATV manufacturers
may therefore certify using the J1088 engine test cycle per the California off-highway motorcycle
and ATV program for the model years 2006 through 2008. After 2008, this option expires and
the FTP becomes the required test cycle. If manufacturers can develop an alternate transient test
cycle (engine or chassis) that shows correlation with the FTP or demonstrates representativeness
of actual ATV operation greater than the FTP, then, through rulemaking, we would consider
allowing the option of an alternative test cycle in place of the FTP.
4.4.4 Small Displacement Engines
For small displacement ATVs of 70 cc or less, we proposed that they would have the
permanent option to certify to the proposed FTP-based ATV standards or meet the Phase 1 Small
SI emission standards for non-handheld Class 1 engines. These standards are 16.1 g/kW-hr
HC+NOx and 610 g/kW-hr CO. Manufacturers argued that ATVs with engine displacements
between 70 cc and 99 cc also should be allowed to certify to the Small SI standards, since the
differences between a 70 cc and 99 cc engine is very small and the ATVs equipped with 99 cc
engines face the same obstacles with the FTP test cycle as the 70 cc and below ATVs. They also
argued that the Phase 1 Small SI standards are too stringent for these engines and recommended
that EPA adopt the Phase 2 standards for Class IB engines of 40 g/kW-hr for HC+NOx and 610
g/kW-hr for CO.
We recognize that the vast majority of engine families, including 4-stroke engines, below
100 cc are not certified to the California standards, which is an indication to us that the standards
proposed may not be feasible for most engines in this size range given the lead time provided.
However, manufacturers did not provide supporting data and we do not have data to confirm that
the level recommended by the manufacturers would result in an appropriate level of control. We
examined the 2002 model year certification data for non-handheld Small SI engines certified to
the Phase 2 Class I-A and I-B engine standards (engines below 100 cc)and found that the five
engine families certified to these standards had average emissions for HC+NOx of about 25
g/kW-hr (see Table 4.4-6). All of these engine families had CO emissions below 500 g/kW-hr
and well below the 610 g/kW-hr level recommended by manufacturers.
4-66
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.4-6
2002 Certification Data for Non-Handheld Small SI Phase 2 Class I-A and I-B Engines
Manufacturer
Honda
MTD Southwest
Honda
Honda
Honda
Engine Family
2HNXS.0224AK
2MTDS.0264Y2
2HNXS.0314AK
2HNXS.0574AK
2HNXS.0991AK
Displacement
22.2
26.2
31.1
49.4
98.5
Average
HC+NOx
(g/kW-hr)
31.6
14.7
41.0
25.4
13.4
25.2
CO
(g/kW-hr)
329.8
483.2
391.4
372.1
445.3
404.4
We believe these levels are more representative of the levels that can be achieved with the
lead time provided through the use of 4-stroke engines than the standards recommended by the
manufacturers. Since we are offering averaging with the HC+NOx standard, a standard based on
the average of 25.0 g/kW-hr for the five engine families is appropriate for ATVs with an engine
displacement under 99 cc. Since we are not offering an averaging program for CO emissions, it
is apparent from the above data that a standard of 400 g/kW-hr would be very difficult for these
smaller ATV engines to achieve. Therefore, based on the above data, we believe that a standard
of 500 g/kW-hr can be achieved with engines under 99 cc. We believe these standards can be
meet through the use of the various technologies described above.
4.4.5 Impacts on Noise, Energy, and Safety
The Clean Air Act directs us to consider potential impacts on noise, energy, and safety
when establishing the feasibility of emission standards for nonroad engines.
As automotive technology demonstrates, achieving low emissions from spark-ignition
engines can correspond with greatly reduced noise levels. Virtually all ATVs are equipped sound
suppression systems or mufflers. The four-stroke engines used in ATVs are considerably more
quiet than two-stroke engines. Electronically controlled fuel systems are able to improve
management of the combustion event which can further help lower noise levels.
Adopting new technologies for controlling fuel metering and air-fuel mixing will lead to
substantial improvements in fuel consumption rates for four-stroke engines. Four-stroke engines
have far less fuel consumption than two-stroke engines. Average mileage for a baseline two-
stroke ATV is 20-25 mpg, while the average four-stroke ATV gets 30-50 mpg.
4-67
-------
Draft Regulatory Support Document
We believe the technology discussed here would have no negative impacts on safety.
Four-stroke engine technology has been utilized on ATVs for numerous years without any
incident. Secondary air and catalysts have been utilized in highway motorcycles and lawn and
garden equipment without any safety concerns.
4.4.6 Conclusion
We expect that the ATV emission standards will largely be met through the conversion
of two-stroke engines to four-stroke engines and with some minor carburetor calibration
modifications and air-fuel ratio enleanment, combined with some use of pulse air injection for
the four-stroke engines which now dominate this market. Our test data indicates that ATVs can
have a wide variety of emissions performance. Some models are very clean and will require a
relatively minor improvement to meet our standards. Other ATVs, especially larger heavier
utility models, will require substantially more work. Our development testing indicates that
control strategies such as carburetor enleanment and pulse air injection can significantly reduce
emissions. In particular, these strategies are a path to allow most ATV models to meet a
HC+NOx standard of 1.5 g/km with due consideration to useful life requirements and
compliance margins most manufacturers adopt for various reasons. The other main control
strategy that we examined was the use of catalysts. While it is well known that catalysts can
significantly reduce exhaust emissions, the results that we had in our testing program fell short of
complete success. For numerous reasons, including lack of time and hardware, we were
unsuccessful at getting all of our test ATVs to meet our proposed HC+NOx standard of 1.0 g/km.
We believe further investigation is warranted. However, due to scheduling concerns, we did not
have the time to complete this investigation. As a result, we have decided to postpone the setting
of phase 2 standards at this time. We plan to continue to investigate the emission reduction
capabilities of ATVs and may establish a second phase of standards in the future.
We are confident that control strategies such as the use of a four-stroke engine with
carburetor enleanment and pulse air injection can easily meet our HC+NOx emission standard of
1.5 g/km even with a 20-percent headroom to accommodate production variability and
deterioration by the 2006 model year. That is why we are, for now, establishing a single set of
standards for ATVs of 1.5 g/km HC+NOx and 25 g/km CO. These technologies have been
utilized in a number of different applications, such as highway motorcycles, personal watercraft,
lawn and garden equipment, and small scooters. These technologies also have potential benefits
beyond emission reductions (e.g., improved fuel economy, reliability and performance, and
reduced noise).
4.5 Off-Highway Motorcycles
The following paragraphs summarize the data and rationale supporting the emission
standards for off-highway motorcycles, which are listed in the Executive Summary.
4-68
-------
Chapter 4: Feasibility of Proposed Standards
4.5.1 Baseline Technology and Emissions
Off-highway motorcycles are similar in appearance to highway motorcycles, but there are
several important distinctions between the two types of machines. Off-highway motorcycles are
not street-legal and are primarily operated on public and private lands over trails and open land.
Off-highway motorcycles tend to be much smaller, lighter and more maneuverable than their
larger highway counterparts. They are equipped with relatively small-displacement single-
cylinder two- or four-stroke engines ranging from 50 to 650 cubic centimeters (cc). The exhaust
systems for off-highway motorcycles are distinctively routed high on the frame to prevent
damage from brush, rocks, and water. Off-highway motorcycles are designed to be operated over
varying surfaces, such as dirt, sand, and mud, and are equipped with knobby tires which provide
better traction in off-road conditions. Unlike highway motorcycles, off-highway motorcycles
have fenders mounted far from the wheels and closer to the rider to keep dirt and mud from
spraying the rider and clogging between the fender and tire. Off-highway motorcycles are also
equipped with a more advanced suspension system than those for highway motorcycles. This
allows the operator to ride over obstacles and make jumps safely. This advanced suspension
system tends to make off-highway motorcycles much taller than highway motorcycles, in some
cases up to a foot taller.
Thirty percent of off-highway motorcycle sales are generally considered to be competition
motorcycles. The vast majority of competition off-highway motorcycles are two-strokes. The
CAA requires us to exempt from our regulations vehicles used for competition purposes. The
off-highway motorcycles that remain once competition bikes are excluded are recreational trail
bikes and small-displacement youth bikes. The majority of recreational trail bikes are equipped
with four-stroke engines. Youth off-highway motorcycles are almost evenly divided between
four-stroke and two-stroke engines.
The fuel system used on off-highway motorcycles, whether two- or four-stroke, are
almost exclusively carburetors, although at least one manufacturer has introduced a four-stroke
off-highway motorcycle with electronic fuel injection. Although many off-highway motorcycles
are four-stroke equipped, they still can have relatively high levels of HC and extremely high
levels of CO, because many of them operate with a "rich" air and fuel mixture, which enhances
performance and allows engine cooling which promotes longer engine life. This is also true for
two-stroke equipped off-highway motorcycles. Rich operation results in high levels of HC, CO,
and PM emissions. In addition, two-stroke engines lubricate the piston and crankshaft by mixing
oil with the air and fuel mixture. This is accomplished by most contemporary two-stroke engines
with a pump that sends two-cycle oil from a separate oil reservoir to the carburetor where it is
mixed with the air and fuel mixture. Some less expensive two-stroke engines require that the oil
be mixed with the gasoline in the fuel tank. Because two-stroke engines tend to have high
scavenging losses, where up to a third of the unburned air and fuel mixture goes out of the
exhaust, lubricating oil particles are also released into the atmosphere, becoming HC particles or
particulate matter (PM). The scavenging losses also result in high levels of raw HC. This is in
contrast to four-stroke engines that use the crankcase as an oil sump and a pump to distribute oil
throughout the engine, resulting in virtually no PM.
4-69
-------
Draft Regulatory Support Document
We tested six high-performance two-stroke motorcycles and four high-performance four-
stroke motorcycles over the FTP. Tables 4.5-1 and 4.5-2 shows that the HC emissions for the
four-stroke bikes is significantly lower than for the two-stroke bikes, whereas the NOx emissions
from the two-strokes were a bit lower. The CO levels were also considerably lower for the four-
stroke bikes.
Table 4.5-1
Four-Stroke Off-Highway Motorcycles Emissions (g/km)
Make
Yamaha
Yamaha
KTM
Husaberg
Model
WR250F
WR400F
400EXC
FE501
Model Year
2001
1999
2001
2001
Eng. Displ.
249 cc
398 cc
398 cc
498 cc
Average
HC
1.46
1.07
1.17
1.30
1.25
CO
26.74
20.95
28.61
25.81
25.52
NOx
0.110
0.155
0.050
0..163
0.109
Table 4.5-2
Two-Stroke Off-Highway Motorcycles Emissions (g/km)
Make
KTM
KTM
KTM
KTM
KTM
KTM
Model
125SX
125SX
200EXC
250SX
250EXC
300EXC
Model Year
2001
2001
2001
2001
2001
2001
Eng. Displ.
124 cc
124 cc
198 cc
249 cc
249 cc
398 cc
Average
HC
33.77
61.41
53.09
62.89
59.13
47.39
52.95
CO
31.00
32.43
39.89
49.29
40.54
45.29
39.74
NOx
0.008
0.011
0.025
0.011
0.016
0.012
0.060
4.5.2 Potentially Available Off-Highway Motorcycle Technologies
A variety of technologies are currently available or in stages of development to be
available for use on two-stroke off-highway motorcycles, such as engine modifications,
improvements to carburetion (improved fuel control and atomization, as well as improved
production tolerances), enleanment strategies for both carbureted and fuel injected engines, and
semi-direct and direct fuel injection. However, it is our belief that manufacturers will, in most
cases, choose to convert their two-stroke engines to four-stroke applications, because of the cost
and complexity of the above mentioned technologies necessary to make a two-stroke engine meet
4-70
-------
Chapter 4: Feasibility of Proposed Standards
our standards. For our standards, we believe that a four-stroke engine with minor improvements
to carburetion and enleanment strategies will be all that is required. Each of these is discussed in
the following sections.
4.5.2.1 Engine Modifications
There are a variety of engine modifications that could reduce emissions from two-stroke
and four-stroke engines. The modifications generally either increase trapping efficiency (i.e.,
reduce fuel short-circuiting) or improve combustion efficiency. Those modifications for two-
stroke engines that increase trapping efficiency include optimizing the intake, scavenge and
exhaust port shape and size, and port placement, as well as optimizing port exhaust tuning and
bore/stroke ratios. Optimized combustion charge swirl, squish and tumble would serve to
improve the combustion of the intake charge for both two- and four-stroke engines.
Manufacturers have indicated that these modifications for two-stroke engines have the potential
to reduce emissions by up to 40 percent, depending on how well the unmodified engine is
optimized for these things, but would be insufficient alone to meet our standardsaa.
4.5.2.2 Carburetion Improvements
There are several things that can be done to improve carburetion in off-highway
motorcycle engines. First, strategies to improve fuel atomization would promote more complete
combustion of the fuel/air mixture. Additionally, production tolerances could be improved for
more consistent fuel metering. Both of these things would allow for more accurate control of the
air/fuel ratio. In conjunction with these improvements in carburetion, the air/fuel ratio could be
leaned out some. Off-highway motorcycle engines are currently calibrated with rich air/fuel
ratios for durability and performance reasons. According to manufacturers, leaner calibrations
would serve to reduce CO and HC emissions by up to 20 percent, depending on how lean the
unmodified engine is prior to recalibrationbb. Small improvements in fuel economy could also be
expected with recalibration.
The calibration changes just discussed (as well as some of the engine modifications
previously discussed) could create concerns about off-highway motorcycle engine durability.
There are many engine improvements that could be made to regain any lost durability that occurs
with leaner calibration. These include changes to the cylinder head, pistons, pipes and ports for
two-stroke and valves for four-stroke, to reduce knock. In addition, critical engine components
could be made more robust with improvements such as better metallurgy to improve durability.
Carburetion improvements alone will not allow manufacturers to meet our standards,
aa See "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
bb See "Memo to Docket on Technical Discussions with Recreational Vehicle
Manufacturers," from Line Wehrly. Docket A-2000-01, IV-B-43.
4-71
-------
Draft Regulatory Support Document
especially for two-stroke engines. Carburetion improvements with four-stroke engines may be
necessary.
The same calibration changes to the air/fuel ratio just discussed for carbureted engines
could also be employed, possibly with more accuracy, with the use of fuel injection. At least one
off-highway motorcycle manufacturer currently employs electronic fuel injection on one of its
models.
4.5.2.3 Direct and Semi-Direct Fuel Injection
In addition to rich air/fuel ratios, one of the main reasons that two-stroke engines have
such high levels of HC emissions is scavenging losses, as described above. One way to reduce or
eliminate such losses is to inject the fuel into the cylinder after the exhaust port has closed. This
can be done by injecting the fuel into the cylinder through the transfer port (semi-direct injection)
or directly into the cylinder (direct injection). Both of these approaches are currently being used
successfully in two-stroke personal watercraft engines and some are showing upwards of 70
percent reductions in emissions. Direct injection is also being used by some motorcycle
manufacturers (e.g., Aprilla) on small mopeds, scooters, and motorcycles to meet emission
standards for two-strokes in Europe and Asia. As discussed above, a small start-up company
called Rev! Motorcycles is planning in the near future to manufacture two-stroke high-
performance recreational and competition off-highway motorcycles utilizing direct fuel injection.
Rev! claims they will be able to meet our optional HC+NOx standard of 4.0 g/km. They have
provided limited data based on computer simulation of what they expect their technology to
achieve.26
Substantial improvements in fuel economy could also be expected with direct injection.
However, there are some issues with off-highway motorcycle operation (larger displacement
engines that experience more transient operation than watercraft and small mopeds) that make
the application of the direct injection technologies somewhat more challenging for motorcycles
than for personal watercraft and small displacement scooters. The biggest obstacle for this
technology is that the many of the two-stroke equipped off-highway motorcycles are youth
models which emphasize low price. Rev! acknowledges that direct injection is expensive and
their motorcycle will have a premium price, but they expressed confidence that the success of
their system would attract customers and the cost of the system would eventually go down.
4.5.2.4 Four-Stroke Engines
Four-stroke engines produce significantly lower levels of HC emissions than two-stroke
engines. This is primarily due to the fact that two-stroke engines experience high scavenging
losses that allow up to a third of the unburned air and fuel mixture to escape into the atmosphere
during the combustion process. Since four-stroke engines have a valve-train system and
introduce the air and fuel mixture into the combustion chamber when the exhaust valve is closed
or almost closed, there is very little scavenging of unburned fuel. Thus, four-stroke engines have
superior HC control to conventional two-stroke engines. Four-stroke engines have comparable
4-72
-------
Chapter 4: Feasibility of Proposed Standards
CO performance to two-stroke engines. CO emissions result from incomplete combustion due to
an excess of fuel in the air and fuel mixture. Thus, CO emissions are a function of air and fuel
mixture. Current unregulated four-stroke and two-stroke engines both operate with a rich air and
fuel mixture, resulting in high levels of CO emissions. Therefore, four-stroke engines do not
have inherently low CO emission levels. Four-stroke engines also generate higher NOx emission
levels than two-stroke engines. This is because NOx emissions are a function of temperature.
Higher combustion temperatures generate higher NOx emission levels. Four-stroke engines have
more complete combustion than conventional two-stroke engines, which results in higher
combustion temperatures and higher NOx emission levels. Thus, four-stroke engines are an
excellent choice for significantly reducing HC emissions. However, to reduce CO emissions, a
four-stroke engine may need some fuel system calibration changes, engine modifications, or the
use of secondary air or a catalyst. To reduce Nox emissions from a four-stroke engine would
require fuel system calibration changes, engine modifications, exhaust gas recirculation (EGR),
or a catalyst.
We expect that the conversion of off-highway motorcycle models utilizing two-stroke
engines to four-stroke engines will be the main method of achieving our off-highway motorcycle
standards. As with ATVs, the question of feasibility for four-stroke engines in off-highway
motorcycles is moot, since more than half of the existing off-highway models are already four-
stroke and, in some cases, have been for a long time. Honda has used four-stroke engines in all
of their off-highway motorcycles (except for their competition motocross bikes) for over thirty
years. In fact, over the last 5 to 10 years, the trend has been to slowly replace two-stroke models
with four-stroke engines. Although the California emission standards have had some impact on
this trend, it has been minor. Four-stroke engines are more durable, reliable, quieter and get far
better fuel economy than two-stroke engines. But probably the single most important factor in
the spread of the four-stroke engine has been major advances in weight reduction and
performance.
Four-stroke engines typically weigh more than two-stroke engines because they need a
valve-train system, consisting of intake and exhaust valves, camshafts, valve springs, valve
timing chains and other components, as well as storing lubricating oil in the crankcase. Since a
four-stroke engine produces a power-stroke once every four revolutions of the crankshaft,
compared to a two-stroke which produces one once every two revolutions, a four-stroke engine
of equal displacement to a two-stroke engine produces less power, on the average of 30 percent
less. So in the past, off-highway motorcycles that used four-stroke engines tended to use very
heavy, large displacement engines, but yet had average power and performance. However, recent
breakthroughs in technologies have allowed manufacturers to design off-highway motorcycles
that use lighter and stronger materials for the engine and the motorcycle frame. The advanced
four-stroke technologies, such as multiple valves, used in some of the high-performance four-
stroke highway motorcycles, have found their way onto off-highway motorcycles, resulting in
vastly improved performance. The newer four-stroke bikes also tend to have an engine power
band or range that is milder of more forgiving than a typical two-stroke bike. Two-stroke bikes
4-73
-------
Draft Regulatory Support Document
tend to run poorly at idle and during low load situations. They also typically generate low levels
of torque at low to medium speeds, whereas four-stroke bikes traditionally generate a great deal
of low-end and mid-range torque. This is important to off-highway motorcycle riders because it
is common when riding off-highway motorcycles on trails or other surfaces to come across
obstacles that require slow speed maneuverability. A two-stroke engine that idles poorly and has
poor low-end torque can easily stall during these maneuvers, whereas a four-stroke bike excels
under these conditions. Current sales figures, as well as articles in off-highway motorcycle trade
magazines, indicate that four-stroke off-highway motorcycles are more popular than ever.
4.5.2.5 Air Injection
Secondary pulse air injection involves the introduction of fresh air into the exhaust pipe
immediately after the gases exit the engine. The extra air causes further combustion to occur as
the gases pass through the exhaust pipe, thereby controlling more of the hydrocarbons that escape
the combustion chamber. This type of system is relatively inexpensive and uncomplicated
because it does not require an air pump; air is drawn into the exhaust through a one-way reed
valve due to the pulses of negative pressure inside the exhaust pipe. Secondary pulse-air
injection is one of the most effective non-catalytic, emissions control technologies; compared to
engines without the system, reductions of 10-40% for HC are possible with pulse-air injection.
This technology is fairly common on highway motorcycles and is used on some off-
highway motorcycle models in California to meet the California off-highway motorcycle and
ATV emission standards. We believe that secondary air injection should not be necessary to
meet our standards, however, some manufacturers may choose to use it on some four-stroke
engine models.
4.5.2.6 Catalyst Technology
We do not believe catalysts will be necessary to meet our standards of 2.0 g/km HC+NOx
and 25.0 g/km CO. We did not pursue standards that would require catalyst technology for off-
highway motorcycles because we do not believe that potential safety and durability issues with
catalysts for off-highway motorcycle applications have been adequately addressed. As discussed
above in Section 4.4.2.6, to meet our proposed Phase 2 ATV standard of 1.0 g/km HC+NOx
would require a design goal of 0.6 to 0.7 g/km HC+NOx to account for certification compliance
margin and emission system deterioration. Although we did not perform any testing of off-
highway motorcycles with catalysts, the results from our ATV testing gave us additional concern
over the viability of catalysts with off-highway motorcycles. For the Polaris Sportsman (HO), a
large 500 cc utility ATV model, we were unable to successfully reduce HC+NOx emissions
below 1.3 g/km using a production three-way catalyst from a federally certified 900 cc highway
motorcycle. The catalysts were larger in volume, precious metal loading, and physical size than
we had initially projected would be necessary for ATVs. The physical size of these catalysts
were well beyond what would be considered acceptable for off-highway motorcycle applications.
4-74
-------
Chapter 4: Feasibility of Proposed Standards
The highway motorcycle that the production catalysts were from weighs around 450 pounds.
Typical four-stroke off-highway motorcycles weigh between 225 and 280 pounds. The exhaust
system, and thus the catalyst, were routed low to the ground where the extra weight would be
least noticeable. For a four-stroke off-highway motorcycle, the exhaust pipe is routed high on the
frame to provide a better center of gravity and keep the exhaust pipe away from water, rocks,
logs, and other items that could damage the pipe. Placing such a large catalyst in a four-stroke
off-highway motorcycle would pose problems of extra weight and packaging, since it is difficult
to find locations in the exhaust pipe to place a large catalyst so that it wouldn't interfere with the
rider.
We have concerns about the safety and durability of catalysts in off-highway motorcycle
applications. As discussed above, off-highway motorcycles operate in very harsh conditions.
They experience extreme shock and jarring that can easily damage a catalyst. It is very common
for off-highway motorcycles to come into contact with rocks, logs, stumps, and trees through the
course of regular riding activities or accidentally in the form of a crash. The substrate of a
catalyst can be very fragile, depending on the material used. We are unaware of any data on the
durability of a catalyst under such harsh operating conditions. There currently are no off-
highway motorcycle models equipped with a catalyst and we know of no studies performed on
the long term durability of a catalyst in an off-highway motorcycle application.
Catalysts operate at very high temperatures which can be a concern for burning the rider
or potentially starting a fire in the riding environment that they frequent, such as forests and
grassy fields. While heat shields may possibly prevent the rider from burns, there is the problem
of where to locate the catalyst so that the catalyst is not in the way of the rider adding concern
over potential burns. Off-highway motorcycles are much taller than highway motorcycles. In
fact, for some shorter riders they are unable to touch the ground with both feet when straddling
their off-highway motorcycle. This can be an additional concern for potential catalyst burns and
where to locate the catalyst. Because the motorcycle is so tall, the rider often has to lean to one
side or another of the bike to keep their balance when the motorcycle is not moving. It is
imperative that the catalyst not be located in a manner that would exacerbate the possibility of
burning the rider or interfering with the riders balance when standing still on the motorcycle.
There is also a question over the durability of heat shields in these harsh applications. Heat
shields used for many highway vehicle applications are not designed for the extreme conditions
that these vehicles operate in. Again, we are not aware of any data that demonstrates the
effectiveness of catalyst heat shields for off-highway motorcycles.
4.5.3 Test Procedure
For off-highway motorcycles, we specify the current highway motorcycle test procedure
for measuring emissions. The highway motorcycle test procedure is the same test procedure as
that used for light-duty vehicles (i.e., passenger cars and trucks) and is referred to as the Federal
Test Procedure (FTP). The FTP for a particular class of engine or equipment is actually the
4-75
-------
Draft Regulatory Support Document
aggregate of all of the emissions tests that the engine or equipment must meet to be certified.
However, the term FTP has also been used traditionally to refer to the exhaust emission test
based on the Urban Dynamometer Driving Schedule (UDDS), also referred to as the LA4 (Los
Angeles Driving Cycle #4). The UDDS is a chassis dynamometer driving cycle that consists of
numerous "hills" which represent a driving event. Each hill includes accelerations, steady-state
operation, and decelerations. There is an idle between each hill. The FTP consists of a cold
start UDDS, a 10 minute soak, and a hot start. The emissions from these three separate events
are collected into three unique bags. Each bag represents one of the events. Bag 1 represents
cold transient operation, bag 2 represents cold stabilized operation, and bag 3 represents hot
transient operation.
In the California program, highway motorcycles are divided into three classes based on
engine displacement, with Class I (50 to 169 cc) being the smallest and Class HI (280 cc and
over) being the largest. The highway motorcycle regulations allow Class I motorcycles to be
tested on a less severe UDDS cycle than the Class II and HI motorcycles. This is accomplished
by reducing the acceleration and deceleration rates on some the more aggressive "hills." We are
applying this same class/cycle distinction for off-highway motorcycles. In other words, off-
highway motorcycles with an engine displacement between 50 and 279 cc (Class I and n) must
be tested over the Class I highway motorcycle FTP test cycle. Off-highway motorcycles with
engine displacements greater than 280 cc would be tested over the Class HI highway motorcycle
FTP test cycle.
4.5.4 Impacts on Noise, Energy, and Safety
The Clean Air Act directs us to consider potential impacts on noise, energy, and safety
when establishing the feasibility of emission standards for nonroad engines.
As automotive technology demonstrates, achieving low emissions from spark-ignition
engines can correspond with greatly reduced noise levels. Virtually all recreational off-highway
motorcycles are equipped with sound suppression systems or mufflers. The four-stroke engines
used in off-highway motorcycles are considerably more quiet than the two-stroke engines used.
Adopting new technologies for controlling fuel metering and air-fuel mixing will lead to
substantial improvements in fuel consumption rates for four-stroke engines. Four-stroke engines
have far less fuel consumption than two-stroke engines. Average mileage for a baseline two-
stroke off-highway motorcycle is 20-25 mpg, while the average four-stroke off-highway
motorcycle gets 45-50 mpg.
We believe the technology discussed here would have no negative impacts on safety.
Four-stroke engine technology has been utilized on off-highway motorcycles for numerous years
without any incident. Secondary air and catalysts have been utilized in highway motorcycles and
lawn and garden equipment without any safety concerns.
4-76
-------
Chapter 4: Feasibility of Proposed Standards
4.5.5 Conclusion
We expect that the off-highway motorcycle emission standards will largely be met
through the conversion of two-stroke engines to four-stroke engines with some minor carburetor
calibration modifications and air-fuel ratio enleanment for some four-strokes. Four-stroke
engines are common in many off-highway motorcycles and have been used for many years.
Certification data from California's off-highway program presented below in Table 4.5-3, as well
as data from our own testing (see Table 4.5-1 above) suggest that four-stroke engines with some
minor fuel system calibration modifications will be capable of meeting our emission standards
even when considering production variability and deterioration. We believe the current sales
volumes of two-stroke off-highway motorcycles, combined with the cost to modify two-stroke
engines for significant emission reductions, will discourage the use of two-stroke engine
technology.
Table 4.5-3
2001 Model Year California Off-highway Motorcycle Certification Data (g/km)
Manufacturer
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Kawasaki
Yamaha
Yamaha
Yamaha
Yamaha
Model*
XR650R
XR400R
XR200R
XR100R
XR80R
XR70R
XR50R
KLX300
TT-R250
TT-R225
TT-R125
TT-R90
Engine Disp.
650 cc
400 cc
200 cc
100 cc
80 cc
70 cc
50 cc
300 cc
250 cc
225 cc
125 cc
90 cc
HC
1.0
0.5
0.7
0.8
0.6
0.8
1.0
1.0
0.7
0.7
0.8
0.8
CO
11.7
6.2
6.8
4.9
6.3
8.2
8.6
5.1
10.9
12.4
5.1
4.9
* All models are four-stroke
4-77
-------
Draft Regulatory Support Document
4.6 Permeation Control from Recreational Vehicles
The following paragraphs summarize the data and rationale supporting the permeation
emission standards for recreational vehicles, which are listed in the Executive Summary.
4.6.1 Baseline Technology and Emissions
4.6.1.1 Fuel Tanks
Recreational vehicle fuel tanks are generally blow-molded or injection-molded using high
density polyethylene (HDPE). Data on the permeation rates of fuel through the walls of
polyethylene fuel tanks shows that recreational vehicle HDPE fuel tanks have very high
permeation rates compared to those used in automotive applications. We tested four ATV fuel
tanks in our lab for permeation. We also tested three portable marine fuel tanks and two portable
gas cans which are of similar construction. This testing was performed at 29°C (85°F) with
gasoline. Prior to testing, the fuel tanks had been stored with fuel in them for more than a month
to stabilize the permeation rate. The permeation rates are presented in Table 4.6-1. The average
for these ten fuel tanks is 1.32 grams per gallon per day.
Table 4.6-1: Permeation Rates for Plastic Fuel Tanks Tested by EPA at 29°C
Tank Capacity
[gallons]
1.3
1.3
1.8
2.1
5.3
6.0
6.0
6.0
6.6
6.6
Permeation Loss
[g/gal/day]
1.66
2.90
1.29
2.28
1.00
0.61
1.19
0.78
0.77
0.75
Tank Type
all terrain vehicle
all terrain vehicle
all terrain vehicle
all terrain vehicle
all terrain vehicle
portable marine
portable marine
portable marine
portable fuel container
portable fuel container
The California Air Resources Board (ARE) investigated permeation rates from portable
fuel containers and lawn & garden equipment fuel tanks. Although this testing was not on
recreational vehicle fuel tanks, the fuel tanks tested are of similar construction. The ARE data is
compiled in several data reports on their web site and is included in our docket.27'28'29'30'31 Table
4.6-2 presents a summary of this data which was collected using the ARE test procedures
described in Section 4.6.3. Although the test temperature 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 used in the EPA testing. The average for these 36 fuel tanks is 1.07
grams per gallon per day.
4-78
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.6-2: Permeation Rates for
Plastic Fuel Containers Tested by ARE Over a 18-41°C Diurnal
Tank Capacity
[gallons]
.0
.0
.0
.0
.0
.0
o
.J
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.4
1.7
2.1
2.1
2.1
2.1
2.5
2.5
3.9
3.9
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.6
7.5
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
1.27
0.67
1.88
1.95
1.91
1.78
1.46
1.09
0.77
0.88
0.89
0.62
0.99
0.55
0.77
0.64
1.39
1.46
1.41
1.47
1.09
0.35
Tank Type
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
lawn & garden
lawn & garden
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
lawn & garden
lawn & garden
portable fuel container
portable fuel container
portable fuel container
lawn & garden
lawn & garden
lawn & garden
portable fuel container
portable fuel container
portable fuel container
portable fuel container
portable fuel container
lawn & garden
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.32
Based on data collected on HDPE samples at four temperatures,33'34 we estimate 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 4.6-1, and the numeric data can be found in
4-79
-------
Draft Regulatory Support Document
Section 4.6.2.3.
Figure 4.6-1: Effect of Temperature on HDPE Permeation
450
400
350
Data on HDPE samples
80% increase in permeation per
10 C increase in temperature
0
10 20 30 40 50
degrees Celsius
60
70
Based on the data from 46 fuel tanks in Tables 8.4-1 and 8.4-2, the average permeation
rate at 29°C is 1.12 grams per gallon per day. However, the standard is based on units of grams
per square meter per day at 28°C. Based on measurements of cut away fuel tanks of this size, we
have found that the wall thickness ranges from 4 to 5 mm. Using an average wall thickness of
4.5 mm and a permeation rate for HDPE of 47 g mm/m2/day at 28°C (Figure 4.6-1) we estimate
that the baseline permeation rate is about 10.4 g/m2/day. Data presented later in this chapter (see
Section 4.2.8.3) shows that the permeation rate of fuel through HDPE is fairly insensitive to the
amount of alcohol in the fuel.
4.6.1.2 Fuel Hoses
Fuel hoses produced for use in recreational vehicles are generally extruded nitrile rubber
with a cover for abrasion resistance. These hoses are generally designed to meet the
requirements under SAE J3035 for an R7 classification. R7 hose has a maximum permeation rate
of 550 g/m2/day at 23°C on ASTM Fuel C (50% toluene, 50% iso-octane). On a fuel containing
an alcohol blend, permeation would likely be higher from these fuel hoses. R7 hose is made
primarily of nitrile rubber (NBR). Based on the data presented in Section 4.2.8.3, permeation
through NBR is about 50 percent higher when tested on Fuel CE10 (10% ethanol) compared to
4-80
-------
Chapter 4: Feasibility of Proposed Standards
testing on Fuel C.
4.6.2 Permeation Reduction Technologies
4.6.2.1 Fuel Tanks
As discussed in Chapter 3, 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.
4.6.2.1.1 Sulfonation
We tested one sulfonated, 6 gallon, HOPE, portable marine fuel tank at 29°C (85°F) with
gasoline. Prior to testing, the fuel tank had been stored with gasoline in it for more than 10
weeks to stabilize the permeation rate. We measured a permeation rate of 0.08 g/gallon/day
which represents more than a 90 percent reduction from baseline.
The California Air Resources Board (ARB) collected test data on permeation rates from
sulfonated portable fuel containers using California certification fuel.36 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 30°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.
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. In this test program, the sulfonater was not aware of
the chemical make up of the fuel tanks. This is the likely reason for the variation in the data even
when the obvious outliers are removed. In support of this theory, the permeation rates were
consistently low for tanks provided by two of the four tank manufacturers. For these 11 fuel
tanks, the average permeation rate was 0.07 which represents more than a 90 percent reduction
from baseline. 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 largely resolved.37 For
this reason we do not include the earlier data in this analysis. Table 4.6-3 includes all of the
permeation data, including the outliers.
4-81
-------
Draft Regulatory Support Document
Table 4.6-3: Permeation Rates for Sulfonated
Plastic Fuel Containers Tested by ARE Over a 18-41°C Diurnal
Tank Capacity
[gallons]
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
ARB also investigated the effect of fuel slosh on the durability of sulfonated surfaces.
Three sulfonated fuel tanks were tested for permeation before and after being rocked with fuel in
them 1.2 million times.38 The results of this testing show that an 85% reduction in permeation
was achieved on average even after the slosh testing was performed. Table 4.6-4 presents these
results which were recorded in units of g/m2/day. The baseline level is an approximation based
on testing of similar fuel tanks.
As with earlier tests performed by ARB, the sulfonater was not aware of the materials
used in the fuel tanks sulfonated for the slosh testing. After the tests were performed, the
4-82
-------
Chapter 4: Feasibility of Proposed Standards
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% reduction in permeation would have been seen.39
list of resins and additives that are compatible with the sulfonation process is included in the
docket.40'41
A
Table 4.6-4: Permeation Rates for Sulfonated Fuel Tanks
with Slosh Testing by ARE Over a 18-41°C Diurnal
Technology Configuration
Approximate Baseline
Sulfonated
Sulfonated & Sloshed
Units
g/m2/day
g/m2/day
% reduction
g/mVday
% reduction
Tankl
10.4
0.73
93%
1.04
90%
Tank 2
10.4
0.82
92%
1.17
89%
Tank3
10.4
1.78
83%
2.49
76%
Average
10.4
1.11
89%
1.57
85%
An in-use durability testing program was also completed for Sulfonated HDPE fuel tanks
and bottles.42 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
HDPE 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
HDPE with 3% carbon black. These bottles were shown to retain over a 99 percent barrier 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.43 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.
4.6.2.1.2 Fluorination
We tested one fluorinated, 6 gallon, HDPE, portable marine fuel tank at 29°C (85°F) with
gasoline. Prior to testing, the fuel tank had been stored with gasoline in it for about 20 weeks to
stabilize the permeation rate. We measured a permeation rate of 0.05 g/gallon/day which
represents more than a 95 percent reduction from baseline.
4-83
-------
Draft Regulatory Support Document
The California Air Resources Board (ARB) collected test data on permeation rates from
fluorinated portable fuel containers using California certification fuel.44'45 The results, presented
in Table 4.6-5, 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. Four different levels of
fluorination treatment were tested. The average permeation rate for the 87 fluorinated fuel tanks
is 0.21 g/gal/day which represents about a 75 percent reduction from baseline. However, 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.46 For this reason we do not include the earlier data in this analysis.
Table 4.6-5: Permeation Rates for Fluorinated
Plastic Fuel Containers Tested by ARB Over a 18-41°C Diurnal
Barrier Treatment*
Level 3
(average = 0.27 g/gal/day)
Tank Capacity
[gallons]
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Permeation Loss
[g/gal/day]
0.04
0.06
0.25
0.12
0.15
0.17
0.09
0.15
0.12
0.18
0.17
0.14
0.18
0.34
0.41
0.41
0.36
0.41
0.23
0.29
0.31
4-84
-------
Chapter 4: Feasibility of Proposed Standards
Level 4
(average =0.09 g/gal/day)
2
2
2
2
2
2
2
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
1
1
1
5
5
5
0.24
0.32
0.16
0.19
0.20
0.11
0.20
0.06
0.06
0.07
0.09
0.10
0.11
0.15
0.23
0.31
0.33
0.24
0.33
0.33
0.51
0.47
0.41
0.45
0.45
0.35
0.37
0.28
0.26
0.35
0.35
0.37
0.28
0.35
0.41
0.47
0.43
0.39
0.47
0.55
0.05
0.05
0.06
0.11
0.11
0.15
4-85
-------
Draft Regulatory Support Document
Level 5
(average =0.07 g/gal/day)
SPAL
(average =0.04 g/gal/day)
2.5
2.5
2.5
2.5
2.5
5
5
5
5
5
5
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.47
ARB investigated the effect of fuel slosh on the durability of fluorinated surfaces. Three
fluorinated fuel tanks were tested for permeation before and after being rocked with fuel in them
1.2 million times.48 The results of this testing show that an 80% reduction in permeation was
achieved on average even after the slosh testing was performed. However, this data also shows
that an 89 percent reduction is feasible. Table 4.6-6 presents these results which were recorded
in units of g/m2/day. The baseline level is an approximation based on testing of similar fuel
tanks.
4-86
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.6-6: Permeation Rates for Fluorinated Fuel Tanks
with Slosh Testing by ARE Over a 18-41°C Diurnal
Technology Configuration
Approximate Baseline
Fluorinated
Fluorinated & Sloshed
Units
g/mVday
g/mVday
% reduction
g/m2/day
% reduction
Tankl
10.4
1.17
89%
2.38
77%
Tank 2
10.4
1.58
85%
2.86
73%
Tank3
10.4
0.47
96%
1.13
89%
Average
10.4
1.07
90%
2.12
80%
One study looked at the effect of alcohol in the fuel on permeation rates from fluorinated
fuel tanks.49 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.
4.6.2.1.3 Barrier Platelets
We tested four portable gas cans molded with low permeation non-continuous barrier
platelets 29°C (85°F) with gasoline. Prior to testing, the fuel tanks had been stored with gasoline
in it for more than 10 weeks to stabilize the permeation rate. Table 4.6-7 presents the emission
results which represent an average of nearly an 85 percent reduction from baseline.
Table 4.6-7: Permeation Rates for Plastic Fuel Containers
with Barrier Platelets Tested by EPA at 29°C
Percent Selar®*
4%
4%
4%
4%
Tank Capacity
[gallons]
5
5.3
6.6
6.6
Permeation Loss
[g/gal/day]
0.34
0.10
0.14
0.13
*trade name for barrier platelet technology used in test program
The California Air Resources Board (ARB) collected test data on permeation rates from
portable fuel containers molded with low permeation non-continuous barrier platelets using
California certification fuel. 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. Five different percentages of the barrier material were tested. The average permeation
rate for the 67 fuel tanks is 0.24 g/gal/day; however, there was a wide range in variation in the
effectiveness of the barrier platelets 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
4-87
-------
Draft Regulatory Support Document
in large emission increases due to pressure built up with temperature variation over the diurnal
cycle. Removing these six outliers, the average permeation rate is 0.15 g/gal/day with a
minimum of 0.04 g/gal/day and a maximum of 0.47 g/gal/day. This represents more than an 85
percent reduction from the average baseline. Table 4.6-8 includes all of the ARE test data,
including the outliers.
Table 4.6-8: Permeation Rates for Plastic Fuel Containers
with Barrier Platelets Tested by ARE Over a 18-41°C Diurnal
Percent Selar®*
4%
(average =0.12 g/gal/day)
6%
(average =0.16 g/gal/day)
Tank Capacity
[gallons]
5.00
5.00
5.00
5.00
5.00
6.00
6.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
5.00
5.00
5.00
5.00
5.00
5.00
6.00
6.00
Permeation Loss
[g/gal/day]
0.08
0.09
0.13
0.16
0.17
0.08
0.10
0.06
0.07
0.10
0.10
0.11
0.11
0.28
0.44
0.45
0.47
0.07
0.07
0.07
0.08
0.12
0.17
0.06
0.07
-------
Chapter 4: Feasibility of Proposed Standards
8%
(average =0.32 g/gal/day)
10%
(average =0.28 g/gal/day)
12%
(average =0.21 g/gal/day)
.00
.00
.00
.00
.00
.00
.00
.00
.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
5.00
5.00
6.00
6.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
.00
.00
.00
.00
.00
.00
0.14
0.17
0.21
0.21
0.21
0.65
0.85
0.98
1.66
0.04
0.05
0.07
0.09
0.12
0.16
0.44
0.08
0.10
0.05
0.06
0.15
0.19
0.19
0.21
0.23
0.26
0.79
0.83
0.88
0.06
0.06
0.07
0.08
0.13
0.14
0.23
0.13
0.14
0.20
0.21
0.23
0.35
*trade name for barrier platelet technology used in test program
The fuel containers tested by ARB used a technology known as Selar® which uses nylon
as the barrier resin. Dupont, who manufacturers Selar®, has recently developed a new resin
4-89
-------
Draft Regulatory Support Document
(Selar KB®) that uses ethylene vinyl alcohol (EVOH) as the barrier resin. EVOH has much lower
permeation than nylon, especially with alcohol fuel blends (see Section 4.6.2.3). Table 4.6-9
presents permeation rates for HDPE and three Selar KB® blends when tested at 60°C on
xylene.50 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% on naptha and 98% on toluene for 8%
Selar KB®.
Table 4.6-9: Xylene Permeation Results for Selar RB® at 60°C
Composition
100% HDPE
10%RB215/HDPE
10% RB 300/HDPE
15%RB421/HDPE
Permeation, g mm/mVday
285
0.4
3.5
0.8
% Reduction
99.9%
98.8%
99.7%
4.6.2.1.4 Coextruded barrier
One study looks at the permeation rates, using AKB test procedures, through multi-layer
fuel tanks.51 The fuel tanks in this study were 6 layer coextruded plastic tanks with EVOH as the
barrier layer (3% 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 4.6-10.
Table 4.6-10: 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%
4.6.2.1.5 Alternative Materials
Permeation can also be reduced from fuel tanks by constructing them out of a lower
permeation material than HDPE. For instance, an that would reduce permeation is the use of
metal fuel tanks because gasoline does not permeate through metal. In addition, 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
Table 4.6-14, nylon would result in about a 98 percent reduction in permeation compared to
HDPE for gasoline. However, for a 10 percent ethanol blend, this reduction would only be about
40-60 percent depending on the grade of nylon. For a 15 percent methanol blend, the permeation
4-90
-------
Chapter 4: Feasibility of Proposed Standards
would actually be several times higher through nylon than HDPE.
Other materials, which have excellent permeation 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.52 As shown in Table 4.6-14, 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 being considered for fuel tank construction
and are being evaluated for permeation resistance by the manufacturer.
4.6.2.2 Fuel Hoses
Thermoplastic fuel lines for automotive applications are generally built to SAE J2260
specifications.53 Category 1 fuel lines under this specification have permeation rates of less than
25 g/m2/day at 60°C on CM15 fuel. One thermoplastic used in automotive fuel line construction
is polyvinylidene fluoride (PVDF). Based on the data presented in Section 4.6.2.3, 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, recreational vehicle manufacturers 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. In addition, using plastic fuel line rather than rubber
hose would require the additional cost of changing hose fittings on the vehicles.
Manufacturers recommended using R9 fuel hose as a low permeation requirement. This
hose is designated under SAE recommended practice J3054 for fuel injection systems and has a
maximum permeation rate of 15 g/m2/day on ASTM Fuel C. On a fuel containing an alcohol
blend, permeation would likely be much higher from these fuel hoses. SAE J30 specifically
notes that "exposure of this hose to gasoline or diesel fuel which contain high levels, greater than
5% by volume, of oxygenates, i.e., ethanol, methanol, or MTBE, may result in significantly
higher permeation rates than realized with ASTM Fuel C." R9 hose is made with a thin low
permeation barrier sandwiched between layers of rubber. A typical barrier material used in this
construction is FKM. Based on the data presented in Section 4.2.8.3 for FKM, the permeation
rate is 3-5 times higher on Fuel CE10 than Fuel C. Therefore, a typical R9 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.
SAE J30 also designates Rl 1 and R12 hose which are intended for use as low permeation
fuel feed and return hose. Rl 1 has thee classes known as A, B, and C. Of these, Rl 1-A has the
lowest permeation specification which is a maximum of 25 g/m2/day at 40°C on CM15 fuel.
Because permeation rates are generally higher on CM15 than CE10 and because they are 2-4
times higher at 40°C than at 23°C, hose designed for this specification would likely meet our
4-91
-------
Draft Regulatory Support Document
permeation requirement. R12 hose has a permeation requirement of 100 g/m2/day at 60°C on
CM15 fuel. This is roughly equivalent in stringency as the Rl 1-A permeation requirement.
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 hose55 is
labeled by General Motors as "construction 6" which is a multilayer hose with an inner layer of
THV sandwiched in inner and outer layers of a rubber known as ECO.CC A hose of this
construction would have less than 8 g/m2/day at 40°C when tested on CE10. In look and
flexibility, this hose is not significantly different than the SAE J30 R7 hose generally used in
recreational vehicle applications.
Permeation data on several low permeation hose designs were provided to EPA by an
automotive fuel hose manufacturer.56 This hose, which is as flexible as R9 hose, was designed
for automotive applications and is available today. Table 4.6-11 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 wieking. The
permeation testing was performed using the ARB 18-41°C diurnal cycle using a fuel with a 10
percent ethanol blend (E10).
Table 4.6-11: 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 E10 fuel. Although hose using THV 800 is available, it is produced for
automobiles that will need to meet the tighter evaporative emission requirements in the
upcoming Tier 2 standards. Hose produced in mass quantities today uses THV 500. This hose is
less expensive and could be used to meet the recreational vehicle permeation requirements.
Table 4.6-12 presents information comparing hose using THV 500 with the hose described above
using THV 800 as a barrier layer.57 In addition, this data shows that permeation rates more than
double when tested on CE10 versus Fuel C. One recreational vehicle manufacturer has
expressed concern to EPA that this hose may be too stiff to stay on the fuel line and fuel tank
connectors without clamps as does their current fuel line. If a manufacturer opts to use this or a
THV = tetrafluoroethylene hexafluoropropylene, ECO = epichlorohydrin/ethylene
oxide
4-92
-------
Chapter 4: Feasibility of Proposed Standards
similar line, this problem will need to be resolved either through further testing, a change to the
connector geometry, the use of an adhesive, or the use of one of any of several of different types
of clamps.
Table 4.6-12: 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 a section of R9 hose and a
section of automotive vent line hose for permeation.58 These 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. We
purchased the R9 hose (which was labeled as such) from a local auto parts store. According to
this testing, the R9 hose is well below the SAE specification of 15 g/m2/day. In fact, it meets this
limit on Fuel CE10 as well. The automotive vent line showed similar results. This data is
presented in Table 4.6-13.
Table 4.6-13: Test Results on Commercially Available Hose Samples (g/m2/day)
Hose Sample
R9
Automotive vent line
FuelC
10.1
10.9
Fuel CE10
12.1
9.0
4.6.2.3 Material Properties
This section 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 4.6-14 gives an overview
of the fuel systems materials included in the data set. Tables 4.6-15 through 4.6-18 present
permeation rates using Fuel C, a 10% ethanol blend (CE10), and a 15% methanol blend (CE15)
for the test temperatures of 23, 40, 50, and 60°C.
4-93
-------
Draft Regulatory Support Document
Table 4.6-14: Fuel System Materials
Material Name
HOPE
Nylon 12
EVOH
Polyacetal
PBT
PVDF
NBR
HNBR
FVMQ
FKM
FEE
PFA
Carilon
HOPE
LDPE
Celcon
THV
E14659
E14944
ETFE
GFLT
FEP
PTFE
FPA
Composition
high-density polyethylene
thermoplastic
ethylene vinyl alcohol, thermoplastic
thermoplastic
polybutylene terephthalate, thermoplastic
polyvinylidene fluoride, fluorothermoplastic
nitrile rubber
hydrogenated nitrile rubber
flourosilicone
fluoroelastomer
fluorothermoplastic
fluorothermoplastic
aliphatic poly-ketone thermoplastic
high density polyethylene
low density polyethylene
acetal copolymer
tetra-fluoro-ethylene, hexa-fluoro-propylene, vinyledene fluoride
fluoropolymer film
fluoropolymer film
ethylene-tetrafluoro-ethylene, fluoroplastic
fluoroelastomer
fluorothermoplastic
polytetrafluoroethylene, fluoroplastic
copolymer of tetrafluoroethylene and perfluoroalkoxy monomer
4-94
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.6-15: Fuel System Material Permeation Rates at 23°C by Fuel Type
59,60,61,62,63
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 FEE 1000L
Teflon PFA1000LP
Tefzel ETFE 1000LZ
Nylon 12 (GM grade)
Nitrite
FKM
FE 5620Q (65.9% fluorine)
FE 5840Q (70.2% fluorine)
PTFE
ETFE
PFA
THV 500
FuelC
g-mm/mVday
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/mVday
_
-
-
-
-
-
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/mVday
35
64
10
3.1
0.4
0.2
1188
828
635
36
32
60
12
3.0
14
44
8
0.03
0.13
0.20
83
1150
-
-
-
0.08*
0.04*
0.05*
0.3
tested on CM20.
4-95
-------
Draft Regulatory Support Document
Table 4.6-16: Fuel System Material Permeation Rates at 40°C by Fuel Type
64,65
Material Name
Carilon
EVOH-F101
EVOH-XEP380
HOPE
LDPE
Nylon 12 (L2101F)
Nylon 12 (L2140)
Celcon
DyneonE14659
DyneonE14944
ETFE Aflon COP
m-ETFE
ETFE Aflon LM730 AP
FKM-70 16286
GFLT 19797
Nitrite
FKM
FE 5620Q (65.9% fluorine)
FE 5840Q (70.2% fluorine)
THV-310X
THV-500
THV-610X
FuelC
g-mm/nf/day
0.06
0.0001
O.OOOl
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/nf/day
1.5
0.013
-
69
350
28
44
2.7
—
0.67
-
0.79
35
38
1540
86
40
12
—
-
CM15
g-mm/nf/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 4.6-17: Fuel System Material Permeation Rates at 50°C by Fuel Type
66
Material Name
Carilon
HOPE
Nylon 12 (L2140)
Celcon
ETFE Afcon COP
FKM-70 16286
GFLT 19797
FuelC
g-mm/nf/day
0.2
190
4.9
0.76
-
25
28
Fuel CE10
g-mm/nf/day
3.6
150
83
5.8
1.7
79
77
CM15
g-mm/nf/day
_
—
-
-
-
—
-
4-96
-------
Chapter 4: Feasibility of Proposed Standards
Table 4.6-18: Fuel System Material Permeation Rates at 60°C by Fuel Type 6768 697°
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/nf/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/nf/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/nf/day
_
-
-
-
-
-
-
-
-
38
35
20
22
9.0
6.5
6.0
1.1
4.6.3 Test Procedures
4.6.3.1 Fuel Tanks
Essentially, two options may be used to test fuel tanks for certification. The first option is
to perform all of the durability tests on a fuel tank and then test the permeation rate. The second
option is to test a fuel tank that has been preconditioned and adjust the results using a
deterioration factor. The deterioration factor would need to be based on testing of that tank or a
similar tank unless you can use good engineering judgment to apply the results of previous
durability testing with a different fuel system. Figure 4.6-2 provides flow charts for these two
options.
4.6.3.1.1 Option 1: full test procedure
Under the first option, the fuel tank is tested both before and after a series of durability
tests. We estimate that this test procedure would take about 49 weeks to complete. Prior to the
first 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 baseline 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 2 weeks.
4-97
-------
Draft Regulatory Support Document
To determine a permeation emission deterioration factor, we are specifying 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.71 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 final permeation 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.
For example, a fuel tank that is only used in vehicles where an outer shell prevents the tank from
being exposed to sunlight may not benefit from UV testing.
After the durability testing, once the permeation rate has stabilized, the fuel tank is
drained and refilled with E10, 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 would be used for the emission rate from this fuel
tank. The difference between the baseline and final permeation rates would be used to determine
a deterioration factor for use on subsequent testing of similar fuel tanks.
4.6.3.1.2 Option 2: base test with DF
Under the second option, the fuel tank is tested for baseline permeation only, then a
deterioration factor (DF) is applied. We estimate that this test procedure would take about 22
weeks to complete. As with Option 1 baseline testing, 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 baseline 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 2 weeks.
The final permeation rate is then determined by applying a DF to the baseline permeation
4-98
-------
Chapter 4: Feasibility of Proposed Standards
rate. The DF, in units of g/m2/day, is added to the baseline permeation rate. This DF must be
determined with testing on a fuel tank in the same emission family.
4.6.3.2 Fuel Hoses
The permeation rate from fuel hoses would be measured at a temperature of 23 °C ± 2 °C
over a period of at least 2 weeks. A longer period may be necessary for an accurate measurement
for hose with low permeation rates. Permeation would be measured through the weight loss
technique described in SAE J30.72 The hose must be preconditioned with a fuel soak to ensure
that the permeation rate has stabilized. Based on times to achieve equilibrium for permeation
measurement described in SAE J226073 for automotive fuel lines, and adjusting for temperature
and test fuel type, we estimate a minimum soak time of 4 weeks. The fuel used for this testing
would be a blend of 90 percent gasoline and ten percent ethanol. This fuel is consistent with the
test fuel used for on-highway evaporative emission testing.
4.6.4 Conclusion
We believe that manufacturers will be able to meet the fuel tank permeation requirements
through several design strategies that include sulfonation, fluorination, barrier platelets, and
coextruded barriers. Our cost analysis, presented in Chapter 5, indicates that sulfonation would
likely be the most attractive technology. However, conversations with manufacturers have
revealed interest in each of these low permeation strategies. We believe the data presented above
supports a final standard which requires about an 85% reduction in permeation, compared
baseline FtDPE fuel tanks, throughout the useful life of the recreational vehicle.
As discussed above, fuel hose is available today that meets the permeation requirements
for recreational vehicles. Low permeation hose was generally developed for automotive
applications; however, we believe that this fuel hose can be used in recreational vehicle
applications. Even assuming that new hose clamps would be required, our analyses in Chapters 5
and 6 show that the low permeation hose would be inexpensive yet effective.
4.6.5 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 permeation standards for recreational vehicles. In
this case, we would not expect evaporative emission controls to have any impact on noise from a
vehicle because noise from the fuel system is insignificant.
We anticipate that permeation emission standards will have a positive impact on energy.
By capturing or preventing the loss of fuel through permeation, we estimate that the average
lifetime fuel savings will be 11.8 gallons for snowmobiles, 5.4 gallons for off-highway
motorcycles and 6.5 gallons for all-terrain vehicles. This translates to a fuel savings of about 12
million gallons in 2030 when most recreational vehicles used in the U.S. are expected to have
permeation emission control.
4-99
-------
Draft Regulatory Support Document
We believe that permeation emission standards will have no negative impacts on safety,
and may even have some benefits due to the reduction of fuel vapor around a recreational
vehicle.
Figure 4.6-2: Flow Chart of Fuel Tank Permeation Certification Test Options
1: Full Test Procedure
2: Base Test with DP
baseline
permeation
test
ElOfuel
28±2C
2 weeks
Pressure Cycling
1000x
-0.5 to 2.0 psi
1 week**
UV Testing
0.4
mW-hr/m2/min
4 weeks
Slosh Testing
1 million cycles
±15 degree angle
ElOfuel
7 weeks**
adjust baseline
test result with DF
to determine
certificaiton level
use final permeation
test result for
certificaiton
* The deterioration factor
(DF) is the difference
between the baseline and
final permeation tests in
Option 1
** EPA estimate
4-100
-------
Chapter 4: Feasibility of Proposed Standards
Appendix to Chapter 4: Emission Index For Recreational Vehicle Hangtags
Sectionl051.135(g) specifies that recreational vehicles should have consumer labels that
show the emission characteristics of the vehicle using a normalized zero to ten index. The index
is called a nonroad emission rating (NER). This appendix describes the derivation of those
indices. The primary indices were derived based on four general principles:
The index should be simple for the consumer to use.
A vehicle with the highest emissions allowed or expected under the regulations should
have a value often.
A vehicle with emissions equal to the average standard should be in the middle of the
range. (For categories with two phases, a vehicle with emissions equal to the average
Phase 2 standard under should be approximately five.)
Each index should allow for vehicles that are significantly cleaner than the average. The
indices should also work without adjustment if we were to establish more stringent
standards in the future.
As described below, we applied these principles separately to each of the categories, considering
the baseline emissions, PEL caps, average standards, and current and future technology options.
In general, since the recreational vehicle programs are designed to allow different technology
options, we believe that a logarithmic scale in generally appropriate. However, in some cases, a
linear scale is more appropriate for all or part of the index. In some cases, it may be possible to
have emissions high enough to calculate the NER as eleven or higher. In those cases, the
regulations specify that the vehicle should be labeled as a ten.
4A.I Snowmobiles
The index for snowmobiles uses a single log-linear curve to convert HC and CO
emissions into normalized values between zero and ten. HC and CO emissions are weighted
based on baseline values so that a 50 percent reduction in HC emissions is equivalent to a 50
percent reduction in CO emissions. (The ratio of baseline CO emissions to baseline HC
emissions is 400:150, or 2.667.) The following equation gives a value often for vehicles with
HC emissions of 150 g/kW-hr and CO emissions 400 g/kW-hr; and a value of five for vehicles
with HC emissions of 75 g/kW-hr and CO emissions 200 g/kW-hr:
NER = 16.61xlog(2.667#C +C<9)-38.22
4-101
-------
Draft Regulatory Support Document
Snowmobile Equation
800
4A.2 Off-highway Motorcycles
The index for off-highway motorcycles uses a combination of a linear curve and a log-
linear curve to convert HC+NOx emissions into normalized values between zero and ten. The
following linear equation, which applies for vehicles with below average emissions gives a value
of five for vehicles with HC+NOx emissions of 2.0 g/km:
NER =2.500(#C + NOx)
The following log-linear equation, which applies for vehicles with above average emissions gives
a value often for vehicles with HC+NOx emissions of 20 g/km; and a value of five for vehicles
with HC+NOx emissions of 2.0 g/km:
NER = 5.000 xlog(#C + NOx) + 3.495
It was necessary to use a linear equation for the lower part of the curve to allow for more
gradations just below the average, and fewer for very low levels. For example, using the log
equation, it would have been necessary to have emission below 1.0 g/km to get an emission
rating that would round to three, while with the linear equation, it would only be necessary to
have emissions below 1.4 g/km to get an emission rating that would round to three.
4-102
-------
Chapter 4: Feasibility of Proposed Standards
Motorcycle Equation
100
10
6
O
I
0.1
4A.3 ATVs (g/km)
The primary index for ATVs uses a combination of a linear curve and a log-linear curve
to convert HC+NOx emissions into normalized values between zero and ten. The following
linear equation, which applies for vehicles with below average emissions gives a value of five for
vehicles with HC+NOx emissions of 1.5 g/km:
NER = 3.333(HC + NOx)
The following log-linear equation, which applies for vehicles with above average emissions gives
a value often for vehicles with HC+NOx emissions of 20 g/km; and a value of five for vehicles
with HC+NOx emissions of 1.5 g/km:
NER = 4.444 x\og(HC + NOx} + 4.2\1
It was necessary to use a linear equation for the lower part of the curve to allow for more
gradations just below the average, and fewer for very low levels. For example, using the log
equation, it would have been necessary to have emission below 0.7 g/km to get an emission
rating that would round to three, while with the linear equation, it would only be necessary to
have emissions below 1.1 g/km to get an emission rating that would round to three.
where HC +NOx is the cycle-weighted emission rates for hydrocarbons plus oxides of nitrogen in
g/km.
4-103
-------
Draft Regulatory Support Document
ATV Equation
100
10
X
O
o
X
0.1
10
4A.4 ATVs (g/kW)
There are two cases in which we allow ATVs to certify to g/kW emission standards based
on engine testing: ATVs less than 100 cc, and ATVs built before 2009. We developed separate
equations for these cases, based on the same general principles as for other ATVs. In developing
these equations, we considered FEL caps, average standards, test cycle issues, and the available
technology options. The following linear equation, applies for ATV with engine smaller than
lOOcc:
0.250
The following log-linear equation, applies for larger ATVs certified under the interim engine
testing option:
NER = 9.898 x log(#C + N(9x)-4.898
4-104
-------
Chapter 4: Feasibility of Proposed Standards
Chapter 4 References
1. "Emission Testing of Nonroad Compression Ignition Engines," prepared by Southwest
Research Institute for the U.S. EPA, SwRI 6886-802, September 1995, Docket A-2000-01,
Document II-A-20.
2. Memorandum from Mike Brand, Cummins, to Bill Charmley, U.S. EPA, "Draft Report on
Emission Testing oj'Nonroad Compression Ignition Engines" November 13, 1995, Docket A-
2000-01, Document II-A-22.
3. Letter from Jeff Carmody, Santa Barbara County Air Pollution Control District, to Mike
Samulski, U.S. EPA, "Marine Engine Replacement Programs," July 21, 1997, Docket A-2000-
01, Document U-A-21.
4. Facsimile from Eric Peterson, Santa Barbara County Air Pollution Control District, to Mike
Samulski, U.S. EPA, "Data on Mercury 4.2 and 2.8 Liter Engines," April 1, 1998, Docket A-
2000-01, Document II-A-24.
5. Smith, M., "Marine Diesel Engine Testing," Prepared by Southwest Research Institute for the
U.S. EPA, Contract # 68-C-98-169, WA 0-7, September 1999, Docket A-2000-01, Document II-
A-26.
6. Data Submission by the Engine Manufacturers Association, July 19, 2000, Docket A-2000-
01, Document H-B-11.
7. International Organization for Standardization, 8178-4, "Reciprocating internal combustion
engines—Exhaust emission measurement—Part 4: Test cycles for different engine applications,"
Docket A-2000-01, Document II-A-19.
8. "Control of Emissions of Air Pollution from New Marine Compression-Ignition Engines at or
Above 37 kW; Final Rule," 64 FR 73300, December 29, 1999.
9. Wilbur, C., "Marine Diesel Engines," Butterworth & Heinemann Ltd, 1984.
10. "Data Collection and Analysis of Real-World Marine Diesel Transient Duty-Cycles," EPA
memo from Matt Spears to Mike Samulski, October 15, 1999, Docket A-2000-01, Document U-
B-09.
11. Memorandum from Mark Wolcott to Charles Gray, "Ambient Temperatures Associated with
High Ozone Concentrations," U.S. Environmental Protection Agency, September 6, 1984,
Docket A-2000-01, Document U-B-06.
4-105
-------
Draft Regulatory Support Document
12. Smith, M., "Marine Diesel Engine Testing," Prepared by Southwest Research Institute for
the U.S. EPA, Contract # 68-C-98-169, WA 0-7, September 1999, Docket A-2000-01,
Document II-A-26.
13. SAE J1937 (reaffirmed JAN1995), "Engine Testing with Low-Temperature Charge Air-
Cooler Systems in a Dynamometer Test Cell," SAE Recommended Practice, Docket A-2000-01,
Document II-A-62.
14. Annex VI of MARPOL 73/78, "Technical code of control of Emissions of Nitrogen Oxides
for Marine Diesel Engines," October 22, 1997, Docket A-2000-01, Document U-A-25.
15. "Emission Factors for Compression Ignition Nonroad Engines Operated on No. 2 Highway
and Nonroad Diesel Fuel," U.S. Environmental Protection Agency, EPA420-R-98-001, March
1998, Docket A-2000-01, Document IV-A-73.
16. "Emission Data and Procedures for Large SI Engines," EPA memorandum from Alan Stout
and Chuck Moulis to Docket A-2000-01, January 2, 2001, EPA420-F-00-050, Document
II-B-05.
17. "Evaluation of Emissions Durability of Off-Road LPG Engines Equipped with Three-Way
Catalysts," by Vlad Ulmet, Southwest Research Institute, SwRI 08.03661 & 08.03377,
November 2000, (Docket A-2000-01, document U-A-07). This document is available through
the National Technical Information Society at (703) 605-6000 (order number PB2002-101191).
18. The Mazda and GM engines are from SwRI 08.03661 & 08.03377, November 2000 (Docket
A-2000-01, document U-A-07). Engine B and Engine E are from "Three-Way Catalyst
Technology For Off-Road Equipment Powered by Gasoline and LPG Engines," by Jeff White et
al, Southwest Research Institute, SwRI 8778, April 1999, (Docket A-2000-01, document
II-A-08).
19."Durability Experience with Electronic Controlled CNG and LPG Engines," A. Lawson et al.,
February 2, 2000, Docket A-2000-01, Document U-D-2.
20."Exhaust Controls Available to Reduce Emissions from Nonroad Heavy-Duty Engines," by
Kevin Brown, Engine Control Systems, in Clean Air Technology News, Winter 1997, Docket A-
2000-01, Document II-A-2.
21. "Case Study: The Results of EVIPCO's GM 3.0 liter Certified Engine Program," presented by
Josh Pietak, February 2, 2000, A-2000-01, Document II-D-11.
22. See SwRI 08.03661 & 08.03377, November 2000 for a further description of the catalyst
damage observed, (Docket A-2000-01, document U-A-07).
23."Emission Modeling for Recreational Vehicles," EPA memorandum from Line Wehrly to
Docket A-98-01, November 13, 2000, Document IV-B-06.
4-106
-------
Chapter 4: Feasibility of Proposed Standards
24. "Development and Validation of a Snowmobile Engine Emission Test Procedure," Jeff J.
White, Southwest Research Institute and Christopher W. Wright, Arctic Cat, Inc., SAE paper
982017, September, 1998, Docket A-2000-01, Document U-A-66.
25. 61 FR 52088, October 4, 1996.
26. Letter to Line Wehrly from Phillip D. McDowell, Docket A-2000-01, Document IV-D-184.
27. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data
reports available in Docket A-2000-01, Document IV-A-09.
28. "Permeation Rates of Small Off Road Engine High-Density Polyethylene Fuel Tanks (April
2001 Testing), June 8, 2001, California Air Resources Board, Docket A-2000-01, Document IV-
A-101.
29. "Permeation Rates of Small Off Road Engine High-Density Polyethylene Fuel Tanks
(February 2001 Testing), June 8, 2001, California Air Resources Board, Docket A-2000-01,
Document IV-A-100.
30. "Permeation Rates of High-Density Polyethylene Fuel Tanks (June 2001), June 12, 2001,
California Air Resources Board, Docket A-2000-01, Document IV-A-99.
31. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early
Container Data," August 29, 2002, Docket A-2000-01, Docket No. IV-A-103.
32. Lockhart, M., Nulman, M., Rossi, G., "Estimating Real Time Diurnal Permeation from
Constant Temperature Measurements," SAE Paper 2001-01-0730, 2001, Docket A-2000-01,
Document IV-A-21.
33. Hopf, G., Ries, H., Gray, E., "Development of Multilayer Thermoplastic Fuel Lines with
Improved Barrier Properties," SAE Paper 940165, 1994, Docket A-2000-01, Document IV-A-22.
34. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of
Polymeric Materials," SAE Paper 2001-01-1999, 2001, Docket A-2000-01, Document IV-A-23.
35. SAE Recommended Practice J30, "Fuel and Oil Hoses,"June 1998, Docket A-2000-01,
Document IV-A-92.
36. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data
reports available in Docket A-2000-01, Document IV-A-09.
37. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early
Container Data," August 29, 2002, Docket A-2000-01, Docket No. IV-A-103.
38. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine
Fuel Tanks," California Air Resources Board, June 21, 2002, Docket A-2000-01, Document IV-
4-107
-------
Draft Regulatory Support Document
A-77.
39. Conversation between Mike Samulski, U.S. EPA and Tom Schmoyer, Sulfo Technologies,
June 17, 2002.
40. "Resin and Additives - SOS Compatible," Email from Tom Schmoyer, Sulfo Technologies
to Mike Samulski and Glenn Passavant, U.S. EPA, June 19, 2002, Docket A-2000-01, Document
IV-A-40.
41. Email from Jim Watson, California Air Resources Board, to Mike Samulski, U.S. EPA,
"Attachment to Resin List," August 30, 2002, Docket A-2000-01, Document IV-A-102.
42. 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 A-2000-01, Document IV-A-76.
43. 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 A-2000-01, Document II-A-60.
44. www.arb.ca.gov/msprog/spillcon/reg.htm, Updated March 26, 2001, Copy of linked data
reports available in Docket A-2000-01, Document IV-A-09.
45. "Permeation Rates of Blitz Fluorinated High Density Polyethylene Portable Fuel
Containers," California Air Resources Board, April 5, 2002, Docket A-2000-01, Document IV-
A-78.
46. Email from Jim Watson, California Air Resources Board, to Phil Carlson, U.S. EPA, "Early
Container Data," August 29, 2002, Docket A-2000-01, Docket No. IV-A-103.
47. www.pensteel.co.uk/light/smp/fluorination.htm. A copy of this site is available in Docket
A-2000-01, Document IV-A-86.
48. "Durability Testing of Barrier Treated High-Density Polyethylene Small Off-Road Engine
Fuel Tanks," California Air Resources Board, June 21, 2002, Docket A-2000-01, Document IV-
A-77.
49. 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 A-2000-01, Document II-A-60.
50. "Selar RB Technical Information," Faxed from David Zang, Dupont, to Mike Samulski, U.S.
EPA on May 14, 2002, Docket A-2000-01, Document IV-A-88.
51. Fead, E., Vengadam, R., Rossi, G., Olejnik, A., Thorn, J., "Speciation of Evaporative
Emissions from Plastic Fuel Tanks," SAE Paper 981376, 1998, Docket A-2000-01, Document
4-108
-------
Chapter 4: Feasibility of Proposed Standards
IV-A-89.
52. E-mail from Alan Dubin, Ticona, to Mike Samulski, U.S. EPA, "Fuel Permeation Chart and
Aggressive Fuels Brochure," July 31, 2002, Docket A-2000-01, Document IV-A-97.
53. SAE Recommended Practice J2260, "Nonmetallic Fuel System Tubing with One or More
Layers," 1996, Docket A-2000-01, Document IV-A-18.
54. SAE Recommended Practice J30, "Fuel and Oil Hoses," June 1998, Docket A-2000-01,
Document IV-A-92.
55. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, June 17, 2002, Docket A-2000-01, Document IV-E-31.
56. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, August 6, 2002, Docket A-2000-01, Document IV-E-33.
57. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, August 6, 2002, Docket A-2000-01, Document IV-E-33.
58. Akron Rubber Development Laboratory, "TEST REPORT; PN# 49503," Prepared for the
U.S. EPA, September 3, 2002, Docket A-2000-01, Document IV-A-106.
59. Hopf, G., Ries, H., Gray, E., "Development of Multilayer Thermoplastic Fuel Lines with
Improved Barrier Properties," SAE Paper 940165, 1994, Docket A-2000-01, Document IV-A-22.
60. Stahl, W., Stevens, R., "Fuel-Alcohol Permeation Rates of Fluoroelastomers, Fluoroplastics,
and Other Fuel Resistant Materials," SAE Paper 920163, 1992, Docket A-2000-01, Document
IV-A-20.
61. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, June 17, 2002, Docket A-2000-01, Document IV-E-31.
62. Goldsberry, D., "Fuel Hose Permeation of Fluoropolymers," SAE Paper 930992, 1993,
Docket A-2000-01, Document IV-A-91.
63. Tuckner, P., Baker, J., "Fuel Permeation Testing Using Gravimetric Methods," SAE Paper
20001-01-1096, 2000, Docket A-2000-01, Document IV-A-96.
64. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of
Polymeric Materials," SAE Paper 2001-01-1999, 2001, Docket A-2000-01, Document IV-A-23.
65. Duchesne, D., Hull, D., Molnar, A., "THV Fluorothermoplastics in Automotive Fuel
Management Systems," SAE Paper 1999-01-0379, 1999, Docket A-2000-01, Document IV-A-
90.
4-109
-------
Draft Regulatory Support Document
66. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of
Polymeric Materials," SAE Paper 2001-01-1999, 2001, Docket A-2000-01, Document IV-A-23.
67. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of
Polymeric Materials," SAE Paper 2001-01-1999, 2001, Docket A-2000-01, Document IV-A-23.
68. "Visit to Dyneon on June 12, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, June 17, 2002, Docket A-2000-01, Document IV-E-31.
69. 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, Document JJ-B-30.
70. Duchesne, D., Hull, D., Molnar, A., "THV Fluorothermoplastics in Automotive Fuel
Management Systems," SAE Paper 1999-01-0379, 1999, Docket A-2000-01, Document IV-A-
90.
71. Draft SAE Information Report J1769, "Test Protocol for Evaluation of Long Term
Permeation Barrier Durability on Non-Metallic Fuel Tanks," Docket A-2000-01, Document
IV-A-24.
72. SAE Recommended Practice J30, "Fuel and Oil Hoses," June 1998, Docket A-2000-01,
Document IV-A-92.
73. SAE Recommended Practice J2260, "Nonmetallic Fuel System Tubing with One or More
Layers," 1996, Docket A-2000-01, Document IV-A-18.
4-110
-------
Chapter 5: Costs of Control
Chapter 5: Costs of Control
This chapter describes our approach to estimating the cost of complying with emission
standards. We start with a general description of the approach to estimating 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.
5.1 Methodology
We developed the costs for individual technologies using information provided by ICF,
Incorporated and Arthur D. Little, as cited below with further consideration to any information
provided in the public comments. The technology characterization and cost figures reflect our
current best judgment based on engineering analysis, information from manufacturers, and the
published literature. The analysis combines cost figures including markups to the retail level.
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 manufacturers' overhead and profit.1
For technologies sold by a supplier to the engine manufacturers, an additional 29 percent markup
is included for the supplier's overhead and profit. All costs are in 2001 dollars.
The analysis presents an estimate of costs that will occur in the first year 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
therefore no longer part of the cost calculation. Second, manufacturers are expected to learn over
time to produce the engines with the new technologies at a lower cost. Because of relatively low
sales volumes, manufacturers are less likely to put in the extra R&D effort for low-cost
manufacturing. Learning will occur in two basic ways. As manufacturers produce more units,
they will make improvements in production methods to improve efficiency. One example of this
is automation. The second way learning occurs is materials learning where manufacturers reduce
scrap. Scrap includes units that are produced but rejected due to 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. Consistent with analyses from other programs, we reduce estimated
variable costs by 20 percent beginning with the third year of production and an additional 20
percent beginning with the sixth year of production.2
We believe it is appropriate to apply this factor here, given that the industries are facing
emission regulations 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
5-1
-------
Draft Regulatory Support Document
the standards contained in the Final Rule. In cases where manufacturers have used certain
technologies, such as with 4-stroke engines, they have not been required to meet standards. They
will be manufacturing new 4-stroke engines or purchasing and installing 4-stroke engines in new
models. Learning will likely occur for these models. Some manufacturers, especially in the
youth ATV market do not have experience with 4-stroke engines. Also, the 4-strokes will need
to be made to meet emissions standards. We believe that learning for these models will continue
to take place.
Many of the engine technologies available to 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, as we describe for each type of engine below.
Even though the analysis does not reflect all the possible technology variations and
options that are available to manufacturers, we believe the projections presented here provide
cost estimates 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.
5.2 Cost of Emission Controls by Engine/Vehicle Type
5.2.1 Recreational Marine Diesel Engines
We have developed cost estimates for diesel engine technologies for several different
applications in a series of reports.3'4'5 This analysis adapts these existing cost estimates for
recreational marine diesel engines with separate estimates for three different sizes of engines.
Recreational marine diesel engines invariably have counterpart engine models used for
commercial application. Manufacturers will design, certify, and manufacture these commercial
models to meet emission standards. The analysis projects that manufacturers will comply with
the new emission standards generally by applying the same technologies for both commercial and
recreational engines. The remaining effort to meet emission standards with the recreational
models is therefore limited to applying new or improved hardware and conducting sufficient
R&D to integrate the new technologies into marketable products. The analysis therefore does not
consider fixed costs to develop the individual technologies separately.
One area where recreational engine designs differ is in turbocharging and aftercooling.
To reach peak performance, recreational engines typically already use optimized turbochargers
and seawater aftercooling, which offer the greatest potential for controlling NOx emissions.
5-2
-------
Chapter 5: Costs of Control
We estimate the total cost impact of new emission standards by considering the cost of
each of the anticipated technologies. The following paragraphs describe these technologies and
their application to recreational marine engines. The analysis then combines these itemized costs
into a composite estimate for the range of marine engines affected by the rulemaking.
Table 5.2.1-1 also includes information on product offerings and sales volumes, which is
needed to calculate amortized fixed costs for individual engines. Estimated sales and product
offerings were compiled from the PSR database based on historical 1997 information.
Table 5.2.1-1
Recreational Marine Diesel Engine Categories for Estimating Costs
Engine Power
Ranges (kW)
37 - 225
225 - 560
560 +
Nominal Engine
Power (kW)
100
400
750
Annual
Sales
11,600
3,560
397
Models
17
15
6
Average Sales
per Model
675
250
70
Manufacturers are expected to develop engine technologies not only to reduce emissions,
but also to improve engine performance. While it is difficult to take into account the effect of
ongoing technology development, EPA is concerned that assessing the full cost of the anticipated
technologies as an impact of new emission standards inappropriately excludes from consideration
the expected benefits for engine performance, fuel consumption, and durability.dd Short of
having sufficient data to predict the future with a reasonable degree of confidence, we face the
need to devise an alternate approach to quantifying the true impact of the new emission
standards. As an attempt to take this into account, we present the full cost of the control
technologies in this chapter, then apply an adjustment to some of these costs for calculating the
cost-per-ton of the emission standards, as described in Chapter 7.
5.2.1.1 Fuel Injection Improvements
All engines are expected to see significant improvements in their fuel injection systems.
The smaller engines will likely undergo incremental improvements to existing unit injector
designs. The analysis projects that engines rated over 600 kW will use common rail injection
technology, which greatly increases the flexibility of tailoring the injection timing and profile to
varying modes of operation. Better control of injection timing and increased injection pressure
contribute to reduced emissions. Table 5.2.1-2 shows the estimated costs for these fuel injection
improvements.
ddWhile EPA does not anticipate widespread, marked improvements in fuel consumption,
small improvements on some engines may occur.
5-3
-------
Draft Regulatory Support Document
Table 5.2.1-2: Fuel Injection Improvements
Component costs
Assembly, markup, and warranty
Composite Unit Cost
lOOkW
$63
$32
$95
400 kW
$98
$46
$144
750 kW
$205
$59
$264
5.2.1.2 Engine Modifications
Manufacturers will be optimizing basic engine parameters to control emissions while
maintaining performance. Such variables include routing of the intake air, piston crown
geometry, and placement and orientation of injectors and valves. Most of these variables affect
the mixing of air and fuel in the combustion chamber. Small changes in injection timing are also
considered in this set of modifications. We expect, however, that manufacturers will complete
this work for commercial marine diesel engines, so that the remaining effort will be focused on
fine-tuning designs for turbocharger matching and other calibration-related changes. Fixed costs
are amortized over a five-year period, using the sales volumes developed in Table 5.2.1-1, with
forward discounting incorporated to account for manufacturers incurring these costs before the
emission standards begin to apply. Table 5.2.1-3 shows the estimated per-engine costs for these
modifications. These costs include the consideration manufacturers must give to offsetting any
crankcase emissions routed to the exhaust. There is no estimated long-term cost to the engine
modifications because manufacturers can fully recover the fixed costs, and we don't expect any
increase in variable costs as a result of these improvements.
Table 5.2.1-3: Engine Modifications
Total fixed costs
Fixed cost per engine
Composite Unit Cost
100 kW
$200,000
$72
$72
400 kW
$200,000
$195
$195
750 kW
$200,000
$697
$697
As described in the preamble to the final rule, the manufacturers are responsible to
comply with emissions at any speed and load that can occur on a vessel. We believe that is not
appropriate to consider additional costs for manufacturers to comply with these "off-cycle"
requirements. This is because we expect that manufacturers can manage engine operation to
avoid unacceptable variation in emission levels by more effectively using the technologies that
will be used to meet the emission limits more broadly, rather than by use of additional hardware.
For example, manufacturers can adjust fuel injection parameters to avoid excessive emissions.
The split-zone approach described in Chapter 4 is designed to accommodate normal variation in
5-4
-------
Chapter 5: Costs of Control
emission levels at different operating points. This approach involves no additional variable cost.
The estimated R&D expenditures reflect the time needed to address this.
5.2.1.3 Certification and Compliance
We have significantly reduced certification procedural requirements in recent years, but
manufacturers are nevertheless responsible for generating the necessary test data and other
information to demonstrate compliance with emission standards. Table 5.2.1-4 lists the expected
costs for different sizes of engines, including the amortization of those costs over five years of
engine sales. Estimated certification costs are based on two engine tests and $10,000 worth of
engineering and clerical effort to prepare and submit the required information.
Until engine designs are significantly changed, engine families can be recertified each
year using carryover of the original test data. Since these engines are currently not subject to any
emission requirements, the analysis includes a cost to recertify an upgraded engine model every
five years.
Costs for production line testing are summarized in Table 5.2.1-5. These costs are based
on testing 1 percent of total estimated sales, then distributing costs over the fleet. Listed costs for
engine testing presume no need to build new test facilities, since we may waive production-line
testing requirements for small-volume production. Few manufacturers, if any, will therefore
need to build new test facilities.
Table 5.2.1-4: Certification
Total fixed costs
Fixed cost per engine
Composite Unit Cost
100 kW
$30,000
$12
$12
400 kW
$30,000
$29
$29
750 kW
$40,000
$139
$139
Table 5.2.1-5: Costs for Production Line Testing
Cost per test
Testing rate
Cost per engine
100 kW
$10,000
1%
$100
400 kW
$10,000
1%
$100
750 kW
$15,000
1%
$150
5-5
-------
5.2.1.4 Total Engine Costs
These individual cost elements can be combined into a calculated total for new emission
standards by assessing the degree to which the different technologies will be deployed. As
shown in Table 5.2.1-6, estimated costs for complying with the emission standards increase with
increasing power ratings. We expect each of the listed technologies to apply to all the engines
that need to meet the new emission standards. Estimated first-year cost impacts range from $300
to $1,300 for the different engine sizes, while long-term cost estimates range from $170 to $460.
Characterizing these estimated costs in the context of their fraction of the total purchase
price and life-cycle operating costs is helpful in gauging the economic impact of the new
standards. The estimated first-year cost increases for all engines are at most 2 percent of
estimated engine prices, with even lower long-term effects, as described above.
Table 5.2.1-6: Diesel Engine Costs
Fuel injection upgrade
Engine modifications
Certification + PLT
Total Engine Cost, year 1
Total Engine Cost, year 6
lOOkW
$95
$72
$111
$278
$172
400 kW
$144
$195
$129
$468
$221
750 kW
$264
$697
$289
$1,251
$459
5.2.1.5 Marine Diesel Aggregate Costs
The above analyses developed incremental per-vessel cost recreational marine diesel
engines. Using these per-engine costs and projections of future annual sales, we have estimated
total aggregate annual costs for emission standards. The aggregate costs are presented on a cash-
flow basis, with hardware and fixed costs incurred in the year the vehicle is sold. Table 5.2.1-7
presents a summary of this analysis. As shown in the table, aggregate net costs stay between $3
million and $6 million.
Table 5.2.1-7
Summary of Annual Aggregate Costs for Marine Diesel Engines (millions of dollars)
Total Costs
2006
$6.2
2010
$7.6
2015
$2.8
2020
$3.1
2025
$3.4
To project annual sales, we started with the 1998 population estimates presented in
Chapter 6. We then used the engine turnover rates and growth estimates to calculate annual
sales. Table 5.2.1-8 provides a summary of the sales estimates used in the aggregate cost
analysis.
-------
Chapter 5: Costs of Control
Table 5.2.1-8
Estimated Annual Sales of Recreational Marine Diesel Engines
Engine Power Range (kW)
37 - 225
225 - 560
560 +
2000
11,600
3,560
397
2006
13,700
4,200
469
2010
15,200
4,620
517
2020
18,700
5,690
636
To calculate annual aggregate costs, the sales estimates have been multiplied by the per-
unit costs discussed above. These calculations take into consideration vehicle sales and
scrappage rates. The year-by-year results of the analysis are provided in Chapter 7.
5.2.2 Large Industrial Spark-Ignition Engines
We estimated the cost of upgrading LPG-fueled and gasoline-fueled Large SI engines.
We developed the costs for individual technologies in cooperation with ICF, Incorporated and
Arthur D. Little.6 The analysis combines these individual figures into a total estimated cost for
each type of engine, including markups to the retail level. A composite cost based on the mix of
engine types provides an estimated industry-wide estimate of the per-engine cost impact.
Gasoline-fueled Large SI engines continue to rely on traditional carburetor designs rather
than incorporating the automotive technology innovations introduced to address emission
controls. Since natural gas- and LPG-fueled engines use comparable technologies, the analysis
presents a single set of costs for both fuels.
The anticipated technology development is generally an outgrowth of automotive
technologies. Over the last thirty years, engineers in the automotive industry have made great
strides in developing new and improved approaches to achieve dramatic emission reductions
with high-performing engines. In more recent years, companies have started to offer these same
technologies for industrial applications. Fundamental to this technology development is the
electronically controlled fuel system and catalytic converters.
Electronically controlled fuel systems allow manufacturers to more carefully meter fuel
into the combustion chambers. This gives the design engineer an important tool to better control
power and emission characteristics over the whole range of engine operation. Careful control of
air-fuel ratio is also essential for effective catalyst conversion. The catalyst reduces the
concentration of pollutant gases in the exhaust stream. We also consider development time to
redesign the combustion chamber and intake air routing, as well as to combine the new control
technologies and optimize engine calibrations. We include these efforts under the total R&D
costs for each engine.
Gasoline engines can use either throttle-body or port-fuel injection. Manufacturers can
5-7
-------
Draft Regulatory Support Document
likely reach the targeted emission levels using simpler throttle-body systems. However, the
performance advantages and the extra assurance for full-life emission control from the more
advanced port-fuel injection systems offer a compelling advantage. The analysis therefore
projects that all gasoline engines will use port-fuel injection. The analysis does not take into
account the performance advantages of port-fuel injection and therefore somewhat overestimates
the cost impact of adopting new emission standards.
Gaseous-fuel engines have very different fuel metering systems due to the fact that LPG
and natural gas evaporate readily at typical ambient temperatures and pressures. Manufacturers
of these engines face a choice between continuing with conventional mixer technology and
upgrading to injection systems. We are aware that manufacturers are researching gaseous
injection systems, but we believe mixer technology will be sufficient to meet the standards. All
the data supporting the feasibility of emission standards for LPG engines is based on engines
using mixer technology.
5.2.2.1 Engine Technology
Tables 5.2.2-1 and 5.2.2-2 show the estimated costs of upgrading each of the engine
types. The cost figures are in the form of retail-price equivalent for an individual engine. The
tables include individual cost estimates of the various components involved in converting a
baseline engine to comply with emission standards. The cost of the catalyst is based on a
precious metal loading of 2.8 g/liter (primarily palladium, with small amounts of platinum and
rhodium) and a catalyst volume 60 percent of total engine displacement.
The analysis incorporates a cost for potential warranty claims related to the new
technologies by adding 5 percent of the increase in hardware costs. The industry has gained
enough experience with electronic fuel systems that we expect a relatively low rate of warranty
claims for them. Catalysts have been used for many years, but not in Large SI applications, so
these technologies may cause a somewhat higher rate of warranty claims.
Even without EPA emission standards, manufacturers will conduct the research and
development needed to meet the 2004 emission standards in California. The R&D impact of new
EPA standards is therefore limited to the additional burden of complying with the 2007
requirements. Estimated costs for research and development are $175,000 for each engine
family. This is based on about six months of time for an engineer and a technician on each fuel
type for each engine family. We expect initial efforts to be more extensive, but cumulative
learning should reduce per-family development costs for subsequent models. These fixed costs
are increased by 7 percent to account for forward discounting, since manufacturers incur these
costs before the new standards apply. Redesigning the first engine model will likely require
significantly more time than this, but we expect the estimated level of R&D to be appropriate as
an average level for the range of models in a manufacturer's product line.
Table 5.2.2-2 presents separate costs for water-cooled and air-cooled gasoline engines.
While many of the components are the same, the main differences include (1) a single fuel
5-8
-------
Chapter 5: Costs of Control
injector and simpler intake manifold for throttle-body injection, (2) smaller sales volume for
amortizing fixed costs, and (3) substantial fixed costs for meeting the 2004 standards. Air-
cooled engines are generally not certified already in California, largely because most applications
involving air-cooled Large SI engines are preempted from California ARB's emission standards.
To take this into account, we have added an estimate of $500,000 for R&D and $100,000 for
tooling costs per engine family. Discounting these costs forward two years and amortizing over
five years of sales results in an additional cost of $166 per air-cooled engine.
Table 5.2.2-1
Estimated Costs for an LPG-fueled Large SI Engine
Hardware Cost to Manufacturer
Regulator/throttle body
Intake manifold
Positive crankcase ventilation
Fuel filter w/ lock-off system
LPG vaporizor
Governor
Converter temperature control valve
Oxygen sensor
ECM
Wiring/related hardware
7uel system total
Catalyst/muffler
Vluffler
Total Hardware Cost
Vlarkup @ 29%
Warranty markup @5%
Total component costs
1004 Fixed costs
1004 Incremental costs
?ixed Cost to Manufacturer
1007 R&D costs
Jnits/yr.
\mortization period (7 % discounting)
1007 Fixed cost/unit
1007 Evap costs
1007 Incremental costs
Baseline
$50
$37
$15
$75
$40
$217
$45
$262
$76
$338
$0
$0
Controlled
$65
$37
$3
$15
$75
$60
$15
$19
$100
$42
$431
$229
$0
$660
$191
$20
$871
$0
$533
$175,000
2,000
5
$26
$0
$0
5-9
-------
Draft Regulatory Support Document
Table 5.2.2-2
Estimated Per-Engine Costs for Gasoline-Fueled Large SI Engines
Hardware Cost to Manufacturer
Carburetor
Injectors (each)
Number of injectors
Pressure Regulator
Fuel filter
Intake manifold
Positive crankcase ventilation
Fuel rail
Throttle body/position sensor
Fuel pump
Oxygen sensor
ECM
Governor
Air intake temperature sensor
Manifold air pressure sensor
Injection timing sensor
Wiring/related hardware
"uel system total
Catalyst/muffler
Vluffler
Total Hardware Cost
Vlarkup @ 29%
Warranty markup @5%
Total Component Costs
2004 Fixed costs
2004 Fixed cost/unit
2004 Incremental costs
Fixed Cost to Manufacturer
2007 R&D Costs
Jnits/yr.
Amortization period (7 % discounting)
2007 Fixed cost/unit
2007 Evap costs
2007 Incremental costs
Water-cooled
Baseline Controlled
$51 $0
$17
4
$11
$3 $4
$35 $50
$3
$13
$60
$15 $30
$19
$150
$40 $60
$5
$11
$12
$42
$144 $538
$229
$45
$189 $767
$55 $222
$29
$244 $1,018
$0
$0
$775
$175,000
1,750
5
$30
$0 $13
$43
Air-cooled
Baseline Controlled
$51 $0
$19
1
$11
$3 $4
$35 $37
$3
—
$76
$15 $26
$19
$140
$40 $60
$5
$11
$12
$42
$144 $465
$229
$45
$189 $694
$55 $201
$25
$244 $920
$600,000
$166
$842
$175,000
1,000
5
$52
$0 $13
$65
In addition to these estimated costs for addressing exhaust emissions, we have analyzed
5-10
-------
Chapter 5: Costs of Control
the costs associated with reducing evaporative emissions from gasoline-fueled engines and
vehicles. This effort consists of three primary areas—permeation, diurnal, and boiling.
To reduce permeation losses, we expect manufacturers to upgrade plastic or rubber fuel
lines to use automotive-grade materials. These fuel lines are readily available at a cost premium
of about $1 per linear foot. If an installed engine has an average of four feet of fuel line, this
translates into an increased cost of $4 per engine.
The standard related to diurnal emissions can be met with a fuel cap that seals the fuel
tank, relieving pressure as needed to prevent the tank from bursting or collapsing. The estimated
cost of upgrading to such a fuel cap is conservatively set at $8, based on the aftermarket cost of
comparable automotive fuel caps. Such caps would be expected to cost much less as an original
equipment upgrade of an existing cap.
Many Large SI engines are installed in equipment in a way that poses little or no risk of
fuel boiling during engine operation. A few models are configured in a way that causes this to be
a possibility, at least under extreme conditions. Preventing fuel boiling is primarily a matter of
isolating the fuel tank from heat sources, such as the engine compartment and the exhaust pipe.
Some additional material may be needed to reduce heat exposure, such as a simple metal shield
or a fiberglass panel. Given several years to redesign engines and equipment, we believe that
manufacturers can readily incorporate such changes into their ongoing R&D programs. To
account for several hours of engineering effort and a small amount of material, we estimate that
these costs averaged over the whole set of gasoline-fueled engines will come to about $1 per
engine.
5.2.2.2 Operating Cost Savings
Introducing electronic closed-loop fuel control will significantly improve engine
operation, with corresponding cost savings, in three areas— reduced fuel consumption, less
frequent oil changes and tuneups, and delayed time until rebuild.
It may also be appropriate to quantify the benefit of longer total engine lifetimes. For
example, passenger cars with low-emission engine technologies last significantly longer than
they did before manufacturers developed and applied these technologies. In addition, engine
performance (responsiveness, reliability, engine warm-up, etc.) will also improve with the new
technologies. However, these benefits are more difficult to quantify and the analysis therefore
does not take them into account.
Fuel consumption rates will improve as manufacturers no longer design engines for
operation in fuel-rich conditions. Some current systems already operate at somewhat leaner air-
fuel ratios than in previous years, but even in these cases, engines generally revert to richer
mixtures when accelerating. Closed-loop fuel systems generally operate close to stoichiometry,
which improves the engine's efficiency of converting the fuel energy into mechanical work.
Information in the docket, including development testing, engineering projections, and user
5-11
-------
Draft Regulatory Support Document
testimony, indicates an estimated 20-percent reduction in fuel consumption rates.7'8'9 Table 5.2.2-
3 shows the value of the estimated fuel savings. These values and calculations are generally
based on our NONROAD emissions model. Since the NONROAD model does not account
separately for air-cooled engines, calculated fuel savings are based on information we received
during the comment period.
Table 5.2.2-3: Estimated Fuel Savings from Large SI Engines
Horsepower
Load factor
Annual operating hours, hr/yr
Lifetime, yr
Baseline bsfc, Ib/hp-hr
Improved bsfc, Ib./hp-hr
Fuel density
Fuel cost
Annual fuel saved (gal/yr)
Annual fuel savings ($/yr)
Lifetime Fuel Savings (NPV)
LPG
66
0.39
1,368
12
0.507
0.406
4.2 Ib./gal
$0.60/gal
845
$507
$4,333
Natural gas
64
0.49
1,164
13
0.507
0.406
0.05 g./ft3
$2. 17/1000 ft3
—
$160
$1,427
Gasoline-
water-cooled
52
0.58
534
12
0.605
0.484
6.1 Ib./gal
$1.10/gal
321
$353
$3,038
Gasoline-
air-cooled
60
0.58
1,000
3
1.10
0.88
6.1 Ib./gal
$1.10/gal
1,233
$1,357
$3,810
In addition to the fuel savings, we expect Large SI engines to see significant
improvements in reliability and durability. Open-loop fueling systems in uncontrolled engines
are prone to drifting calibrations as a result of varying fuel quality, wear in engine components,
changing ambient conditions, and other factors. Emission-control systems that operate with a
feedback loop to compensate for changing conditions for a near-constant air-fuel ratio
significantly reduces the following problems.
-incomplete (and eventually unstable) combustion
-absorption of fuel in lubricating oil
-deposits on valves, spark plugs, pistons, and other engine surfaces
-increased exhaust temperatures
Automotive engines clearly demonstrate that modern fuel systems reduce engine wear and the
need for repairs.
This analysis incorporates multiple steps to take these anticipated improvements into
account. First, oil change intervals are estimated to increase by 15 percent. Reduced fuel loading
in the oil (and other improvements such as piston ring design) can significantly extend its
working life. Similarly, tune-up intervals are estimated to increase by 15 percent. This results
largely from avoiding an accumulation of deposits on key components, which allows for longer
operation between regularly scheduled maintenance. Third, we estimate that engines will last 15
percent longer before needing overhaul. The reduced operating temperatures and generally
reduced engine wear associated with closed-loop fuel systems account for this extended lifetime
5-12
-------
Chapter 5: Costs of Control
to rebuild. These quantitative estimates of maintenance-related savings are derived from
observed changes in automotive performance when upgrading from carburetion to fuel injection.
Table 5.2.2-4 summarizes the details of the methodology for converting these maintenance
improvements into estimated cost savings over the lifetime of the engines.
Table 5.2.2-4: Maintenance
Baseline oil change interval (hrs)
Improved oil change interval (hrs)
Cost per oil change ($)
Baseline tune-up interval (hrs)
Improved tune-up interval (hrs)
Cost per tune -up ($)
Baseline rebuild interval (hrs)
Improved rebuild interval (hrs)
Rebuild cost ($)
3aseline lifetime maintenance cost
mproved lifetime maintenance cost
^ifetime maintenance savings (TSIPV)
LPG/
natural gas
200
230
$30
400
460
$75
7,000
8,050
$800
$2,902
$2,681
$221
Gasoline
150
172.5
$30
400
460
$75
5,000
5,750
$800
$2,573
$2,354
$219
These large estimated fuel and maintenance savings relative to the estimated incremental
cost of producing low-emitting engines raise the question of why normal market forces have
failed to induce manufacturers to design and sell engines with emission-control technologies on
the basis of the expected performance improvements. Since forklifts are the strongly dominant
application using Large SI engines, this question effectively applies specifically to forklifts. We
have observed that forklift users generally see their purchase as an expense that doesn't add value
to a company's product, whether that applies to manufacturing, warehouse, or retail facilities.
While operating expenses require less internal justification or decision-making, purchasing new
equipment involves extensive review and oversight by managers who are very sensitive to capital
expenditures. This is reinforced by an April 2000 article in a trade publication, which quotes an
engineering estimate of 20- to 40-percent improvement in fuel economy while stating that it is
unclear whether purchasers will tolerate any increase in the cost of the product.10 Market theory
would predict that purchasers select products with technologies that result in the lowest net cost
(with some appropriate discount for costs incurred over time). It seems that companies have
historically focused on initial costs to the exclusion of potential cost savings over time, which
would account for the lack of emission-control technologies on current sales of Large SI engines.
This priority given to initial cost therefore affects the competitive decisions of engine
manufacturers, who will be less willing to take the business risk of developing a more costly
product than its competitors, even if the product would eventually provide substantial savings to
the purchaser. Also, the initial costs of changing designs and using new technologies can serve
5-13
-------
Draft Regulatory Support Document
as a deterrent to including newer cost-efficient technologies in established engine types.
In addition to the engine improvements described above, the costs associated with
controlling evaporative emissions would be offset by savings from retaining more fuel that can
be used to power the engine. To estimate these costs, we compare the total emission reductions
from diurnal, running loss, hot soak, and refueling emissions with the total gasoline-fueled
engine population in 2030. The resulting reduction of 0.04 tons hydrocarbon per engine
translates into estimated annual savings of $11. Spread over 13 years and discounted to the point
of sale leads to a net present value of $98 saved.
5.2.2.3 Compliance Costs
We estimate that certification costs come to $70,000 per engine family. We expect
manufacturers to combine similar engines using different fuels in the same family. This expands
the size of engine families, but calls for several tests to complete the certification process for
each family. This includes six engine tests and $10,000 worth of engineering and clerical effort
to prepare and submit the required information. Until engine designs are significantly changed,
engine families can be recertified each year using carryover of the original test data. This cost is
therefore amortized over five years of engine sales, with an assumed volume of 3,000 engines per
year from each engine family. This engine-family sales volume is larger than those presented for
amortizing fixed costs above, because engine families will include multiple fuel types. The
resulting cost for certification is $6 per engine. Since these engines are currently not subject to
any EPA emission requirements, the analysis includes a cost to recertify an upgraded engine
model every five years. Since manufacturers already need to submit data for California
certification, they will incur most of these costs independent of EPA requirements.
Manufacturers must generally do production-line testing on a quarterly basis, but reduced
testing rates apply if engine testing shows consistently good test results. Manufacturers must
generate and submit this test data to comply with the requirements adopted by California ARB.
The EPA requirement for production-line testing therefore adds no test burden to manufacturers.
Even with a transient duty cycle for certification, manufacturers may rely on steady-state test
procedures at the production line. We therefore fully expect that manufacturers will need only to
send the "California" test data to EPA to satisfy requirements for production-line testing. The
analysis therefore includes no cost for additional routine testing of production engines. In fact,
manufacturers may pursue alternate methods to show that production engines comply with
emission standards, which may lead to lower testing costs.
We may select up to 25 percent of a manufacturers' s engine families for in-use testing.
This means that a manufacturer would need to have eight engine families for us to be able to
select two engine families in a given year. Since this is likely to be a rare scenario, we project an
annual testing rate of one engine family per year for each manufacturer to assess the cost of the
in-use testing program. The analysis includes the cost of testing in-use engines on a
dynamometer, which requires:
- engine removal and replacement ($4,000)
5-14
-------
Chapter 5: Costs of Control
transport ($1,000)
- steady-state and transient testing ($ 15,000)
Testing six engines and adding costs for administration and reporting of the testing program
leads to a total cost of about $125,000 for an engine family. These costs can be spread over a
manufacturer's total annual sales, which averages about 15,000 units for most companies. The
resulting cost per engine is about $8.
As with production-line testing, we expect in-use emission testing to simultaneously
satisfy California ARB and EPA requirements. In certain circumstances, however, we may use
our discretion to direct a manufacturer to do in-use testing on an engine family separately from
California ARB. Since we expect this to be the exception, this analysis likely overestimates the
cost impact of adopting federal requirements to do in-use testing. In fact, manufacturers may
reduce their compliance burden with the optional field-testing procedures. Table 5.2.2-5 shows
the estimated costs from the various compliance programs.
In addition, we expect several manufacturers to upgrade testing facilities to allow for in-
house measurement of emissions during transient engine operation. We generally expect each
major manufacturer to equip one test cell with a new dynamometer and the associated controllers
and analyzers. Installation of transient test cell would cost about $500,000. This consists of
about $225,000 each for an electric dynamometer and the associated controllers, and $50,000 for
a battery of sampling equipment and analyzers. An additional capital cost of $80 is estimated for
precision calipers with digital readout to ensure dimensional accuracy of catalyst diameters.
Dividing these costs over six engine families for five years leads to a calculated per-engine cost
under $10.
Table 5.2.2-5
Cost of Compliance Programs
Compliance Program
Element
Certification
In-use testing
Facility upgrade
Total
Estimated Per-
Engine Costs
$6
$8
$7
$21
5.2.2.4 Total Costs
Table 5.2.2-6 presents the combined cost figures for the different engine types and
calculates a composite cost based on their estimated distribution. The estimated 2004 costs are
based on the adding component costs and compliance costs. No R&D cost is estimated for
manufacturers to do additional development work beyond what is necessary to comply with
5-15
-------
Draft Regulatory Support Document
California ARB standards. Conversely, the estimated 2007 costs are based on R&D (and
ongoing compliance costs), with no anticipated increase in component costs, except those related
to reducing evaporative emissions. The estimated cost of complying with the emission standards
is sizable, but the lifetime savings from reduced operating costs nevertheless more than
compensate for the increased costs. Costs for gasoline engines are presented as a composite of
air-cooled models (estimated 3 percent of total sales) and water-cooled models (estimated 20
percent of total sales).
Table 5.2.2-6
Estimated First-Year Cost Impacts of New Emission Standards
Standards
2004
Engine Type
LPG
natural gas
gasoline
Composite
Sales Mix of
Engine Types
68%
9%
23%
—
Increased Production
Cost per Engine*
$550
$550
$800
$605
Lifetime Operating Costs
per Engine (NPV)
$-4,330
$-1,650
$-3,140
$-3,815
2007
LPG
natural gas
gasoline
Composite
68%
9%
23%
—
$40
$40
$60
$50
—
—
$-100
$-20
*The estimated long-term costs decrease by about 35 percent.
5.2.2.5 Large SI Aggregate Costs
The above analyses developed incremental per-vessel cost estimates for Large SI engines.
Using these per-engine costs and projections of future annual sales, we have estimated total
aggregate annual costs for the exhaust and evaporative emission standards. The aggregate costs
are presented on a cash-flow basis, with hardware and fixed costs incurred in the year the vehicle
is sold and fuel savings occurring as the engines are operated over their lifetimes. Table 5.2.2-7
presents a summary of this analysis. As shown in the table, aggregate costs generally range from
$70 million to $90 million. Net costs decline as fuel savings continue to ramp-up as more
vehicles meeting the standards are sold and used. Fuel savings are projected to more than offset
the costs of the program starting by the second year of the program.
5-16
-------
Chapter 5: Costs of Control
Table 5.2.2-7: Summary of Annual Aggregate Costs and
Fuel Savings for Large SI Engines (millions of dollars)
Total Costs
Fuel Savings
Net Costs
2004
$89
($53)
$36
2005
$91
($103)
($12)
2010
$71
($326)
($255)
2015
$73
($421)
($348)
2020
$81
($472)
($391)
To project annual sales, we started with the number of model year 2000 engines estimated
by the NONROAD model for the 2000 calendar year. We then applied a growth rate of 3 percent
of year 2000 sales (increasing by 3,900 units annually) to estimate future sales. Table 5.2.2-8
provides a summary of the sales estimates used in the aggregate cost analysis.
Table 5.2.2-8
Estimated Annual Sales of Large SI Engines
2000
130,000
2004
145,600
2010
169,000
2020
208,000
To calculate annual aggregate costs, the sales estimates have been multiplied by the per-
unit costs. Annual fuel savings have been calculated based on the reduction in fuel consumption
expected from the standards (as described in section 5.2.2.2 of this chapter) as calculated by the
NONROAD model. The model takes into consideration vehicle sales and scrappage rates. The
year-by-year results of the analysis are provided in Chapter 7.
5.2.3 Recreational Vehicles
5.2.3.1 Technologies and Estimated Costs
We estimated costs separately for snowmobiles, ATVs, and off-highway motorcycles.
Individual technology costs were developed in cooperation with EPA by ICF Incorporated and
Arthur D. Little - Acurex Environmental.11 Any comments received on the rule were also
evaluated and included where appropriate. Costs were prepared for a typical engine that falls
within the displacement ranges noted below. Costing out multiple engine sizes allowed us to
estimate significant differences in costs for smaller vs. larger engines. The costs include a mark-
up to the retail level. This Chapter also provides a brief overview of the technologies, with more
information provided in Chapter 4. Costs are provided for both the baseline technology and the
new technology (e.g., a two-stroke engine and a four-stroke engine), with the cost of the change
in technology due to the new standards being the increment between the two costs.
5-17
-------
Draft Regulatory Support Document
The R&D costs shown are average costs. The first engine line R&D cost is expected to
be significantly higher but the costs would be distributed across the manufacturer's entire product
line.12 To account for any additional warranty cost associated with a change in technology, we
have added 5 percent of the incremental hardware cost.13
As noted in section 5.1, fixed costs are spread over the first five years of sales for
purposed of the cost analysis, with the exception of new facility costs for ATV testing which are
spread over 10 years. We have used 10 years for amortization rather than 5 years because we
believe it is more representative for a capital investment that will be used for at least that long a
time period. We estimated that R&D and facility costs will be incurred three years prior to
production on average and tooling and certification costs will be incurred one year prior to
production. These fixed costs were then increased seven percent for each year prior to the start
of production to reflect the time value on money.
To approximate average annual sales per engine line, we divided the total 2001 annual
unit sales by estimated total number of engines lines industry-wide.ee Based on limited sales data
from individual manufacturers provided to EPA on a confidential basis, there appears to be a
large distinction in sales volume between small engine and large engine displacements for ATVs.
The cost analysis accounts for this difference by using a larger annual sales rate per engine line
for larger displacement ATVs, as shown below.
As noted below, the fuel savings over the life of the vehicle due to some of the projected
technology changes can be substantial and for snowmobiles are projected to offset the cost of the
emission controls. As discussed below, these fuel savings will occur because 2-stroke
powerplants are inefficient and the changes needed to reduce hydrocarbons from these engines
also improve fuel consumption. Because the fuel savings outweigh up front costs, one might
question why manufacturers have continued to use 2-stroke engines. Manufacturers have not
made these changes in the absence of emission standards for several likely reasons. Since fuel
costs are not a significant portion of the overall price of ownership, customers may not place a
high value on fuel economy compared to initial cost and engine simplicity. Especially in the case
of snowmobiles and off-road motorcycles, manufacturers have built a customer base over many
years using 2-stroke technology; ATVs which are dominantly 4-stroke are relatively new to the
recreational vehicle market.. The engines are relatively simple and the production costs are
relatively low because the manufacturers have been building the engines for many years. To
capture the fuel economy benefits, manufacturers would have to invest substantially in R&D and
more complex powerplants in the face of uncertainty with regard to market acceptance of the new
product. Such a move could also lower profits per vehicle. Considering all these factors,
manufacturers have historically chosen to focus improvements in other areas such as increasing
horsepower and overall vehicle design.
ee Based on publicly available product information for the large manufacturers, we
estimated 32 engine lines for snowmobiles, 43 lines for ATVs, and 42 lines for off-highway
motorcycles for the 2001 model year.
5-18
-------
Chapter 5: Costs of Control
However, manufacturers are now introducing 4-strokes and direct injection 2-stroke
engines into the snowmobile market. For model year 2003, all manufacturers will have at least
one 4-stroke snowmobile model available and one manufacturer is introducing direct injection 2-
stroke technology. This may mean that manufacturers are adjusting their perspectives on
potential marketplace acceptance of advanced technologies.
5.2.3.1.1 Snowmobiles
Phase 1
Snowmobiles are currently almost exclusively powered by carbureted 2-stroke engines.
However, as noted above, manufacturers are beginning to introduce 4-strokes and 2-stroke direct
fuel injection. Manufacturers have also provided comment that they plan to rely more heavily on
these technologies to meet Phase 1 standards than originally thought prior to proposal. For these
reasons, we have adjusted our projected baseline technology mix as well as our projected
technology mix for the Phase 1 standards for purposes of the cost analysis. Based on discussions
with manufacturers, we believe that up to 10 percent of production will be 4-stroke and 10
percent will be direct fuel injection for Phase 1. We believe manufacturers will be ramping up
the introduction of these technologies in order to obtain experience with them prior to the start of
the program. These technologies will provide surplus emissions reductions which will allow the
manufacturers to use lesser technologies on other models under the averaging program.
For cost purposes, we are projecting that 4-stroke engines are likely to be equipped with
electronic fuel injection systems to optimize emissions and overall performance of these engines.
Therefore we are including electronic fuel injection costs for 4-strokes. Tables 5.2.3-1 through
5.2.3-4 provide costs for direct injection systems (both air assisted direct injection and pump
assisted direct injection) and for converting from a 2-stroke to 4-stroke engine with electronic
fuel injection.
We have estimated the incremental cost of going from carbureted 2-stroke to direct
injection to range from $262 to $342 per engine and conversion to 4-stroke to be about $454 to
$770. Electronic fuel injection for snowmobiles is estimated to incrementally cost $174 to $119.
Note that the overall consumer costs for these advanced technologies are substantially lower after
the fuel economy improvements are taken into account. Estimates of the fuel savings are
provided below. For 4-stroke snowmobiles, where possible, we have examined available price
information on manufacturer web sites for the various 4-stroke models and comparable 2-stroke
models and found price differences to be similar to our cost estimates in most cases. We did not
receive detailed public comments on our cost estimates for the various snowmobile technologies.
5-19
-------
Draft Regulatory Support Document
Table 5.2.3-1: Air Assisted Direct Injection System Costs for Snowmobiles
< 500 cc
Baseline
Modified
> 500cc
Baseline
Modified
Hardware Costs
Carburetor
Number Required
Fuel Metering Solenoid (each)
Number Required
Air Pump
Air Pump Gear
Air Pressure Regulator
Throttle Body/Position Sensor
Intake Manifold
Electric Fuel Pump
Fuel Pressure Regulator
ECM
Air Intake Temperature Sensor
Manifold Air Pressure Sensor
Injection Timing Sensor/Timing Wheel
Wiring/Related Hardware
Hardware Cost to Manufacturer
Labor @ $28 per hour
Labor overhead @ 40%
OEM mark-up @ 29%
Royalty @ 3%
Warranty Mark-up @ 5%
Total Component Costs
$60
2
$5
$125
$1
$1
$37
$164
$15
2
$25
$5
$5
$35
$30
$5
$3
$140
$5
$11
$10
$20
$324
$14
$6
$100
$10
$10
$464
$60
3
$5
$185
$2
$1
$55
$243
$15
3
$25
$5
$5
$35
$30
$5
$3
$140
$5
$11
$10
$20
$339
$21
$8
$107
$10
$8
$493
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
4,400
5
$0
$164
$178,500
$25,000
4,400
5
$13
$476
$312
$0
$0
4,400
5
$0
$243
$178,500
$25,000
4,400
5
$13
$505
$263
5-20
-------
Chapter 5: Costs of Control
Table 5.2.3-2: Pump-Assisted Direct Fuel Injection System Costs for Snowmobiles
< SOOcc
Baseline
Modified
> SOOcc
Baseline
Modified
Hardware Costs
Carburetor
Number Required
Nozzle/ Accumulator (each)
Number Required
High-Pressure Cam Fuel Pump
Cam Pump Gear
Throttle Body/Position Sensor
Intake Manifold
Fuel Transfer Pump
ECM
Air Intake Temperature Sensor
Manifold Air Pressure Sensor
Injection Timing Sensor/Timing Wheel
Wiring/Related Hardware
Hardware Cost to Manufacturer
Labor @ $28 per hour
Labor overhead @ 40%
OEM mark-up @ 29%
Royalty @ 3%
Warranty Mark-up @ 5%
Total Component Costs
$60
2
$5
$125
$1
$1
$37
$164
$33
2
$20
$5
$35
$30
$5
$140
$5
$11
$10
$20
$347
$14
$6
$106
$10
$11
$494
$60
o
3
$5
$185
$2
$1
$55
$243
$33
3
$25
$5
$35
$30
$5
$140
$5
$11
$10
$20
$385
$21
$8
$120
$12
$10
$556
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
4,400
5
$0
$164
$178,500
$25,000
4,400
5
$13
$506
$343
$0
$0
4,400
5
$0
$243
$178,500
$25,000
4,400
5
$13
$568
$327
5-21
-------
Draft Regulatory Support Document
Table 5.2.3-3: Two-Stroke to Four Stroke Conversion Costs for Snowmobiles
< 500 cc
2-Stroke
4-Stroke
> 500 cc
2-Stroke
4-Stroke
Engine
Clutch
Labor @ $28 per hour
Labor overhead @ 40%
Markup @ 29%
Warranty Mark up @ 5%
Total Component Costs
$400
$50
$14
$6
$136
$606
$700
$75
$21
$8
$233
$16
$1,053
$650
$80
$14
$6
$217
$967
$1,170
$120
$21
$8
$OOO
JoJ
$28
$1,730
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
4,400
5
$0
$606
$94,416
$20,000
4,400
5
$7
$1,060
$455
$0
$0
4,400
5
$0
$967
$94,416
$20,000
4,400
5
$7
$1,737
$770
5-22
-------
Chapter 5: Costs of Control
Table 5.2.3-4: Electronic Fuel Injection Costs for Snowmobiles
Fuel Injection Costs
400cc
Baseline
Modified
700cc
Baseline
Modified
Hardware Costs
Carburetor
Number Required
Injectors (each)
Number Required
Pressure Regulator
Intake Manifold
Throttle Body/Position Sensor
Fuel Pump
ECM
Air Intake Temperature Sensor
Manifold Air Pressure Sensor
Injection Timing Sensor
Wiring/Related Hardware
Hardware Cost to Manufacturer
Labor @ $28 per hour
Labor Overhead @ 40%
Manufacturer Mark-up @ 29%
Warranty Mark-up3 @ 5%
Total Component Costs
$60
2
$5
$125
$1
$1
$37
$164
$12
2
$10
$30
$35
$20
$100
$5
$10
$5
$10
$249
$4
$2
$72
$6
$333
$60
3
$5
$185
$2
$1
$54
$242
$12
3
$10
$35
$35
$20
$100
$5
$10
$5
$10
$266
$6
$3
$77
$4
$356
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs ($)
Incremental Total Cost ($)
$0
$0
4,400
5
$0
$164
$69,417
$10,000
4,400
5
$5
$338
$175
$0
$0
4,400
5
$0
$242
$69,417
$10,000
4,400
5
$5
$361
$119
In addition to the advanced technologies, we are also basing the cost analysis for Phase 1
standards on some use of engine modifications, carburetor improvements, and recalibration. We
are projecting lower usage of this approach compared to the proposal (60% compared to 100%)
based on the comments we received concerning the use of advanced technology to meet Phase 1
standards. Manufacturers are likely to be able to reduce emissions for some models by leaning
out the air/fuel mixture, improving carburetors for better fuel control and less production
5-23
-------
Draft Regulatory Support Document
variation, and modifying the engine to withstand higher temperatures and potential misfire
episodes attributed to enleanment. Engine modifications are also likely to be made to improve
air/fuel mixing and combustion. The cost estimates for engine modifications and carburetor
improvements are provided in Tables 5.2.3-5 and 5.2.3-6. Recalibration work is included as part
of the R&D for the technologies. The incremental cost per unit for engine modifications is
estimated to be $18 to $25, with modifications to the carburetor estimated to cost an additional
$18 to $24 per engine.
Table 5.2.3-5: Snowmobile Engine Modification Costs for Two-Stroke Engines
< 500 cc
Baseline
Modified
> 500 cc
Baseline
Modified
Hardware Costs
Improved Pistons
Number Required
Hardware Cost to Manufacturer
Labor @ $28 per hour
Labor Overhead @ 40%
Manufacturer Mark-up @ 29%
$10
2
$20
$6
$2
$6
Warranty Mark-up @ 5%
Total Component Costs
$34
$12
2
$24
$6
$2
$7
$0
$39
$12
3
$36
$8
$3
$10
$57
$15
3
$45
$8
$3
$13
$0
$69
Fixed Cost to Manufacturer
R&D Costs per line
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
4,400
5
$0
$34
$178,500
$25,000
4,400
5
$13
$51
$18
$0
$0
4,400
5
$0
$57
$178,500
$25,000
4,400
5
$13
$81
$25
5-24
-------
Chapter 5: Costs of Control
Table 5.2.3-6: Modified Carburetor Costs for Snowmobiles
< 500 cc
Baseline
Modified
> 500 cc
Baseline
Modified
Hardware Costs
Carburetor
Number Required
Hardware Cost to Manufacturer
Labor @ $28 per hour
Labor Overhead @ 40%
Manufacturer Mark-up @ 29%
Warranty Mark-up @ 5%
Total Component Costs
$60
2
$120
$1
$1
$35
$157
$65
2
$130
$1
$1
$o o
38
$1
$171
$60
o
5
$180
$2
$1
$53
$236
$65
3
$195
$2
$1
$57
$1
$256
Fixed Cost to Manufacturer
R&D Costs per line
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
4,400
5
$0
$157
$61,875
$5,000
4,400
5
$4
$175
$18
$0
$0
4,400
5
$0
$236
$61,875
$5,000
4,400
5
$4
$260
$24
Phase 2 and Phase 3
We have based the cost analysis for the Phase 2 and Phase 3 standards primarily on the
expanded use of direct fuel injection 2-stroke engines and 4-stroke engines. We expect that by
the 2010 time frame these two technologies will be fully developed and able to be used on a
larger fraction of the fleet. Our projections that these later Phases will be met primarily through
the expanded use of these technologies is consistent with our discussions with manufacturers.
This chapter provides a cost analysis for the primary Phase 2 program which calls for a 50
percent reduction from baseline levels for both HC and a 30 percent reduction for CO emissions
in 2010. The Phase 3 standard begins in 2012 and requires a further reduction in CO from 30
percent to 50 percent. Manufacturers have some flexibility in meeting the Phase 3 standards
which allows them to meet less stringent CO requirements if additional HC reductions are
achieved. We would expect the same technologies to be used to meet these all of these programs
but in somewhat different combinations. For example, some manufacturers may rely on 4-stroke
technology more so than direct injection 2-stroke technology. This is discussed in detail in
Chapter 4. With averaging, manufacturers, will optimize their technology paths for each phase
of standards and each manufacturer will have somewhat different mixes of technology.
5-25
-------
Draft Regulatory Support Document
For Phase 2 and Phase 3, we are projecting that 50 and 70 percent of models,
respectively, will be equipped with either direct injection 2-stroke or 4-stroke engines. We
anticipate that remaining models will consist of 2-stroke technologies with some further
optimization. One additional technology that may be used is pulse air. We are projecting the use
of pulse air systems with recalibration on a portion of the snowmobile engines that are not
equipped with advanced technology systems. Pulse air provides a small incremental emission
reduction for these engines and would help manufacturers meet the Phase 2 and Phase 3 average
HC and CO standards. As shown in Table 5.2.3-7, we have estimated pulse air to cost about $40.
Catalysts are also a potential option for snowmobiles but would entail a significant R&D effort
and may not be available for snowmobile applications in the 2010 time frame. However, we
believe manufacturers are more likely to focus on developing the advanced technologies noted
above, which provide the consumer with benefits in addition to lower emissions. Therefore, we
have not included catalyst costs in our cost estimates.
Table 5.2.3-7: Calibration/Pulse-Air Costs for Snowmobiles
Baseline
Modified
Hardware Costs
Pulse Air Valve
Labor @ $28 per hour
Labor overhead @ 40%
Markup @ 29%
Warranty Mark up @ 5%
Total Component Costs
$0
$18
$1
$0
$5
$0
$25
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$54,750
$200,000
4,400
5
$15
$40
$40
5.2.3.7.2 All-terrain Vehicles (ATVs)
ATVs are equipped primarily with carbureted 4-strokes, with 2-stroke engines used
mostly in small displacement and sport models. We expect manufacturers to take several steps in
response to the standards and test cycle requirements. Beginning in 2006, we expect most
manufacturers will take some advantage of the transitional interim test procedures and standards
offered from 2006-2008 but will need to phase out the use of 2-stroke engines. In addition, for
the 4-stroke ATVs, we are also projecting that as manufacturers transition to the chassis test
5-26
-------
Chapter 5: Costs of Control
cycle, recalibration will be needed and that pulse air systems will be used on about 50 percent of
the models to ensure that the fleet meets the standards on average. Pulse air systems are
currently used on a few ATV and off-highway motorcycles models to meet California standards.
We do not believe that the level of the standards will require the use of pulse air beyond 50
percent, given that only a few models in California are currently equipped with the technology.
Using pulse air may give the manufacturer more flexibility in calibrating for performance on
some models. Technological feasibility is discussed in Chapter 4.
We are basing our technology projection on what manufacturers have done to meet the
California emissions standards. We believe this to be the most likely technology path for
manufacturers, because 4-strokes are accepted in the market and provide consumers with fuel
economy and reliability benefits. Beyond using 4-stroke engines, we expect manufacturers to
undertake an R&D effort to recalibrate models and select and optimize pulse air systems. Some
recalibration is likely, due to the change in test procedures. We received comments that we
underestimated the amount of R&D necessary for ATVs and, upon evaluation, have adjusted the
estimates upwards. We continue to believe manufacturers will approach this effort in an orderly
manner and we would expect them to focus R&D on a first engine line and then apply what they
learn to subsequent lines.ff Table 5.2.3-8 provides the estimated R&D for ATVs. We believe the
increased level of R&D shown below is substantial considering the technological difficulty of the
final standards. We believe the estimated amounts also are sufficient because manufacturers
have already invested in R&D and technology to meet the California program which contains
standards that are similar in stringency.
Table 5.2.3-8: R&D Cost Estimate for ATVs
Base R&D Costs for 1st engine line
Engine lines per manufacturer
Base R&D per line
Individual Engine Line R&D
Total R&D per line
Units/yr.
Years to recover
R&D Fixed cost/unit
< 200 cc
$724,000
8
$90,500
$238,000
328,500
5,600
5
$16.40
> 200 cc
$724,000
8
$90,500
$238,000
$328,500
20,000
5
$4.59
Tables 5.2.3-9 and 5.2.3-10 provide cost estimates for the ATV technologies discussed
above. We estimate the incremental cost per unit of replacing a 2-stroke engine with a 4-stroke
engine to be about $219 to $349, depending on engine size. Costs for a mechanical pulse air
system is estimated to be about $27 to $33 per unit. As shown in the tables below, fixed costs
ff We have estimated a base R&D effort of 12 months for the first engine line and 6
additional months for subsequent lines and have used the costing methodology provided in the
Arthur D. Little - Acurex cost report to calculate the increased R&D cost.
5-27
-------
Draft Regulatory Support Document
for larger displacement models are spread over a significantly larger annual unit sales volume to
account for the relatively high average number of unit sales per engine line for these products.
Table 5.2.3-9: Two-Stroke to Four Stroke Conversion Costs for ATVs
< 200 cc
2-Stroke
4-Stroke
> 200 cc
2-Stroke
4 Stroke
Hardware Costs
Engine
Labor @ $28 per hour
Labor overhead @ 40%
Markup @ 29%
Warranty Mark up @ 5%
Total Component Costs
$400
$14
$6
$122
$542
$550
$21
$8
$168
$8
$755
$500
$14
$6
$151
$671
$750
$21
$8
$226
$13
$1,018
Fixed Cost to Manufacturer
R&D Costs
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$0
5,6200
5
$0
$541
$94,416
$15,000
5,600
5
$5
$760
$219
$0
$0
20,000
5
$0
$670
$94,416
$18,000
20,000
5
$2
$1,019
$349
5-28
-------
Chapter 5: Costs of Control
Table 5.2.3-10: Pulse-Air Costs for Four-Stroke ATVs
< 200 cc
Baseline
Modified
> 200 cc
Baseline
Modified
Hardware Costs
Pulse Air Valve
Labor @ $28 per hour
Labor overhead @ 40%
Markup @ 29%
Warranty Mark up @ 5%
Total Component Costs
$0
$18
$1
$0
$5
$0
$25
$0
$18
$1
$0
$5
$0
$25
Fixed Cost to Manufacturer
Tooling Costs
Units/yr.
Years to recover
Fixed cost/unit
Total Costs
Incremental Total Cost
$0
$159,091
5,600
5
$7
$33
$33
$0
$159,091
20,000
5
$2
$27
$27
5.2.3.1.3 Off-highway Motorcycles
Currently, off-highway motorcycles are about 65 percent 2-stroke, with many of the 2-
stroke engines used in competition and youth models. As with ATVs, we expect that
manufacturers will meet standards primarily by using 4-stroke engines. Manufacturers may also
use pulse air systems and recalibration on a relatively small fraction of their models to ensure
their overall fleet meets the standards. We have estimated their use for off-highway motorcycles
at about 25 percent for purposes of the cost analysis. The R&D efforts will likely be lower for
off-highway motorcycles than for ATVs because the level of the standard is less stringent and
there is no change in the test procedure from what is now required in California. We do not
believe the standards will require pulse air technology in more than 25 percent of models, given
that only a few models in California are currently equipped with this technology. As discussed in
5.2.3.4 below, vehicles used solely for competition are exempt from standards and we expect
some 2-stroke competition models to remain in the market.
Tables 5.2.3-11 and 5.2.3-12 provide cost estimates for off-highway motorcycle
technologies for three engine displacement ranges. We estimate the incremental cost per unit of
replacing a 2-stroke engine with a 4-stroke engine to be about $219 to $353, depending on engine
size. Costs for a mechanical pulse air valve system and recalibration is estimated to be about $39
per unit.
5-29
-------
^
^
«
o
"o
E
g
.=
.ŁP
'A
la
o
.
o
o
U
=
^2
K
&•
4^
^
=
o
U
4*
o
Ł
s
o
ta
o
^5
O)
o
•^
op
o
H
fO
—
S
H
U
0
Al
U
U
0
fS
v
U
U
^H
V
OJ
^
o
b
1
Tf
OJ
o
b
C-)
$Ł
2
-^^
op
•4
-------
„
^
"w
w
o
"o
s
1
IS
la
o
Ł
o
U
•^
i
W5
"s
^
^
•^
W
Ł
13
U
jg
o
+j
s
o
ta
fO
Ł!
—
S
H
^
o
^M
•3
o
^
_g
"S
ffl
.-
•3
o
"3
sS
CO
•a
:M
•3
o
^^
_g
"3
ffl
K
0
U
4>
e«
"S
sS
I
OO
^}
oo
oo
^^
Air Valve
PH
^^-
s
^^
o
a
oo
r,
n
>n
>n
to recover
1
^I
&5
S
TT
^^
cost/unit
1
ON
^^
0
ON
0
ON
^^
0
@
O
U
13
"o
H
ON
5X}
ON
ON
^^
®
O
U
13
§
"a
I
hH
-------
Draft Regulatory Support Document
5.2.3.1.4 Crankcase Controls
The proposal included a requirement for crankcase emission controls for recreational
vehicles. Crankcase controls have been required on passenger cars for more than 30 years, and it
is normally a simple process of routing crankcase exhaust emissions to the engine intake to be
burned as part of normal engine operation. Most current 4-stroke recreational vehicle engines
use positive crankcase ventilation systems today; crankcase emissions are not significant in
current 2-stroke engines. For those converting to 4-stroke in the future, crankcase controls will
be required at a cost of about $3 per engine. These are included in the 2-stroke to 4-stroke
conversion and replacement costs.
5.2.3.1.5 Permeation Control from Recreational Vehicles
As discussed in earlier chapters, we believe that there are several technologies that could
be used to meet the permeation emission standards. Table 5.2.3-13 presents our best estimates of
the costs of applying various evaporative emission control technologies to recreational vehicles
using the average fuel tank sizes and hose lengths discussed in Chapter 6.
The cost for including low permeation barrier platelets in blow-molded fuel tanks
(generally known as Selar®) is based on increased material costs. No changes should be
necessary to the blow-molding equipment. We used 10 percent EVOH which is about $3 per
pound and 90 percent HDPE which is about $0.50 per pound. This equates to a price increase of
about $0.30 per pound. Depending on the shape of the fuel tank and the wall thickness,
recreational vehicle fuel tanks weigh about 1-1.3 pounds per gallon of capacity. Costs for multi-
layer fuel tanks with continuous barriers are not included, but would be expected to be higher
because two additional injection screws would be necessary for the barrier and adhesion layers.
Another option would be to mold the entire fuel tank of a low permeation material such as nylon,
an acetal copolymer, or a thermoplastic polyester. These materials have list prices of about $2.00
per pound; therefore, the cost of using these alternative materials would be about 7 times higher
than presented below for barrier platelets with 10% EVOH.
Surface treatment costs are based on price quotes from a companies that specialize in this
fluorination14 and sulfonation.15 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 our fluorination prices assumes rectangular shaped containers. For irregular
shaped fuel tanks, the costs would be higher because they would have to be fit into baskets with
volumes larger than the volume of the fuel tanks. Therefore, we consider a 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 in recreational vehicles.
These prices do not include the cost of transporting the tanks; we estimated that shipping,
handling and overhead costs would be an additional $0.22 to $0.81 per fuel tank depending on
tank size.16
5-32
-------
Chapter 5: Costs of Control
Barrier fuel hose incremental costs estimates are based on costs of existing products used
in marine and automotive applications.17'18'19 We estimate that the cost increment compared to
R7 hose used in most recreational applications today is about $0.60 per foot. Some
manufacturers have commented that they do not use hose clamps today, but would need them if
they use barrier hose. Other manufacturers already use hose clamps, but may need to upgrade
them in some applications. To be conservative, we consider the cost of adding hose clamps to all
applications. These hose clamps cost about $0.20 each.20 For ATVs and OHMCs, we include
the costs of two hose clamps for each vehicle (one for each end of the hose). Snowmobiles can
require 4 to 8 hose clamps depending on the fuel pump configuration, number of carburetors, and
if a fuel return line is included. We include the cost of 6 hose clamps for snowmobiles in this
analysis.
Table 5.2.3-13: Permeation Control Technologies and Incremental Costs
Technology
barrier platelets (10% EVOH)
sulfonation treatment*
shipping/handling
fluorination treatment*
shipping/handling
1/4" I.D. hose barrier fuel hose*
hose clamps*
Snowmobiles
11 gallon tank
3.5ft. hose
$3.30
$1.50
$0.81
$8.39
$0.81
$2.71
$1.55
ATVs**
4 gallon tank
1 ft. hose
$1.50
$1.20
$0.30
$3.23
$0.30
$0.77
$0.52
OHMCs
3 gallon tank
1.5ft. hose
$1.20
$1.20
$0.22
$2.42
$0.22
$1.16
$0.52
* includes a 29% markup for overhead and profit
** includes utility vehicles
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 would be
approximately $800,000.21 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. Based on a production capacity of 150,000 units per year, and a 10 year life, the
average sulfonation cost per fuel tank would be about $0.60. These costs would be lower for
higher production volumes. In addition, if a manufacturer were to sulfonate their fuel tanks in-
house, they would not need to pay shipping and handling costs.
To determine the total costs per recreational vehicle we use the scenario that all
manufacturers use sulfonation to reduce permeation from their fuel tanks and use barrier fuel
hose. For this analysis, we consider the cost of shipping fuel tanks to an outside vendor for
5-33
-------
Draft Regulatory Support Document
treatment rather than using the lower cost of in-house sulfonation. For competition off-highway
motorcycles, which make up about 29 percent of OHMC sales, we assume that no low
permeation technology would be used. We estimate the total per vehicle costs to be $6.56 for
snowmobiles, $2.79 for ATVs, and $3.10 for non-competition OHMCs. Weighting a cost of $0
for competition OHMCs, we get an average cost of $2.14 per off-highway motorcycle. These
costs do not include the fuel savings associated with a reduction permeation which is discussed
below in section 5.2.3.2.3.
As a sensitivity analysis, we estimated what the costs would be if the fuel tank
permeation control technology applied by manufacturers were equally distributed by barrier
platelets, sulfonation, and fluorination. Not considering fuel costs, the estimated fuel tank costs,
under this scenario, would be $4.93 for snowmobiles, $2.18 for ATVs, and $1.75 for non-
competition OHMCs. This represents about a 20-100% increase in the cost estimates for fuel
tanks (no change in fuel hose costs). However, we believe that manufacturers are likely to use
sulfonation to meet the fuel tank permeation standards because it appears to be the most cost
effective strategy in most cases. Although barrier platelets and fluorination could likely be
applied earlier, we believe that we are providing adequate lead time for manufacturers to
incorporate sulfonation into their commercial processes.
5.2.3.2 Operating Cost Savings
5.2.3.2.1 Snowmobiles
Both direct injection and conversion from two-stroke to 4-stroke yield substantial fuel
economy benefits. Typical 2-stroke engines have relatively poor fuel economy performance
because a portion of the combustion mixture passes through the engines unburned. Because 4-
stroke and direct injection 2-stroke engine designs essentially do not allow this to occur, they
provide better fuel economy as well as substantially lower HC emissions. We have estimated
fuel savings based on a 25 percent reduction in fuel consumption, based on typical performance
of these technologies. Lifetime fuel costs are provided in Table 5.2.3-14.22'23
5-34
-------
Chapter 5: Costs of Control
Table 5.2.3-14: Fuel Cost for Snowmobiles
Engine
Engine power
Load Factor
Annual Operating Hours, hr/yr
Lifetime, yr
BSFC, Ib/bhp-hr
Fuel Density (Ibs/gal)
Fuel Cost ($/gal)*
Yearly Fuel Consumption (gal/yr)
Yearly Fuel Cost ($/yr)
Lifetime Fuel Cost (NPV)
Baseline 2-Stroke
small
45
0.34
57
12
1.66
6.17
$1.10
235
$258
$2,050
large
100
0.34
57
12
1.25
6.17
$1.10
521
$574
$4,556
Advanced Technology
Engines (25% savings)
small
45
0.34
57
12
1.66
6.17
$1.10
176
$194
$1,537
large
100
0.34
57
12
1.25
6.17
$1.10
391
$430
$3,417
* Excluding taxes
5.2.3.2.2 A TVs and Off-highway Motorcycles
Conversion from 2-stroke to 4-stroke engines yields a fuel economy improvement for
ATVs and off-highway motorcycles as well. Tables 5.2.3-15 and 5.2.3-16 provide estimates of
fuel consumption for both 2-stroke and 4-stroke engines. We have estimated that switching from
a 2-stroke to a 4-stroke engine reduces fuel consumption by about 25 percent. Lifetime fuel
savings for ATVs resulting from switching from a 2-stroke to a 4-stroke engine is estimated to be
$124. For off-highway motorcycles, the projected lifetime fuel savings is $140.
Table 5.2.3-15: Fuel Cost for ATVs
Engine
Annual Miles
Lifetime, yr
BSFC, Ib/mile
Fuel Density (Ibs/gal)
Fuel Cost ($/gal)*
Yearly Fuel Consumption (gal/yr)
Yearly Fuel Cost ($/yr)
Lifetime Fuel Cost (NPV)
2-Stroke
1,570
13
0.213
6.17
$1.10
54
$60
$498
4-Stroke
1,570
13
0.160
6.17
$1.10
41
$45
$374
* Excluding taxes
5-35
-------
Draft Regulatory Support Document
Table 5.2.3-16: Fuel Cost Savings for Off-highway Motorcycles
Engine
Annual Miles
Lifetime, yr
BSFC, Ib/mile
Fuel Density (Ibs/gal)
Fuel Cost ($/gal)*
Yearly Fuel Consumption (gal/yr)
Yearly Fuel Cost ($/yr)
Lifetime Fuel Cost (NPV)
2-Stroke
1,600
12
0.268
6.17
$1.10
68
$75
$594
4-Stroke
1,600
12
0.201
6.17
$1.10
52
$57
$454
* Excluding taxes
5.2.3.2.3 Permeation Control Fuel Savings
Evaporative emissions are essentially fuel that is lost to the atmosphere. Over a the
lifetime of a typical recreational vehicle, this can result in a significant loss in fuel. The
anticipated reduction in evaporative emissions due to the permeation standards will result in
significant fuel savings. Table 5.2.3-17 presents the value of the fuel savings for control of
permeation emissions. These numbers are calculated using an estimated fuel cost of $1.10 per
gallon and fuel density of 6 Ibs/gallon (for lighter hydrocarbons which evaporate first). The
figures in Table 5.2.3-17 are based on the per vehicle emissions described in Chapter 6.
Table 5.2.3-17: Fuel Savings Per Vehicle Due to the Proposed Standards
Average Parameters
Evaporative HC reduced [tons/life]
Fuel savings [gallons/life]
Undiscounted savings [$/life @$1.10/gal]
Lifetime fuel savings (NPV, 7%)
Snowmobiles
0.0396
13
$14
$11
ATVs
0.0221
7
$8
$6
OHMCs
0.0177
6
$6
$5
5.2.3.3 Compliance Costs
We estimate ATV and off-highway motorcycle chassis-based certification to cost about
$25,000 per engine line, including $10,000 for engineering and clerical work and $15,000 for
durability and certification testing. For snowmobile engine-based certification, we estimate costs
to be about $30,000, recognizing that engine testing is somewhat more expensive than vehicle
testing due to the time needed to set up the engine on the test stand. As with other fixed costs,
we amortized the cost over 5 years of engine sales to calculate per unit certification costs shown
in Table 5.2.3-18. The actual certification costs for ATVs and off-highway motorcycles are
likely to be lower than those shown in the table above because manufacturers are likely to use
5-36
-------
Chapter 5: Costs of Control
certification data generated for the California program.
Table 5.2.3-18: Estimated Per Unit Certification Costs
units/year/family
certification costs
Snowmobiles
4,400
$1.78
ATVs
5,600
$1.17
20,000
$0.21
Off-highway
Motorcycles
6,000
$1.09
We have estimated that manufacturers must test about 0.2 percent of their production to
meet production-line testing requirements. Using per test costs of $2,500 for vehicle testing and
$5,000 per test for engine testing, we estimate a per unit cost for production line testing of $5 for
off-road motorcycles and ATVs and $10 for snowmobiles.
In general, we expect manufacturers to use existing test facilities. For manufacturers with
insufficient chassis testing capabilities for ATVs, we expect them to carry over engine-based
certifications from the California program during the transition period, but to phase-in chassis-
based certification during the transition time frame. Because the option of engine-based testing
is available for only three years, manufacturers will need to do chassis testing of ATVs by 2009.
We have therefore estimated the cost of new chassis testing facilities to be included in the cost of
the standards. The costs are based on an estimate provided by one manufacturer that a full test
cell would cost $2 million to build. We have estimated that on average manufacturers will need
two such facilities to conduct testing. The costs will vary somewhat among manufacturers
depending on the state of their existing facilities and the number of vehicle families that must be
certified. However, we believe that this is a generous estimate because some manufacturers will
likely be able to upgrade existing test facilities instead of building new facilities.
By estimating $4 million per manufacturer, with 7 manufacturers, and amortizing the
costs over 10 years (10 years x 729,000 units), we estimate an average per unit cost of $6.70. We
have used 10 years for amortization rather than 5 years because we believe it is more
representative for a capital investment that will be used at least that long.
5.2.3.4 Recreational Vehicle Total Costs
The analysis below combines the costs estimated above for various technologies into a
total composite or average cost for each vehicle type. The composite analysis weights the costs
by projecting the percentage of the use of various technologies, both in the baseline and control
scenario, to project industry-wide average per vehicle costs. The technologies and the mix
projections are discussed in Chapter 4 and are based largely on discussions with individual
manufacturers and in some cases on confidential business information.
A summary of the estimated near-term and long-term per unit average incremental costs
5-37
-------
Draft Regulatory Support Document
and fuel savings for recreational vehicles is provided in Table 5.2.3-19. Long-term costs do not
include fixed costs, which are retired, and include cost reductions due to the learning curve.
Table 5.2.3-19: Total Average Per Unit Costs and Fuel Savings
near-term costs
long-term costs
fuel savings (NPV)
Snowmobile
Phase 1
$80
$47
($67)
Snowmobile
Phase 2
$131
$77
($286)
Snowmobile
Phase 3
$89
$54
($191)
ATV
$87
$45
($29)
Off-
highway
Motorcycle
$158
$98
($53)
Tables 5.2.3-20 through 5.2.3-24 provide the detailed average, or composite, per unit
costs for snowmobiles, ATVs, and off-highway motorcycles. For snowmobiles, where there are
three phases of standards, the costs are incremental to the previous standard. The composite
costs are based on the estimated distribution of the different engine displacement ranges. We
estimated an approximate distribution of sales among the displacement ranges using limited sales
data provided by some manufacturers on a confidential basis and production data from Power
Systems Research. Incremental costs are shown both for the near-term and long-term. Long
term costs reflect the retirement of fixed costs and the affect of the learning curve, described in
section 5.1.
5-38
-------
Table 5.2.3-20: Estimated Average Costs For Snowmobiles (Phase 1)
< 500 cc
(30%)
> 500 cc
(70%)
engine
modifications
modified
carburetor
direct injection*
electronic fuel
injection
4-stroke engine
permeation
control
compliance
total
engine
modifications
modified
carburetor
direct injection*
electronic fuel
injection
4-stroke engine
permeation
control
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$18
$18
$328
$175
$455
$7
$12
--
$25
$24
$295
$119
$770
$7
$12
--
--
--
Lifetime
Fuel
Savings
$0
$0
($512)
$0
($512)
($11)
--
--
$0
$0
($1,139)
$0
($1,139)
($11)
$0
--
--
--
Baseline
0%
0%
7%
12%
7%
0%
0%
--
0%
0%
7%
12%
7%
0%
0%
--
--
--
Phase 1
60%
60%
10%
15%
10%
100%
100%
--
60%
60%
10%
15%
10%
100%
100%
--
--
--
Incrementa
ICost
$11
$11
$10
$5
$14
$7
$12
$69
$15
$14
$9
$4
$23
$7
$12
$84
$80
$47
Incremental
Fuel Savings
$0
$0
($15)
$0
($15)
($11)
$0
($41)
$0
$0
($34)
$0
($34)
($11)
$0
($79)
($67)
($67)
-------
Table 5.2.3-21: Estimated Average Costs For Snowmobiles For Phase 2 Incremental to
Phase 1
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$175
$455
$2
--
$41
$295
$119
$770
$2
--
--
--
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
--
--
$0
($1,139)
$0
($1,139)
--
--
--
--
Phase 1
0%
10%
15%
10%
0%
--
0%
10%
15%
10%
0%
--
--
--
Phase 2
30%
35%
20%
15%
100%
--
30%
35%
20%
15%
100%
--
--
--
Incrementa
ICost
$12
$82
$9
$23
$2
$128
$12
$74
$6
$39
$2
$132
$131
$77
Incremental
Fuel Savings
$0
($128)
$0
($26)
$0
($154)
$0
($285)
$0
($57)
$0
($342)
($286)
($286)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
-------
Chapter 5: Costs of Control
Table 5.2.3-22: Estimated Average Costs For Snowmobiles Phase 3 Incremental to Phase 2
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$175
$455
$2
--
$41
$295
$119
$770
$2
--
--
--
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
--
--
$0
($1,139)
$0
($1,139)
--
--
--
--
Phase 2
30%
35%
20%
15%
0%
--
30%
35%
20%
15%
0%
--
--
--
Phase 3
30%
50%
25%
20%
100%
--
30%
50%
25%
20%
100%
--
--
--
Incrementa
ICost
$0
$49
$9
$23
$2
$83
$0
$44
$6
$39
$2
$91
$89
$54
Incremental
Fuel Savings
$0
($77)
$0
($26)
$0
($103)
$0
($171)
$0
($57)
$0
($228)
($191)
($191)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
5-41
-------
Table 5.2.3-23: Estimated Average Costs For ATVs
< 200 cc
(15%)
> 200 cc
(85%)
4-stroke engine
pulse air
R&D for
exhaust
including
recalibration
permeation
control
compliance
total
4-stroke engine
pulse
air/recalibration
R&D for
exhaust
including
recalibration
permeation
control
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$219
$33
$16
$3
$13
--
$349
$27
$5
$3
$12
--
--
--
Lifetime
Fuel
Savings
(NPV)
($124)
$0
$0
($6)
--
--
($124)
$0
$0
($6)
--
--
--
--
Baseline
8%
0%
0%
0%
0%
--
93%
0%
0%
0%
0%
--
--
--
Control
100%
50%
100%
100%
100%
--
100%
50%
100%
100%
100%
--
--
--
Incrementa
ICost
$202
$17
$16
$3
$13
$251
$24
$14
$5
$3
$12
$58
$87
$45
Incremental
Fuel Savings
(NPV)
($114)
$0
$0
($6)
--
($119)
($9)
$0
$0
($6)
--
($14)
($29)
($29)
-------
Table 5.2.3-24: Estimated Average Costs For Off-highway Motorcycles (Non-competition
models only)
< 125 cc
(37%)
125<250cc
(21%)
> 250 cc
(42%)
4-stroke engine
pulse
air/recalibration
permeation
control
compliance
total
4-stroke engine
pulse
air/recalibration
permeation
control
compliance
total
4-stroke engine
pulse
air/recalibration
permeation
control
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$219
$39
$3
$7
-
$286
$39
$3
$7
-
$353
$39
$3
$7
-
-
Lifetime
Fuel
Savings
(NPV)
($140)
$0
($5)
-
-
($140)
$0
($5)
-
-
($140)
$0
($5)
-
-
-
Baseline
82%
0%
0%
0%
-
30%
0%
0%
0%
-
45%
0%
0%
0%
-
-
Control
100%
25%
100%
100%
-
100%
25%
100%
100%
-
100%
25%
100%
100%
-
-
Incrementa
ICost
$39
$10
$3
$7
$59
$200
$10
$3
$7
$220
$194
$10
$3
$7
$214
$158
$98
Incremental
Fuel Savings
(NPV)
($11)
$0
($5)
-
($16)
($98)
$0
($5)
-
($103)
($77)
$0
($5)
-
($82)
($53)
($53)
-------
Draft Regulatory Support Document
The above table for off-highway motorcycles shows the anticipated split between two-
stroke and 4-stroke models in the various engine size categories. Currently, off-highway
motorcycles are about 63 percent 2-stroke with many of the 2-stroke engines used in competition
and youth models. In recent years, more high performance and competition models have been
successfully introduced with 4-stroke engines and there appears to be a trend toward increased
use of 4-stroke engines. Models used solely for competition are exempt from emission
standards. We expect some 2-stroke competition models to continue to be available under this
exemption. For purposes of the cost analysis, we have estimated that 29 percent of all off-
highway motorcycles will be exempt as competition models and that these models will be
equipped with 2-stroke engines. We have based the estimate of exempt models on the our
estimate of the current use of 2-strokes in the motocross market. We believe the emissions
standards will be achievable for 4-stroke engines, especially with averaging, and that
manufacturers would elect to certify all 4-stroke models to market them to the widest possible
consumer base.
To account for the competition model exemption in the calculation of average costs, we
have adjusted the percentage of 2-stroke engines from the overall baseline percentage of off-
highway motorcycle sales using the 29 percent estimate noted above. This adjustment is
necessary to determine average costs only for those off-highway motorcycles covered by the
program. Table 5.2.3-25 provides our estimate of the baseline percentage of 2-strokes in overall
sales and the percentage of the non-competition model sales.
Table 5.2.3-25: Estimated Off-highway Motorcycle Percent 2-stroke Engine Usage
Displacement
< 125 cc
125 to 249 cc
> 250 cc
Overall Baseline
2-stroke percentage
42%
79%
68%
Baseline 2-stroke
percentage Excluding
Competition Models
18%
70%
55%
5.2.3.5 Recreational Vehicle Aggregate Costs
The above analyses developed incremental per vehicle cost estimates for snowmobiles,
ATVs, and off-highway motorcycles. Using these per vehicle costs and projections of future
annual sales, we have estimated total aggregate annual costs for the recreational vehicles
standards. The aggregate costs are presented on a cash flow basis, with hardware and fixed costs
incurred in the year the vehicle is sold and fuel savings occurring as the vehicle is operated over
its life. This may understate the time-value of the fixed costs because they are likely to be
incurred before the vehicle is sold; however, this has a negligible effect on the results of this
5-44
-------
Chapter 5: Costs of Control
analysis. Table 5.2.3-26 presents a summary o f the results of this analysis. As shown in the
table, aggregate net costs increase from about $65 million in 2006 to about $129 million in 2010.
Net costs are projected then to decline as fuel savings continue to ramp-up as more vehicles
meeting the standards are sold and used and fixed costs are amortized. Fuel savings are projected
to more than offset the costs of the program starting in 2015.
Table 5.2.3-26
Summary of Annual Aggregate Costs and Fuel Savings (millions of dollars)
Snowmobiles
ATVs
Off-highway
Motorcycles
Permeation control
Total
Fuel Savings
Net Costs
2006
$6.58
$42.46
$16.27
—
$65.31
($1.60)
$63.71
2010
$37.55
$62.55
$24.24
$4.59
$128.93
($39.90)
$89.03
2015
$41.91
$49.69
$21.53
$4.72
$117.85
($121.70)
($3.85)
2020
$41.56
$44.81
$22.63
$4.83
$113.83
($187.00)
($73.17)
2025
$41.56
$44.81
$23.79
$4.86
$115.02
($212.60)
($97.58)
To project annual sales, we started with 2001 sales estimates provided by industry
organizations. We then adjusted the numbers and applied sales growth estimates consistent with
the modeling performed to estimate total emissions (see Section 6.2.4.1.1). For ATVs, we added
70,000 units to account for sales from companies not included in the industry organization
estimates. Sales growth for snowmobiles and off-highway motorcycle sales is projected to be
about one percent per year. The off-road motorcycle sales were reduced by 29 percent to account
for the exemption of competition models. ATVs are modeled differently because recent sales
growth rates have been significantly higher than one percent but are at rates not likely to be
sustained indefinitely. We project that ATV sales will continue to grow at a higher rate over the
next few years but will level off by 2006. Table 5.2.3-27 provides a summary of the sales
estimates used in the aggregate cost analysis.
5-45
-------
Draft Regulatory Support Document
Table 5.2.3-27: Estimated Annual Recreational Vehicle Sales
Snowmobiles
ATVs
Off-highway
motorcycles*
2001
140,629
880,000
195,250
2006
189,497
985,754
205,210
2010
210,367
985,754
213,542
2020
240,162
985,754
235,883
Non-competition only
To calculated annual aggregate costs, the sales estimates have been multiplied by the per
unit costs. Fuel savings have been calculated using the NONROAD model to calculate the shift
in use from 2-stroke to 4-stroke vehicles, and also direct injection 2-strokes for snowmobiles,
over time. The model takes into consideration vehicle sales and scrappage rates. The standards
phase-in schedule for off-highway motorcycles and ATVs (50/100% in 2006/2007) has also been
taken into account. The detailed year-by-year analysis is provided in Chapter 7.
5-46
-------
Chapter 5: Costs of Control
Chapter 5 References
1. "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 A-2000-01, document II-A-54).
2.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, EPA420-R-99-023,
December 1999. A copy of this document is included in Air Docket A-2000-01, at Document
No. U-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).
3."Estimated Economic Impact of New Emission Standards for Heavy-Duty Highway Engines,"
by Lou Browning, Acurex Environmental Corporation Final Report (FR 97-103), March 31,
1997, Docket A-2000-01, document II-A-51.
4."Incremental Costs for Nonroad Engines: Mechanical to Electronic," Memorandum from Lou
Browning, Acurex Environmental, to Alan Stout, EPA, April 1, 1997 (Docket A-2000-01,
document U-A-52).
5."Incremental Cost Estimates for Marine Diesel Engine Technology Improvements,"
Memorandum from Louis Browning and Kassandra Genovesi, Arcadis Geraghty & Miller, to
Alan Stout, EPA, September 30, 1998 (Docket A-2000-01, document II-A-53).
6."Large SI Engine Technologies and Costs," by Lou Browning, Arthur D. Little - Acurex
Environmental, Final Report, September 2000, Docket A-2000-01, document II-A-15.
7. "Exhaust Controls Available to Reduce Emissions from Nonroad Heavy-Duty Engines," in
Clean Air Technology News, Winter 1997, p. 1 (Docket A-98-01; item II-A-02).
8. "It's Not Easy Being Green," Modern Materials Handling, April 2000 (Docket A-2000-01;
item II-A-06).
9. Letter from William Platz, Western Propane Gas Association, January 24, 2001 (Docket
A-2000-01; item II-D-40).
10. "It's Not Easy Being Green," Modern Materials Handling, April 2000 (Docket A-2000-01;
item II-A-06).
5-47
-------
Draft Regulatory Support Document
11. "Nonroad Recreational Vehicle Technologies and Costs", Arthur D. Little - Acurex
Environmental, Draft Final Report, July 2001 (Docket A-2000-01, document U-A-31)..
12. "Nonroad Recreational Vehicle Technologies and Costs", Arthur D. Little - Acurex
Environmental, Draft Final Report, July 2001 (Docket A-2000-01, document U-A-31).
13. "Nonroad Recreational Vehicle Technologies and Costs", Arthur D. Little - Acurex
Environmental, Draft Final Report, July 2001 (Docket A-2000-01, document U-A-31).
14. "Information on Costs and Effectiveness of Fluorination Received from Fluoroseal,"
Memorandum from Mike Samulski to Docket A-2000-1, March 27, 2002, Docket A-2000-01,
Document IV-B-03.
15. "Visit to Sulfo Technologies LLC on April 18, 2002," Memorandum from Mike Samulski,
U.S. EPA to Docket A-2000-01, April 22, 2002, Document IV-B-07.
16. "Shipping Costs," Memorandum from Glenn Passavant, U.S. EPA to Docket A-2000-01,
March 27, 2002, Document IV-B-01.
17. Trident Marine Hose, "Retail Price List 2001," Docket A-2000-01, Document No. IV-A-15.
18. 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 A-2000-01, Document No. IV-B-05.
19. "Meeting with Avon on June 27, 2002," Memorandum from Mike Samulski, U.S. EPA to
Docket A-2000-01, August 6, 2002, Docket A-2000-01, Document IV-E-33.
20. 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 A-2000-01, Document No. IV-B-05.
21. "Visit to Sulfo Technologies LLC on April 18, 2002," Memorandum from Mike Samulski,
U.S. EPA to Docket A-2000-01, April 22, 2002, Document IV-B-07.
22. Brake-specific fuel consumption (BSFC) based on 4-stroke BSFC estimates provided by
Power Systems Research.
23. "Monthly Energy Review", Calendar year 2000 average Refiner Prices of Petroleum
Products to End Users (Cents per gallon, excluding taxes), Energy Information Administration.
5-48
-------
Chapter 6: Emissions Inventory
Chapter 6: Emissions Inventory
6.1 Methodology
The following chapter presents our analysis of the emission impact of the standards for
recreational marine, large spark-ignition equipment, snowmobiles, all-terrain vehicles, and off-
highway motorcycles. We first present an overview of the methodology used to generate the
emissions inventories, followed by a discussion of the specific information used in generating the
inventories for each of the regulated categories of engines as well as the emission inventories.
Emissions from a typical piece of equipment are also presented.
6.1.1 Off-highway Exhaust Emissions
We are in the process of developing an emission model that will calculate emissions
inventories for most off-highway vehicle categories, including those in this rule. This draft
model is called NONROAD. For this effort we use the most recent version of the draft
NONROAD model publicly available with some updates that we anticipate will be included in
the next draft release. This section gives a brief overview of the calculation methodology used in
NONROAD for calculating exhaust emission inventories. Inputs and results specific to each of
the off-highway categories in this rule are discussed in more detail later in this chapter. For more
detailed information on the draft NONROAD model, see our website at
www. epa.gov/otaq/nonrdmdl. htm.
For the inventory calculations in this rule, each class of off-highway engines was divided
into power ranges to distinguish between technology or usage differences in each category. Each
of the engine applications and power ranges were modeled with distinct annual hours of
operation, load factors, and average engine lives. The basic equation for determining the exhaust
emissions inventory, for a single year, from off-highway engines is shown below:
Emissions = ^ population* power* load* annualuse * emissionfactor} (Eq. 6-1)
This equation sums the total emissions for each of the power ranges for a given calendar
year. "Population" refers to the number of engines estimated to be in the U.S. in a given year.
"Power" refers to the population-weighted average rated power for a given power range. Two
usage factors are included; "load" is the ratio between the average operational power output and
the rated power, and "annual use" is the average hours of operation per year. Emission factors
are applied on a brake-specific basis (g/kW-hr) and represent the weighted value between levels
from baseline and controlled engines operating in a given calendar year. Exhaust emission
6-1
-------
Draft Regulatory Support Document
inventories were calculated for HC, CO, and NOx from all engines and additionally for PM from
compression-ignition engines. Although some of the emission standards combine HC and NOx,
it is useful to consider the HC and NOx emission impacts separately. (As described throughout
this document, the standards for all-terrain vehicles (ATVs) and off-highway motorcycles are
based on a chassis test, with the standards in grams per kilometer. For these two categories of
equipment, the equation used by the NONROAD model for calculating emissions is similar to
Equation 6-1 except that the "load factor" and "power" terms are not included in the calculation,
the "annual use" is input on a miles/year basis, and the "emission factors" are entered on a gram
per mile basis.)
To be able to determine the mix between baseline and controlled engines, we need to
determine the turnover of the fleet. Through the combination of historical population and
scrappage rates, historical sales and retirement of engines can be estimated. We use a normalized
scrappage rate and fit it to the data for each engine type on average operating life. Figure 6.1.1-1
presents the normalized scrappage curve used in the draft NONROAD model. For further
discussion of this scrappage curve, see our report titled "Calculation of Age Distributions —
Growth and Scrappage," (NR-007).
Figure 6.1.1-1: Normalized Scrappage Curve
1
o
1 0.8
w
c
% 0.6
c
'O5
iS o.4
M—
O
c
|0.2
2.
LJ_
0
(
— --^
•^
\
>
V
\
\
) 0.5 1 1.5 2
Engine Age Normalized by Average Useful Life
6.1.2 Off-highway Evaporative Emissions
Evaporative emissions refer to hydrocarbons released into the atmosphere when gasoline,
or other volatile fuels, evaporate from a vehicle. For this analysis, we model three types of
evaporative emissions:
- permeation: These emissions are due to fuel that works its way through the material
used in the fuel system. Permeation is most common through plastic fuel tanks and
rubber hoses.
6-2
-------
Chapter 6: Emissions Inventory
- diurnal: These emissions are due to temperature changes throughout the day. As the
day gets warmer, the fuel heats up and begins to evaporate.
- refueling: These emissions are the vapors displaced from the fuel tank when fuel is
dispensed into the tank.
We are currently in the process of revising the inputs to the calculations for evaporative
emissions in the draft NONROAD model. The analysis for this rule includes the inputs that we
anticipate will be used in the draft NONROAD model. The evaporative emission calculations
are available in spreadsheet form in the docket.1
Because diurnal and refueling emissions are dependent on ambient temperatures and fuel
properties which vary through the nation and through the year, we divided the nation into six
regions and modeled each region individually for each day of the year. The daily temperatures by
region are based on a report which summarizes a survey of dispensed fuel and ambient
temperatures in the United States.2
6.1.2.1 Permeation Emissions
For our permeation emissions modeling, we used the emission data presented in Chapter
4 to determine the mass of hydrocarbons permeated through plastic fuel tanks and rubber fuel
hoses on recreational vehicles. No permeation occurs through metal fuel tanks. Because
permeation is very sensitive to temperature, we used Arrhenius' relationship3 to adjust the
emission factors by temperature:
P(T) = P0 x EXP(-a / T) (Eq. 6-2)
where:
T = absolute temperature
P(T) = permeation rate at T
P0 and a are constants
We determined the constants by relating the equation to the known properties of materials
used in fuel tanks and hoses (presented in Chapter 4). Based on data presented in Chapter 4,
permeation increases by about 80 percent with each 10°C increase in temperature for high
density polyethylene (HDPE). We do not have similar data for nitrile rubber used in hoses;
however, in general, permeation doubles with every 10°C increase in temperature.4 In addition,
we have data on the effect of temperature on permeation through FKM which is a
fluoroelastomer commonly used as a permeation barrier in hoses. This data, presented in
Chapter 4, supports using the general relationship, in our modeling, of doubling permeation
through hoses for every 10°C increase in temperature.
6-3
-------
Draft Regulatory Support Document
6.1.2.2 Diurnal Emissions
For diurnal emission estimates, we used the Wade equations5'6'7 to calculate grams of
hydrocarbons emitted per day per volume of fuel tank capacity. The Wade equations are well
established and are used in both the MOBILE and draft NONROAD models with an adjustment
based on empirical data. These calculations are a function of vapor space, fuel vapor pressure,
and daily temperature variation and are as follows:
Vapor space (ft3) = ((1.15 - tank fill) x tank size) / 7.841 (Eq. 6-3)
where:
tank fill = fuel in tank/fuel tank capacity
tank size = fuel tank capacity in gallons
TI (°F) = (Tmax - Tmin) x 0.922 + Tmin (Eq. 6-4)
where:
Tmax = maximum diurnal temperature (°F)
Tmin = minimum diurnal temperature (°F)
V100 (psi) = 1.0223 x RVP + [(0.0357 X RVP)/(1-0.0368 x RVP)] (Eq. 6-5)
where:
V100 = vapor pressure at 100°F
RVP = Reid Vapor Pressure of the fuel
E100 (%) = 66.401-12.718 x V100 +1.3067 x V1002 - 0.077934 x V1003
+ 0.0018407 x V1004 (Eq. 6-6)
X) = E100 + [(262 / (0.1667 * E100 + 560) - 0.113] x (100 - Tmin) (Eq. 6-7a)
Dmax (%) = E100 + [(262 / (0.1667 * E100 + 560) - 0.113] x (100 - TJ (Eq. 6-7b)
where:
Dmin/max = distillation percent at the max/min temperatures in the fuel tank
E100 = percent of fuel evaporated at 100°F from equation 6-6
P! (psi) = 14.697 - 0.53089 x D^H- 0.0077215 x Dmin2 - 0.000055631 x Dmin3
+ 0.0000001769 x Dmin4 (Eq. 6-8a)
PF (psi) = 14.697 - 0.53089 x Dmax + 0.0077215 x Dmax2 - 0.000055631 x Dmax3
+ 0.0000001769 x Dmax4 (Eq. 6-8b)
6-4
-------
Chapter 6: Emissions Inventory
Density (Ib/gal) = 6.386 - 0.0186 x RVP (Eq. 6-9)
MW (Mb mole) = (73.23 - 1.274 x RVP) + [0.5 x( Tmin + Tx) - 60] x 0.059 (Eq. 6-10)
Diurnal emissions (grams) = vapor space x 454 x density x [520 / (690 - 4 x MW)]
x 0.5 x [Pj / (14.7 - P!> + PF / (14.7 - PF)]
x [(14.7 - Pj) / (Tmin + 460) - (14.7 - PF) / (Tx + 460)] (Eq. 6-11)
where:
MW = molecular weight of hydrocarbons from equation 6-10
PI/F = initial and final pressures from equation 6-8
We use these same equations in our modeling of evaporative emissions from on-highway
vehicles. However for on-highway applications we make a correction of 0.78 based on empirical
data.8 Because this correction is based on automotive applications we do not apply this
correction factor here. Instead we use a correction factor of 0.65 which is based on the data we
collected on exposed fuel tanks vented through a hose. This test data is presented in Table 6.1.2-
1 compared to calculated theoretical results.
Table 6.1.2-1
Baseline Diurnal Evaporative Emission Results (varied temperature)
Fuel Tank Capacity
17 gallons
24 gallons
Evaporative HC
[g/gallon/day]
1.39
1.5
Wade HC
[g/gallon/day]
2.3
2.3
ratio of measured to
Wade
0.6
0.65
Title 40, Section 80.27 of the Code of Federal Regulations specifies the maximum
allowable fuel vapor pressure allowed for each state in the U.S. for each month of the year. We
used these limits as an estimate of fuel vapor pressure in our calculations.
6.1.2.3 Refueling Vapor Displacement
We used the draft NONROAD model to determine the amount of fuel consumed by
recreational vehicles. To calculate refueling emissions, we used an empirical equation to
calculate grams of vapor displaced during refueling events. This equation was developed based
on testing of 22 highway vehicles under various refueling scenarios and in the benefits
calculations for our onboard refueling vapor recovery rulemaking for cars and trucks.9 These
calculations are a function of fuel vapor pressure, ambient temperature, and dispensed fuel
temperature. The refueling vapor generation equation is as follows:
Refueling vapor (g/gal) = EXP(-1.2798 - 0.0049 x (Td - Ta) + 0.0203 x Td
0.1315 xRVP)
(Eq. 6-12)
6-5
-------
Draft Regulatory Support Document
where:
Td = dispensed fuel temperature (°F)
Ta = ambient fuel temperature (°F)
RVP = Reid Vapor Pressure of the fuel
6.2 Effect of Emission Controls by Engine/Vehicle Type
The remainder of this chapter discusses the inventory results for each of the classes of
engines/vehicles included in this document. These inventory projections include both exhaust
and evaporative emissions. Also, this section describes inputs and methodologies used for the
inventory calculations that are specific to each engine/vehicle class.
6.2.1 Compression-Ignition Recreational Marine
We projected the annual tons of exhaust HC, CO, NOx, and PM from CI recreational
marine engines using the draft NONROAD model discussed above. This section describes
inputs to the calculations that are specific to CI recreational marine engines then presents the
results. These results are for the nation as a whole and include baseline and control inventory
projections.
6.2.1.1 Inputs for the Inventory Calculations
Several usage inputs are specific to the calculations for CI recreational marine exhaust
emissions. These inputs are load factor, annual use, average operating life, and population.
Based on data collected in developing the draft NONROAD model, we use a load factor of 35
percent, an annual usage factor of 200 hours, and an average operating life of 20 years. The draft
NONROAD model includes current and projected engine populations. Table 6.2.1-1 presents
these population estimates for selected years. These population estimates have been updated
since the NPRM using new data collected from the boating industry discussed in Chapter 2.
Table 6.2.1-1
Projected CI Recreational Marine Population by Year
Year
population
2000
261,000
2005
301,000
2010
340,000
2020
419,000
2030 I
497,000 1
We used the data presented in Chapter 4 to develop the baseline emission factors. For the
control emission factors, we projected that the manufacturers will design their engines to meet
the standard at regulatory useful life with a small compliance margin. (The regulatory useful life
is the period of time for which a manufacturer must demonstrate compliance with the emission
standards.) To determine the HC and NOx split for the standards, we used the HC and NOx data
presented in Chapter 4 from CI recreational marine engines near the standards. Consistent with
6-6
-------
Chapter 6: Emissions Inventory
our modeling of heavy-duty highway emissions, we assumed a compliance margin of 8 percent.
This compliance margin is based on historical practices for highway and nonroad engines with
similar technology. Engine manufacturers give themselves some cushion below the certification
level on average so that engine-to-engine variability will not cause a significant number of
engines to exceed the standard. Also, we used the deterioration factors in the draft NONROAD
model which have been updated since the NPRM; the only significant update is to the PM
deterioration factor which is now larger. Table 6.2.1-2 presents the emission factors used in this
analysis for new engines and for engines deteriorated to the regulatory useful life (10 years).
Table 6.2.1-2
Emission Factors for CI Recreational Marine Engines
Engine Technology
baseline
controlled:
< 0.9 liters/cylinder
0.9-1.2 liters/cylinder
> 1.2 liters/cylinder
HC [g/kW-hr]
new 10 yrs
0.295 0.300
0.181 0.184
0.181 0.184
0.182 0.184
NOx [g/kW-hr]
new 10 yrs
8.94 9.05
6.69 6.72
6.41 6.44
6.42 6.44
CO [g/kW-hr]
new 10 yrs
1.27 1.39
1.27 1.39
1.27 1.39
1.27 1.39
PM [g/kW-hr]
new 10 yrs
0.219 0.270
0.219 0.270
0.219 0.270
0.181 0.184
In our analysis of the CI recreational marine engine emissions inventory, we may
underestimate emissions, especially PM, due to engine deterioration in-use. We believe that
current modeling only represents properly maintained engines, but may not be representative of
in-use tampering or malmaintenance. However, we have not fully evaluated the limited data
currently available and we are in the process of collecting more data on in-use emission
deterioration. Once this has been completed we will decide whether or not we need to update our
deterioration rates both in this analysis and in the Draft NONROAD model.
6.2.1.2 Reductions Due to the Standard
We anticipate that the standards will result in a 28 percent reduction in HC+NOx and a
25 percent reduction in PM in 2030. We are not claiming any benefits from the cap on CO
emissions. The following tables present our projected exhaust emission inventories for CI
recreational marine engines and the anticipated emission reductions.
Table 6.2.1-3
Projected HC Reductions for CI Recreational Marine Engines [short tons]
Calendar Year
2000
2005
2010
2020
2030
Baseline
1,270
1,460
1,650
2,030
2,410
Control
1,270
1,460
1,490
1,450
1,510
Reduction
0
0
159
575
899
% Reduction
0%
0%
10%
28%
37%
6-7
-------
Draft Regulatory Support Document
Table 6.2.1-4
Projected NOx Reductions for CI Recreational Marine Engines [short tons]
Calendar Year
2000
2005
2010
2020
2030
Baseline
38,000
43,600
49,400
60,800
72,200
Control
38,000
43,600
45,800
48,000
52,200
Reduction
0
0
3,550
12,800
20,000
% Reduction
0%
0%
7%
21%
28%
Table 6.2.1-5
Projected PM Reductions for CI Recreational Marine Engines [short tons]
Calendar Year
2000
2005
2010
2020
2030
Baseline
,000
,150
,300
,600
,900
Control
1,000
1,150
1,230
1,310
1,420
Reduction
0
0
75
294
478
% Reduction
0%
0%
6%
18%
25%
6.2.1.3 Per Vessel Emissions from CI Recreational Marine Engines
This section describes the development of the HC plus NOx emission estimates on a per
engine basis over the average lifetime of typical CI recreational marine engines. As in the cost
analysis in Chapter 5, we look at three engine sizes for this analysis (100, 400, and 750 kW) as
well as a composite of all engine sizes. The emission estimates were developed to estimate the
cost per ton of the standards as presented in Chapter 7.
The new and deteriorated emission factors used to calculate the HC and NOx emissions
from typical CI recreational marine engines were presented in Table 6.2.1-2. A brand new
engine emits at the zero-mile level presented in the table. As the engine ages, the emission levels
increase based on the pollutant-specific deterioration factor. The load factor for these engines is
estimated to be 0.35, the annual usage rate is estimated to be 200 hours per year, and the average
lifetime is estimated to be 20 years.
Using the information described above and the equation used for calculating emissions
from nonroad engines (see Equation 6-1), we calculated the lifetime HC+NOx emissions from
typical marine engines both baseline and controlled engines. Table 6.2.1-6 presents these results
with and without the consideration of a 7 percent per year discount on the value of emission
reductions.
6-8
-------
Chapter 6: Emissions Inventory
Table 6.2.1-6
Lifetime HC+NOx Emissions from Typical CI Recreational Marine Engines (tons)
Engine
Size
100 kW
400 kW
750 kW
Composite
Baseline
Undiscounted
1.44
5.78
7.18
2.58
Discounted
0.82
3.26
4.53
1.47
Control
Undiscounted
1.01
4.06
5.08
1.81
Discounted
0.57
2.30
3.20
1.03
Reduction
Undiscounted
0.43
1.72
2.10
0.77
Discounted
0.24
0.97
1.32
0.44
6.2.1.4 Crankcase Emissions from CI Recreational Marine Engines
We anticipate some benefits in HC, NOx, and PM from the closed crankcase
requirements for CI recreational marine engines. Based on limited engine testing, we estimate
that crankcase emissions of HC and PM diesel engines are each about 0.013 g/kW-hr.10 NOx
data varies, but crankcase NOx emissions may be as high as HC and PM. Therefore, we use the
same crankcase emission factor of 0.013 g/kW-hr for each of the three constituents.
For this analysis, we assume that manufacturers will use the low cost option of routing
crankcase emissions to the exhaust and including them in the total exhaust emissions when the
engine is designed to the standards. Because exhaust emissions must be reduced slightly to
offset any crankcase emissions, the crankcase emission control is functionally equivalent to a 100
percent reduction in crankcase emissions.
The engine data we use to determine crankcase emission levels is based on new heavy-
duty engines. We do not have data on the effect of in-use deterioration of crankcase emissions.
However, we expect that these emissions increase as the engine wears. Therefore, this analysis
may underestimate the benefits that would result from our crankcase emission requirements.
Table 6.2.1-7 presents our estimates of the fleetwide reductions crankcase emissions from CI
recreational marine engines.
Table 6.2.1-7
Crankcase Emissions Reductions from CI Recreational Marine Engines [short tons]
Calendar Year
2000
2005
2010
2020
2030
HC+NOx
0
0
39
145
260
PM
0
0
19
73
130
6-9
-------
Draft Regulatory Support Document
6.2.2 Large Spark-Ignition Equipment
6.2.2.1 Exhaust Emissions from Large SI Equipment
We projected the annual tons of exhaust HC, CO, and NOx from large industrial spark-
ignition (SI) engines using the draft NONROAD model described above. This section describes
inputs to the calculations that are specific to these engines then presents the results of the
modeling.
6.2.2.1.1 Inputs for Exhaust Inventory Calculations
Several usage inputs are specific to the calculations for Large SI engines. These inputs
are load factor, annual use, average operating life, and population. Because the Large SI category
is made up of many applications, the NONROAD model contains application-specific
information for each of the applications making up the Large SI category. Table 6.2.2-1 presents
the inputs used in the NONROAD model for each of the Large SI applications. (The average
operating life for a given application can vary within an application by power category. In such
cases, the average operating life value presented in Table 6.2.2-1 is based on the average
operating life estimate for the engine with the average horsepower listed in the table.)
The NONROAD model generally uses population data based on information from Power
Systems Research, which is based on historical sales information adjusted according to survival
and scrappage rates. We are, however, using different population estimates for forklifts based on
a recent market study.11 That study identified a 1996 population of 491,321 for Class 4 through 6
forklifts, which includes all forklifts powered by internal combustion engines. Approximately 80
percent of those were estimated to be fueled by propane, with the rest running on either gasoline
or diesel fuel. Assuming an even split between gasoline and diesel for these remaining forklifts
leads to a total population of spark-ignition forklifts of 442,000. The NONROAD model
therefore uses this estimate for the forklift population, which is significantly higher than that
estimated by Power Systems Research. Table 6.2.2-1 shows the estimated population figures
used in the NONROAD model for each application, adjusted for the year 2000.
The split between LPG and gasoline in various applications warrants further attention.
Engines are typically sold without fuel systems, which makes it difficult to assess the distribution
of engines sales by fuel type. Also, engines are often retrofitted for a different fuel after a period
of operation, making it still more difficult to estimate the prevalence of the different fuels. The
high percentage of propane systems for forklifts, compared with about 60 percent estimated by
Power Systems Research, can be largely attributed to expenses related to maintaining fuel
supplies. LPG cylinders can be readily exchanged with minimal infrastructure cost as compared
to gasoline storage. Natural gas systems typically offer the advantage of pipeline service, but the
cost of installing high-pressure refueling equipment is an obstacle to increased use of natural gas
systems.
Some applications of nonroad SI equipment face much different refueling situations.
6-10
-------
Chapter 6: Emissions Inventory
Lawn and garden equipment is usually not centrally fueled and therefore operates almost
exclusively on gasoline, which is more readily available. Agriculture equipment is
predominantly powered by diesel engines. Most of these operators likely have storage tanks for
diesel fuel. For those who use spark-ignition engines in addition to, or instead of, the diesel
models, we expect them in many cases to be ready to invest in gasoline storage tanks as well,
resulting in little or no use of LPG or natural gas for those applications. For construction, general
industrial, and other equipment, there may be a mix of central and noncentral fueling, and motive
and portable equipment. We therefore believe that estimating an even mix of LPG and gasoline
for these engines is most appropriate. The approximate distribution of fuel types for the
individual applications used in the NONROAD model are listed in Table 6.2.2-1.
Table 6.2.2-1
Operating Parameters and Population Estimates for Various Large SI Applications
Application
Forklift
Generator
Commercial turf
Aerial lift
Pump
Welder
Baler
Air compressor
Scrubber/sweeper
Chipper/grinder
Swathers
Leaf blower/vacuum
Sprayers
Specialty vehicle/cart
Oil field equipment
Skid/steer loader
Other agriculture equipment
Irrigation set
Avg. Rated
HP
69
59
28
52
45
67
44
65
49
66
95
79
66
66
44
47
162
97
Load
Factor
0.30
0.68
0.60
0.46
0.69
0.68
0.62
0.56
0.71
0.78
0.52
0.94
0.65
0.58
0.90
0.58
0.55
0.60
Hours
per Year
1800
115
682
361
221
408
68
484
516
488
95
282
80
65
1104
310
124
716
Average
Operating
Life (yrs)
8.3
25.0
3.7
18.1
9.8
12.7
25.0
11.1
4.1
7.9
25.0
11.3
25.0
25.0
1.5
8.3
25.0
7.0
2000
Population
499,693
143,705
55,433
38,637
35,541
19,006
18,635
17,261
13,272
13,000
12,030
11,797
9,429
9,145
7,855
7,427
5,488
5,176
Percent
LPG/CNG
95
100
0
50
50
50
0
50
50
50
0
0
0
50
100
50
0
50
6-11
-------
Draft Regulatory Support Document
Application
Trencher
Rubber-tired loader
Other general industrial
Terminal tractor
Bore/drill rig
Concrete/industrial saw
Rough terrain forklift
Other material handling
Ag. tractor
Paver
Roller
Other construction
Crane
Pressure washer
Paving equipment
Aircraft support
Gas compressor
Front mowers
Other lawn & garden
Tractor/loader/backhoe
Hydro power unit
Surfacing equipment
Railway maintenance
Crushing/processing equip
Refrigeration/AC
Dumpers/tenders
Combines
Avg. Rated
HP
54
71
82
93
78
46
66
67
82
48
55
126
75
39
39
99
110
32
61
58
50
40
33
63
55
66
123
Load
Factor
0.66
0.71
0.54
0.78
0.79
0.78
0.63
0.53
0.62
0.66
0.62
0.48
0.47
0.85
0.59
0.56
0.85
0.65
0.58
0.48
0.56
0.49
0.62
0.85
0.46
0.41
0.74
Hours
per Year
402
512
713
827
107
610
413
386
550
392
621
371
415
115
175
681
6000
86
61
870
450
488
184
241
605
127
125
Average
Operating
Life (yrs)
11.3
8.8
7.8
4.7
25.0
3.2
11.5
7.3
8.8
5.8
7.8
16.8
15.4
15.3
14.5
7.9
0.8
25.0
25.0
7.2
6.0
6.3
13.1
14.6
10.8
25.0
25.0
2000
Population
3,622
3,172
2,922
2,698
2,604
2,264
1,923
1,594
1,597
1,365
1,360
1,275
1,239
1,212
1,107
904
783
658
402
359
331
313
276
235
169
124
31
Percent
LPG/CNG
50
50
50
50
50
50
50
50
0
50
50
50
50
50
50
50
100
0
0
50
50
50
50
50
100
0
0
6-12
-------
Chapter 6: Emissions Inventory
An additional issue related to population figures is the level of growth factored into
emission estimates for the future. The NONROAD model incorporates application-specific
growth figures based on projections from Power Systems Research. The model projects growth
rates separately for the different fuels for each application. Table 6.2.2-2 presents the population
estimates of Large SI engines (rounded to the nearest 1,000 units) by fuel type for selected years.
Table 6.2.2-2
Projected Large SI Population by Year
Category
Gasoline LSI
LPG LSI
CNG LSI
Total LSI
2000
224,000
645,000
88,000
957,000
2005
232,000
766,000
97,000
1,095,000
2010
240,000
890,000
108,000
1,238,000
2020
261,000
1,132,000
132,000
1,525,000
2030
294,000
1,364,000
155,000
1,813,000
Southwest Research Institute recently compiled a listing of test data from past and current
testing projects.12 These tests were all conducted on new or nearly new engines and are used in
the NONROAD model as zero-mile levels (ZML). Table 6.2.2-3 summarizes this test data by
fuel type. (The emission levels for gasoline engines are a population-weighted average of the
water-cooled and air-cooled average emission levels, assuming air-cooled engines are 3 percent
of all large spark-ignition engines, or 13 percent of gasoline large spark-ignition engines.) All
engines were operated on the steady-state ISO C2 duty cycle, except for two engines that were
tested on the steady-state D2 cycle. The results from the different duty cycles were comparable.
Lacking adequate test data for engines fueled by natural gas, we model those engines to have the
same emission levels as those fueled by liquefied petroleum gas (LPG), based on the similarity
between engines using the two fuels (in the case of hydrocarbon emissions, the equivalence is
based on non-methane hydrocarbons).
Emission levels often change as an engine ages. In most cases, emission levels increase
with time, especially for engines equipped with technologies for controlling emissions. We
developed deterioration factors for uncontrolled Large SI engines based on measurements with
comparable highway engines.13 Table 6.2.2-3 also shows the deterioration factors that apply at
the median lifetime estimated for each type of equipment. For example, a deterioration factor of
1.26 for hydrocarbons multiplied by the emission factor of 6.2 g/hp-hr for new gasoline engines
indicates that modeled emission levels increase to 7.8 g/hp-hr when the engine reaches its median
lifetime. The deterioration factors are linear multipliers, so the modeled deterioration at different
points can be calculated by simple interpolation.
Emissions during transient operation can be significantly higher than during steady-state
operation. Based on emission measurements from highway engines comparable to uncontrolled
6-13
-------
Draft Regulatory Support Document
Large SI engines, we have measured transient emission levels that are 30 percent higher for HC
and 45 percent higher for CO relative to steady-state measurements.14 The NONROAD model
therefore multiplies steady-state emission factors by a transient adjustment factor (TAP) of 1.3
for HC and 1.45 for CO to estimate emission levels during normal, transient operation. Test data
do not support adjusting NOx emission levels for transient operation and so a TAP of 1.0 is used
for NOx emissions. Also, the model applies no transient adjustment factor for generators,
pumps, or compressors, since engines in these applications are less likely to experience transient
operation.
Table 6.2.2-3
Zero-Mile Level Emission Factors (g/hp-hr), Deterioration Factors (at Median Life)
and Transient Adjustment Factors for Pre-Control Large SI Engines
Fuel Category
Gasoline
LPG
CNG
THC
ZML
3.9
1.7
24.6
DF
1.26
1.26
1.26
TAF
1.3
1.3
1.3
CO
ZML
107.2
28.2
28.2
DF
1.35
1.35
1.35
TAF
1.45
1.45
1.45
NOx
ZML
8.4
12.0
12.0
DF
1.03
1.03
1.03
TAF
1.0
1.0
1.0
As manufacturers comply with the Phase 1 emission standards for Large SI engines, we
expect the emission factors, deterioration factors and transient adjustment factors will be
affected. To estimate the Phase 1 deterioration factors, we relied upon deterioration information
for current Class lib heavy-duty gasoline engines developed for the MOBILE6 emission model.
Class lib engines are the smallest heavy-duty engines and are comparable in size to many Large
SI engines. They also employ catalyst/fuel system technology similar to the technologies we
expect to be used on Large SI engines. To estimate the Phase 1 emission factors at zero miles,
we back-calculated the emission levels based on the standards and the estimated deterioration
factors, assuming manufacturers will design to meet a level 10 percent below the standard to
account for variability. (The emission levels for Phase 1 gasoline engines were back-calculated
from a population-weighted average of the Phase 1 standards for water-cooled and air-cooled
engines, assuming 13 percent of gasoline engines are air-cooled.) Given that these engines will
employ a catalyst to meet the standards, we believe a 10 percent compliance margin is
appropriate. (Including a margin of compliance below the standards is a practice that
manufacturers have followed historically to provide greater assurance that their engines meet
emission standards in the event of a compliance audit.) Because the standards include an
HC+NOx standard, we assumed the HC/NOx split would stay the same as pre-control engines at
the end of the regulated useful life. Table 6.2.2-4 presents the zero-mile levels, deterioration
factors used in the analysis of today's Phase 1 standards for Large SI engines. The Phase 1
standards are to take effect in 2004 for all engines.
The transient adjustment factors for Phase 1 engines were based on testing performed at
Southwest Research Institute on engines that are similar to those expected to be certified under
6-14
-------
Chapter 6: Emissions Inventory
the Phase 1 standards. The testing was performed on one gasoline fueled engine and two LPG-
fueled engines. A complete description of the testing performed and the results of the testing is
summarized in the docket for the rulemaking.15 Because we did not have any test results for
CNG-fueled engines, the same transient adjustment factors for LPG-fueled engines were used.
Table 6.2.2-4
Zero-Mile Level Emission Factors (g/hp-hr), Deterioration Factors (at Median Life)
and Transient Adjustment Factors for Phase 1 Large SI Engines
Fuel Category
Gasoline
LPG
CNG
THC
ZML
0.59
0.25
3.7
DF
1.64
1.64
1.64
TAF
1.7
2.9
2.9
CO
ZML
29.9
24.5
24.5
DF
1.36
1.36
1.36
TAF
1.7
1.45
1.45
NOx
ZML
1.5
2.1
2.1
DF
1.15
1.15
1.15
TAF
1.4
1.5
1.5
In a similar manner, as manufacturers comply with the Phase 2 emission standards for
Large SI engines, we expect the emission factors, deterioration factors and transient adjustment
factors will be affected. To estimate the Phase 2 deterioration factors, we relied upon the same
information noted above for Phase 1 engines. The technologies used to comply with the Phase 2
standards are expected to be further refinements of the technologies we expect to be used on
Phase 1 Large SI engines. For that reason, we are applying the Phase 1 deterioration factors to
the Phase 2 engines. To estimate the Phase 2 emission factors at zero miles, we back-calculated
the emission levels based on the standards and the estimated deterioration factors, assuming
manufacturers will design to meet a level 10 percent below the standard to account for
variability. Given that these engines will employ a catalyst to meet the standards, we believe a
10 percent compliance margin is appropriate. (Including a margin of compliance below the
standards is a practice that manufacturers have followed historically to provide greater assurance
that their engines meet emission standards in the event of a compliance audit.) As noted in
Chapter 4, the Phase 2 CO standard for all engines (except air-cooled gasoline engines) is
dependent on the HC+NOx level of the engine. For modeling purposes, we have assumed that
all engines (except air-cooled gasoline engines) will certify at an equivalent HC+NOx standard
of 1.7 g/kW-hr, yielding a CO standard of 7.9 g/kW-hr. Again, because the standards include an
HC+NOx standard, we assumed the HC/NOx split would stay the same as pre-control engines at
the end of the regulated useful life. (As with the Phase 1 emission factors, the emission levels for
Phase 2 gasoline engines were back-calculated from a population-weighted average of the Phase
2 standards for water-cooled and air-cooled engines, assuming 13 percent of gasoline engines are
air-cooled.) Table 6.2.2-5 present the zero-mile levels, deterioration factors used in the analysis
of today's Phase 2 standards for Large SI engines. The Phase 2 standards are to take effect in
2004 for all engines.
Under the Phase 2 program for Large SI engines, the test procedure will be switched from
a steady-state test to a transient test. Therefore, the in-use emission performance of Phase 2
6-15
-------
Draft Regulatory Support Document
engines should be similar to the emissions performance over the test cycle. For this reason, the
transient adjustment factors for Phase 2 engines is set at 1.0 for all pollutants.
Table 6.2.2-5
Zero-Mile Level Emission Factors (g/hp-hr), Deterioration Factors (at Median Life)
and Transient Adjustment Factors for Phase 2 Large SI Engines
Fuel Category
Gasoline
LPG
CNG
THC
ZML
0.3
0.1
1.6
DF
1.64
1.64
1.64
TAF
1.0
1.0
1.0
CO
ZML
11.9
3.9
3.9
DF
1.36
1.36
1.36
TAF
1.0
1.0
1.0
NOx
ZML
0.7
0.9
0.9
DF
1.15
1.15
1.15
TAF
1.0
1.0
1.0
6.2.2.1.2 Exhaust Emission Reductions Due to the Standards
Tables 6.2.2-6 through 6.2.2-8 present the projected HC, CO, and NOx exhaust emissions
inventories respectively, assuming engines remain uncontrolled and assuming we adopt the Phase
1 and Phase 2 standards. The tables also contain estimated emission reductions for each of the
pollutants. We anticipate that the standards will result in a 92 percent reduction in exhaust HC,
91 percent reduction in NOx, and a 88 percent reduction in CO by 2020
Table 6.2.2-6
Projected HC Inventories and Reductions for Large SI Engines (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
166,000
180,000
197,000
235,000
274,000
Control
166,000
136,000
59,000
19,000
17,000
Reduction
0
44,000
138,000
216,000
257,000
% Reduction
0%
24%
70%
92%
94%
6-16
-------
Chapter 6: Emissions Inventory
Table 6.2.2-7
Projected CO Inventories and Reductions for Large SI Engines (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
1,734,000
1,873,000
2,022,000
2,336,000
2,703,000
Control
1,734,000
1,712,000
945,000
277,000
265,000
Reduction
0
161,000
1,077,000
2,059,000
2,438,000
% Reduction
0%
9%
53%
88%
90%
Table 6.2.2-8
Projected NOx Inventories and Reductions for Large SI Engines (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
308,000
348,000
389,000
472,000
553,000
Control
308,000
273,000
118,000
43,000
44,000
Reduction
0
75,000
271,000
429,000
509,000
% Reduction
0%
21%
70%
91%
92%
6.2.2.2 Evaporative and Crankcase Emission Control from Large SI Equipment
We projected the annual tons of hydrocarbons evaporated into the atmosphere from Large
SI gasoline engines using the methodology discussed above in Section 6.1.2. These evaporative
emissions include diurnal and refueling emissions. Although the standards do not specifically
require the control of refueling emissions, we have included them in the modeling for
completeness. We have also calculated estimates of hot-soak and running losses for Large SI
gasoline engines using separate information on those emissions. Finally, we present crankcase
emissions for all Large SI engines based on the NONROAD model. This section describes
inputs to the calculations that are specific to Large SI engines and presents our baseline and
controlled national inventory projections for evaporative and crankcase emissions.
6.2.2.2.1 Inputs for the Inventory Calculations
Several usage inputs are specific to the evaporative emission calculations for Large SI
engines. These inputs are fuel tank sizes, population, and distribution throughout the nation.
The draft NONROAD model includes current and projected engine populations for each state
and we used this distribution as the national fuel tank distribution. Table 6.2.2-9 presents the
population of Large SI gasoline engines for 1998.
6-17
-------
Draft Regulatory Support Document
Table 6.2.2-9
1998 Population of Large SI Gasoline Engines by Region
Region
Northeast
Southeast
Southwest
Midwest
West
Northwest
Total
Total
87,200
38,300
22,700
35,000
28,600
9,200
221,000
The draft NONROAD model breaks this engine distribution further into ranges of engine
sizes. For each of these power ranges we apply a fuel tank size for our evaporative emission
calculations based on the fuel tank sizes used in the NONROAD model.
Table 6.2.2-10 presents the baseline diurnal emission factors for the certification test
conditions and a typical summer day with low vapor pressure fuel and a half-full tank.
Table 6.2.2-10
Diurnal Emission Factors for Test Conditions and Typical Summer Day
Evaporative Control
baseline
72-96°F, 9 RVP* Fuel, 40% fill
1.5 g/gallon/day
60-84°F, 8 RVP* Fuel, 50% fill |
0.55 g/gallon/day |
* Reid Vapor Pressure
We used the draft NONROAD model to determine the amount of fuel consumed by Large
SI gasoline engines. As detailed earlier in Table 6.2.2-1, the NONROAD model has annual
usage rates for all Large SI applications. Table 6.2.2-11 presents the fuel consumption estimates
we used in our modeling. For 1998, the draft NONROAD model estimated that Large SI
gasoline engines consumed about 300 million gallons of gasoline.
Table 6.2.2-11
Fuel Consumption Estimates used in Refueling Calculations for Large SI Gasoline Engines
Technology
Pre -control
Tier I/Tier 2
BSFC, Ib/hp-hr
0.605
0.484
To estimate inventories of hot-soak and running loss emissions from Large SI gasoline
engines, we applied a factor to the diurnal emissions inventory estimates based on evaporative
emission inventories prepared for the South Coast Air Quality Management District.16 The hot
soak inventory was estimated to be 3.9 times as high as the diurnal inventory, and the running
6-18
-------
Chapter 6: Emissions Inventory
loss inventory was estimated to be two-thirds of the diurnal inventory. Finally, crankcase
emissions (from all Large SI engines) were generated using the draft NONROAD model.
Table 6.2.2-12 contains the baseline evaporative emission and crankcase emission
inventories for Large SI engines.
Table 6.2.2-12
Baseline Evaporative and Crankcase Emissions from Large SI Equipment [short tons]
Calendar
Year
2000
2005
2010
2020
2030
Diurnal
700
720
750
810
920
Refueling
,400
,430
,520
,680
,900
Hot-Soak
2,720
2,820
2,920
3,171
3,577
Running Loss
470
480
500
540
610
Crankcase
54,550
59,100
64,950
77,340
90,180
6.2.2.2.2 Evaporative and Crankcase Emission Reductions Due to the Requirements
We anticipate that the evaporative emission requirements for Large SI engines will result
in about a 90 percent reduction in diurnal, running loss emissions, and hot soak emissions. The
new requirements for Large SI equipment includes an evaporative emission standard of 0.2
grams per gallon of fuel tank capacity for 24-hour day when temperatures cycle between 72° and
96° F. In our modeling, we consider a 3.0 psi pressure relief valve. In this case, the model only
accounts for hydrocarbon emissions generated at pressures greater than 3.0 psi (see Equation 7).
The evaporative emission requirements are scheduled to take effect in 2007 with the Tier 2
requirements, except for the hot-soak requirements which will take effect in 2004 with the Tier 1
requirements. In addition, because the fuel consumption of Large SI engines will be reduced by
20 percent, the refueling emissions will be reduced proportionally as well. The refueling benefits
will be realized beginning in 2004 as the Tier 1 standards take effect. Finally, the standards also
require that engines have a closed crankcase. We expect the crankcase emissions will generally
be routed to the engine and combusted, nearly eliminating crankcase emissions. For modeling
purposes, we have assumed that the crankcase emissions are reduced by 90 percent. The
crankcase requirements are schedule to take effect in 2004 with the Tier 1 requirements.
Table 6.2.2-13 present the evaporative emission inventories and crankcase emissions
inventories for Large SI engines based on the reductions in emissions noted above. The
reductions are achieved over time as the fleet turns over to Tier 1 or Tier 2 engines. Table 6.2.2-
14 presents the corresponding reductions in evaporative and crankcase emissions for Large SI
engines due to the requirements.
6-19
-------
Draft Regulatory Support Document
Table 6.2.2-13
Control Case Evaporative and Crankcase
Emissions from Large SI Equipment [short tons]
Calendar
Year
2000
2005
2010
2020
2030
Diurnal
700
720
550
150
70
Refueling
,400
,380
,360
,360
,520
Hot-Soak
2,720
2,440
1,600
410
260
Running Loss
470
480
370
100
50
Crankcase
54,550
44,930
25,170
12,880
9,020
Table 6.2.2-14
Reductions in Evaporative and Crankcase
Emissions from Large SI Equipment [short tons]
Calendar
Year
2000
2005
2010
2020
2030
Diurnal
0
0
200
670
850
Refueling
0
50
160
320
380
Hot-Soak
0
380
1,320
2,760
3,316
Running Loss
0
0
130
450
570
Crankcase
0
14,200
39,800
64,500
81,200
6.2.2.3 Per Equipment Emissions from Large SI Equipment
The following section describes the development of the HC+NOx emission estimates on
a per piece of equipment basis over the average lifetime or typical Large SI piece of equipment.
The emission estimates were developed to estimate the cost per ton of the standards as presented
in Chapter 7. The estimates are made for an average piece of Large SI equipment for each of the
three fuel groupings (gasoline, LPG, and CNG). Although the emissions vary from one nonroad
application to another, we are presenting the average numbers for the purpose of determining the
emission reductions associated with the standards from a typical piece of Large SI equipment
over its lifetime.
In order to estimate the emission from a piece of Large SI equipment, information on the
emission level of the engine, the power of the engine, the load factor of the engine, the annual
hours of use of the engine, and the lifetime of the engine are needed. The values used to predict
the per piece of equipment emissions for this analysis and the methodology for determining the
values are described below.
The information necessary to calculate the HC and NOx emission levels of a piece of
equipment over the lifetime of a typical piece of Large SI equipment were presented in Table
6-20
-------
Chapter 6: Emissions Inventory
6.2.2-3 through Table 6.2.2-5. A brand new piece of equipment emits at the zero-mile level
presented in the tables. As the equipment ages, the emission levels increase based on the
pollutant-specific deterioration factor. Deterioration, as modeled in the NONROAD model,
continues until the equipment reaches the median life of that equipment type. The deterioration
factors presented in Table 6.2.2-3 through Table 6.2.2-5 when applied to the zero-mile levels
presented in the same tables, represent the emission level of the engine at the end of its median
life. The emissions at any point in time in between can be determined through interpolation.
(For this analysis, the HC emissions from CNG engines is calculated on an NMHC+NOx basis,
with NMHC emissions estimated to be 4.08 percent of THC emissions.)
To estimate the average power for equipment in each of the Large SI fuel groupings, we
used the population estimates contained in the NONROAD model and the average horsepower
information presented in Table 6.2.2-1. To simplify the calculations, we used the most common
applications within each category that represent 80 percent or more of the fuel grouping
population. For gasoline engines, the top ten applications with the highest populations were
used. For LPG and CNG, the top four applications with the highest populations were used.
Table 6.2.2-15 lists the applications used in the analysis.
Table 6.2.2-15
Large SI Applications Used in Per Equipment Analysis
Gasoline
LPG
CNG
Commercial Turf Equipment
Balers
Forklifts
Aerial Lifts
Pumps
Swathers
Leafblowers/Vacuums
Sprayers
Welders
Air Compressors
Forklifts
Generator Sets
Aerial Lifts
Pumps
Forklifts
Generator Sets
Other Oil Field Equipment
Irrigation Sets
Based on the applications noted above for each fuel, we calculated the population-
weighted average horsepower for Large SI equipment to be 51.6 hp for gasoline equipment, 65.7
hp for LPG equipment, and 64.5 hp for CNG equipment.
To estimate the average load factor for equipment in each of the Large SI fuel groupings,
we used the population estimates contained in the NONROAD model and the load factors as
presented in Table 6.2.2-1. As noted above, to simplify the calculations, we used the most
common applications within each category that represent 80 percent or more of the fuel grouping
population. Based on the most populous applications noted above, we calculated the population-
weighted average load factor for Large SI equipment to be 0.58 for gasoline equipment, 0.39 for
LPG equipment, and 0.49 for CNG equipment.
6-21
-------
Draft Regulatory Support Document
To estimate the average annual hours of use for equipment in each of the Large SI fuel
groupings, we used the population estimates contained in the NONROAD model and the hours
per year levels as presented in Table 6.2.2-1. As noted above, to simplify the calculations, we
used the most common applications within each category that represent 80 percent or more of the
fuel grouping population. Based on the most populous applications noted above, we calculated
the population-weighted average annual hours of use for Large SI equipment to be 534 hours for
gasoline equipment, 1368 hours for LPG equipment, and 1164 hours for CNG equipment.
Finally, to estimate the average lifetime for equipment in each of the Large SI fuel
groupings, we used the population estimates contained in the NONROAD model and the average
operating life information as presented in Table 6.2.2-1. As noted above, to simplify the
calculations, we used the most common applications within each category that represent 80
percent or more of the fuel grouping population. Based on the most populous applications noted
above, we calculated the population-weighted average lifetime for Large SI equipment to be 12.3
years for gasoline equipment, 12 years for LPG equipment, and 13 years for CNG equipment.
Using the information described above and the equation used for calculating emissions
from nonroad equipment (see Equation 6-1), we calculated the lifetime HC+NOx emissions from
typical Large SI equipment for both pre-control engines and engines meeting the Tier 1 and Tier
2 standards. Table 6.2.2-16 presents the lifetime HC+NOx emissions for Large SI equipment on
both an undiscounted and discounted basis (using a discount rate of 7 percent). Table 6.2.2-17
presents the corresponding lifetime HC+NOx emission reductions for the Tier 1 and Tier 2
standards.
Table 6.2.2-16
Lifetime HC+NOx Emissions from Typical Large SI Equipment (tons)*
Control
Level
Pre-control
Tierl
Tier 2
Gasoline
Un-
discounted
3.05
0.74
0.24
Discounted
2.13
0.51
0.17
LPG
Un-
discounted
6.81
1.86
0.49
Discounted
4.79
1.30
0.34
CNG
Un-
discounted
7.06
1.83
0.55
Discounted
4.85
1.24
0.37
* For CNG engines only, the emissions are calculated on the basis of NMHC+NOx.
6-22
-------
Chapter 6: Emissions Inventory
Table 6.2.2-17
Lifetime HC+NOx Emission Reductions from Typical Large SI Equipment (tons)*
Control
Increment
Pre-control
to Tier 1
Tier 1 to Tier
2
Gasoline
Un-
discounted
2.31
0.50
Discounted
1.62
0.34
LPG
Un-
discounted
4.94
1.37
Discounted
3.50
0.95
CNG
Un-
discounted
5.24
1.28
Discounted
3.61
0.87
* For CNG engines only, the reductions are calculated on the basis of NMHC+NOx.
We also calculated per equipment lifetime evaporative emission reductions using an
average lifetime of 13 years. For this analysis, we only consider gasoline powered equipment.
We determine annual per vehicle evaporative emissions by dividing the total annual evaporative
emissions for 2000 by the recreational vehicle populations shown in Table 6.2.2-9 (grown to
2000). Per vehicle emission reductions are based on the modeling described above. Table 6.2.2-
18 presents these results with and without the consideration of a 7 percent per year discount on
the value of emission reductions.
Table 6.2.2-18
Typical Lifetime Evaporative Emissions Per Large SI Gasoline Equipment(tons)
Evaporative
Component
Diurnal
Refueling
Hot Soak
Running Loss
Total
Baseline
Undiscounted
0.041
0.081
0.158
0.027
0.307
Discounted
0.028
0.056
0.109
0.019
0.211
Control
Undiscounted
0.003
0.065
0.011
0.002
0.081
Discounted
0.002
0.045
0.008
0.001
0.056
Reduction
Undiscounted
0.038
0.016
0.147
0.025
0.225
Discounted
0.026
0.011
0.101
0.017
0.155
6.2.3 Snowmobile Exhaust Emissions
We projected the annual tons of exhaust HC, CO, NOx, and PM from snowmobiles using
the draft NONROAD model discussed above. This section describes inputs to the calculations
that are specific to snowmobiles then presents the results. These results are for the nation as a
whole and include baseline and control inventory projections.
6-23
-------
Draft Regulatory Support Document
6.2.3.1 Inputs for the Inventory Calculations
Several usage inputs are specific to the calculations for snowmobile exhaust emissions.
These inputs are load factor, annual use, average operating life, and population. Based on data
developed for our Final Finding for recreational equipment and Large SI equipment, we use a
load factor of 34 percent and an annual usage factor of 57 hours.17 Using historical snowmobile
sales information for 1970 through 2001 and nationwide snowmobile registrations, both provided
by ISMA, and the scrappage curve used in the NONROAD model, we have updated our estimate
of average life from 9 years (as used in the proposal) to 13 years for this analysis.18 The draft
NONROAD model includes current and projected engine populations. The growth rates used in
the NONROAD model have been updated based on historical sales information (provided by
ISMA) and sales projections (developed by NERA in an analysis of the proposed snowmobile
standards for ISMA).19'20 Table 6.2.3-1 presents the snowmobile population estimates (rounded
to the nearest 1,000 units) for selected years.
Table 6.2.3-1
Projected Snowmobile Populations by Year
Year
Population
2000
1,622,000
2005
2,000,000
2010
2,407,000
2020
3,089,000
2030 1
3,377,000 1
The emission factors and deterioration factors for pre-control 2-stroke engines were
developed for the Final Finding as noted above. For the control case emission factors (i.e.,
engines designed to comply with the Phase 1, Phase 2, or Phase 3 standards), we are projecting
that manufacturers will use a mix of several different technologies that have significantly
different emission characteristics. The three control technologies we believe will be used are a
modified 2-stroke design, a direct injection 2-stroke engine, and a 4-stroke engine.
For the modified 2-stroke engine we assumed that manufacturers will design their engines
to meet the Phase 1 standards at regulatory useful life with a small compliance margin. (Because
we are not adopting a NOx standard for snowmobiles, we have assumed that NOx levels will
remain at the pre-control levels for modified 2-stroke engines.) In determining the zero-mile
levels of modified 2-stroke engines, we assumed a compliance margin of 20 percent to account
for variability. (The standards for snowmobiles are not based on the use of catalysts. Engine out
emissions tend to have more variability than the emissions coming from an engine equipped with
a catalyst. For this reason, we are using a compliance margin of 20 percent. As noted earlier,
including a margin of compliance below the standards is a practice that manufacturers have
followed historically to provide greater assurance that their engines meet emission standards in
the event of a compliance audit.) We have assumed that the deterioration rates of modified 2-
strokes will stay the same as the deterioration rates for pre-control 2-stroke engines. Table 6.2.3-
2 presents the emission factors used in this analysis for new engines and the maximum
deterioration factors applied to snowmobiles operated out to their median lifetime. (For the
calculations, the zero-mile levels were determined based on the pro-rated amount of deterioration
6-24
-------
Chapter 6: Emissions Inventory
expected at the regulatory lifetime, which is 300 hours for snowmobiles. As noted earlier, the
regulatory useful life is the period of time for which a manufacturer must demonstrate
compliance with the emission standards. The median lifetime of in-use equipment is longer than
the regulatory life.)
Table 6.2.3-2
Zero-Mile Level Emission Factors (g/hp-hr) and Deterioration Factors (at Median
Lifetime) for Snowmobile Engines
Engine Category/
Technology
Pre-control 2-stroke
Modified 2-stroke
Direct Injection 2-stroke
4-stroke
THC
ZML
111
53.7
21.8
7.8
Max
DF
1.2
1.2
1.2
1.15
CO
ZML
296
147
90
123
Max
DF
1.2
1.2
1.2
1.17
NOx
ZML
0.9
0.9
2.8
9.2
Max
DF
1.0
1.0
1.0
1.0
PM
ZML
2.7
2.7
0.57
0.15
Max
DF
1.2
1.2
1.2
1.15
Table 6.2.3-2 contains the zero-mile level and deterioration factors for direct injection 2-
stroke engines and 4-stroke engines as well. The emission levels were based on the results of
testing of prototype snowmobile engines employing these technologies or other similarly sized
engines employing these technologies.21
The Phase 1 standards are phased-in with 50% of engines for 2006 and 100% of enignes
for 2007. The Phase 2 standards take effect in 2010 for all engines. The Phase 3 standards take
effect in 2012 for all engines. For modeling purposes, we estimated the percent of engines that
will employ each of the control technologies to comply with the Phase 1, Phase 2, and Phase 3
standards. Table 6.2.3-3 contains the technology assumptions for the base case and under the
Phase 1, Phase 2, and Phase 3 standards. Currently, all engines are 2-strokes. Based on
discussions with manufacturers, we have assumed that manufacturers will begin introducing a
limited number of direct injection 2-strokes and some 4-strokes in the coming years.
6-25
-------
Draft Regulatory Support Document
Table 6.2.3-3
Snowmobile Engine Technology Mix Under the Base and Control Cases
Scenario
Current Baseline
2006 Baseline
Phase 1 (2006)
Phase 1 (2007)
Phase 2
Phase 3
Uncontrolled
2-strokes
100%
86%
53%
20%
20%
10%
Modified
2-stroke
-
-
30%
60%
30%
20%
Direct Injection
2-stroke
-
7%
8.5%
10%
35%
50%
4-stroke
-
7%
8.5%
10%
15%
20%
6.2.3.2 Reductions Due to the Standards
We anticipate that the standards for snowmobiles will result in a 57 percent reduction in
HC, a 46 percent reduction in CO, and a 42 percent reduction in PM by the year 2020. As
manufacturers adopt advanced technologies that result in significant HC, CO and PM emissions,
we expect the relatively limited amount of NOx from snowmobiles to increase under the
program. Tables 6.2.3-4 through 6.2.3.-7 present our projected HC, CO, NOx, and PM exhaust
emission inventories for snowmobiles and the anticipated emission reductions from the Phase 1,
Phase 2 and Phase 3 standards.
Table 6.2.3-4
Projected HC Inventories and Reductions for Snowmobiles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
205,000
250,000
286,000
345,000
375,000
Control
205,000
250,000
243,000
148,000
133,000
Reduction
0
0
43,000
197,000
242,000
% Reduction
0%
0%
15%
57%
65%
6-26
-------
Chapter 6: Emissions Inventory
Table 6.2.3-5
Projected CO Inventories and Reductions for Snowmobiles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
546,000
668,000
775,000
950,000
1,035,000
Control
546,000
668,000
670,000
508,000
497,000
Reduction
0
0
105,000
442,000
538,000
% Reduction
0%
0%
14%
46%
52%
Table 6.2.3-6
Projected NOx Inventories and Reductions for Snowmobiles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
1,400
1,900
3,000
5,000
5,500
Control
1,400
1,900
3,500
10,000
12,100
Reduction
0
0
(500)
(5,000)
(6,600)
% Reduction
0%
0%
-16%
-101%
-121%
Table 6.2.3-7
Projected PM Inventories and Reductions for Snowmobiles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
5,000
6,100
7,000
8,400
9,100
Control
5,000
6,100
6,700
4,900
4,400
Reduction
0
0
300
3,500
4,700
% Reduction
0%
0%
4%
42%
52%
6.2.3.3 Per Equipment Emissions from Snowmobiles
The following section describes the development of the HC and CO emission estimates
on a per piece of equipment basis over the average lifetime or a typical snowmobile. The
emission estimates were developed to estimate the cost per ton of the standards as presented in
6-27
-------
Draft Regulatory Support Document
Chapter 7.
In order to estimate the emission from a snowmobile, information on the emission level
of the engine, the power of the engine, the load factor of the engine, the annual hours of use of
the engine, and the lifetime of the engine are needed. The values used to predict the per piece of
equipment emissions for this analysis and the methodology for determining the values are
described below.
The information necessary to calculate the HC and CO emission levels of a piece of
equipment over the lifetime of a typical snowmobile were presented in Table 6.2.3-2. A brand
new snowmobile emits at the zero-mile level presented in the table. As the snowmobile ages, the
emission levels increase based on the pollutant-specific deterioration factor. Deterioration, as
modeled in the NONROAD model, continues until the equipment reaches the median life. The
deterioration factors presented in Table 6.2.3-2 when applied to the zero-mile levels presented in
the same table, represent the emission level of the snowmobile at the end of its median life. The
emissions at any point in time in between can be determined through interpolation.
To estimate the average power for snowmobiles, we used the population and power
distribution information contained in the NONROAD model and determined the population-
weighted average horsepower for snowmobiles. The population-weighted horsepower for
snowmobiles was calculated to be 48.3 hp.
As described earlier in this section, the load factor for snowmobiles is estimated to be
0.34, the annual usage rate is estimated to be 57 hours per year, and the average lifetime is
estimated to be 13 years.
Using the information described above and the equation used for calculating emissions
from nonroad equipment (see Equation 6-1), we calculated the lifetime HC and CO emissions
from a typical snowmobile for both pre-control engines and engines meeting the Phase 1, Phase
2, and Phase 3 standards. (The per vehicle estimates are a weighted-average of the different
technologies assumed under the base and control cases as presented earlier in Table 6.2.3-3.)
Table 6.2.3-8 presents the lifetime HC and CO emissions for a typical snowmobile on both an
undiscounted and discounted basis (using a discount rate of 7 percent). Table 6.2.3-9 presents
the corresponding lifetime HC and CO emission reductions for the Phase 1, Phase 2 and Phase 3
standards.
6-28
-------
Chapter 6: Emissions Inventory
Table 6.2.3-8
Lifetime HC and CO Emissions from a Typical Snowmobile (tons)
Control Level
Pre-control
Phase 1
Phase 2
Phase 3
HC
Undiscounted
1.45
0.85
0.70
0.51
Discounted
0.98
0.57
0.47
0.34
CO
Undiscounted
3.99
2.50
2.27
1.90
Discounted
2.71
1.70
1.54
1.29
Table 6.2.3-9
Lifetime HC and CO Emission Reductions from a Typical Snowmobile (tons)
Control Increment
Pre-control to Phase 1
Phase 1 to Phase 2
Phase 2 to Phase 3
HC
Undiscounted
0.60
0.15
0.19
Discounted
0.40
0.10
0.14
CO
Undiscounted
1.49
0.23
0.37
Discounted
1.01
0.16
0.25
6.2.4 All-Terrain Vehicle Exhaust Emissions
We projected the annual tons of exhaust HC, CO, NOx, and PM from all-terrain vehicles
(ATVs) using the draft NONROAD model discussed above. This section describes inputs to the
calculations that are specific to ATVs then presents the results. These results are for the nation
as a whole and include baseline and control inventory projections.
6.2.4.1 Inputs for the Inventory Calculations
Several usage inputs are specific to the calculations for ATV exhaust emissions. These
inputs are annual use, average operating life, and population. Based on data developed for our
Final Finding for recreational equipment and Large SI equipment, we use an average operating
life of 13 years for ATVs.22 Based on several surveys of ATV operators, we have revised the an
annual usage factor for ATVs for this analysis to 1,570 miles per year.23 The updated mileage
analysis for ATVs is presented in detail in the appendix to this chapter. (Because the ATV
standards are chassis-based standards instead of engine-based, the NONROAD model has been
revised to model ATVs on the basis of gram per mile emission factors and annual mileage
accumulation rates. Load factor is not needed for such calculations.)
The draft NONROAD model includes current and projected engine populations. Table
6-29
-------
Draft Regulatory Support Document
6.2.4-1 presents these population estimates (rounded to the nearest 1,000 units) for selected
years. The ATV population growth rates used in the NONROAD model have been updated for
this analysis to reflect the expected growth in ATV populations based on updated ATV sales
information and sales growth projections supplied by the Motorcycle Industry Council (MIC), an
industry trade organization. The growth rates were developed separately for 2-stroke and 4-
stroke ATVs. Based on the sales information from MIC, sales of ATVs have been growing
substantially throughout the 1990s, averaging 25 percent growth per year over the last 6 years.
MIC estimates that growth in sales will continue for the next few years, although at lower levels
often percent or less, with no growth in sales projected by 2005. Combining the sales history,
growth projections, and information on equipment scrappage, we have estimated that the
population of ATVs will grow significantly through 2010, and then grow at much lower levels.24
(The population of 2-stroke ATVs presented in Table 6.2.4-1 are for baseline population
estimates. Under the ATV standards, 2-stroke designs are expected to be phased-out as they are
converted to 4-stroke designs.)
Table 6.2.4-1
Projected ATV Populations by Year
Category
4-stroke ATVs
2-stroke ATVs*
All ATVs
2000
3,919,000
690,000
4,609,000
2005
6,240,000
1,678,000
7,918,000
2010
8,453,000
2,461,000
10,914,000
2020
10,080,000
3,001,000
13,081,000
2030
10,188,000
3,036,000
13,224,000
* - The projected population estimates for 2-stroke ATVs are for baseline calculations only.
Under the Phase 1 standards, we expect all 2-stroke engines will be converted to 4-stroke
designs.
The baseline HC, CO, and NOx emission factors used in the NONROAD model for
ATVs have been updated based on recent testing of ATVs and off-highway motorcycles as
presented in Chapter 4. PM emissions were not measured in the test program. Therefore,
baseline PM emission factors were based on testing of both off-highway motorcycles and pre-
control on-highway motorcycles.25 The baseline deterioration factors (for pre-control engines)
were developed for the Final Finding as noted above. For the control emission factors (i.e.,
engines complying with the Phase 1 standards), we assumed that the manufacturers will design
their engines to meet the standards at regulatory useful life with a small compliance margin.
Because we are adopting a HC+NOx standard for ATVs, we have assumed that the Phase 1
HC/NOx split will remain the same as the pre-control HC/NOx split. For the Phase 1 standards
for ATVs, we assumed a compliance margin of 20 percent to account for variability. (As noted
earlier, including a margin of compliance below the standards is a practice that manufacturers
have followed historically to provide greater assurance that their engines will meet emission
standards in the event of a compliance audit.) Because the standards for ATVs are expected to
be met by 4-stroke designs, we assumed that the deterioration rates will stay the same as the
deterioration rates for pre-control 4-stroke ATVs. Table 6.2.4-2 presents the emission factors
6-30
-------
Chapter 6: Emissions Inventory
used in this analysis for new ATVs and the maximum deterioration factors for ATVs which
applies at the median lifetime. (For the calculations, the zero-mile levels were determined based
on the pro-rated amount of deterioration expected at the regulatory lifetime, which is 6,214 miles
(10,000 kilometers) for ATVs. As noted earlier, the regulatory useful life is the period of time
for which a manufacturer must demonstrate compliance with the emission standards. The
median lifetime of in-use equipment is longer than the regulatory life. As noted earlier, the
regulatory useful life is the period of time for which a manufacturer must demonstrate
compliance with the emission standards. The median lifetime of in-use equipment is longer than
the regulatory life.)
Table 6.2.4-2
Zero-Mile Level Emission Factors (g/mi) and Deterioration Factors (at Median Lifetime)
for ATVs
Engine Category
Baseline/Pre -control
2-stroke
Baseline/Pre -control
4-stroke
Control/Phase 1 -
4-stroke
THC
ZML
53.9
2.4
1.6
Max
DF
1.2
1.15
1.15
CO
ZML
54.1
48.5
42.9
Max
DF
1.2
1.17
1.17
NOx
ZML
0.15
0.41
0.26
Max
DF
1.0
1.0
1.0
PM
ZML
2.1
0.06
0.06
Max
DF
1.2
1.2
1.15
The Phase 1 standards are to be phased in at 50 percent in 2007 and 100 percent in 2008.
However, because there are a significant number of small volume manufacturers that produce 2-
stroke ATVs, and because we have compliance flexibilities for such manufacturers, we have
modeled the phase in of the standards for the current 2-stroke ATVs based on the schedule
contained in Table 6.2.4-3.
Table 6.2.4-3
Assumed Phase-In Schedule for Current 2-Stroke ATVs Used in the Modeling Runs
Model Year
2005
2006
2007
2008
2009
Pre-control 2-stroke
100%
65%
30%
15%
0%
Phase 1 4-stroke
0%
35%
70%
85%
100%
6-31
-------
Draft Regulatory Support Document
6.2.4.2 Reductions Due to the Standards
We anticipate that the standards for ATVs will result in a 86 percent reduction in HC, a
37 percent reduction in CO, and a 86 percent reduction in PM by the year 2020. As
manufacturers convert their engines from 2-stroke to 4-stroke design and achieve these
significant reductions, we expect there may be a minimal increase in NOx. Tables 6.2.4-4
through 6.2.4-7 present our projected HC, CO, NOx, and PM exhaust emission inventories for
ATVs and the anticipated emission reductions from the Phase 1 standards.
Table 6.2.4-4
Projected HC Inventories and Reductions for ATVs (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
89,000
200,000
291,000
353,000
357,000
Control
89,000
200,000
198,000
49,000
40,000
Reduction
0
0
92,000
304,000
317,000
% Reduction
0%
0%
32%
86%
89%
Table 6.2.4-5
Projected CO Inventories and Reductions for ATVs (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
437,000
755,000
1,042,000
1,250,000
1,263,000
Control
437,000
755,000
989,000
1,085,000
1,092,000
Reduction
0
0
53,000
165,000
171,000
% Reduction
0%
0%
5%
13%
14%
6-32
-------
Chapter 6: Emissions Inventory
Table 6.2.4-6
Projected NOx Inventories and Reductions for ATVs (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
3,000
4,900
6,600
7,900
8,000
Control
3,000
4,900
5,900
5,900
6,000
Reduction
0
0
(700)
(2,000)
(2,000)
% Reduction
0%
0%
-11%
-25%
-26%
Pro
Table 6.2.4-7
ected PM Inventories and Reductions for ATVs (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
3,200
7,400
10,800
13,100
13,300
Control
3,200
7,400
7,400
1,800
1,500
Reduction
0
0
3,400
11,300
11,800
% Reduction
0%
0%
32%
86%
89%
6.2.4.3 Per Equipment Emissions from All-Terrain Vehicles
The following section describes the development of the HC+NOx emission estimates on
a per piece of equipment basis over the average lifetime or a typical ATV. The emission
estimates were developed to estimate the cost per ton of the standards as presented in Chapter 7.
In order to estimate the emissions from an ATV, information on the emission level of the
vehicle, the annual usage rate of the engine, and the lifetime of the engine are needed. The
values used to predict the per piece of equipment emissions for this analysis and the methodology
for determining the values are described below.
The information necessary to calculate the HC and NOx emission levels of a piece of
equipment over the lifetime of a typical ATV were presented in Table 6.2.4-2. A brand new
ATV emits at the zero-mile level presented in the table. As the ATV ages, the emission levels
increase based on the pollutant-specific deterioration factor. Deterioration, as modeled in the
NONROAD model, continues until the equipment reaches the median life. The deterioration
factors presented in Table 6.2.4-2 when applied to the zero-mile levels presented in the same
table, represent the emission level of the ATV at the end of its median life. The emissions at any
6-33
-------
Draft Regulatory Support Document
point in time in between can be determined through interpolation.
As described earlier in this section, the annual usage rate for an ATV is estimated to be
1,570 miles per year and the average lifetime is estimated to be 13 years.
Using the information described above and the equation used for calculating emissions
from nonroad equipment modified to remove the power and load variables (see Equation 6-1),
we calculated the lifetime HC+NOx emissions from a typical ATV for both pre-control engines
(shown separately for 2-stroke and 4-stroke engines and a composite weighted value) and engines
meeting the Phase 1 standards. Table 6.2.4-8 presents the lifetime HC+NOx emissions for a
typical ATV on both an undiscounted and discounted basis (using a discount rate of 7 percent).
Table 6.2.4-9 presents the corresponding lifetime HC+NOx emission reductions for the Phase 1.
Table 6.2.4-8
Lifetime HC+NOx Emissions from a Typical ATV (tons)
Control Level
Pre-control (2-stroke)
Pre-control (4-stroke)
Pre-control (Composite)
Phase 1
HC+NCh
Undiscounted
1.37
0.07
0.35
0.05
Discounted
0.93
0.05
0.24
0.03
Table 6.2.4-9
Lifetime HC+NOx Emission Reductions from a Typical ATV (tons)
Control Increment
Pre-control (Composite) to Phase 1
HC+NOx
Undiscounted
0.30
Discounted
0.21
6.2.5 Off-highway Motorcycle Exhaust Emissions
We projected the annual tons of exhaust HC, CO, NOx, and PM from off-highway
motorcycles using the draft NONROAD model discussed above. This section describes inputs to
the calculations that are specific to off-highway motorcycles then presents the results. These
results are for the nation as a whole and include baseline and control inventory projections.
6.2.5.1 Inputs for the Inventory Calculations
Several usage inputs are specific to the calculations for off-highway motorcycles exhaust
emissions. These inputs are annual use, average operating life, and population. Based on an
6-34
-------
Chapter 6: Emissions Inventory
updated analysis of fuel consumption and fuel use, we have revised our estimate of annual usage
for off-highway motorcycles to 1,600 miles per year.26 (The updated mileage analysis for off-
highway motorcycles is presented in detail in the appendix to this chapter.) We have also revised
our estimate of the average operating life of off-highway motorcycles to 12 years based on
historical sales and population information provided by the Motorcycle Industry Council.27
(Because the off-highway motorcycle standards are chassis-based standard instead of engine-
based, the NONROAD model has been revised to model off-highway motorcycles on the basis of
gram per mile emission factors and annual mileage accumulation rates. Load factor is not needed
for such calculations.)
The draft NONROAD model includes current and projected engine populations. Table
6.2.5-1 presents these population estimates (rounded to the nearest 1,000 units) for selected
years. (The population of 2-stroke off-highway motorcycles presented in Table 6.2.5-1 are for
baseline population estimates. Under the off-highway motorcycle standards, non-competition 2-
stroke designs are expected to be phased-out as they are converted to 4-stroke designs.
Competition models will remain 2-stroke designs.) The population growth rates used in the
NONROAD model have been updated based on historical sales information provided by MIC
and a projected one percent growth in sales.28
Table 6.2.5-1
Projected Off-Highway Motorcycle Populations by Year
Category
4-stroke
Off-highway
Motorcycles
2-stroke
Off-highway
Motorcycles*
All
Off-highway
Motorcycles
2000
444,000
902,000
1,346,000
2005
656,000
1,333,000
1,989,000
2010
862,000
1,750,000
2,612,000
2020
1,038,000
2,108,000
3,146,000
2030
1,133,000
2,300,000
3,433,000
* - The projected population estimates for 2-stroke off-highway motorcycles are for baseline
calculations only. To meet the standards, we expect all non-competition 2-strokes will be
converted to 4-stroke designs. All 2-stroke competition models are assumed to remain 2-strokes.
The baseline HC, CO, and NOx emission factors used in the NONROAD model for off-
highway motorcycles have been updated based on recent testing of ATVs and off-highway
motorcycles as presented in Chapter 4. PM emissions were not measured in the test program.
Therefore, baseline PM emission factors were based on testing of both off-highway motorcycles
and pre-control on-highway motorcycles.29 The baseline deterioration factors (for pre-control
engines) were developed for the Final Finding as noted above. For the control emission factors
(i.e., Phase 1 off-highway motorcycles), we assumed that the manufacturers will design their
6-35
-------
Draft Regulatory Support Document
engines to meet the standards at regulatory useful life with a small compliance margin. Because
we are adopting a HC+NOx standard for off-highway motorcycles, we have assumed that the
Phase 1 HC/NOx split will remain the same as the pre-control HC/NOx split. For the Phase 1
standards for off-highway motorcycles, we assumed a compliance margin of 20 percent to
account for variability. (Including a margin of compliance below the standards is a practice that
manufacturers have followed historically to provide greater assurance that their engines will meet
emission standards in the event of a compliance audit.) Because the standards for off-highway
motorcycles are expected to be met by 4-stroke designs, we assumed that the deterioration rates
will stay the same as the deterioration rates for pre-control 4-stroke off-highway motorcycles.
Table 6.2.5-2 presents the emission factors used in this analysis for new off-highway motorcycles
and the maximum deterioration factors applied to off-highway motorcycles operated out to their
median lifetime. (For the calculations, the zero-mile levels were determined based on the pro-
rated amount of deterioration expected at the regulatory lifetime, which is 6,210 miles (10,000
kilometers) for off-highway motorcycles. As noted earlier, the regulatory useful life is the period
of time for which a manufacturer must demonstrate compliance with the emission standards.
The median lifetime of in-use equipment is longer than the regulatory life.)
Table 6.2.5-2
Zero-Mile Level Emission Factors (g/mi) and Deterioration Factors (at Median Lifetime)
for Off-Highway Motorcycles
Engine Category
Baseline/Pre -control
2-stroke*
Baseline/Pre -control
4-stroke
Control/Phase 1
4-stroke
THC
ZML
53.9
2.4
2.1
Max
DF
1.2
1.15
1.15
CO
ZML
54.1
48.5
30.6
Max
DF
1.2
1.17
1.17
NOx
ZML
0.15
0.41
0.34
Max
DF
1.0
1.0
1.0
PM
ZML
2.1
0.06
0.06
Max
DF
1.2
1.15
1.15
* - Competition models are assumed to remain at pre-control levels under the final program for
off-highway motorcycles.
The Phase 1 standards phase in at 50 percent in 2007 and 100 percent in 2008. However,
because there are a significant number of small volume manufacturers that produce off-highway
motorcycles (who can take advantage of compliance flexibilities), and because competition off-
highway motorcycles are exempt from the standards, we have modeled the phase in of the
standards for off-highway motorcycles based on the schedule contained in Table 6.2.5-3.
6-36
-------
Chapter 6: Emissions Inventory
Table 6.2.5-3
Assumed Phase-In Schedule for Current Off-Highway Motorcycles
Used in the Modeling Runs
Model Year
2005
2006
2007
2008
2009+
Current 4-stroke
Off-highway Motorcycles
Pre -control
100%
56%
12%
6%
0%
Phase 1
0%
44%
88%
94%
100%
Current 2-stroke
Off-highway Motorcycles
Pre -control
100%
76%
53%
49%
46%
Phase 1
0%
24%
47%
51%
54%
6.2.5.2 Reductions Due to the Standards
We anticipate that the standards for off-highway motorcycles will result in a 49 percent
reduction in HC, a 26 percent reduction in CO, and a 50 percent reduction in PM by the year
2020. As manufacturers convert their engines from 2-stroke to 4-stroke design and achieve these
significant emission reductions, we project there may be a small increase in NOx inventories.
Tables 6.2.5-4 through 6.2.5.-7 present our projected HC, CO, NOx, and PM exhaust emission
inventories for off-highway motorcycles and the anticipated emission reductions from the Phase
1 standards. (The emission inventories presented below for off-highway motorcycles include
competition motorcycles that will be exempt from the standards.)
Table 6.2.5-4
Projected HC Inventories and Reductions for Off-Highway Motorcycles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
97,000
143,000
188,000
226,000
246,000
Control
97,000
143,000
151,000
115,000
121,000
Reduction
0
0
36,000
111,000
126,000
% Reduction
0%
0%
19%
49%
51%
6-37
-------
Draft Regulatory Support Document
Table 6.2.5-5
Projected CO Inventories and Reductions for Off-Highway Motorcycles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
137,000
203,000
226,000
321,000
350,000
Control
137,000
203,000
239,000
236,000
254,000
Reduction
0
0
27,000
84,000
96,000
% Reduction
0%
0%
10%
26%
27%
Table 6.2.5-6
Projected NOx Inventories and Reductions for Off-Highway Motorcycles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
600
800
1,100
1,300
1,400
Control
600
800
1,200
1,500
1,700
Reduction
0
0
(100)
(200)
(300)
% Reduction
0%
0%
-8%
-19%
-19%
Table 6.2.5-7
Projected PM Inventories and Reductions for Off-Highway Motorcycles (short tons)
Calendar Year
2000
2005
2010
2020
2030
Baseline
3,700
5,500
7,300
8,700
9,500
Control
3,700
5,500
5,900
4,400
4,600
Reduction
0
0
1,400
4,300
4,900
% Reduction
0%
0%
20%
50%
52%
6.2.5.3 Per Equipment Emissions from Off-highway Motorcycles
The following section describes the development of the HC+NOx emission estimates on
a per piece of equipment basis over the average lifetime or a typical off-highway motorcycle.
The emission estimates were developed to estimate the cost per ton of the standards as presented
in Chapter 7.
6-38
-------
Chapter 6: Emissions Inventory
In order to estimate the emissions from an off-highway motorcycle, information on the
emission level of the vehicle, the annual usage rate of the engine, and the lifetime of the engine
are needed. The values used to predict the per piece of equipment emissions for this analysis and
the methodology for determining the values are described below.
The information necessary to calculate the HC and NOx emission levels of a piece of
equipment over the lifetime of a typical off-highway motorcycle were presented in Table 6.2.5-2.
A brand new off-highway motorcycle emits at the zero-mile level presented in the table. As the
off-highway motorcycle ages, the emission levels increase based on the pollutant-specific
deterioration factor. Deterioration, as modeled in the NONROAD model, continues until the
equipment reaches the median life. The deterioration factors presented in Table 6.2.5-2 when
applied to the zero-mile levels presented in the same table, represent the emission level of the
off-highway motorcycle at the end of its median life. The emissions at any point in time in
between can be determined through interpolation.
As described earlier in this section, the annual usage rate for an off-highway motorcycle
is estimated to be 1,600 miles per year and the average lifetime is estimated to be 12 years.
Using the information described above and the equation used for calculating emissions
from nonroad equipment modified to remove the power and load variables (see Equation 6-1),
we calculated the lifetime HC+NOx emissions from a typical off-highway motorcycle for both
pre-control engines (shown separately for 2-stroke and 4-stroke engines and a composite
weighted value) and engines under the Phase 1 standards. (Competition bikes, which are exempt
from the standards, are not included in the calculations.) Table 6.2.5-8 presents the lifetime
HC+NOx emissions for a typical off-highway motorcycle on both an undiscounted and
discounted basis (using a discount rate of 7 percent). Table 6.2.5-9 presents the corresponding
lifetime HC+NOx emission reductions for the Phase 1 standards.
Table 6.2.5-8
Lifetime HC+NOx Emissions from a Typical Off-highway Motorcycle (tons)*
Control Level
Pre-control (2-stroke)
Pre-control (4-stroke)
Pre-control (Composite)
Phase 1
HC+NO>
Undiscounted
1.27
0.06
0.60
0.06
Discounted
0.89
0.04
0.42
0.04
* The emission estimates do not include competition off-highway motorcycles that remain at pre-
control emission levels.
6-39
-------
Draft Regulatory Support Document
Table 6.2.5-9
Lifetime HC+NOx Emission Reductions from a Typical Off-highway Motorcycle (tons)*
Control Increment
Pre-control (Composite) to Phase 1
HC+NOx
Undiscounted
0.54
Discounted
0.38
* The reduction estimates do not include competition off-highway motorcycles that remain uncontrolled,
and therefore do not realize any emission reductions under the new standards.
6.2.6 Evaporative Emissions from Recreational Vehicles
We projected the annual tons of hydrocarbons evaporated into the atmosphere from
snowmobiles, ATVs, off-highway motorcycles using the methodology discussed above in
Section 6.1.2. These evaporative emissions include permeation, diurnal and refueling emissions.
Although the standards do not specifically require the control of diurnal and refueling emissions,
we have included them in the modeling for completeness. This section describes inputs to the
calculations that are specific to each of the recreational vehicle types and presents our baseline
and controlled national evaporative inventory projections.
6.2.6.1 General Inputs for the Inventory Calculations
Several usage inputs are specific to the calculations of evaporative emissions from ATVs.
These inputs are fuel tank sizes, population, and distribution throughout the nation. The draft
NONROAD model includes current and projected engine populations for each state and we used
this distribution as the national fuel tank distribution. Table 6.2.6-1 presents the population of
recreational vehicles for 1998.
Table 6.2.6-1
1998 Population of Recreational Vehicles by Region
Region
Northeast
Southeast
Southwest
Midwest
West
Northwest
Total
Snowmobiles
954,000
0
11,000
419,000
40,000
140,000
1,560,000
ATVs
1,420,000
1,010,000
363,000
457,000
423,000
249,000
3,930,000
Off-Highway Motorcycles
427,000
304,000
109,000
137,000
127,000
75,000
1,180,000
We based average fuel tank sizes on sales literature for recreational vehicles.
Snowmobile fuel tanks range from 10 gallons to about 12 gallons. For ATVs, fuel tanks range
from one gallon for the smaller youth models to five gallons for the larger utility models.
6-40
-------
Chapter 6: Emissions Inventory
Finally, off-highway motorcycle fuel tanks range in capacity from approximately one gallon on
some smaller youth models to about three gallons on some enduro motorcycles. For this
analysis, we used average fuel tank sizes of 11 gallons for snowmobiles, 4 gallons for ATVs, and
3 gallons for off-highway motorcycles.
Based on our examination of recreational vehicles, we have found that fuel hoses
generally have an inside diameter of about 6 mm (1/4 inch). For ATVs, we estimate one foot of
fuel line on average. For off-highway motorcycles, we estimate that they use approximately one
to two feet of fuel line on average. We use 1.5 feet in our analysis. Snowmobiles are a little
more complex because they use multi-cylinder engines (either two or three cylinders). For two
cylinder engines we estimate two to three feet of fuel line and for three cylinder engines we
estimate three to four feet of fuel line. We use 3.5 feet in our analysis.
6.2.6.2 Permeation Emissions Inventory and Reductions
Based on the data presented in Chapter 4, we developed the emission factors presented in
Table 6.2.6-2. For the purposes of this modeling, fuel tank permeation rates are expressed in
terms of g/gallon/day because the defining characteristic of the fuel tanks in our model is
capacity. The standard requires that the fuel tanks meet an 85 percent reduction in permeation
throughout its useful life. For this modeling, we assume that manufacturers will strive to achieve
a 95 percent reductions from new tanks and that the permeation control will deteriorate to 85
percent by the end of the life of an average tank. Hose permeation rates are based on g/m2/day.
We believe that hoses designed to meet the 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
this hose to be about half of the permeation from fuel hose designed to meet 15 g/m2/day on
gasoline.88 To show the effect of temperature on permeation rates, we present emission rates at
three temperatures.
Table 6.2.6-2
Fuel Tank and Hose Permeation Emission Factors
Material
Polyethylene fuel tanks
New barrier treated HDPE fuel tank
Aged barrier treated HDPE fuel tank
SAE R7 fuel hose
SAE R9 barrier fuel hose
Alcohol resistant barrier fuel hose
23°C (73°F)
0.78 g/gal/day
0.04 g/gal/day
0.11 g/gal/day
550 g/mVday
15 g/m2/day
7.5 g/mVday
29°C (85°F)
1.12 g/gal/day
0.06 g/gal/day
0.17 g/gal/day
873 g/mVday
24 g/mVday
12 g/mVday
40°C (104°F)
2.08 g/gal/day
0.10 g/gal/day
0.31 g/gal/day
1800 g/mVday
49 g/nf/day
25 g/mVday
8g This is appropriate because the baseline emissions are modeled based on the use of
gasoline as a fuel. If we were to consider that a fraction of the fuel contains oxygenates, both the
baseline and control emission inventory projections would increase.
6-41
-------
Draft Regulatory Support Document
Using the vehicle populations and temperature distributions discussed above, we
calculated baseline and controlled permeation emission inventories for recreational vehicles.
Tables 6.2.6-3 and 6.2.6-4 present our projected permeation reductions from fuel tanks and
hoses.
Table 6.2.6-3
Projected Fuel Tank Permeation Emissions from Recreational Vehicles [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
3,389
3,389
0
3,985
3,985
0
882
882
0
8,255
8,255
0
2005
4,181
4,181
0
6,751
6,751
0
1,303
1,303
0
12,234
12,234
0
2010
5,032
3,586
1,446
9,275
7,388
1,887
1,710
1,370
340
16,016
12,343
3,673
2020
6,456
901
5,555
11,109
2,602
8,507
2,061
834
1,227
19,626
4,337
15,288
2030
7,061
746
6,315
11,231
1,249
9,982
2,248
857
1,391
20,539
2,851
17,688
Table 6.2.6-4
Projected Fuel Hose Permeation Emissions from Recreational Vehicles [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
4,471
4,471
0
4,243
4,243
0
1,878
1,878
0
10,592
10,592
0
2005
5,516
5,516
0
7,189
7,189
0
2,774
2,774
0
15,478
15,478
0
2010
6,638
4,361
2,007
9,876
7,771
2,105
3,642
2,880
762
20,156
15,282
4,873
2020
8,517
452
8,065
11,829
1,931
9,898
4,389
1,513
2,876
24,735
3,896
20,838
2030
9,315
127
9,188
11,959
245
11,714
4,787
1,520
3,268
26,061
1,891
24,169
6-42
-------
Chapter 6: Emissions Inventory
6.2.6.3 Per Vehicle Permeation Emissions
In developing the cost per ton estimates in Chapter 7, we need to know the lifetime
emissions per recreational vehicle. The lifetime emissions are based on the projected lives of 9
years for snowmobiles, 13 years for ATVs, and 9 years for off-highway motorcycles. We
determine annual per vehicle evaporative emissions by dividing the total annual evaporative
emissions for 2000 by the recreational vehicle populations shown in Table 6.2.6-1 (grown to
2000). Competition motorcycles, which are exempt form the standards, are not included in these
calculations. Per vehicle emission reductions are based on the modeling described above. Table
6.2.6-5 presents these results with and without the consideration of a 7 percent per year discount
on the value of emission reductions.
Table 6.2.6-5
Typical Lifetime Permeation Emissions Per Recreational Vehicle (tons)
1 Baseline
Undiscounted
Snowmobiles
Tank
Hose
Total
0.0180
0.0238
0.0418
Discounted
0.0140
0.0184
0.0324
Control
Undiscounted
0.0019
0.0003
0.0022
Discounted
0.0015
0.0003
0.0017
Reduction
Undiscounted
0.0161
0.0235
0.0396
Discounted
0.0125
0.0182
0.0307
All Terrain Vehicles
Tank
Hose
Total
0.0114
0.0121
0.0234
0.0078
0.0083
0.0161
0.0012
0.0002
0.0014
0.0008
0.0001
0.0009
0.0102
0.0119
0.0221
0.0070
0.0082
0.0152
Off-Highway Motorcycles
Tank
Hose
Total
0.0059
0.0126
0.0184
0.0046
0.0097
0.0143
0.0006
0.0002
0.0008
0.0005
0.0001
0.0006
0.0053
0.0124
0.0177
0.0041
0.0096
0.0137
6.2.6.4 Other Evaporative Emissions
We calculated diurnal and refueling vapor loss emissions using the general inputs in
section 6.2.6.1 and the methodology described in sections 6.1.2.2 and 6.2.1.3. Although we are
not regulating these emissions, we present the inventory projections for comparison. Table 6.2.6-
6 presents the baseline diurnal emission factors for the certification test conditions and a typical
summer day with low vapor pressure fuel and a half-full tank. (This comparison is for
6-43
-------
Draft Regulatory Support Document
illustrative purposes; as discussed above, we modeled daily temperature for 365 days over 6
regions of the U.S.) Decreasing temperature and fuel RVP and increasing fill level all have the
effect of reducing the diurnal emission factor. Table 6.2.6-7 presents our diurnal emission
projections.
Table 6.2.6-6
Diurnal Emission Factors for Test Conditions and Typical Summer Day
Evaporative Control
72-96°F, 9 RVP* Fuel, 40% fill
60-84°F, 8 RVP* Fuel, 50% fill
baseline
1.5 g/gallon/day
0.55 g/gallon/day
* Reid Vapor Pressure
Table 6.2.6-7
Projected Diurnal Emissions from Recreational Vehicles [short tons]
Calendar Year
2000
2005
2010
2020
2030
Snowmobiles
2,223
2,743
3,301
4,235
4,632
ATVs
3,079
5,216
7,167
8,584
8,678
Off-Highway Motorcycles
681
1,006
1,321
1,592
1,737
To calculate the refueling vapor displacement emissions from recreational vehicles, we
needed to know the amount of fuel added to the fuel tank per year. Therefore, we used the draft
NONROAD model to determine the amount of fuel consumed by recreational vehicles. We then
used the amount of fuel consumed as the amount of fuel added to the fuel. Table 6.2.6-8
contains the projected refueling emission inventories for recreational vehicles.
Table 6.2.6-8
Projected Refueling Emissions from Recreational Vehicles [short tons]
Calendar Year
2000
2005
2010
2020
2030
Snowmobiles
1,814
2,230
2,596
2,922
3,120
ATVs
928
1,620
1,185
2,510
2,532
Off-Highway Motorcycles
368
544
684
773
840
6-44
-------
Chapter 6: Emissions Inventory
Appendix to Chapter 6: ATV and Off-highway Motorcycle Usage Rates
This appendix presents the analyses used to determine the annual average usage rates for
ATVs and off-highway motorcycles.
6A.1 ATV Usage
On October 5, 2001, EPA published proposed emission regulations for nonroad land-
based recreational vehicles. These regulations covered snowmobiles, off-highway motorcycles,
and all-terrain vehicles (ATVs). The Motorcycle Industry Council, Inc. (MIC) and the Specialty
Vehicle Institute of America (SVIA) submitted comments suggesting that the EPA estimates for
ATV usage had been substantially overestimated. They stated that our mileage estimate of 7,000
miles per year was too high and that based on some additional information that they had
obtained, a more reasonable estimate was a lifetime average of 350 miles per year. As a result of
these comments and the subsequent new information, EPA has revised it's estimate of annual
ATV usage.
Background
On November 20, 2000 EPA published a Final Finding of Contribution and Advance
Notice of Proposed Rulemaking (ANPRM) for large nonroad spark-ignition engines and land-
based recreational vehicles. In this process, we developed emission inventories for the various
engine and vehicle categories covered by both these documents. EPA developed inventories
using NONROAD model, which computes emission estimates for nonroad engines at selected
geographic and temporal scales. The model incorporates data on emission rates, usage rates, and
vehicle population to determine annual emission levels of various pollutants.
For recreational vehicles, and more specifically ATVs, data on emission rates and usage rates
was extremely limited. We approached members of the ATV industry to provide us with any
data that they had on emission and usage rates. Unfortunately, all of the emission data industry
had for ATVs was collected on the J1088 steady state engine test cycle rather than the FTP
transient vehicle test cycle that we proposed. Industry also indicated that they didn't have any
data on ATV usage rates. MIC provided survey data on off-highway motorcycle usage, but did
not provide any information on ATV usage. Through our literature search, we ultimately found a
study by the United States Consumer Product Safety Commission (CPSC) published in April of
1998 titled, "All-Terrain Vehicle Exposure, Injury, Death, and Risk Studies" that provided
information on ATV usage. This study provided the basis for our estimate of ATV usage for the
NPRM.
We did not receive any comments on our estimate of ATV usage during the comment
period for the Final Finding and ANPRM . In fact, we did not receive any comments until after
the Notice of Proposed Rulemaking (NPRM) was published in October of 2001.
ATV Usage in the ANPRM and NPRM
6-45
-------
Draft Regulatory Support Document
Because we received no comment or additional information for the ANPRM and NPRM,
we determined that the CPSC study was the best source of information available. After
converting hours of use to miles ridden, we estimated an annual average of 7,000 miles/year. A
complete description of the modeling parameters for ATVs used in the NPRM is contained in an
EPA memorandum entitled "Emission Modeling for Recreational Vehicles."30
New Information
Since the publication of the October 2001 NPRM, several new pieces of information on
ATV usage have become available. These new sources consist of:
• Nationwide sources
- ATV manufacturer warranty data
- A Honda owner survey
- ATV Industry Panel Survey (consisting of five ATV manufacturers)
• State studies on economic impact of ATV operation on their respective states
California31
- Colorado32
- Maine33
- Michigan34
I. Utah35
• Instrumented ATV Usage Data (CE-CERT)
- Speed information
Each of these sources is discussed in more detail below.
Warranty Data
One ATV manufacturer supplied ATV mileage and hour data from some its warranty
claims submitted over a period of four years. The data was substantial and represented a good
cross section of the country. The data is proprietary and was provided to us as confidential
business information. This manufacturer does not have odometers or hour-meters on all of their
ATV models, but provided data on those models equipped with an odometer or hour-meter,
which happens to be only their utility models. Thus, there is no data for any of their sport
models.
Intuitively, we were concerned about using data from warranty claims because of the
possibility that usage data for machines that have been experiencing problems may not be
reflective of how someone actually operates an ATV. Depending on the nature of the warranty
claim, the ATV owner may decide to not operate their machine as much as they want because of
a mechanical problem that doesn't allow the ATV to work or concern that the problem could be
6-46
-------
Chapter 6: Emissions Inventory
exacerbated by continued operation. Ultimately, because of the size of the data set, we felt we
couldn't dismiss the data simply based on the fact that the data is from warranty claims. We did
however have another concern with the data. The manufacturer indicated to us that they require
mileage to be reported on the warranty claim form. However, discussions with several local
dealers indicated something different. One dealer stated that the manufacturer had told them to
record hours instead of mileage, so that they either didn't include hours or only casually added it
when they remembered. Another dealer said that the manufacturer had indicated to them that
neither input was important, since the warranty is based on time after purchase (e.g., six months)
rather than usage and that they, therefore, entered data somewhat haphazardly, if at all. These
inconsistencies raised concerns over the accuracy of the mileage and hour data. If dealerships
don't pay close attention to what numbers they enter into the warranty claim forms, then the
warranty data could be suspect.
To eliminate this concern and more in general as a means to provide a degree of
validation to the data set used, we decided to only use data which contained both odometer and
hour meter readings. This way we could compare the values and make sure that they appeared to
be consistent with each other. Of the data points supplied, almost half of the data had only
odometer readings, while the other half had only hour readings. There was, however, a smaller
subset of data that included both types of data (approximately 3,000 data points). This data was
further screened as discussed below.
Honda Study
Honda hired a contractor to perform a phone survey of Honda ATV owners to inquire as
to how many total hours and miles were on their machines. The surveyor asked the owner if the
odometer and hour meter on their ATV was functional. If so, they asked them to read the
mileage and hour reading directly from their ATV. Honda only contacted people who had
purchased utility models since they are the only ATV models Honda sells that are equipped with
odometer and hour meters. The Honda survey does not contain data for sport models. Honda
used the odometer and hour meter readings combined with the model year of each model to
determine what the yearly mileage and hour usage was for each ATV in the survey. They had a
sample size of 611 ATVs that were mostly distributed evenly and randomly across the country,
thus the survey results appear to provide a national perspective.
The survey did not include any ATVs newer than 13 months or older than four years.
Honda wanted data for ATVs older than 13 months because in order to determine the number of
miles and hours ridden per year, they simply took the odometer or hour meter reading and
divided it by the machines age. For example, a machine that had 2,000 miles and was two years
old would average 1,000 miles per year. If they selected data from machines newer than a year
old, they would have to extrapolate to at least a year to get the average yearly usage. They felt
that extrapolating the data would be improper since it could either overestimate or underestimate
the usage depending on how the owner rode their machine during the months involved. If the
data was for a machine was only six months old, then the simplest way to extrapolate would be to
double the mileage or hours from the first six months. There is no way of knowing whether the
6-47
-------
Draft Regulatory Support Document
owner would have ridden more or less in the following six months, thus the concern with over-
or underestimating the usage.
Industry Panel Survey
In 1997, five of the major ATV manufacturers conducted an industry panel survey to
determine how well the survey information from the ATV exposure study performed by CPSC in
the same year would correlate with their own independent, but similar survey. The purpose of
the industry panel survey was to use a similar methodology and format as the CPSC study but to
survey an independent random sample of ATV owners to replicate the CPSC survey . They
aimed for the same approximate sample size gathered randomly from across the country.
Relevant survey questions used phrasing almost identical to that used in the CPSC survey. The
survey and data were provided to us on a confidential basis and cannot be shared here. However,
it can be stated that the yearly hour usage results from the industry panel survey are very
consistent with the CPSC study results.
State Studies
All of the state studies were done in 2000 or later and were not available at the time we
originally developed our ATV usage estimates for the proposal, with the exception of the
California study which was done in 1994. Three of the studies (Colorado, Maine, and Utah)
were provided to us by MIC. The Michigan study was obtained by EPA after a literature search
on ATV activity and usage. We were made aware of the California study through comments
from the Blue Ribbon Coalition. The purpose of the state studies was to measure the economic
impact of ATV and other recreational vehicle operation on the state economy. One of the results
from the studies was an estimate of how often ATVs were used in the respective state for that
particular year. The studies were based on user surveys that were typically mailed to registered
ATV owners. Mileage estimates were typically based off a single question posed in the survey
that asked the participant "How many miles did you ride your ATV in the past year?" All of the
studies measured usage in miles per year. Maine also recorded information on hours per year.
Average annual ATV usage from the state studies ranged from 320 mi/yr in Michigan to 1,270
mi/yr in Utah. It should be noted that according to the NONROAD model, these four states only
represent approximately four percent of the total U.S. ATV population and only Michigan is in
the top 20 states in ATV population.
The state studies were good for their intended purpose but since they weren't designed
specifically to answer the questions at hand, they each have some shortcomings that limit their
value to us. For example, all four states are cold climate states with cold winters and snow
accumulation that may limit the amount of annual operation, especially compared to some of the
warmer states that have higher ATV populations (e.g., Texas, Georgia, Tennessee, Alabama,
etc.). The ATV industry has indicated that ATV operation is becoming very prevalent in
agricultural use. Two of the states, Utah and Maine, are not large agricultural states, thus
potentially resulting in a lower usage estimate than could be expected from a national study. All
four of the state studies focused only on registered ATV owners. This has the potential for
6-48
-------
Chapter 6: Emissions Inventory
underestimating the number of miles ridden, since it does not provide a broad spectrum of all
ATV riders in the respective state. In some states, registration is only required for use on public
lands. Mileage estimates from three of the four studies were based on a single question inquiring
about ATV use. There was no attempt made to verify with the respondent the accuracy of their
estimate, as was done in the CPSC and Industry Panel studies. Four of the studies had
discrepancies between their estimates of mileage and fuel usage. In almost each of the studies,
the amount of fuel the respondents estimated they used for their ATV in one year would result in
mileage results far higher than the actual mileage estimates provided by the respondents, creating
a level of uncertainty about the viability of the mileage estimates. Finally, the California study
combined data for ATVs with off-highway motorcycles, making it impossible to discern the
mileage or fuel consumption for only ATVs.
We also obtained data from a separate report done by the State of California on ATV
activity data collection. California hired the University of California, College of Engineering -
Center for Environmental Research and Technology (CE-CERT) to instrument 41 ATVs and
have the owners operate them in several California off-road parks and measure vehicle and
engine speed.36 This work was done to help California better estimate ATV in-use operation and
emissions inventories within California. At this time, California has not completed their analysis
of the data, nor have they started to develop any new modeling, so their work is unavailable as a
source for ATV inventories. However, the CE-CERT draft report provides a summary of ATV
activity work. They focused on measuring vehicle speed and fuel consumption.
ATV Usage Derivation Methodology for the Final Rulemaking
Criteria
In attempting to reconcile the results from the various data sets, we established three
guiding criteria. The ideal data set would have all of these characteristics: 1) national scope; 2)
"real" data (actual measurement readings as opposed to survey results based on recollection); and
3) a broad spectrum of ATV use (sport and utility operation). None of the existing data sets meet
all three criteria. Therefore, we decided that it was important to select data sets that met two of
the three criteria. Four of the data sets meet two of the above criteria. The CPSC and Industry
Panel Survey data have a national scope and broad spectrum of ATV use. The warranty data and
the Honda survey data are both real data that provide a national scope. The state studies,
however, only provide a broad spectrum of use and many have a bias towards use on public
lands. They do not provide a national scope, nor are they generally based on "real" data.
Therefore, our methodology to determine ATV usage is based on the CPSC, Industry Panel
Survey, warranty, and Honda data. The state studies were not used because they did not meet
two of out three criteria, and as was briefly summarized above, had some shortcomings we could
not resolve. Of the three criteria, we felt that data which provide a national scope was the most
important, since it would remove any possible regional or state bias in ATV usage that could
exist. For example, some states may have higher usage levels because of unique or appealing
terrain, a large amount of public and private land available for riding on, an extended riding
season due to warmer climate, or greater potential for agricultural, ranching, and hunting usage,
6-49
-------
Draft Regulatory Support Document
that may not be reflected if we only use data from the four states that have performed studies on
ATV usage.
Utility vs. Sport ATVs
Utility ATVs are designed for multiple purposes and are most often used for hunting and
fishing, camping, yard work, farm work, as well as recreational trail riding. Sport ATVs are
designed for aggressive recreational riding over rough terrain and closed courses, where higher
speeds and performance are desired. According to Kawasaki, currently 75% of all ATV sales are
for utility models and 25% are for sport models. Ideally, we would want the population
percentage of sport and utility usage rather than sales, but this data is not available.
Hours vs. Miles
The NONROAD model uses miles per year of operation, rather than hours per year of
operation, as one of the main inputs in calculating the inventory estimates for HC, CO, NOx,
and PM emissions. Thus, to be consistent with the needs of the model, we were required to make
sure all of the data used was in miles per year of operation. Only the Honda and warranty data
had mileage data. However, all four data sets have hour data. In order to convert the hour data
into mileage estimates, we had to multiply the hour values by an average ATV speed estimate.
Average Speed
Ideally, we would want to develop an estimate for the average ATV speed that includes
both of the different types of models (utility and sport). Unfortunately, there wasn't a single data
set that could be used to determine average speed for both types of models. The Honda and
warranty data only included utility models. However, from these data sets we were able to
determine average speed for a utility ATV, since the ATVs in these data sets were equipped with
odometers and hour meters, which allowed us to calculate average speed. From this data we
were able to determine that the average speed for utility ATVs is about 8 mi/hr.
None of the four data sets had information that would allow the calculation of average
speed for sport ATV models. As discussed above, CE-CERT instrumented 41 ATVs and had the
owners operate them in several California off-road parks and measure vehicle and engine speed.
The off-road parks examined allowed operation over trails, desert, and sand dunes. Of the 41
instrumented ATVs, 36 were sport models and five were utility models. For the purposes of our
analysis, we considered all 41 ATVs as indicative of sport operation, since the riding that
occurred in these off-road parks was clearly recreational or sport, rather than utility usage. The
average speed for all 41 ATVs was about 13 mi/hr.
Methodology
The data permitted us to develop a methodology that would determine fleet average miles
per year by weighting separate mileage estimates for utility and sport ATVs based on average
6-50
-------
Chapter 6: Emissions Inventory
use, average speed and sales. The equation looks like this:
Utility A TVs Sport A TVs
(0.75)(hours/yr)(miles/hour) + (0.25)(hours/yr)(miles/hour) = Total miles/year for all ATVs
The 0.75 factor represents the percentage of total ATV sales that are for utility models,
while the 0.25 factor represents the remaining percentage of sales which are for sport models.
Population would have been preferable to sales, but that information was not available.
Utility ATV Estimates
To determine the mileage estimate for utility ATV models, we chose to use the data from
the Honda and warranty data sets. We selected these two data sets because they both consisted
entirely of data for utility ATVs. We merged both data sets and calculated the average hours per
year of operation and average speed (mi/hr). Prior to merging the data sets we performed several
quality checks of the data. First, we only used data that had both mileage and hour values. This
was so we could calculate an average speed for utility ATVs. All of the Honda data had both
values (approximately 605 data points). The warranty data had only a relatively small subset of
data that contained both mileage and hours (approximately 3,000 data points). Next, we
eliminated any of the warranty data that was for ATVs newer than 30 days and older than three
years, consistent with MIC's analysis. We found that for the warranty data, there appeared to a
significant number of data points that were duplicates (number of instances where same entry
was made twice). Since some of these duplicates were for usage rates that were either very high
or very low, we decided to remove all duplicates so that they would not bias the data. We also
deleted any samples that had identical miles and hours figures, on the basis that these readings
were probably mistakes, since it was unlikely that a rider would ride the exact same number of
miles and hours per year (e.g., 500 mi/yr and 500 hr/yr). Finally, we deleted any data from both
data sets that had an average speed greater than 25 mph, since information provided by the
American Motorcycle Association (AMA) on ATV race track statistics indicates that for
professional ATV racers, the average speed is 24 mph. Therefore, it did not seem reasonable to
include data for speeds in excess of those achieved by professional ATV racers.
The combined sample size of the merged data set was 2,531. The average speed for
utility ATVs from the merged data set was 8 miles per hour and the average hours of use was 151
hours per year. Our hours per year estimate for utility ATV use is corroborated by the CPSC
study and information from MIC. A discussion of nonrecreational or utility use in the CPSC
study states "..high use nonrecreational (utility) drivers tend to be older (36 years and up).." (See
page 14 of CPSC study). MIC has stated that the average age of individuals buying utility ATV
models is between 40 and 50 years old. The CPSC study indicates that for riders in the 40 to 50
year old age range, the average hourly usage was 158 hours per year (see page 27 of CPSC
study).
Sport ATV Estimates
6-51
-------
Draft Regulatory Support Document
To determine the mileage estimate for sport ATV models, we used the data from the
CPSC and Industry Panel Survey data sets. Since we were unable to determine average speed
from these data sets, we used the average speed of 13 mph derived from the CE-CERT data for
the 41 instrumented ATVs.
The CPSC and Industry Panel studies were done in 1997. Based on information from
these studies, between 50%-75% of the ATVs in both studies were from the 1980-1995 model
years. Between 1980 and 1990, sport ATVs were the predominant ATVs sold in the U.S.
Although their sales were starting to decline in favor of utility models, sport models were still
responsible for approximately 50% of all ATV sales from 1990 through 1995 and were the
majority of the ATV population. Therefore, both of these studies are most likely biased towards
operation with sport ATV models and should, therefore, be most representative of sport ATV
operation.
The annual riding hours from both data sets was determined by multiplying results of
three survey questions concerning riding patterns: (1) the number of months during which ATVs
were ridden during the previous year, (2) the number of days of riding in an average month, and
(3) the number of hours of riding in an average day. The total hours per year were then
calculated from the following equation.
hours months days hours
year year month day
We averaged annual rider hours from the CPSC and industry panel surveys, due to their
similarities in approach and results. In deriving average estimates from each, we reviewed
results for the questions used in the calculation, and modified some results that we considered
implausible. Specifically, for those records where the respondent claimed more than 10 hours of
use on an average day of riding, we limited daily usage at a maximum of 10 hours. The resulting
annual average usage rate was 216 hours per year.
In relation to their study objectives, the CPSC and Industry Panel studies both presented
usage results for the average rider, rather than for the average ATV. In other words, results are
presented as hours/rider/year, rather than hours/ATV/year. For the NPRM, we attempted to
correct hours/rider to hours/ATV using the ratio of the national rider population to the total ATV
population, as follows1*:
** In the NPRM analysis, we also applied an adjustment to subtract "inactive" riders from the total rider
population. In subsequent correspondence, the author of the CPSC study indicated that such an adjustment was
unnecessary, as the national population estimated in the report was intended to represent only "active riders," defined
as riders who had reported using their ATVs in the previous year. Thus, the "inactive rider" adjustment is not
presented here.
6-52
-------
Chapter 6: Emissions Inventory
hours hours I national rider population (riders) |
ATV-year rider-year {national ATV population (ATVs) j
In this analysis, we recalculated the average usage rate (i.e., hours per rider-year) using a data set
of results for individual respondents, which enabled review of individual responses, as mentioned
above. To be consistent with this approach, it would be appropriate to recalculate the
"correction" using individual responses, as opposed to gross national averages, as in the equation
above. However, several pieces of data needed for this calculation were unavailable, specifically,
the numbers of riders and ATVs in each respondent household. Accordingly, for purposes of this
analysis, we assumed that rider hours as reported in the CPSC and industry panel studies were
equivalent to ATV operating hours.
Mileage Estimate
By plugging in the above values derived for utility and sport ATVs average hourly
operation and average speed into the equation discussed above, we were able to determine a
mileage estimate for ATVs of 1,608 mile per year.
Utility A TVs Sport A TVs
(0.75)(151 hr/yr)(8 mi/hr) + (0.25)(216 hr/yr)(13 mi/hr) = 1,608 mi/yr
Conclusion
It is informative to consider the outcome from our methodology to the results of the
studies we did not use, or the alternative application of some of the individual studies that we did
use. The state studies do not have the strength of the national studies and were not used in our
analysis. The state studies represent only 4% of U.S. ATV registrations and all four states are
cold weather states that may not reflect winter use in warmer states. State methodologies give
results of mixed value. For example, two state studies had low mileage estimates: Michigan had
an estimate of 320 mi/yr and Colorado had an estimate of 610 mi/yr, while Utah had an estimate
of 1,270 mi/yr which is closer to our estimate. Maine had even more mixed results. Their
estimate ranged from 535 mi/yr to 1,646 mi/yr depending on which methodology they used to
determine mileage, the direct question or the multiple questions. The Honda survey data had an
estimate of 560 mi/yr. The warranty data had an estimate of 1,340 mi/yr. Both of these data sets
included only utility ATVs. The CPSC and Industry Panel studies had hour estimates of
approximately 250 hr/yr, which depending on the average speed used, can have a mileage range
of 1,900 mi/yr (for the average utility ATV speed of 8 mph) to 3,150 mi/yr (for the average sport
ATV speed of 13 mph). Therefore, we believe that our estimate of 1,608 miles per year is
reasonable and the best estimate considering all of the available data.
There is currently no data set which alone can be characterized as providing the best
estimate of ATV annual usage. All of the available data sets have some shortcomings. Looking
across all of the studies considered in the analysis yields mileage estimates from 320 mi/yr to
6-53
-------
Draft Regulatory Support Document
3,150 mi/yr. It is impossible to reconcile all eight data sets and it is not analytically appropriate
to average all of the data sets because they aren't all of equal strength or value. The methodology
we've developed is the best way to reconcile broadly ranging data of the highest value.
6A.2 Off-Highway Motorcycle Usage
On October 5, 2001, EPA published proposed emission regulations for nonroad land-
based recreational vehicles. These regulations covered snowmobiles, off-highway motorcycles,
and all-terrain vehicles (ATVs). The Motorcycle Industry Council, Inc. (MIC) submitted
comments suggesting that the EPA estimates for off-highway motorcycle (OHMC) usage had
been overestimated. They stated that our mileage estimate of 2,400 miles per year was too high
and that based on some additional information that they had obtained, a more reasonable estimate
was a lifetime average of 600 miles per year. As a result of these comments and the subsequent
new information, EPA has revised it's estimate of annual OHMC usage.
Background
On November 20, 2000 EPA published a Final Finding of Contribution and Advance
Notice of Proposed Rulemaking (ANPRM) for large nonroad spark-ignition engines and land-
based recreational vehicles. We had to develop emission inventories for the various engine and
vehicle categories covered by both of these documents. EPA has developed an emissions model
named NONROAD, which computes nationwide emission levels for nonroad engines. The
model incorporates data on emission rates, usage rates, and vehicle population to determine
annual emission levels of various pollutants. For recreational vehicles, and more specifically
OHMCs, data on emission rates and usage rates was extremely limited. Because of the lack of
data, we initially grouped OHMCs and ATVs together. However, as we performed literature
searches and attempted to uncover additional data on OHMC emissions and activity, it became
apparent that OHMCs and ATVs were used differently and unique emission rates, usage rates,
and populations should be established. We approached members of the OHMC industry to
provide us with any data that they had on emission and usage rates. MIC provided survey data
on off-highway motorcycle usage. We also found a study done in 1999 by the Oak Ridge
National Laboratory (ORNL) titled, "Fuel Used for Off-Road Recreation: A Reassessment of the
Fuel Use Model" that provided information on OHMC usage. We examined these two studies to
develop our estimate of OHMC usage for the November 2000, ANPRM and the October 2001,
NPRM.
Off-Highway Motorcycle Usage as developed for ANPRM and NPRM
For OHMC, there were two sources of information on activity or usage rates that we
examined. The first source was information provided by the motorcycle industry. MIC
periodically conducts surveys to obtain diverse information on motorcycle facts, such as number
of motorcycles per rider, types and makes of bikes, on-road or off-road, bike education, etc. The
survey also gathers information on motorcycle usage. MIC used two methods of estimating
OHMC usage from the survey results. Method one was based on the results of a single question
6-54
-------
Chapter 6: Emissions Inventory
that asks the respondent how many miles they rode their OHMC in the last year. Method two is
based on the compilation of the response from three questions: 1) how many months do you ride
per year, 2) how many days do you ride per month, and 3) how many miles do ride per day. The
MIC estimate for method one was 222 miles per year and 1,260 miles per year for method two.
MIC suggested that method one was the more appropriate estimate because method two may
compound any error that exists in the results of each of the three questions. We had concerns
with the results of the MIC survey because the values for method one and two were so
dramatically different.
The second source of information was the 1999 ORNL study. In their study, ORNL
estimated total average fuel usage for off-highway motorcycles. They provided a medium
estimate of average fuel usage for OHMCs of 59 gallons per year. Data from California and
some older SwRI work on OHMC emission testing suggested that the average fuel economy for
OHMCs was approximately 50 miles per gallon (mpg), as tested over the FTP (a relatively non-
aggressive driving cycle when compared to some OHMC uses). We determined that this
estimate could be too high for actual in-use off-road operation, so we derived from the data an
estimate of 40 mpg. By multiplying the average fuel used per year by the average fuel economy,
we arrived at an estimate of approximately 2,400 miles per year.
OHMC Usage = (59 gallons/year)(40 miles/gallon) = 2,400 miles/year
We also found another ORNL study published in 1994 where MIC also estimated average
fuel usage in their survey with a resulting mean value of 214 gallons per year.37 If we used our
estimate of 40 mpg, 214 gallons per year would yield 8,560 miles. Because of the large
discrepancies in the three MIC based values, we chose to use the estimate of 2,400 miles per
year.
New Information on Off-Highway Motorcycle Usage
Since the publication of the NPRM in October 2001, several new pieces of information
on OHMC usage have become available. These new sources consist of state studies from
California38, Michigan39, Oregon40, and Utah41 on OHMC usage (the California and Oregon
studies were used in both of the ORNL studies). These studies present information on the
number of miles OHMC's are ridden per year and/or the number of gallons of fuel used per year
riding OHMCs. We also received information from the American Motorcycle Association
(AMA) on rider surveys which attempt to quantify the number of miles ridden per year by the
average OHMC rider.
Finally, we obtained new information on the fuel consumption of OHMCs. The state of
California hired the University of California, College of Engineering - Center for Environmental
Research and Technology (CE-CERT) to instrument a number of OHMCs that were operated in
several California off-road parks and motocross tracks and measure vehicle and engine speed.42
This work was done to help California better estimate OHMC in-use operation and emissions
inventories within California. At this time, California has not completed their analysis of the
6-55
-------
Draft Regulatory Support Document
data, nor have they started to develop any new modeling, so their work is unavailable as a source
for OHMC emissions inventories. However, they have shared with us data on fuel consumption
from the OHMC testing. We also had updated emission and fuel economy test results for 10
OHMCs tested by EPA over the FTP.
State Studies
All four of the state studies included estimates of average yearly total fuel consumption
for OHMCs, but only the Michigan and Utah studies also provided estimates for average yearly
mileage for OHMCs. The average yearly total fuel consumption for the four studies ranges from
32 gallons per year for Michigan to 89 gallons per year for Oregon. The average for the four
studies is 57 gallons per year. Table 6A.2-1 lists the average yearly total fuel consumption for
the four studies. The two states that provided estimates for average yearly mileage were
Michigan and Utah. Michigan listed a yearly mileage of 494 miles per year, while Utah had a
value more than twice that with 1,067 miles per year.
Table 6A.2-1
Off-Highway Motorcycle Average Gallons of Fuel Consumed and Mileage Ridden Per Year
State Study
Michigan
California
Utah
Oregon
Average
Average Gallons Per Year
32
44
62
89
57
Average Mileage Per Year
494
n/a
1,067
n/a
781
AMA Survey
AMA presented survey results from 1994, 1996, 1998, & 2000 on how many miles AMA
members rode OHMCs in each of these years. The data indicates a trend toward increased
mileage each year. The survey was based on a mailing to AMA members listing questions as to
riding habits. AMA broke the survey results into six bins based on miles ridden in the last 12
months:
0-499 mi/yr
500 - 999 mi/yr
1,000-1,499 mi/yr
1,500-1,999 mi/yr
2,000 or more
- No answer
They determined the total number of miles ridden by taking the median value of each bin
6-56
-------
Chapter 6: Emissions Inventory
and multiplying it by the number of responses in that bin. They did this for each bin. They then
summed the results for all of the bins. The summation was then divided by the total number of
responses. For the bin categorizing responses of 2,000 miles or more, rather than using the
median, as with the other bins, they capped the mileage at 2,000 miles. This is problematic since
19% of all responses fell into this bin. By capping the values in this bin at 2,000 miles, the
estimate for this bin is too low. This would indicate that their estimate for average total OHMC
miles ridden per year is also probably too low. They estimated that in 2000, the average AMA
member rode 1,158 miles.
New Fuel Economy Estimates
We have tested nine OHMCs at our National Vehicle and Fuel Emissions Laboratory
(NVFEL) in Ann Arbor, Michigan. We also have the fuel economy results from a test done by
California on a 1999 Yamaha WR400. All of the tests are over the transient highway motorcycle
FTP test cycle. Table 6A.2-2 lists the results for the 4-stroke OHMCs. Table 6A.2-3 lists the
results for the 2-stroke OHMCs.
Table 6A.2-2
FTP Fuel Economy for 4-Stroke Off-Highway Motorcycles
Manufacturer
Yamaha
Yamaha
Husaberg
KTM
Model
WR250F
WR400
FE501
400EXC
Model Year
2001
1999
2001
2001
Average
Fuel Economy
(mpg)
39
55
53
54
50
6-57
-------
Draft Regulatory Support Document
Table 6A.2-3
FTP Fuel Economy for 2-Stroke Off-Highway Motorcycles
Manufacturer
KTM
KTM
KTM
KTM
KTM
KTM
Model
125 SX
125 SX
200 EXC
250 SX
250 EXC
300 EXC
Model Year
2001
2001
2001
2001
2001
2001
Average
Fuel Economy
(mpg)
21
31
22
18
20
21
22
The CE-CERT data developed for the State of California was based on actual in-use fuel
consumption measurements made on numerous OHMCs operated by the owners at several off-
road motorcycle parks and a motocross track. The parks consisted of trail riding, desert riding,
sand dune riding, and a mixture of all three. These riding scenarios could be considered closer to
worst case conditions that may not be reflective of average in-use operation nationally. The
results were 24 mpg for the 2-stroke machines and 27 mpg for the 4-stroke machines.
Off-Highway Motorcycle Usage Derivation Methodology for the Final Rule
Based on the new information we have received, there are two approaches we could
choose to estimate annual average OHMC usage. The first would be to base the estimate on the
mileage estimates presented in the Michigan, Utah, and AMA studies. The second would be to
use the same methodology we used for the ANPRM and NPRM, which uses total fuel
consumption from four state studies and fuel economy measurements from the California survey
and EPA FTP results to estimate mileage.
The first approach appears to be limited, since the AMA study under predicts the annual
mileage and since we do not have the raw data, there doesn't appear to be a method to upgrade
the estimate that wouldn't be somewhat arbitrary. This leaves only the mileage per year
estimates from the two state studies. There were two concerns with using the mileage estimates
from the two state studies. First of all, many OFDVIC models are not equipped with odometers,
which would make it difficult for participants responding to the state surveys to recall how many
miles they actually rode. Secondly, the average gallons per year and miles ridden per year
reported result in average fuel economy estimates of 15 and 17 miles per gallon. These values
are considerably lower than values from the CE-CERT and EPA testing. This means that either
the gallons per year estimates are high or the mileage per year estimates are low. Since we had
more sources for total fuel consumption and fuel economy values based on emissions test results
6-58
-------
Chapter 6: Emissions Inventory
and actual in-use operation, it appears to be more appropriate to use the second methodology
(which is based on fuel consumption), rather than the first methodology (which is based on
mileage) with only two questionable data points.
The equation for estimating average annual OHMC mileage based on fuel consumption
is:
OHMC Usage in miles per year = (gallons/year)(miles/gallon)
The gallons per year value is based on the average of the four state studies which is 57 gallons
per year. We are not including the ORNL study directly. The ORNL study consisted of data that
they had obtained from the California and Oregon studies and the MIC survey. ORNL agrees
with us that they thought the MIC survey information was of limited value for the same reasons
that we pointed put. To address their concern over using this data, they decided to give each of
the three studies a weighted value, with the MIC and Oregon studies having lower weightings
than the California study. We decided that it was more prudent to just use the California and
Oregon studies in combination with the other two new state studies from Utah and Michigan,
rather than include the MIC data.
For the fuel economy we had FTP results from EPA testing and in-use results from CE-
CERT. Since there is no way of knowing which of these set of values are the most correct (in-
use data was for relatively extreme operation) we chose to take the average of the two data sets.
However, before we did this, we decided to determine the overall fuel economy for each data set
based on the weighted impact of the two different types of engines, 2-stroke and 4-stroke. The
current break-down of 2-stroke and 4-stroke engines in OHMCs is 67% for 2-stroke engines and
33% for 4-stroke engines. Thus, we used the following equation to estimate fuel economy:
Fuel Economy (FE) = (0.67)(2-stroke FE (mpg)) + (0.33)(4-stroke FE (mpg))
For the EPA FTP testing, the average weighted fuel economy results are the following:
FE = (0.67)(22 mpg) + (0.33)(50 mpg) = 31 mpg
For the CE-CERT in-use measurements, the average weighted fuel economy results are the
following:
FE = (0.67)(24 mpg) + (0.33)(27 mpg) = 25 mpg
The average of these two data sets is 28 mpg. Combining the value of 28 mpg with the fuel
consumption value of 57 gallons per year results in an average of 1,600 miles per year for
OHMCs.
OHMC Usage = (57 gallons/year)(28 miles/gallon) = 1,600 miles/year
6-59
-------
Draft Regulatory Support Document
6-60
-------
Chapter 6: Emissions Inventory
Chapter 6 References
1. "Evaporative Emission Calculations for Recreational and Large SI Vehicles," Memorandum
from Mike Samulski, U.S. EPA to Docket A-2000-01, September 6, 2002, Document IV-B-38.
2. API Publication No. 4278, "Summary and Analysis of Data from Gasoline Temperature
Survey Conducted at Service Stations by American Petroleum Institute," Prepared by Radian
Corporation for American Petroleum Institute, November 11, 1976, Docket A-2000-01,
Document II-A-16.
3. Nulman, M., Olejnik, A., Samus, M., Fead, E., Rossi, G., "Fuel Permeation Performance of
Polymeric Materials," SAE Paper 2001-01-1999, 2001, Docket A-2000-01, Document No. IV-A-
23.
4. Lockhart, M., Nulman, M, Rossi, G., "Estimating Real Time Diurnal Permeation from
Constant Temperature Measurements," SAE Paper 2001-01-0730, 2001, Docket A-2000-01,
Document No. IV-A-21.
5. D. T. Wade, "Factors Influencing Vehicle Evaporative Emissions," SAE Paper 670126, 1967,
Docket A-2000-01, Document II-A-59.
6. Wade et. al., "Mathematical Expressions Relating Evaporative Emissions from Motor
Vehicles without Evaporative Loss-Control Devices to Gasoline Volatility," SAE Paper 72070,
1972, Docket A-2000-01, Document II-A-58.
7. S. Raghuma Reddy, "Prediction of Fuel Vapor Generation from a Vehicle Fuel Tank as a
Function of Fuel RVP and Temperature," SAE Paper 892089, 1989, Docket A-2000-01,
Document II-A-61.
8. S. Raghuma Reddy, "Prediction of Fuel Vapor Generation from a Vehicle Fuel Tank as a
Function of Fuel RVP and Temperature," SAE Paper 892089, 1989, Docket A-2000-01,
Document II-A-61.
9. "Final Regulatory Impact Analysis: Refueling Emission Regulations for Light Duty Vehicles
and Trucks and Heavy Duty Vehicles," U.S. EPA, January 1994, Docket A-2000-01, Document
II-A-79.
10. Pagan, Jaime, "Investigation on Crankcase Emissions from a Heavy-Duty Diesel Engine,"
U.S. Environmental Protection Agency, March, 1997, Docket A-2000-01, Document II-A-70.
1 l."The Role of Propane in the Fork Lift/Industrial Truck Market: A Study of its Status, Threats,
and Opportunities," Robert E. Myers for the National Propane Gas Association, December 1996,
Docket A-2000-01, Document II-A-86.
6-61
-------
Draft Regulatory Support Document
12."Three-Way Catalyst Technology for Off-Road Equipment Powered by Gasoline and LPG
Engines—Final Report" Jeff J. White, et al, Southwest Research Institute, SwRI 8778, April
1999, p. 45, Docket A-2000-01, Document II-A-08.
13."Revisions to the June 2000 Release of NONROAD to Reflect New Information and Analysis
on Marine and Industrial Engines," EPA memorandum from Mike Samulski to Docket A-98-01,
November 2, 2000, Docket A-2000-01, Document II-B-08.
14."Regulatory Analysis and Environmental Impact of Final Emission Regulations for 1984 and
Later Model Year Heavy Duty Engines," U.S. EPA, December 1979, p. 189, Docket A-2000-01,
Document U-A-71.
15. "Transient Adjustment Factors for Large SI Engines," EPA memorandum from Alan Stout to
Docket A-2000-01, September 27, 2001, Docket A-2000-01, Document U-B-29.
16. "Measurement of Evaporative Emissions from Off-Road Equipment," prepared for South
Coast Air Quality Management District by Southwest Research, November 1998, Docket A-
2000-01, Document II-A-10.
17. Emission Modeling for Recreational Vehicles," EPA memorandum from Line Wehrly to
Docket A-98-01, November 13, 2000, Docket A-2000-01, Document U-B-19.
18. "Median Life of Snowmobiles," note from Phil Carlson, EPA to Ed Klim, ISMA, May 10,
2002, Docket A-2000-01, Document IV-B-28.
19. "Updated Snowmobile Sales Projections from National Economic Research Associates,",
EPA memo from Phil Carlson to Docket A-2000-01, August 1, 2002, Docket A-2000-01,
Document IV-B-27.
20. "Updated Population Growth Projections for Snowmobiles, ATVs, and OHMCs," EPA
memo from Phil Carlson to Docket A-2000-01, August 7, 2002, Document IV-B-29.
21. "Updated Emission Factors for Snowmobiles, ATVs, and OHMCs," EPA memo from Phil
Carlson to Docket A-2000-01, August 8, 2002, Document IV-B-26.
22. Emission Modeling for Recreational Vehicles," EPA memorandum from Line Wehrly to
Docket A-98-01, November 13, 2000, Docket A-2000-01, Document U-B-19.
23. "Estimate for All-Terrain Vehicle Annual Usage," EPA memo from Line Wehrly to Docket
A-2000-01, September 6, 2002, Document IV-B-35.
24. "Updated Population Growth Projections for Snowmobiles, ATVs, and OHMCs," EPA
memo from Phil Carlson to Docket A-2000-01, August 7, 2002, IV-B-29.
25. "Updated Emission Factors for Snowmobiles, ATVs, and OHMCs," EPA memo from Phil
Carlson to Docket A-2000-01, August 8, 2002, Document IV-B-26.
6-62
-------
Chapter 6: Emissions Inventory
26. "Estimate for Off-Highway Motorcycle Annual Usage," EPA memo from Line Wehrly to
Docket A-2000-01, August 26, 2002, Document IV-B-36.
27. "Revised Median Life for Off-Highway Motorcycles," EPA memo from Phil Carlson to
Docket A-2000-01, August 6, 2002, Document IV-B-30.
28. "Updated Population Growth Projections for Snowmobiles, ATVs, and OHMCs," EPA
memo from Phil Carlson to Docket A-2000-01, August 7, 2002, Document IV-B-29.
29. "Updated Emission Factors for Snowmobiles, ATVs, and OHMCs," EPA memo from Phil
Carlson to Docket A-2000-01, August 8, 2002, Document IV-B-26.
30. "Emission Modeling for Recreational Vehicles," EPA memorandum from Line Wehrly to
Docket A-98-01, November 13, 2000, Docket A-2000-01, Document U-B-19.
31. "1993 statewide Off-Highway Vehicle User Survey Analysis," Prepared for the California
Department of Parks and Recreation, Off-Highway Vehicle Recreation Division by California
State University.
32. " Economic Contribution of Off-Highway Vehicle Use in Colorado," Prepared for the
Colorado Off-Highway Vehicle Coalition by Hazen and Sawyer, July 2001 (submitted as an
attachment to the Motorcycle Industry Council's comments, document IV-D-214).
33. "Gasoline Consumption Attributable to ATVs in Maine," Margaret Chase Smith Center for
Public Policy, The University of Maine, Jonathon Rubin, Suzanne K. Hart, and Charles Morris,
June 2001 (submitted as an attachment to the Motorcycle Industry Council's comments,
document IV-D-214).
34. "Michigan Licensed Off-Road Vehicle Use and Users: 1998-1999," Department of Park,
Recreation and Tourism Resources, Michigan State University, October 25, 2000
35. "Off-Highway Vehicle Uses and Owner Preferences in Utah," Institute for Outdoor
Recreation and Tourism, Department of Forestry Resources, Utah State University, July 22, 2001
(submitted as an attachment to the Motorcycle Industry Council's comments, document IV-D-
214).
36. "Off-Highway Motorcycle/All-Terrain Vehicle Activity-Data Collection; and Personal
Watercraft Activity-Data Collection; Test Cycle Development and Emissions Test," Prepared for
California Air Resource Board by Thomas Durbin, Matthew R. Smith, Ryan D. Wilson, and Ted
Younglove of College of Engineering - Center for Environmental Research and Technology, July
2002.
37. "Fuel Used for Off-Highway Recreation," Patricia Hu, David Trumble, and An Lu, prepared
by Oak Ridge National Laboratory for Federal Highway Administration, July 1994 (ORNL-
6794).
6-63
-------
Draft Regulatory Support Document
38. "A Study To Determine Fuel Tax Attributable To Off-Highway and Street Licensed Vehicles
Used For Recreation Off-Highway," prepared for the California Department of Transportation In
Cooperation with the Department of Parks and Recreation, by Tyler and Associates, November
1990.
39. "Michigan Licensed Off-Road Vehicle Use and Users: 1998-1999," Department of Park,
Recreation and Tourism Resources, Michigan State University, October 25, 2000.
40. "Fuel Used for Off-Highway Recreation," Patricia Hu, David Trumble, and An Lu, prepared
by Oak Ridge National Laboratory for Federal Highway Administration, July 1994 (ORNL-
6794).
41. "Off-Highway Vehicle Uses and Owner Preferences in Utah," Institute for Outdoor
Recreation and Tourism, Department of Forestry Resources, Utah State University, July 22, 2001
(submitted as an attachment to the Motorcycle Industry Council's comments, document IV-D-
214).
42. "Off-Highway Motorcycle/All-Terrain Vehicle Activity-Data Collection; and Personal
Watercraft Activity-Data Collection; Test Cycle Development and Emissions Test," Prepared for
California Air Resource Board by Thomas Durbin, Matthew R. Smith, Ryan D. Wilson, and Ted
Younglove of College of Engineering - Center for Environmental Research and Technology, July
2002.
6-64
-------
Chapter 7: Cost per Ton
Chapter?: Cost Per Ton
7.1 Cost Per Ton by Engine Type
7.1.1 Introduction
This chapter presents our estimate of the cost per ton of the various standards contained
in this rule. The analysis relies on the costs estimates presented in Chapter 5 and the estimated
lifetime emissions reductions using the information presented in Chapter 6. The chapter also
presents a summary of the cost per ton of other recent EPA mobile source rulemakings for
comparison purposes. Finally, this chapter presents the estimated costs and emission reductions
as incurred over the first twenty years after the standards are implemented.
In calculating net present values that were used in our cost-per-ton estimates, we used a
discount rate of 7 percent, consistent with the 7 percent rate reflected in the cost-per-ton analyses
for other recent mobile source programs. OMB Circular A-94 requires us to generate benefit and
cost estimates reflecting a 7 percent rate. Using the 7 percent rate allows us to make direct
comparisons of cost-per-ton estimates with estimates for other, recently adopted, mobile source
programs.
However, we consider that the cost and cost-per-ton estimates for future proposed mobile
source programs could reflect a 3 percent rate. The 3 percent rate is in the 2 to 3 percent range
recommended by the Science Advisory Board's Environmental Economics Advisory Committee
for use in EPA social benefit-cost analyses, a recommendation incorporated in EPA's new
Guidelines for Preparing Economic Analyses (November 2000). 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-per-ton of this rule with future proposed rules which might use the 3
percent rate. The results using both a 3 percent and 7 percent discount rate are provided in this
chapter.
7.1.2 Compression-Ignition Recreational Marine
As described in Chapter 5, several of the anticipated engine technologies will result in
improvements in engine performance that go beyond emission control. While the cost estimates
described in Chapter 5 do not take into account the observed value of performance
improvements, these non-emission benefits should be taken into account in the calculation of
cost-effectiveness. We believe that an equal weighting of emission and non-emission benefits is
justified for those technologies which clearly have substantial non-emission benefits, namely
electronic controls, fuel injection changes, turbocharging, and aftercooling for diesel engines and
upgrading to electronic fuel injection for gasoline engines. For some or all of these technologies,
a greater value for the non-emission benefits could likely be justified. This has the effect of
7-1
-------
Draft Regulatory Support Document
halving the cost for those technologies in the cost-per-ton calculation. The cost-per-ton values in
this chapter are based on this calculation methodology.
Although the rule will also result in PM reductions, we apply the total cost to the ozone
forming gases (HC and NOx) presented in Chapter 6 for these calculations. The estimated per
vessel costs presented in Chapter 5 change over time, with reduced costs in the long term. We
have estimated both a near-term and long-term cost per ton as presented in Table 7.1-1 assuming
a 7 percent discount rate. Table 7.1-2 presents the cost per tons results assuming a 3 percent
discount rate..
Table 7.1-1
Estimated CI Recreational Marine Cost Per Ton of HC + NOx Reduced
(7 percent discount rate)
100 kW near-term
100 kW long-term
400 kW near-term
400 kW long-term
750 kW near-term
750 kW long-term
Composite near-term
Composite long-term
Total Cost per
Vessel (NPV)
$231
$141
$396
$175
$1,118
$374
$291
$155
Lifetime Reductions
(NPV tons)
0.24
0.97
1.32
0.44
Discounted Per Vessel Cost
($/ton)
$954
$583
$409
$181
$844
$282
$669
$356
7-2
-------
Chapter 7: Cost per Ton
Table 7.1-2
Estimated CI Recreational Marine Cost Per Ton of HC + NOx Reduced
(3 percent discount rate)
100 kW near-term
100 kW long-term
400 kW near-term
400 kW long-term
750 kW near-term
750 kW long-term
Composite near-term
Composite long-term
Total Cost per
Vessel (NPV)
$231
$141
$396
$175
$1,118
$374
$291
$155
Lifetime Reductions
(NPV tons)
0.33
1.31
1.69
0.59
Discounted Per Vessel Cost
($/ton)
$703
$429
$301
$133
$661
$221
$495
$263
7.1.3 Large Industrial SI Equipment
This section provides our estimate of the cost per ton of emissions reduced for large SI
engines >19 kW. We have calculated cost per ton on the basis of exhaust HC plus NOx for
gasoline, LPG and CNG engines and evaporative HC for gasoline engines. The analysis relies on
the costs estimates in presented in Chapter 5 and the estimated net present value of the per
vehicle lifetime emissions reductions (tons) presented in Chapter 6.
For the exhaust emission standards, the estimated per vehicle costs presented in Chapter 5
change over time, with reduced costs in the long term. We have estimated both a near-term and
long-term cost per ton. In addition, we have estimated the cost per ton both with and without
estimated fuel/maintenance savings. We have estimated the cost per ton for both the Phase 1 and
Phase 2 standards, with the Phase 2 estimates incremental to Phase 1. The results of the cost per
ton analysis for exhaust emission controls are presented in Tables 7.1.3-1 through 7.1.3-3 for
gasoline, LPG and CNG engines assuming a 7 percent discount rate. The results of the cost-per-
ton analysis for exhaust emission controls using a 3 percent discount rate follow in Tables 7.1.3-
4 through 7.1.3-6.
7-3
-------
Draft Regulatory Support Document
Table 7.1-3
Estimated Large SI Gasoline Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(7 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$802
$487
$60
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($3,247)
-
Lifetime
Reductions
(NPV tons)
1.6
0.3
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$496
$301
$175
$41
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($1,514)
($1,708)
-
.
Table 7.1-4
Estimated Large SI LPG Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(7 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$552
$340
$53
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($4,557)
-
Lifetime
Reductions
(NPV tons)
3.5
1.0
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$158
$97
$56
$15
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($1,146)
($1,206)
-
.
7-4
-------
Chapter 7: Cost per Ton
Table 7.1-5
Estimated Large SI CNG Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(7 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$552
$340
$53
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($1,648)
-
Lifetime
Reductions*
(NPV tons)
3.6
0.9
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$153
$94
$61
$16
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($304)
($363)
-
.
* The reductions are calculated on the basis of NMHC+NOx for CNG engines only.
Table 7.1-6
Estimated Large SI Gasoline Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(3 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$802
$487
$60
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($3,926)
-
Lifetime
Reductions
(NPV tons)
2.0
0.4
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$409
$248
$143
$33
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($1,573)
($1,733)
-
.
7-5
-------
Draft Regulatory Support Document
Table 7.1-7
Estimated Large SI LPG Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(3 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$552
$340
$53
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($5,492)
-
Lifetime
Reductions
(NPV tons)
4.2
1.2
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$131
$81
$46
$12
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($1,162)
($1,212)
-
.
Table 7.1-8
Estimated Large SI CNG Engine >19 kW Cost Per Ton of Exhaust HC+NOx Reduced
(3 percent discount rate)
Standard
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Total Cost
per Vehicle
(NPV)
$552
$340
$53
$14
Lifetime
Fuel/
Maintenance
Cost per
Vehicle
(NPV)
($2,005)
.
Lifetime
Reductions*
(NPV tons)
4.4
1.1
Discounted Per
Vehicle Cost Per Ton
without
Fuel/Maintenance
Savings
($/ton)
$125
$77
$49
$13
Discounted Per
Vehicle Cost Per Ton
with
Fuel/Maintenance
Savings
($/ton)
($321)
($369)
-
_
* The reductions are calculated on the basis of NMHC+NOx for CNG engines only.
For the evaporative emission standards, the estimated per vehicle costs are presented in
7-6
-------
Chapter 7: Cost per Ton
Chapter 5. We have estimated the cost per ton both with and without the estimated fuel savings
which occur as evaporative emissions are reduced. The results of the cost per ton analysis for
evaporative emission controls for gasoline large SI engines >19 kW are presented in Table 7.1-9
based on both a 7 percent and 3 percent discount rate.
Table 7.1-9
Estimated Large SI Gasoline Engine >19 kW Cost Per Ton of Evaporative HC Reduced
Discount
Rate
7%
3%
Total Cost
per Vehicle
(NPV)
$13
$13
Lifetime
Fuel Cost
per Vehicle
(NPV)
($56)
($69)
Lifetime
Evaporative
HC Reductions
(NPV tons)
0.16
0.19
Discounted
Per Vehicle
Cost Per Ton
without Fuel Savings
($/ton)
$84
$68
Discounted
Per Vehicle
Cost Per Ton
with Fuel Savings
($/ton)
($279)
($295)
7.1.4 Recreational Vehicle Exhaust Emissions
This section provides our estimate of the cost per ton of exhaust emissions reduced for
recreational vehicles. We have calculated cost per ton on the basis of HC plus NOx for off-road
motorcycles and ATVs, and both HC and CO for snowmobiles. For snowmobiles, we have
spread costs evenly over HC and CO reductions for purposes of calculating cost per ton. If
reductions in other pollutants were included, the cost per ton estimates would be lower. The
analysis relies on the per vehicle costs estimated in Chapter 5 and the estimated net present value
of the per vehicle lifetime emissions reductions (tons) presented in Chapter 6. These cost per ton
estimates do not include permeation control which is calculated separately for recreational
vehicles, below.
The estimated per vehicle costs presented in Chapter 5 change over time, with reduced
costs in the long term. We have estimated both a near-term and long-term cost per ton. In
addition, we have estimated cost per ton both with and without estimated fuel savings. For
snowmobiles, we have estimated the cost per ton for all three phases of standard incremental to
the previous standards. The results of the analysis using the 7 percent discount rate are presented
in Tables 7.1-10 through Table 7.1-12. The results using the 3 percent discount rate follow in
Tables 7.1-13 through 7.1-15.
7-7
-------
Draft Regulatory Support Document
Table 7.1-10
Estimated Snowmobile Average Cost Per Ton of HC and CO Reduced
(7 percent discount rate)
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Phases
near-term
Phase 3
long-term
Total
Average
Cost per
Vehicle
$73
$40
$131
$77
$89
$54
Lifetime
Average Fuel
Cost per
Vehicle
(NPV)
($57)
($286)
($191)
Lifetime
Average
Reductions
(NPV tons)
HC
0.40
0.10
n/a
CO
1.02
n/a
0.25
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
HC
$90
$50
$1,370
$810
n/a
n/a
CO
$40
$20
n/a
n/a
$360
$220
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
HC
$20
($20)
($1,610)
($2,190)
n/a
n/a
CO
$10
($10)
n/a
n/a
($410)
($550)
Table 7.1-11
Estimated ATV Average Cost Per Ton of HC + NOx Reduced
(7 percent discount rate)
near-term
long-term
Total
Average
Cost per
Vehicle
$84
$42
Lifetime
Average Fuel
Cost per
Vehicle
(NPV)
($24)
Lifetime
Average
Reductions
(NPV tons)
0.21
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$400
$200
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$290
$90
7-8
-------
Chapter 7: Cost per Ton
Table 7.1-12
Estimated Off-highway Motorcycle Average Cost Per Ton of HC + NOx Reduced*
(7 percent discount rate)
near-term
long-term
Total
Average
Cost per
Vehicle
$155
$95
Lifetime
Average Fuel
Cost per
Vehicle
(NPV)
($48)
Lifetime
Average
Reductions
(NPV tons)
0.38
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$410
$250
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$280
$120
* non-competition models only
Table 7.1-13
Estimated Snowmobile Average Cost Per Ton of CO Reduced
(3 percent discount rate)
Phase 1
near-term
Phase 1
long-term
Phase 2
near-term
Phase 2
long-term
Phases
near-term
Phase 3
long-term
Total
Average
Cost per
Vehicle
$73
$40
$131
$77
$89
$54
Lifetime
Average Fuel
Cost per
Vehicle
(NPV)
($57)
($286)
($191)
Lifetime
Average
Reductions
(NPV tons)
HC
0.50
0.12
n/a
CO
1.25
n/a
0.31
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
HC
$70
$40
$1,110
$650
n/a
n/a
CO
$30
$20
n/a
n/a
$290
$180
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
HC
$20
($20)
($1,305)
($1,770)
n/a
n/a
CO
$10
($10)
n/a
n/a
($330)
($450)
7-9
-------
Draft Regulatory Support Document
Table 7.1-14
Estimated ATV Average Cost Per Ton of HC + NOx Reduced
(3 percent discount rate)
Phase 1
near-term
Phase 1
long-term
Total
Average
Cost per
Vehicle
$84
$42
Lifetime
Average
Fuel Cost per
Vehicle
(NPV)
($24)
Lifetime
Average
Reductions
(NPV tons)
0.26
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$330
$160
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$240
$70
Table 7.1-15
Estimated Off-highway Motorcycle Average Cost Per Ton of HC + NOx Reduced*
(3 percent discount rate)
near-term
long-term
Total
Average
Cost per
Vehicle
$155
$95
Lifetime
Average Fuel
Cost per
Vehicle
(NPV)
($48)
Lifetime
Average
Reductions
(NPV tons)
0.46
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$340
$210
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$230
$100
* Non-competition models only
7.1.5 Recreational Vehicle Permeation Emissions
This section provides our estimate of the cost per ton of permeation emissions reduced for
recreational vehicles. The analysis relies on the per vehicle costs estimated in Chapter 5 and the
estimated lifetime emissions reductions (tons) presented in Chapter 6. All costs and emission
reductions are discounted to the year of sale of the boats at a rate of 7 percent. Table 7.1-16
presents the cost per ton with and without consideration of the significant fuel savings that will
result from evaporative emission control assuming a 7 percent discount rate. The cost per ton
results assuming a 3 percent discount rate are presented in Table 7.1-17. As shown in these
tables, the fuel savings more than offset the cost of the evaporative emission control technology.
7-10
-------
Chapter 7: Cost per Ton
Table 7.1-16
Estimated Cost Per Ton of HC Reduced
(7 percent discount rate)
Total
Cost Per
Vehicle
Lifetime Fuel
Savings Per
Vehicle
(NPV)
Lifetime
Reductions
Per Vehicle
(NPV tons)
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
Snowmobiles
tank permeation
hose permeation
total
$2
$4
$7
$5
$7
$11
0.0125
0.0182
0.0307
$185
$234
$214
($178)
($129)
($149)
All Terrain Vehicles
tank permeation
hose permeation
total
$2
$1
$3
$3
$3
$6
0.0070
0.0082
0.0152
$215
$157
$184
($148)
($206)
($179)
Off-Highway Motorcycles
tank permeation
hose permeation
total
$1
$2
$3
$1
$3
$5
0.0041
0.0096
0.0137
$348
$175
$226
($15)
($188)
($137)
7-11
-------
Draft Regulatory Support Document
Table 7.1-17
Estimated Cost Per Ton of HC Reduced
(3 percent discount rate)
Total
Cost Per
Vehicle
Lifetime Fuel
Savings Per
Vehicle
(NPV)
Lifetime
Reductions
Per Vehicle
(NPV tons)
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
Snowmobiles
tank permeation
hose permeation
total
$2
$4
$7
$5
$8
$13
0.0144
0.0209
0.0353
$161
$204
$186
($202)
($159)
($177)
All Terrain Vehicles
tank permeation
hose permeation
total
$2
$1
$3
$3
$4
$7
0.0086
0.0100
0.0186
$175
$128
$150
($188)
($235)
($213)
Off-Highway Motorcycles
tank permeation
hose permeation
total
$1
$2
$3
$2
$4
$6
0.0047
0.0110
0.0157
$302
$152
$197
($61)
($211)
($166)
7.2 Cost Per Ton for Other Mobile Source Control Programs
Because the primary purpose of cost-effectiveness is to compare our program to
alternative programs, we made a comparison between the cost per ton values presented in this
chapter and the cost-effectiveness of other programs. Table 7.2-1 summarizes the cost
effectiveness of several recent EPA actions for controlled emissions from mobile sources. These
values show that the cost-effectiveness of the standards for this rulemaking fall within the range
of these other programs.
7-12
-------
Chapter 7: Cost per Ton
Table 7.2-1
Cost-effectiveness of Previously Implemented
Mobile Source Programs (Costs Adjusted to 1997 Dollars)
Program
Tier 2 vehicle/gasoline sulfur
2007 Highway HD diesel
2004 Highway HD diesel
Off-highway diesel engine
Tier 1 vehicle
NLEV
Marine SI engines
On-board diagnostics
Marine CI engines
$/ton
1,340-2,260
1,458-1,867
212-414
425 - 675
2,054 - 2,792
1,930
1,171 - 1,846
2,313
24 - 176
The primary advantage of making comparisons to previously implemented programs is
that their cost-effectiveness values were based on a rigorous analysis and are generally accepted
as representative of the efficiency with which those programs reduce emissions. Unfortunately,
previously implemented programs can be poor comparisons because they may not be
representative of the cost-effectiveness of potential future programs. In the context of the
Agency's rulemaking to revise the ozone and PM NAAQS", the Agency compiled a list of
additional known technologies that may be considered in devising new emission reductions
strategies.1 Through this broad review, over 50 technologies were identified to reduce NOx,
VOC, or PM. The cost-effectiveness of these technologies averaged approximately $5,000/ton
for VOC, $13,000/ton for NOx, and $40,000/ton for PM.
In summary, given the array of controls that will have to be implemented to make
progress toward attaining and maintaining the NAAQS, we believe that the weight of the
evidence from alternative means of providing substantial NOx + NMHC emission reductions
indicates that our program is cost-effective. This is true from the perspective of other mobile
source control programs or from the perspective of other stationary source technologies that
might be considered.
7.3 20-Year Cost and Benefit Analysis
The following section presents the year-by-year cost and emission benefits associated
with the standards for the 20-year period after implementation of the standards. For the
categories where we expect a reduction in fuel consumption due to the standards, the fuel savings
11 This rulemaking was remanded by the D.C. Circuit Court on May 14, 1999. However,
the analyses completed in support of that rulemaking are still relevant, since they were designed
to investigate the cost-effectiveness of a wide variety of potential future emission control
strategies.
7-13
-------
Draft Regulatory Support Document
are presented separately. The overall cost, incorporating the impact of the fuel savings is also
presented.
Table 7.3-1 presents the year-by-year cost and emission benefits for the compression-
ignition (CI) recreational marine requirements. (The numbers presented in Table 7.3-1 are not
discounted.)
Table 7.3-1
Cost and Emission Benefits of the CI Recreational Marine Requirements
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
HC+NOx
Benefits ('tons')
639
1,310
2,015
2,842
3,705
4,583
5,496
6,424
7,361
8,333
9,313
10,300
11,320
12,345
13,373
14,407
15,416
16,423
17,379
18.190
CO
Benefits ('tons')
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Cost w/o
Fuel Savings
$7,806,010
$8,365,319
$8,573,839
$9,413,530
$9,637,035
$5,213,411
$5,176,672
$5,290,764
$4,958,052
$5,062,713
$5,167,682
$5,272,652
$5,377,623
$5,482,592
$5,587,562
$5,692,532
$5,797,503
$5,902,472
$6,007,442
$6.112.413
Fuel Savings
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Cost w/
Fuel Savings
$7,806,010
$8,365,319
$8,573,839
$9,413,530
$9,637,035
$5,213,411
$5,176,672
$5,290,764
$4,958,052
$5,062,713
$5,167,682
$5,272,652
$5,377,623
$5,482,592
$5,587,562
$5,692,532
$5,797,503
$5,902,472
$6,007,442
$6.112.413
Table 7.3-2 presents the sum of the costs and emission benefits over the twenty year
period after the CI recreational marine requirements take effect, on both a non-discounted basis
and a discounted basis (assuming a seven percent discount rate). The annualized cost and
emission benefits for the twenty-year period (assuming the seven percent discount rate) are also
presented.
7-14
-------
Chapter 7: Cost per Ton
Table 7.3-2
Annualized Cost and Emission Benefits for the Period 2006-2025
due to the CI Recreational Marine Requirements
Undiscounted
20-year Value
Discounted
20-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
181,174
79,294
7,485
CO Benefits
(tons)
0
0
0
Cost w/o
Fuel Savings
(Million $)
$125.9
$75.6
$7.1
Fuel Savings
(Million $)
$0.0
$0.0
$0.0
Cost w/
Fuel Savings
(Million $)
$125.9
$75.6
$7.1
Table 7.3-3 presents the year-by-year cost and emission benefits for the large spark-
ignition (SI) engine exhaust and evaporative requirements. (The numbers presented in Table 7.3-
3 are not discounted.)
7-15
-------
Draft Regulatory Support Document
Table 7.3-3
Cost and Emission Benefits of the Large SI Engine Requirements
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
HC+NOx
Benefits (tons)
77,259
133,247
187,149
265,975
329,756
391,853
451,604
506,031
542,932
576,173
606,048
627,504
646,713
664,729
681,633
697,598
712,638
727,377
741,822
756,116
CO
Benefits (tons)
82,130
161,404
239,617
474,426
678,940
883,333
1,076,572
1,260,180
1,427,950
1,589,734
1,730,897
1,803,389
1,866,433
1,922,727
1,972,496
2,017,393
2,059,586
2,099,624
2,137,602
2,176,504
Cost w/o
Fuel Savings
$88,806,711
$91,185,462
$75,632,060
$84,493,379
$86,588,256
$68,943,347
$70,571,930
$72,200,513
$68,895,067
$70,414,812
$71,934,556
$73,454,300
$74,974,044
$76,493,788
$78,013,532
$79,533,276
$81,053,020
$82,572,765
$84,092,509
$85,612,253
Fuel Savings
$52,725,475
$102,980,886
$152,926,193
$198,943,367
$242,829,040
$285,094,033
$325,741,703
$360,969,773
$379,398,454
$395,033,152
$408,985,187
$421,230,723
$432,435,409
$443,121,586
$453,291,958
$462,975,097
$471,991,726
$480,919,953
$489,742,176
$498,805,313
Cost w/
Fuel Savings
$36,081,236
($11,795,424)
($77,294,133)
($114,449,988)
($156,240,784)
($216,150,686)
($255,169,773)
($288,769,260)
($310,503,387)
($324,618,340)
($337,050,631)
($347,776,423)
($357,461,365)
($366,627,798)
($375,278,426)
($383,441,821)
($390,938,706)
($398,347,188)
($405,649,667)
($413,193,060)
Table 7.3-4 presents the sum of the costs and emission benefits over the twenty year
period after the large SI engine exhaust and evaporative requirements are to take effect, on both a
non-discounted basis and a discounted basis (assuming a seven percent discount rate). The
annualized cost and emission benefits for the twenty-year period (assuming the seven percent
discount rate) are also presented.
7-16
-------
Chapter 7: Cost per Ton
Table 7.3-4
Annualized Cost and Emission Benefits for the Period 2004-2023
due to the Large SI Engine Requirements
Undiscounted
20-year Value
Discounted
20-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
10,324,157
4,945,366
466,808
CO Benefits
(tons)
27,660,937
12,631,259
1,192,303
Cost w/o
Fuel Savings
(Million $)
$1,565.5
$892.4
$84.2
Fuel Savings
(Million $)
$7,060.1
$3,433.5
$324.1
Cost w/
Fuel Savings
(Million $)
($5,494.7)
($2,541.1)
($239.9)
Table 7.3-5 presents the year-by-year cost and emission benefits for the snowmobile
exhaust and permeation requirements. (The numbers presented in Table 7.3-5 are not
discounted.)
7-17
-------
Draft Regulatory Support Document
Table 7.3-5
Cost and Emission Benefits of the Snowmobile Requirements
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
HC+NOx
Benefits (tons)
3,933
12,374
22,502
32,977
45,890
59,319
76,209
93,845
112,031
130,397
148,455
165,914
181,480
194,065
204,737
214,492
222,824
229,775
235,195
239.208
CO
Benefits (tons)
9,941
31,272
54,058
77,582
105,287
134,052
169,882
207,354
245,980
284,962
323,196
360,691
394,252
420,522
442,187
461,929
478,985
493,443
504,816
513.372
Cost w/o
Fuel Savings
$6,583,529
$13,546,439
$13,183,508
$13,455,182
$38,933,137
$38,685,132
$51,957,587
$52,701,157
$45,309,024
$44,402,290
$41,860,214
$41,738,365
$42,211,850
$42,677,612
$43,138,523
$43,138,523
$43,138,523
$43,138,523
$43,138,523
$43.138.523
Fuel Savings
$391,491
$1,225,462
$2,469,788
$3,747,560
$9,545,473
$15,633,653
$25,065,896
$34,856,171
$44,859,909
$54,975,510
$65,045,977
$74,963,244
$84,545,886
$93,597,148
$102,179,264
$110,195,147
$116,664,922
$121,533,783
$125,181,189
$127.680.885
Cost w/
Fuel Savings
$6,192,038
$12,320,977
$10,713,720
$9,707,622
$29,387,664
$23,051,479
$26,891,691
$17,844,987
$449,115
($10,573,219)
($23,185,764)
($33,224,879)
($42,334,036)
($50,919,536)
($59,040,741)
($67,056,624)
($73,526,400)
($78,395,261)
($82,042,667)
($84.542.362^)
Table 7.3-6 presents the sum of the costs and emission benefits over the twenty year
period after the exhaust and permeation requirements for snowmobiles take effect, on both a non-
discounted basis and a discounted basis (assuming a seven percent discount rate). The
annualized cost and emission benefits for the twenty-year period (assuming the seven percent
discount rate) are also presented.
7-18
-------
Chapter 7: Cost per Ton
Table 7.3-6
Annualized Cost and Emission Benefits for the Period 2006-2025
due to the Snowmobile Requirements
Undiscounted
20-year Value
Discounted
20-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
2,625,622
1,141,218
107,723
CO Benefits
(tons)
5,713,763
2,499,999
235,983
Cost w/o
Fuel Savings
(Million $)
$746.1
$379.9
$35.9
Fuel Savings
(Million $)
$1,214.4
$494.6
$46.7
Cost w/
Fuel Savings
(Million $)
($552.9)
($145.8)
($10.8)
Table 7.3-7 presents the year-by-year cost and emission benefits for the exhaust and
permeation requirements for ATVs. (The numbers presented in Table 7.3-7 are not discounted.)
7-19
-------
Draft Regulatory Support Document
Table 7.3-7
Cost and Emission Benefits of the ATV Requirements
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
HC+NOx
Benefits (tons)
6,321
23,496
44,313
69,788
97,132
125,655
154,669
183,543
211,466
238,164
263,043
285,924
304,746
316,793
324,521
329,849
333,031
335,389
337,137
338.413
CO
Benefits (tons)
4,380
14,702
26,267
39,269
53,061
67,377
81,890
96,230
110,237
123,603
136,030
147,442
156,446
161,571
164,444
166,533
167,857
168,858
169,554
170.055
Cost w/o
Fuel Savings
$42,463,856
$79,998,942
$76,517,949
$70,286,998
$65,302,237
$56,379,476
$52,441,476
$52,441,476
$52,441,476
$52,441,476
$49,999,146
$47,556,815
$47,556,815
$47,556,815
$47,556,815
$47,556,815
$47,556,815
$47,556,815
$47,556,815
$47.556.815
Fuel Savings
$933,911
$4,771,537
$9,546,220
$13,556,430
$17,819,539
$22,221,930
$26,654,575
$31,026,962
$35,203,428
$39,163,369
$42,825,354
$46,173,993
$48,949,487
$50,819,932
$52,105,004
$52,985,302
$53,516,650
$53,912,720
$54,215,317
$54.442.855
Cost w/
Fuel Savings
$41,529,945
$75,227,405
$66,971,729
$56,730,568
$47,482,698
$34,157,546
$25,786,901
$21,414,514
$17,238,048
$13,278,107
$7,173,792
$1,382,822
($1,392,672)
($3,263,117)
($4,548,189)
($5,428,487)
($5,959,835)
($6,355,905)
($6,658,502)
($6.886.040^)
Table 7.3-8 presents the sum of the costs and emission benefits over the twenty year
period after the exhaust and permeation requirements for ATVs take effect, on both a non-
discounted basis and a discounted basis (assuming a seven percent discount rate). The
annualized cost and emission benefits for the twenty-year period (assuming the seven percent
discount rate) are also presented.
7-20
-------
Chapter 7: Cost per Ton
Table 7.3-8
Annualized Cost and Emission Benefits for the Period 2006-2025
due to the ATV Requirements
Undiscounted
20-year Value
Discounted
20-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
4,323,393
1,951,668
184,224
CO Benefits
(tons)
2,225,806
1,014,866
95,796
Cost w/o
Fuel Savings
(Million $)
$1,078.7
$641.0
$60.5
Fuel Savings
(Million $)
$710.8
$325.3
$30.7
Cost w/
Fuel Savings
(Million $)
$367.9
$315.7
$29.8
Table 7.3-9 presents the year-by-year cost and emission benefits for the off-highway
motorcycle exhaust and permeation requirements. (The numbers presented in Table 7.3-9 are not
discounted.
7-21
-------
Draft Regulatory Support Document
Table 7.3-9
Cost and Emission Benefits of the Off-Highway Motorcycle Requirements
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
HC+NOx
Benefits (tons)
3,085
9,742
18,028
27,409
37,325
47,542
57,733
67,631
77,400
86,976
96,030
103,553
108,707
112,249
114,994
117,320
119,371
121,137
122,719
124.218
CO
Benefits (tons)
2,330
7,398
13,408
20,236
27,463
34,917
42,364
49,612
56,774
63,810
70,471
76,047
79,882
82,490
84,503
86,207
87,712
89,007
90,173
91.284
Cost w/o
Fuel Savings
$16,269,072
$31,813,960
$29,592,786
$26,871,067
$24,698,975
$21,818,012
$21,366,690
$21,580,357
$21,796,160
$22,014,121
$22,234,263
$22,456,605
$22,681,171
$22,907,983
$23,137,063
$23,368,434
$23,602,118
$23,838,139
$24,076,521
$24.317.286
Fuel Savings
$633,450
$2,061,773
$3,878,230
$5,903,201
$8,016,233
$10,166,886
$12,282,632
$14,311,527
$16,290,860
$18,207,111
$19,981,626
$21,421,145
$22,409,671
$23,107,057
$23,655,679
$24,122,020
$24,532,680
$24,886,440
$25,200,670
$25.496.728
Cost w/
Fuel Savings
$15,635,622
$29,752,187
$25,714,556
$20,967,866
$16,682,742
$11,651,126
$9,084,058
$7,268,830
$5,505,300
$3,807,010
$2,252,637
$1,035,460
$271,500
($199,074)
($518,616)
($753,586)
($930,562)
($1,048,301)
($1,124,149)
($ 1.1 79.442}
Table 7.3-10 presents the sum of the costs and emission benefits over the twenty year
period after the exhaust and permeation requirements for off-highway motorcycles take effect, on
both a non-discounted basis and a discounted basis (assuming a seven percent discount rate).
The annualized cost and emission benefits for the twenty-year period (assuming the seven
percent discount rate) are also presented.
7-22
-------
Chapter 7: Cost per Ton
Table 7.3-10
Annualized Cost and Emission Benefits for the Period 2006-2025
due to the Off-Highway Motorcycle Requirements
Undiscounted
20-year Value
Discounted
20-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
1,573,169
715,044
67,495
CO Benefits
(tons)
1,156,088
525,674
49,620
Cost w/o
Fuel Savings
(Million $)
$470.4
$268.9
$25.4
Fuel Savings
(Million $)
$326.6
$149.1
$14.1
Cost w/
Fuel Savings
(Million $)
$143.9
$119.8
$11.3
Table 7.3-11 presents the year-by-year cost and emission benefits for all of the
requirements. (The numbers presented in Table 7.3-11 are not discounted.)
7-23
-------
Draft Regulatory Support Document
Table 7.3-11
Cost and Emission Benefits of the Requirements for All Equipment Categories
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
HC+NOx
Benefits (tons)
77,259
133,247
201,127
312,897
416,614
524,869
635,656
743,130
837,039
927,616
1,014,306
1,091,374
1,163,554
1,230,420
1,287,886
1,333,050
1,370,263
1,403,445
1,432,464
1,458,840
1,482,773
1.504.484
CO
Benefits (tons)
82,130
161,404
256,268
527,798
772,673
1,020,420
1,262,383
1,496,526
1,722,086
1,942,930
2,143,888
2,275,764
2,396,130
2,506,907
2,603,076
2,681,976
2,750,720
2,814,293
2,872,156
2,927,812
2,980,012
3.028.620
Cost w/o
Fuel Savings
$88,806,711
$91,185,462
$148,754,528
$218,218,038
$214,456,337
$188,970,125
$209,143,314
$194,296,545
$199,837,493
$202,428,566
$196,439,267
$197,374,901
$194,235,348
$193,518,225
$195,840,991
$198,158,277
$200,472,982
$202,329,067
$204,187,466
$206,048,201
$207,911,297
$209.776.777
Fuel Savings
$52,725,475
$102,980,886
$154,885,046
$207,002,139
$258,723,278
$308,301,224
$361,122,948
$408,992,242
$443,401,557
$475,227,812
$505,339,384
$533,576,713
$560,288,366
$585,679,968
$609,197,002
$630,499,234
$649,931,673
$668,222,422
$684,456,428
$699,138,256
$712,465,187
$724.482.067
Cost w/
Fuel Savings
$36,081,236
($11,795,424)
($6,130,518)
$11,215,899
($44,266,941)
($119,331,100)
($151,979,633)
($214,695,697)
($243,564,064)
($272,799,246)
($308,900,116)
($336,201,812)
($366,053,018)
($392,161,743)
($413,356,011)
($432,340,957)
($449,458,691)
($465,893,354)
($480,268,962)
($493,090,055)
($504,553,890)
($514.705.289^)
Table 7.3-12 presents the sum of the costs and emission benefits over the twenty-two year
period after all of the requirements take effect, on both a non-discounted basis and a discounted
basis (assuming a seven percent discount rate). The annualized cost and emission benefits for the
twenty-two year period (assuming the seven percent discount rate) are also presented. (A twenty-
two period is used in this aggregate analysis to cover the first twenty years of each of the
standards which begins in 2004 for large SI engines and concludes in 2006 for the other
categories of equipment.)
7-24
-------
Chapter 7: Cost per Ton
Table 7.3-12
Annualized Cost and Emission Benefits for the Period 2004-2025
due to the Requirements for All Equipment
Undiscounted
22-year Value
Discounted
22-year Value
Annualized
Value
HC+NOx
Benefits
(tons)
22,106,425
9,073,158
789,161
CO Benefits
(tons)
44,300,504
17,971,253
1,561,958
Cost w/o
Fuel Savings
(Million $)
$4,374.0
$2,176.7
$192.5
Fuel Savings
(Million $)
$11,072.1
$4,701.9
$410.1
Cost w/
Fuel Savings
(Million $)
($6,698.1)
($2,525.2)
($217.6)
7-25
-------
Draft Regulatory Support Document
Chapter 7 References
1 ."Regulatory Impact Analyses for the Particulate Matter and Ozone National Ambient Air
Quality Standards and Regional Haze Rule," Appendix B, "Summary of control measures in the
PM, regional haze, and ozone partial attainment analyses," Innovative Strategies and Economics
Group, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC, July 17, 1997, Docket A-2000-01, Document II-A-77.
7-26
-------
Chapter 7: Cost per Ton
7-27
-------
Draft Regulatory Support Document
7-28
-------
Chapter 8: Small Business Flexibility Analysis
Chapter 8: Small Business Flexibility Analysis
This section presents our Small Business Flexibility Analysis (SBFA) which evaluates the
impacts of the rule on small businesses. Prior to issuing our proposal, we analyzed the potential
impacts of our program on small businesses. As a part of this analysis, we convened two Small
Business Advocacy Review (SBAR) Panels, under the requirements of the Regulatory Flexibility
Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act of 1996
(SBREFA), 5 USC 601 et seq. Through the two Panel processes, we gathered advice and recom-
mendations from small entity representatives (SERs) who would be affected by the regulation.
The two Panel reports have been placed in the rulemaking record.
8.1 Requirements of the Regulatory Flexibility Act
The Regulatory Flexibility Act was amended by SBREFA to ensure that concerns re-
garding small entities are adequately considered during the development of new regulations that
affect them. 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 proposed rules will have on small entities. In general, the Regulatory Flexibility
Act calls for determining, to the extent feasible, a rule's economic impact on small entities,
exploring regulatory options for reducing any significant economic impact on a substantial
number of such entities, and explaining the ultimate choice of regulatory approach.
For purposes of assessing the impacts of this final rule on small entities, a small entity is
defined as: (1) a small business that meet the definition for business based on SB A size stan-
dards; (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. This rulemaking will only affect the small businesses.
When proposing rules subject to notice and comment under the Clean Air Act, we are
generally required under the Regulatory Flexibility Act to conduct an Initial Regulatory Flexi-
bility Analysis, unless we certify that the requirements of a regulation will not cause a significant
impact on a substantial number of small entities. Although we are not required to conduct a
Final Regulatory Flexibility Analysis (FRFA), EPA has decided to prepare an assessment of the
impacts of the final rule on small entities. This SBFA would meet the requirements of a FRF A,
were EPA required to prepare one.
In accordance with section 609 of the RFA, EPA conducted an outreach to affected small
entities and convened a Small Business Advocacy Review (SBAR) Panel prior to proposing this
rule, to obtain advice and recommendations of representatives of the small entities that poten-
tially would be subject to the rule's requirements. Through the Panel process, we gathered advice
8-1
-------
Draft Regulatory Support Document
and recommendations from small-entity representatives who would be affected by the regulation,
and published the results in a Final Panel Report, dated July 17, 2001. EPA had previously
convened a separate Panel for marine engines and vessels. This panel also produced a report,
dated August 25, 1999. We also prepared an Initial Regulatory Flexibility Analysis (IRFA) in
accordance with section 603 of the Regulatory Flexibility Act. The IRFA is found in chapter 8 of
the Draft Regulatory Support Document. Both Panel reports and the IRFA have been placed in
the docket for this rulemaking (Public Docket A-2000-01, items U-A-85, U-F-22, and HI-B-01).
We proposed the majority of the Panel recommendations, and took comments on this and
other issues. The information we received during the course of the rulemaking indicated that
fewer small entities than we had first estimated would be significantly impacted by the rule.
During the SBAR Panel process, we were concerned that ATV and off-highway motorcycle
importers would have limited access to certified models for import. We received no comments
confirming this concern and believe that the use of cleaner four-stroke engines in these vehicles
will continue to increase. As a result, we believe all these small companies should be able to find
manufacturers that are able to supply compliant engines for import into the U.S. These importers
incur no development costs, and they are not involved in adding emission-control hardware or
other variable costs to provide a finished product to market. We also expect that importers would
select vehicles for import that have fuel tanks and hoses that comply with the permeation stan-
dards. However, even if they were not able to find such vehicles, the few additional dollars per
vehicle that it would cost to bring them into compliance with the permeation standards is insig-
nificant in comparison with the normal selling prices for these vehicles. They should therefore
expect to buy and sell their products with the normal markup to cover their costs and profit. As
noted below, we expect all 21 known small-business importers to face compliance costs of less
than one percent of their revenues. Thus, EPA has determined that this final rule will not have a
significant economic impact on a substantial number of small entities. Also, as a result of
comments received on the proposal, we are finalizing changes that we believe will further reduce
the level of impact to small entities directly regulated by the rule. These changes and can be
found below in Section 8.6, "Steps Taken to Minimize the Economic Impact on Small Entities."
The key elements of the Small Business Flexibility Analysis include:
• the need for and objectives of the rule;
• the significant issues raised by public comments, a summary of the Agency's
assessment of those issues, and a statement of any changes made to the rule as a
result of those comments;
• the types and number of affected small entities to which this rule will apply;
• the projected reporting, record keeping, and other compliance requirements of the
regulation, including the classes of small entities that would be affected and the
type of professional skills necessary for preparation of the report or record;
8-2
-------
Chapter 8: Small Business Flexibility Analysis
• the steps taken to minimize the economic impacts of the regulation on small
entities, consistent with the stated objectives of the applicable statutes.
8.2 Need For and Objectives of the Rule
The process of establishing standards for nonroad engines began in 1991 with a study to
determine whether emissions of carbon monoxide (CO), oxides of nitrogen (NOx), and volatile
organic compounds (VOCs) from new and existing nonroad engines, equipment, and vehicles are
significant contributors to ozone and CO concentrations in more than one area that has failed to
attain the national ambient air quality standards for ozone and CO.jj In 1994, EPA finalized its
finding that nonroad engines as a whole "are significant contributors to ozone or carbon
monoxide concentrations" in more than one ozone or carbon monoxide nonattainment area.kk
Upon making this finding, the Clean Air Act (CAA or the Act) requires EPA to establish
standards for all classes or categories of new nonroad engines that cause or contribute to air
quality nonattainment in more than one ozone or carbon monoxide (CO) nonattainment area.
Since the finding in 1994, EPA has been engaged in the process of establishing programs to
control emissions from nonroad engines used in many different applications. Nonroad categories
already regulated include:
Land-based compression ignition (CI) engines (e.g., farm and construction equipment),
• Small land-based spark-ignition (SI) engines (e.g., lawn and garden equipment, string
trimmers),
• Marine engines (outboards, personal watercraft, CI commercial, CI engines <37kW)
• Locomotive engines
On December 7, 2000, EPA issued an Advance Notice of Proposed Rulemaking
(ANPRM), and then issued a Notice of Proposed Rulemaking (NPRM) on September 14, 2001.
This final rule continues the process of establishing standards for nonroad engines and vehicles,
as required by CAA section 213(a)(3), with new emission standards for recreational marine
diesel engines, recreational vehicles, and other nonroad spark-ignition engines over 19 kW.
8.3 Issues Raised by Public Comments
The two SBAR Panels considered a wide range of options and regulatory alternatives for
providing small businesses with flexibility in complying with the regulation. As part of the
process, the Panels requested and received comment on several ideas for flexibility that were
suggested by SERs and Panel members. The major options recommended by the Panel can be
jj "Nonroad Engine and Vehicle Emission Study—Report and Appendices," EPA-
21A-201, November 1991 (available in Air docket A-91-24). It is also available through the
National Technical Information Service, referenced as document PB 92-126960.
kk 59 FR 31306 (July 17, 1994).
8-3
-------
Draft Regulatory Support Document
found in Section 9 of the Panel Reports.
Many of the flexible approaches recommended by the Panels can be applied to several of
the equipment categories that may be affected by the regulation. However, during the consul-
tation process, it became evident that, in a few situations, it could be helpful to small entities if
unique provisions were available. Three such provisions are described below.
(a) Snowmobiles: The Panel recommended that EPA seek comment on a provision
allowing small snowmobile manufacturers to request a relaxed standard for one or more
engine families, up to 300 engines per year, until the family is retired or modified, if such
a standard is justifiable based on the criteria described in the Panel report. Based on
comments received, we have adopted this provision, increasing the sales allowance to
600 engines per year.
(b) ATVs and Off-road Motorcycles: The Panel recommended that the hardship
provision for ATVs and off-road motorcycles allow for annual review of the relief for up
to two years for importers to obtain complying products. We are adopting this provision.
(c) Large SI: The Panel recommended that small entities be granted the flexibility
initially to reclassify a small number of their small displacement engines into EPA's
small spark-ignition engine program (40 CFR part 90). Small entities would be allowed
to use those requirements instead of the requirements we adopt for large entities. We are
not adopting this provision, preferring instead to rely on the more flexible approach
provided under the hardship provisions. Since there are only two companies affected, we
believe this approach best addresses these concerns.
The Panel also crafted recommendations to address SERs' concerns that ATV and off-
road motorcycle standards that essentially required manufacturers to switch to four-stroke
engines might increase costs to the point that many small importers and manufacturers could
experience significant adverse effects. The Panel recommended that EPA request comment in its
proposed rule on the effect of the regulation on these small entities, with the specific intent of
developing information—including the extent to which sales of their products would likely to be
reduced in response to changes in product price attributable to the standards—that could be used
to inform a decision in the final rule as to whether EPA should provide additional flexibility
beyond that considered by the Panel. We received no comments addressing this concern and
therefore believe that the use of four-stroke engines for ATVs and off-highway motorcycles will
continue to increase; as a result all these companies should be able to find manufacturers that are
able to supply compliant engines into the U.S. market.
In the NPRM for this rule, we proposed only exhaust emission controls for recreational
vehicles. However, several commenters raised the issue of control of evaporative emissions
related to permeation from fuel tanks and fuel hoses, and indicated that our obligations under
section 213 of the Clean Air Act included control of permeation emissions. The commenters
pointed to work done by the California Air Resources Board (ARE) on permeation emissions
8-4
-------
Chapter 8: Small Business Flexibility Analysis
from plastic fuel tanks and rubber fuel line hoses for various types of nonroad equipment, as well
as portable plastic fuel containers, as evidence of a new emissions concern. Our own investiga-
tion into the hydrocarbon emissions related to permeation of fuel tanks and fuel hoses from
recreational land-based and marine applications supports the concerns raised by the commenters.
Therefore, on May 1, 2002, we published a notice in the Federal Register reopening the comment
period and requesting comment on possible approaches to regulating permeation emissions from
recreational vehicles. The notice provided a detailed analysis of possible approaches to regu-
lating permeation emissions and the expected costs and emission reductions from these ap-
proaches. The notice also cited sample regulation language that could be used if we decided to
finalize such requirements. Commenters had thirty days from May 1, 2002 to provide comments
on the notice. We received comments from several affected manufacturers during the comment
period, including at least one small entity. These comments have been addressed in the final
Summary and Analysis of Comments document, and we have made several changes to the rule in
response to suggestions of the commenters.
We received a number of other comments from engine and equipment manufacturers and
consumers during the comment period after we issued the NPRM. A number of small engine and
equipment manufacturers commented on the financial hardships they would face in complying
with the proposed regulations. Most requested that we consider a number of hardship provisions,
primarily an exemption from or a delay in the implementation of the proposed standards, or
certain flexibilities in the certification process. Due to the wide variety of engines, vehicles, and
equipment covered by this rulemaking, we decided that a variety of provisions were needed to
address the concerns of the small entities involved. A summary of the comments pertaining to
these small entity issues can be found in our Final Summary and Analysis of Comments
document contained in the public docket for this rulemaking. Changes to the proposal as a result
of SER or other comments are noted below in section 8.6 for each of the sectors affected by this
rule.
8.4 Description of Affected Entities
Table 8.4-1 provides an overview of the primary SBA small business categories
potentially affected by this regulation.
8-5
-------
Draft Regulatory Support Document
Table 8.4-1
Primary SBA Small Business Categories Potentially Affected by this Regulation
Industry
Motorcycles and motorcycle parts
manufacturers
Snowmobile and ATV manufacturers
Independent Commercial Importers of
Vehicles and parts
Nonroad SI engines
Internal Combustion Engines
Boat Building and Repairing
NAICSa Codes
336991
336999
421110
333618
333618
336612
Defined by SBA as a
Small Business If:b
<500 employees
<500 employees
<100 employees
< 1,000 employees
<1000 employees
<500 employees
a. North American Industry Classification System
b. According to SBA's regulations (13 CFR part 121), businesses with no more than the listed number of employees or
dollars in annual receipts are considered "small entities" for purposes of a regulatory flexibility analysis.
8.4.1 Recreational Vehicles (ATVs, off-highway motorcycles, and snowmobiles)
The ATV sector has the broadest assortment of manufacturers. There are seven com-
panies, Bombardier, Honda, Polaris, Kawasaki, Yamaha, Suzuki, and Arctic Cat, representing
over 95 percent of total domestic ATV sales. The remaining 5 percent come from one small
manufacturer, IPC, and a number of importers who tend to import inexpensive, youth-oriented
ATVs from China and other Asian nations.. EPA has identified 21 small companies (as defined
in Table 8.4.1, above) that offer off-road motorcycles, ATVs, or both products. Annual unit sales
for these companies can range from a few hundred to several thousand units per year.
We expect all 21 known small-business importers to face compliance costs less than one
percent of their revenues. These companies incur no development costs and they are not invol-
ved in adding emission-control hardware or other variable costs to provide a finished product to
market. As a result, they should expect to buy and sell their products with the normal mark-up to
cover their costs and profit. During the SBAR Panel process, we were also concerned that
importers would have limited access to certified models for import. We received no comments
confirming this concern and believe that the supply of four-stroke engines for ATVs and off-
highway motorcycles will continue to increase; as a result all these companies should be able to
find manufacturers that are able to supply compliant engines into the U.S. market. We also
received no comments regarding the permeation standards issue, and believe that the importers
will simply purchase compliant models and pass the costs on to the ultimate consumers.
8-6
-------
Chapter 8: Small Business Flexibility Analysis
Five large manufacturers, Honda, Kawasaki, Yamaha, Suzuki, and KTM. accounted for
approximately 85 percent of all off-highway motorcycle production for sale in the U.S. There are
three small business manufacturing off-highway motorcycles in the U.S. Two of these companies
make only competition models, so they don't need to certify their products under this regulation.
ATK already offers engines that should be meeting the new emission standards, especially under
our provisions allowing design-based certification, so we estimate that their compliance costs
will be much less than one percent of their revenues.
IPC is the only small business manufacturing ATVs, offering two separate youth ATV
models. IPC already uses four-stroke engines. Moreover, the standards are based on emissions
per kilometer, which are easier to meet for models with small-displacement engines. We
estimate compliance costs of about $50,000 for R&D plus $15,000 for certification, which is
much less than 1 percent of IPC's annual revenues.
We do not believe that compliance with the permeation standards will place a significant burden
on either the small manufacturers or on the importers. We have estimated the cost of compliance
for ATVs and off-highway motorcycles at roughly three dollars per vehicle for the fuel hoses and
surface coating for the fuel tank. This estimate includes shipping, and is based on buying the ne-
cessary hoses and surface treatment for the fuel tanks from outside suppliers. Thus, no capital
outlays are required, and the increase in vehicle cost is insignificant, so that it can easily be pas-
sed along to the ultimate consumer. However, to ensure that these requirements do not adversely
affect small manufacturers, we are implementing, where they are applicable to permeation, the
same flexibility options we proposed for the exhaust emission standards.
Based on available industry information, four major manufacturers, Arctic Cat, Bombar-
dier (also known as Ski-Doo), Polaris, and Yamaha, account for over 99 percent of all domestic
snowmobile sales. The remaining one percent comes from very small manufacturers who tend to
specialize in unique and high performance designs. There is also one potential manufacturer
(Redline), which we have learned is owned by a larger entity (TMAG) and is therefore not a
small business, that hopes to produce snowmobiles within the next year.
We are aware of five small businesses that have been producing snowmobiles. Two of
these have discontinued production since we completed the SBAR panel. Two of the remaining
three manufacturers (Crazy Mountain and Fast, Inc.) specialize in high performance versions of
standard recreational snowmobile types (i.e., travel and mountain sleds). The other manufacturer
(Fast Trax) produces a unique design, which is a scooter-like snowmobile designed to be ridden
standing up. Most of these manufacturers build less than 50 units per year.
Fast, Inc. produces four engine models, one of which is a four-stroke design. The four-
stroke engine will need no development or certification work, since we allow design-based
certification for this situation. We expect the two-stroke engines to qualify for the special stan-
dards that apply to small businesses. As a result, Fast will have only limited development costs
8-7
-------
Draft Regulatory Support Document
to reduce emissions from these engines. We estimate a total of $75,000 in R&D and $15,000 for
certification for each of the three engine families. They are projecting sales of around 1,000 units
for the time when standards would apply. Since this is a substantial increase over their current
volume of 180 per year, we base revenue calculations on projected sales of only 500 per year.
The resulting calculation shows a compliance burden less than one percent.
Fast, Inc. was the only recreational vehicle manufacturer to comment on the permeation
provisions contained in the May 1 notice. Fast stated that, as a small manufacturer of snowmo-
biles, they would undergo additional hardship due to this rule, because they do not have the sales
volume to warrant installing the barrier treatment equipment for fuel tanks. They also commen-
ted that shipping and processing of fuel tanks by an outside vendor could take 3-4 months, and
that as a small business it would be unworkable for them to tie up funds for such a long period.
We agree that it is neither necessary nor cost-effective for a small manufacturer to make
the capital investment necessary for an in-house treatment facility, given the relatively low cost
of the compliance with the requirements and the availability of materials and treatment support
by outside vendors. Low permeation fuel hoses are available from vendors today, and we would
expect that surface treatment would be applied through an outside company. The $5 to $7 per
vehicle incremental cost resulting from the permeation requirements is insignificant compared to
the price of one of these high-end sleds, and should not pose a significant cash-flow problem,
particularly in view of the likely sales volumes involved. These costs are based on vendor costs,
including shipping charges.
Since the costs are low and no capital investment is required, we believe that the per-
meation control requirements should be relatively easy for small businesses to meet. However, to
make sure that these requirements do not adversely affect small entities, we are implementing,
where they are applicable to permeation, the same flexibility options we proposed for the recrea-
tional vehicle exhaust emission standards . These flexibility options included a 2 year delay of
the standards, design-based certification, broader engine families, waiving production line
testing, use of assigned deterioration factors, carryover of certification data, ABT, and hardship
provisions. These are further described below in section 8.6.. Given the low costs and these
flexibilities, there should be no significant economic impact on small entities.
Crazy Mountain produces only about 20 snowmobiles per year in addition to their more
extensive business in aftermarket parts and accessories for snowmobiles from other manufac-
turers. We don't have revenue information for the whole company, but we expect that total costs
of redesigning and certifying their single model will exceed 3 percent of snowmobile revenues.
However, with its low production volume, Crazy Mountain could likely qualify for the special
standards that apply to small businesses.
Fast Trax provided no response to repeated outreach efforts to determine potential
economic effects of the final rule. We expect them to purchase compliant engines, which would
-------
Chapter 8: Small Business Flexibility Analysis
result in a compliance burden of less than one percent. Due to the small engine displacements
used in current models, we would expect these engines to be certified to the Small SI standards.
8.4.2 Large Spark Ignition Engines
The Panel was aware of one engine manufacturer of Large SI engines that qualifies as a
small business. Westerbeke plans to produce engines that meet the standards adopted by CARB
in 2004, with the possible exception of one engine family. If EPA adopts long-term standards,
this would require manufacturers to do additional calibration and testing work. If EPA adopts
new test procedures (including transient operation), there may also be a cost associated with
upgrading test facilities. We expect that Westerbeke will face relatively small compliance costs
as a result of this rule, since the California-compliant engines will need only a small amount of
additional development effort to meet the long-term standards. We estimate that they will need
$200,000 each for two engine families, with a potential need to spend an additional $300,000 for
upgrading test cells. These costs are less than one percent of their annual revenues.
Since we completed the proposal Wisconsin Motors, a small business, bought the assets
of a company that had gone bankrupt. This company did not exist during the SBAR Panel pro-
cess associated with this rule. Through public comments and other outreach efforts, this com-
pany has stated that it faces significant development costs, though much of this effort is required
to improve the engine enough to sustain a market presence as other manufacturers continue to
make improvements to competitive engines. Under the hardship provisions, we expect them to
spread compliance costs over several years to reduce the impact of emission standards. Wiscon-
sin should be able to delay compliance until they are able to retool for production and add
developmental efforts to incorporate emission-control technologies. Substantial tooling expenses
will be necessary independent of emission standards. We estimate a need for $500,000 for
emission-measurement facilities and $500,000 of development costs for each of two engine
models. New testing to certify and show compliance on these models comes to about $50,000
total. These costs are about 4 percent of the projected revenues for the time frame when Wiscon-
sin will be certifying their engines. Since this manufacturer is operating in a niche market with
customers providing public comments citing the need for these engines, we expect that most of
the increased cost of production will be recovered by increased revenues.
8.4.3 Marine Vessels
Marine vessels include the boat, engine, and fuel system. Exhaust emission controls
including NTE requirements, as addressed in the two Panel Reports, would affect the engine
manufacturers and may affect boat builders.
8-9
-------
Draft Regulatory Support Document
8.4.3.1 Small Diesel Engine Marinizers
We have determined that there are at least 16 companies that manufacture diesel engines
for recreational vessels. Nearly 75 percent of diesel engines sales for recreational vessels in 2000
can be attributed to three large companies. Six of the 16 identified companies are considered
small businesses as defined by SB A. Based on sales estimates for 2000, these six companies
represent approximately 4 percent of recreational marine diesel engine sales. The remaining
companies each comprise between two and seven percent of sales for 2000.
We are thus aware of six small businesses that may produce recreational marine diesel
engines. Alaska Diesel and Westerbeke do not offer recreational versions of the marine diesel
engines that are different than their commercial products. The regulations allow manufacturers
to certify all their products under the commercial standards, even if they may be used in recrea-
tional applications. As a result, these companies would likely minimize their costs by certifying
all their products to the commercial standards. We therefore believe that they will experience no
significant new compliance costs for these engines as a result of this regulation. Daytona has, to
the best of our knowledge, discontinued production of their marine product line.
For those companies that will be certifying recreational marine diesel engines, we directly
apply the development and certification costs from Chapter 5. For each engine family, we esti-
mate $200,000 of development costs and $30,000 of certification costs. The variable costs
considered in Chapter 5 are very small relative to the price of the engines, so we would expect
manufacturers to fully recover these costs over time.
American Diesel is a small business for which we were unable to identify gross revenues.
However, based on the fact that they reported an employee count of 17, we can reasonably esti-
mate their business volume. They produce a single engine model, so their total estimated fixed
costs are $230,000. For compliance costs to fall in the range of 1 to 3 percent of annual revenues,
total revenues would need to be between $2.5 and $7.6 million. This is a reasonable estimate
compared to other companies producing these engines with a similar number of employees.
Marine Power also sells only a single model. Comparing fixed costs (spread over three
years) to their estimated annual revenues of $10 million shows that their compliance burden is
0.8 percent of revenues.
Peninsular Diesel has annual revenues of about $2 million from three employees. They
also sell a single engine model. Their estimated compliance burden is 3.8 percent of revenues.
8.4.3.2 Small Recreational Boat Builders
We have less precise information about recreational boat builders than is available about
engine manufacturers. We have utilized several sources, including trade associations and
8-10
-------
Chapter 8: Small Business Flexibility Analysis
Internet sites when identifying entities that build or sell recreational boats. We have also worked
with an independent contractor to assist in the characterization of this segment of the industry.
Finally, we received a list of nearly 1,700 boat builders known to the U.S. Coast Guard to
produce boats using engines for propulsion. More than 90% of the companies identified so far
would be considered small businesses as defined by SBA (NAIC code 336612).
8.4.4 Results for All Small entities
For this regulation as a whole, we expect 32 small businesses to have total compliance
costs less than 1 percent of their annual revenues. We estimate that one company will have com-
pliance costs between 1 and 3 percent of revenues. Three companies will likely have compliance
costs exceeding 3 percent of revenues, but at least one will likely be able to benefit from the
relief provisions outlined below. These estimates include the costs for compliance with the
permeation standards.
8.5 Projected Reporting, Recordkeeping, and Other Compliance
Requirements of the Regulation
For any emission control program, we be sure that the regulated engines will meet the
standards. Historically, EPA programs have included provisions placing manufacturers
responsible for providing these assurances. This final rule includes testing, reporting, and record
keeping requirements. Testing requirements for some manufacturers include certification
(including deterioration testing), and production-line testing. Reporting requirements include test
data and technical data on the engines including defect reporting. Manufacturers keep records of
this information.
8.6 Steps to Minimize Significant Economic Impact on Small Entities
EPA conducted outreach to small entities and convened two Small Business Advocacy
Review Panels to obtain advice and recommendations of representatives of the small entities that
potentially would be subject to the rule's requirements. The first panel covered only marine
engines and vessels. That Panel published its report on August 29, 1999, and where appropriate,
its recommendations have been incorporated into this analysis. In a subsequent Federal Register
notice dated May 2, 2002 (67 FR 21613), EPA sought comment on applying permeation control
standards for fuel tanks and fuel hoses used on recreational vehicles. These provisions would
generally apply to those controls as well.
On May 3, 2001, EPA's Small Business Advocacy Chairperson convened a second Panel
covering all engine/vehicle categories in this rulemaking, under Section 609(b) of the Regulatory
Flexibility Act (RFA) as amended by the Small Business Regulatory Enforcement Fairness Act
of 1996 (SBREFA). In addition to the Chair, the Panel consisted of the Director of the Assess-
8-11
-------
Draft Regulatory Support Document
ment and Standards Division (ASD) within EPA's Office of Transportation and Air Quality, the
Chief Counsel for Advocacy of the Small Business Administration, and the Deputy Adminis-
trator of the Office of Information and Regulatory Affairs within the Office of Management and
Budget. As part of the SBAR process, the Panel met with small entity representatives (SERs) to
discuss the potential emission standards and, in addition to the oral comments from SERs, the
Panel solicited written input. In the months preceding the Panel process, EPA conducted
outreach with small entities from each of the five sectors as described above. On May 18, 2001,
the Panel distributed an outreach package to the SERs. On May 30 and 31, 2001, the Panel met
with SERs to hear their comments on preliminary alternatives for regulatory flexibility and
related information. The Panel also received written comments from the SERs in response to the
discussions at this meeting and the outreach materials. The Panel asked SERs to evaluate how
they would be affected under a variety of regulatory approaches, and to provide advice and
recommendations regarding early ideas for alternatives that would provide flexibility to address
their compliance burden.
SERs representing companies in each of the sectors addressed by the Panel raised con-
cerns about the potential costs of complying with the rules under development. For the most
part, their concerns were focused on two issues: (1) the difficulty (and added cost) that they
would face in complying with certification requirements associated with the standards EPA is
developing, and (2) the cost of meeting the standards themselves. SERs observed that these costs
would include the opportunity cost of deploying resources for research and development, ex-
penditures for tooling/retooling, and the added cost of new engine designs or other parts that
would need to be added to equipment in order to meet EPA emission standards. In addition, in
each category, the SERs noted that small manufacturers (and in the case of one category, small
importers) have fewer resources and are therefore less well equipped to undertake these new
activities and expenditures. Furthermore, because their product lines tend to be smaller, any
additional fixed costs must be recovered over a smaller number of units. Thus, absent any
provisions to address these issues, new emission standards are likely to impose much more
significant adverse effects on small entities than on their larger competitors.
The Panel discussed each of the issues raised in the outreach meetings and in written
comments by the SERs. The Panel agreed that EPA should consider the issues raised by the
SERs and that it would be appropriate for EPA to propose and/or request comment on various
alternative approaches to address these concerns. The Panel's key discussions centered around
the need for and most appropriate types of regulatory compliance alternatives for small busi-
nesses. The Panel considered a variety of provisions to reduce the burden of complying with new
emission standards and related requirements. Some of these provisions would apply to all
companies (e.g., averaging, banking, and trading), while others would be targeted at the unique
circumstances faced by small businesses. A complete discussion of the regulatory alternatives
recommended by the Panel can be found in the Final Panel Report. Summaries of the Panel's
recommended alternatives for each of the sectors subject to this action can be found in their
respective sections of the preamble. The vast majority of the Panel recommendations were
8-12
-------
Chapter 8: Small Business Flexibility Analysis
adopted by the Agency, and are being finalized as part of this rule, either as first-tier or second-
tier flexibilities.
First-tier flexibilities provide the greatest flexibility for many small entities. These
provisions are likely to be most valuable because they either provide more time for compliance
(e.g., additional lead time and hardship provisions) or allow for certification of engines based on
particular engine designs or certification to other EPA programs. We are adopting these pro-
visions essentially as proposed.
Second-tier flexibilities have the potential to reduce near-term and even long-term costs
once a small entity has a product it is preparing to certify. These are important in that the costs
of testing multiple engine families, testing a fraction of the production line, and developing
deterioration factors can be significant. Small businesses may also meet an emission standard on
average or generate credits for producing engines that emit at levels below the standard; these
credits can then be sold to other manufacturers for compliance or banked for use in future model
years. We are adopting these provisions essentially as proposed.
8.6.1 General Provisions
The most universal of the first-tier flexibilities are the hardship provisions. These apply to
all the categories of vehicles and engines covered by this rulemaking. The Panel recommended
that we propose two types of hardship provisions. The first type allows small businesses to pe-
tition EPA for additional lead time (e.g., up to 3 years) to comply with the standards. To qualify,
a small manufacturer must make the case that it has taken all possible business, technical, and
economic steps to comply, but that the burden of compliance costs will have a significant impact
on the company's solvency. A manufacturer must provide a compliance plan detailing when and
how it will achieve compliance with the standards. Hardship relief may include requirements for
reducing emission on an interim basis and/or purchasing and using emission credits. The length
of the hardship relief decided during review of the hardship application may be up to one year,
with the potential to extend the relief as needed. The second hardship program allows companies
to apply for hardship relief if circumstances outside their control cause the failure to comply (i.e.,
supply contract broken by parts supplier) and if the failure to sell the subject engines will have a
major impact on the company's solvency. We would, however, not grant hardship relief if
contract problems with a specific company prevent compliance for a second time.
Since equipment manufacturers who don't manufacture their own engines depend on
engine manufacturers to supply certified engines, there was a concern that these engines would
not be received in time to produce complying equipment by the date emission standards take
effect. We have heard of certified engines being available too late for equipment manufacturers
to redesign their equipment for changing engine size or performance characteristics. To address
this concern, equipment manufacturers may request up to one extra year before using certified
engines if they are not at fault and will face serious economic hardship without an extension.
8-13
-------
Draft Regulatory Support Document
A second-tier of flexibility, the averaging, banking and trading (ABT) program is also
almost universal in its applicability. Averaging programs allow a manufacturer to certify one or
more engine families at emission levels above the applicable emission standards, provided that
the increased emissions are offset by one or more engine families certified below the applicable
standards. Adding an emission-credit program containing banking and trading provisions, allow
manufacturers to generate emission credits for certifying below the standards, and bank them for
future use in their own averaging program or sell them to another entity.
ABT programs are being finalized for all categories of vehicles and engines covered by
this rule, except for Large SI engines. However, a simplified ABT variation, which we are calling
"family banking," will allow Large SI manufactures to certify an engine family early, and then to
delay certification of a comparable engine family to the Phase 1 standards. ABT provisions are
not limited to small entities, but provide another flexibility for reducing the burden on these
entities.
8.6.2 Nonroad recreational vehicles
As described above, the report of the Small Business Advocacy Review Panel addresses
the concerns of small-volume manufacturers of recreational vehicles. To identify representatives
of small businesses for this process, we used the definitions provided by the Small Business
Administration for producers and importers of motorcycles, ATVs, and snowmobiles (fewer than
500 employees for manufacturers, 100 for importers). Eleven small businesses agreed to serve as
small-entity representatives. These companies represented a cross-section of off-highway motor-
cycle, ATV, and snowmobile manufacturers, as well as importers of off-highway motorcycles
and ATVs. We proposed to adopt the provisions recommended by the panel and received
comments on the proposals. We are now finalizing the provisions below essentially as proposed,
with the modifications noted below.
As noted above, permeation standards were not part of the original NPRM for this rule,
which incorporated recommendations from the SBAR Panel process. When we reopened the
comment period on May 1, 2002 to request comment on possible approaches to regulating
permeation emissions from recreational vehicles, we did not specifically discuss small business
issues. However, it was our intent that the proposed flexibilities for exhaust emissions should
carry over to permeation controls for all three vehicle categories, to the extent that they are ap-
plicable, and we are finalizing these flexibilities for the permeation standards as well as for the
exhaust standards. Thus, we are effectively extending the work of the SBAR panel to cover the
permeation requirements in this final rule by including the flexibilities described below.
The following Panel recommendations apply to nonroad motorcycles, ATVs and snow-
mobiles. The Panel recommended that EPA restrict the flexibilities described below for off-road
motorcycle and ATV engines to those produced or imported by small entities with combined
annual sales of less than 5,000 units per model year. Because of the differences, both in numbers
8-14
-------
Chapter 8: Small Business Flexibility Analysis
and production, between small snowmobile manufacturers and small ATV/off-road motorcycle
manufacturers, the Panel recommended no maximum production limits for snowmobiles.
Additional lead time. The Panel recommended that EPA propose at least a two-year
delay, but seek comment on whether a longer time period is appropriate given the costs of com-
pliance for small businesses and the relationship between importers and their suppliers. This
would provide additional time for small-volume manufacturers to revise their manufacturing
process, and would allow importers to change their supply chain to acquire complying products.
The Panel recommended that EPA request comment on the appropriate length for a delay (lead-
time). We are finalizing a two year delay beyond the date that larger businesses must comply
with the standards for the Phase 1, and (in the case of snowmobiles) Phase 2 and Phase 3
standards.
Design-based certification. The Panel recommended that EPA propose to permit small
entities to use design certification. The Panel also recommended that EPA work with the small-
entity representatives and other members of the industry to develop appropriate criteria for such
design-based certification. We are finalizing this recommendation. Small-volume manufacturers
may use design-based certification, which allows us to issue a certificate to a small business for
the emission-performance standard based on a demonstration that engines or vehicles meet de-
sign criteria rather than by emission testing. The intent is to demonstrate that an engine using a
design similar to or superior than that being used by larger manufacturers to meet the emission
standards will ensure compliance with the standards. The demonstration must be based in part
on emission test data from engines of a similar design. Under a design-based certification
program, a manufacturer provides evidence in the application for certification that an engine or
vehicle meets the applicable standards for its useful life based on its design (e.g., the use a four-
stroke engine, advanced fuel injection, or any other particular technology or calibration). Design
criteria might include specifications for engine type, calibrations (spark timing, air /fuel ratio,
etc.), and other emission-critical features, including, if appropriate, catalysts (size, efficiency,
precious metal loading). Manufacturers submit adequate engineering and other information
about their individual designs showing that they will meet emission standards for the useful life.
Broaden engine families. The Panel recommended that EPA request comment on engine
family flexibility, in addition to conducting design-based certification emissions testing. Under
this provision, small businesses may define their engine families more broadly, putting all their
models into one engine family (or more, as needed) for certification purposes. Manufacturers
could then certify their engines using the "worst-case" configuration within the family. A small
manufacturer who might need to conduct certification emission testing, rather than pursuing
design-based certification, would likely find broadened engine families useful
Production-line testing (PLT) waiver. The Panel recommended that EPA propose to
provide small manufacturers and small importers a waiver from manufacturer production line
testing. The Panel also recommended that EPA request comment on whether limits or the scope
8-15
-------
Draft Regulatory Support Document
of this waiver are appropriate. Under PLT, manufacturers must test a small sampling of produc-
tion engines to ensure that production engines meet emission standards. We are waiving pro-
duction-line testing requirements for small manufacturers. This waiver will eliminate produc-
tion-line testing requirements for small businesses.
Use of assigned deterioration factors (DFs) for certification. The Panel recommended that
EPA propose to provide small business with the option to use assigned deterioration factors.
Small manufacturers may use DFs assigned by EPA. Rather than performing a durability
demonstration for each family for certification, manufacturers may elect to use deterioration
factors determined by us to demonstrate emission levels at the end of the useful life, thus
reducing the development and testing burden. This might also be a very useful and cost-bene-
ficial option for a small manufacturer opting to perform certification emission testing instead of
design-based certification.
Using emission standards and certification from other EPA programs. A wide array of
engines certified to other EPA programs may be used in recreational vehicles. For example,
there is a large variety of engines certified to EPA lawn and garden standards (Small SI). The
Panel recommended that EPA propose to provide small business with this flexibility through the
fifth year of the program and request comment on which of the already established standards and
programs are believed to be a useful certification option for the small businesses. We are ac-
cepting that recommendation. Manufacturers of recreational vehicles may use engines certified
to any other EPA standards for five years. Under this approach, engines certified to the Small SI
standards may be used in recreational vehicles, even though the recreational vehicle application
may not be the primary intended application for the engine. These engines would then meet the
Small SI standards and related provisions rather than those adopted in this document for recrea-
tional vehicles. Small businesses using these engines will not have to recertify them, as long as
they do not alter the engines in a way that might cause it to exceed the emission standards it was
originally certified to meet. Naturally, a small manufacturer may also use a comparable certified
engine produced by a large manufacturer, as long as the small manufacturer did not change the
engine in a way that might cause it to exceed the applicable emission standards. This provides a
reasonable degree of emission control. For example, if a manufacturer changed a certified engine
only by replacing the stock exhaust pipes with pipes of similar configuration or the stock muffler
and air intake box with a muffler and air box of similar air flow, the engine would still be eligible
for this flexibility option, subject to our review.
Averaging, banking, and trading (AST). The Panel recommended that EPA propose to
provide small business with the same ABT program flexibilities that would apply for large manu-
facturers and request comment on how the provisions could be enhanced for small business to
make them more useful. For the overall program, we are adopting corporate-average emission
standards with opportunities for banking and trading of emission credits. At first we expect the
averaging provisions to be most helpful to manufacturers with broad product lines. Small manu-
8-16
-------
Chapter 8: Small Business Flexibility Analysis
facturers and small importers with only a few models might not have as much opportunity to take
advantage of these flexibilities. However, we received comment from one small manufacturer
supporting these types of provisions as a critical component of the program. Therefore, we are
adopting corporate-average emission standards with opportunities for banking and trading of
emission credits for small manufacturers.
8.6.2.1 Off-highway motorcycles and ATVs
In addition to ABT, EPA is finalizing other provisions that are not limited to small
entities, but which could prove helpful to small businesses. Small entities could benefit from
harmonization of the ATV standards with California emission standards since only one model,
rather than two, would need to be certified to allow the product to be sold in all 50 states. Simi-
larly, the 2 gram and the optional 4 gram HC +NOx emission standards for off-highway motor-
cycles could make it less costly for small entities to comply with the standards, in addition to
their primary purposes of preventing product shortages and encouraging certification of competi-
tion bikes. The optional 4 gram HC + NOx standard in fact was suggested in the comments sub-
mitted by a small manufacturer. Finally, small ATV producers could benefit from the option of
complying with engine-based emission standards using the SAE J1088 test procedure for three
years. This flexibility could allow small entities to phase in major equipment purchases such as
chassis dynamometers necessary to be able to run the Federal Test Procedure.
As stated earlier, we are applying the flexibilities outlined above in section 8.6.2 to en-
gines produced or imported by small entities with combined off-highway motorcycle and ATV
annual sales of fewer than 5,000 units. The SBAR Panel recommended these provisions to
address the potentially significant adverse effects on small entities of an emission standard that
may require conversion to four-stroke engines. The 5,000-unit threshold is intended to provide
these flexibilities to those segments of the market where the need is likely to be greatest, and to
ensure that the flexibilities do not result in significant adverse environmental effects during the
period of additional lead-time recommended below. For example, some importers with access to
large supplies of vehicles from major overseas manufacturers could substantially increase their
market share by selling less expensive noncomplying products. In addition, we are limiting some
or all of these flexibilities to companies that are in existence or have product sales at the time we
proposed emission standards to avoid creating arbitrary opportunities in the import sector, and to
guard against the possibility of corporate reorganization, entry into the market, or other action for
the sole purpose of circumventing emission standards.
8.6.2.2 Snowmobiles
As in the case of off-highway motorcycles and ATVs, small snowmobile manufacturers
may benefit from provisions set for both large and small manufacturers. Small entities could
benefit from the pull ahead standards provision, whereby a manufacturer could certify to the
Phase 2 standards and bypass the Phase 1 standards. There are special snowmobile ABT
8-17
-------
Draft Regulatory Support Document
provisions that could also be helpful to small entities. The early credit provision, where
manufacturers could generate credits by marketing clean snowmobiles earlier than 2006, and the
elimination of FEL limits for Phase 1 are the prime examples. However, Even with these and the
broad flexibilities for all recreational vehicles described above in section 8.6.2, there may be a
situation where a small snowmobile manufacturer cannot comply. There are only a few small
snowmobile manufacturers, who sell only a few hundred sleds a year, which represents less than
0.5 percent of total annual production. Therefore, the per-unit cost of regulation may be
significantly higher for these small entities because they produce very low volumes.
Additionally, these companies do not have the design and engineering resources to tackle
compliance with emission standard requirements at the same time as large manufacturers and
tend to have limited ability to invest the capital necessary to conduct emission testing related to
research, development, and certification. Finally, some of the requirements of the snowmobile
program may be infeasible or highly impractical because some small-volume manufacturers may
have typically produced engines with unique designs or calibrations to serve niche markets (such
as mountain riding). The new snowmobile emission standards may thus impose significant eco-
nomic hardship on these few manufacturers whose market presence is small. We therefore be-
lieve significant additional flexibility for these small snowmobile manufacturers is necessary and
appropriate, as described below.
Additional lead time. The Panel recommended that EPA propose to delay the standards for small
snowmobile manufacturers by two years from the date when other manufacturers would be re-
quired to comply. The Panel also recommended that EPA propose that emission standards for
small snowmobile manufacturers be phased in over an additional two years (four years to fully
implement the standard). We are adopting these recommendations. The two-year delay noted
above in the general provisions in section 8.6.1 also applies to the timing of the standards for
snowmobiles. In addition, for small snowmobile manufacturers, the emission standards phase in
over an additional two years at a rate of 50 percent, then 100 percent. Phase 1 thus phases in at
50/100 percent in 2008/2009, Phase 2 phases in at 50/100 percent in 2012/2013, and Phase 3
phases in at 50/100 percent in 2014/2015.
Unique snowmobile engines. The Panel recommended that EPA seek comment on an additional
provision, which would allow a small snowmobile manufacturer to petition EPA for relaxed
standards for one or more engine families. The Panel also recommended that EPA allow a
provision for EPA to set an alternative standard at a level between the prescribed standard and
the baseline level until the engine family is retired or modified in such a way as to increase
emission and for the provision to be extended for up to 300 engines per year per manufacturer
would assure it is sufficiently available for those manufacturers for whom the need is greatest.
Finally, the Panel recommended that EPA seek comment on initial and deadline dates for the
submission of such petitions. We received no comments in this area, but for clarity have decided
to require at least nine months lead time by the petitioner.
In response to these recommendations and comments, we are adopting an additional pro-
8-18
-------
Chapter 8: Small Business Flexibility Analysis
vision to allow a small snowmobile manufacturer to petition us for relaxed standards for one or
more engine families. The manufacturer must justify that the engine has unique design charac-
teristics, calibration, or operating characteristics that make it atypical and infeasible or highly
impractical to meet the emission-reduction requirements, considering technology, cost, and other
factors. At our discretion, we may then set an alternative standard at a level between the prescri-
bed standard and the baseline level, which would likely apply until the family is retired or modi-
fied in a way that might alter emissions. These engines will be excluded from averaging calcula-
tions. We proposed that this provision be limited to 300 snowmobiles per year. However, we
received comment that this limit is too restrictive to be of much assistance to small businesses.
Based on this comment we are adopting a limit for this provision of 600 snowmobiles per year.
8.6.3 Nonroad industrial engines
As is the case for nonroad recreational vehicles, some of the provisions not specifically
targeted at small entities may ease the burden of compliance for them. For example, comments
from equipment manufacturers, including small entities, have made it clear that some nonroad
applications involve operation in severe environments that require the use of air-cooled engines,
which rely substantially on enrichment to provide additional cooling relative to water-cooled
engines. Severe-duty applications include concrete saws and concrete pumps, which are exposed
to high levels of concrete dust and highly abrasive particles. At the richer air-fuel ratios,
catalysts are able to reduce NOx emissions but oxidation of CO emissions is much less effective.
As a result, we are adopting less stringent emission standards for these "severe-duty" engines.
Manufacturers may request approval in identifying additional severe-duty applications subject to
these less stringent standards based on the current use of air-cooled engines or some other
engineering arguments showing that air-cooled engines are necessary for these applications. This
arrangement generally prevents these higher-emitting engines from gaining a competitive
advantage in markets that don't already use air-cooled engines.
The SBAR Panel recommended that EPA propose several possible provisions to address
concerns that the new EPA standards could potentially place small businesses at a competitive
disadvantage to larger entities in the industry. Except as noted, we have adopted the specific
Panel recommendations listed below.
Using Certification and Emissions Standards from Other EPA Programs. The Panel
made several recommendations for this provision. First, the Panel recommended that EPA
temporarily expand this arrangement to allow small numbers of constant-speed engines up to 2.5
liters (up to 30kW) to be certified to the Small SI standards. Second, the Panel further
recommended that EPA seek comment on the appropriateness of limiting the sales level of 300.
Third, the Panel recommended that EPA request comment on the anticipated cap of 30 kW on
the special treatment provisions outlined above, or whether a higher cap on power rating is
appropriate. Finally, the Panel recommended that EPA propose to allow small-volume manu-
facturers producing engines up to 30kW to certify to the small SI standards during the first 3
8-19
-------
Draft Regulatory Support Document
model years of the program. Thereafter, the standards and test procedures which could apply to
other companies at the start of the program would apply to small businesses. We are not
adopting this provision and are instead relying on the hardship provisions in the final rule, which
will allow us to accomplish the objective of the proposed provision with more flexibility.
Delay of Emission Standards. The Panel recommended that EPA propose to delay the
applicability of the long-term standards to small-volume manufacturers for three years beyond
the date at which they would generally apply to accommodate the possibility that small com-
panies need to undertake further design work to adequately optimize their designs and to allow
them to recover the costs associated with the near-term emission standards. We are also folding
this provision into the scope of the hardship provision, but believe it would be appropriate to
allow up to four years delay, depending on need.
Production Line Testing. The Panel made several recommendations for this provision.
First, the Panel recommended that EPA adopt provisions allowing more flexibility than is
available under the California Large SI program or other EPA programs generally to address the
concern that production-line testing is another area where small-volume manufacturers typically
face a difficult testing burden. Second, the Panel recommended that EPA allow small-volume
manufacturers to have a reduced testing rate if they have consistently good test results from
testing production-line engines. Finally, the Panel recommended that EPA allow small-volume
manufacturers to use alternative low-cost testing options to show that production-line engines
meet emission standards.
Deterioration Factors. The Panel recommended that EPA allow small-volume manufac-
turers to develop a deterioration factor based on available emission measurements and good
engineering judgement. We are adopting an approach that gives manufacturers wide discretion to
establish deterioration factors for Large SI engines. The general expectation is that manufac-
turers will rely on emission measurements from engines have operated for an extended period,
either in field service or in the laboratory. The manufacturer should do testing as needed to be
confident that their engines will meet emission standards under the in-use testing program. How-
ever, we intend to rely on manufacturers' technical judgment and related data (instead of results
from in-use testing) to appropriately estimate deterioration factors to protect themselves from the
risk of noncompliance.
Hardship Provision. The Panel recommended that EPA propose two types of hardship
provisions for Large SI engines. First the Panel recommended that EPA allow small businesses
to petition EPA for additional lead time (e.g., up to 3 years) to comply with the standards.
Second, the Panel recommended that EPA_allow small businesses to apply for hardship relief if
circumstances outside their control cause the failure to comply (i.e., supply contract broken by
parts supplier) and if the failure to sell the subject engines would have a major impact on the
company's solvency. We are adopting hardship provisions to address the particular concerns of
8-20
-------
Chapter 8: Small Business Flexibility Analysis
small-volume manufacturers, which generally have limited capital and engineering resources.
These hardship provisions are generally described in Section 8.6.1. For Large SI engines, we are
adopting a longer available extension of the deadline, up to three years, for meeting emission
standards for companies that qualify for special treatment under the hardship provisions. We
will, however, not extend the deadline for compliance beyond the three-year period. This
approach considers the fact that, unlike most other engine categories, qualifying small businesses
are more likely to be manufacturers designing their own products. Other types of engines more
often involve importers, which are limited more by available engine suppliers than design or
development schedules.
8.6.4 Recreational marine diesel engines
Prior to the proposal, we conducted a Small Business Advocacy Review Panel. The
panel process gathers input from small entities potentially affected by the new regulations. To
identify small businesses representatives for this process, we used the Small Business Adminis-
tration definitions for engine manufacturers and boat builders. We then contacted companies
manufacturing internal-combustion engines employing fewer than 1,000 people to be small-entity
representatives for the Panel. Companies selling or installing such engines in boats and em-
ploying fewer than 500 people were also considered small businesses for the Panel. Based on
this information, we asked 16 small businesses to serve as small-entity representatives. These
companies represented a cross-section of both gasoline and diesel engine marinizers, as well as
boat builders. With input from small-entity representatives, the Panel drafted a report with
findings and recommendations on how to reduce the potential small-business burden resulting
from this rule. The Panel's recommendation's were proposed by EPA and are now being
finalized essentially as proposed. Commenters generally supported these provisions. The
following sections describe these flexibilities.
8.6.4.1 Engine Dressers
The manufacturers involved include engine dressers, small-volume engine marinizers,
and small-volume boat builders. Many recreational marine diesel engine manufacturers modify
new, land-based engines for installation on a marine vessel. Some of the companies that modify
engines for installation in boats make no changes that might affect emissions. Their modifica-
tions may consist only of adding mounting hardware and a generator or reduction gears for
propulsion. They may involve installing a new marine cooling system that meets original
manufacturer specifications and duplicates the cooling characteristics of the land-based engine,
but with a different cooling medium (i.e., sea water). In many ways, these manufacturers are
similar to nonroad equipment manufacturers who purchase certified land-based nonroad engines
to make auxiliary engines. This simplified approach of producing an engine can more accurately
be described as dressing an engine for a particular application.
To clarify the responsibilities of engine dressers under this rule, we will exempt them
8-21
-------
Draft Regulatory Support Document
from the requirement to certify engines to emission standards, as long as they meet the following
seven conditions.
(1) The engine being dressed (the "base" engine) must be a highway, land-based nonroad,
or locomotive engine, certified pursuant to 40 CFR part 86, 40 CFR part 89, or 40 CFR
part 92, respectively, or a marine diesel engine certified pursuant to this part.
(2) The base engine's emissions, for all pollutants, must meet the otherwise applicable
recreational marine emission limits. In other words, starting in 2005, a dressed nonroad
Tier 1 engine will not qualify for this exemption, because the more stringent standards for
recreational marine diesel engines go into effect at that time.
(3) The dressing process must not involve any modifications that can change engine
emissions. We do not consider changes to the fuel system to be engine dressing, because
this equipment is integral to the combustion characteristics of an engine. However, we are
expanding the small-volume engine dresser definition to include water-cooled turbochar-
gers where the goal is to match the performance of the non-water-cooled turbocharger on
the original certified configuration. We believe this would provide more opportunities
for diesel marinizers to be excluded from certification testing if they operate as dressers
(4) All components added to the engine, including cooling systems, must comply with the
specifications provided by the engine manufacturer.
(5) The original emissions-related label must remain clearly visible on the engine.
(6) The engine dresser must notify purchasers that the marine engine is a dressed
highway, nonroad, or locomotive engine and is exempt from the requirements of
40 CFR part 94.
(7) The engine dresser must report annually to us the models that are exempt pursuant to
this provision and such other information as we deem necessary to ensure appropriate use
of the exemption.
Any engine dresser not meeting all these conditions will be considered an engine manu-
facturer and will accordingly need to certify that new engines comply with this rule's provisions
and label the engine, showing that it is available for use as a marine engine. An engine dresser
violating the above criteria might also be liable under anti-tampering provisions for any change
made to the land-based engine that affects emissions.
8-22
-------
Chapter 8: Small Business Flexibility Analysis
8.6.4.2 Small Diesel Engine Marinizers
The other small entities can be categorized as sterndrive and inboard engine marinizers,
compression-ignition recreational marine engine marinizers, and boat builders that use these
engines. We are providing additional flexibilities listed below for small-volume engine mari-
nizers. The purpose of these flexibilities is to reduce the burden on companies who cannot
distribute their fixed costs over a large number of engines. For this reason, we are defining a
small-volume engine manufacturer based on annual U.S. sales of engines, and are providing the
additional flexibilities on this basis, rather than on business size in terms of the number of em-
ployees, revenue, or other such measures. The production count we will use includes all engines
(automotive, other nonroad, etc.), not just recreational marine engines. We consider recreational
marine diesel engine manufacturers to be small volume for purposes of this provision if they
produce fewer than 1,000 internal combustion engines per year. Based on our characterization of
the industry, there is a natural break in production volumes just above the 500 engine sales mark.
The next smallest manufacturers make tens of thousands of engines. We chose 1,000 engines as
a limit because it groups together all the marinizers most needing relief, while still allowing for
reasonable sales growth.
Delay Standards for Five Years. The Panel recommended that EPA delay the standards
for five years for small businesses. We are concerned about the loss of emission control from part
of the fleet during this time, but we recognize the special needs of small-volume marinizers and
believe the added time may be necessary for these companies to comply with emission standards.
This additional time will allow small-volume marinizers to obtain and implement proven, cost-
effective emission-control technology. We are adopting the five-year delay; the standards will
take effect from 2011 to 2014 for small-volume marinizers, depending on engine size. Marini-
zers may apply this five-year delay to all or just a portion of their production. Thus they may still
sell engines that meet the standards where possible on some product lines, while delaying the
introduction of emission-control technology on other product lines. This option provides more
time for small marinizers to redesign their products, allowing time to learn from the technology
development of the rest of the industry.
Design-Based Certification The Panel recommended that EPA allow manufacturers to
certify by design and to be able to generate credits under this approach. The Panel also recom-
mended that EPA provide adequately detailed design specifications and associated emission
levels for several technology options that could be used to certify. Although we proposed this
approach, we were unable to specify any technology options for diesel engines that could be used
for a design-based certification. We requested comment on such designs and received no com-
ment. Therefore, we are not finalizing a design-based certification option. However, as noted
above, we are finalizing the engine dresser provisions and expanding these provisions to include
water-cooled turbocharging. This will essentially allow some engines to be exempt from the
standards based on design.
8-23
-------
Draft Regulatory Support Document
Broadly Defined Product Certification Families The Panel recommended that EPA take
comment on the need for broadly defined emission families and how these families should be
defined. We have established engine criteria for distinguishing between engine families which
could result in a number of engine families for a manufacturer depending on the make-up of their
product line. We are allowing small-volume marinizers to put all of their models into one
engine family (or more as necessary) for certification purposes. Marinizers would then certify
using the "worst-case" configuration. This approach is consistent with the option offered to post-
manufacture marinizers under the commercial marine regulations. This approach has the advan-
tage of minimizing certification testing, because the marinizer can use a single engine in the first
year to certify their whole product line. As with large companies, the small-volume manufac-
turers could then carry-over certification data from year to year until they change their engine
designs in a way that might significantly affect emissions.
Minimize compliance requirements. The Panel suggested we eliminate the compliance
burden on small entities to the extent possible. As a result, we proposed to eliminate production-
line and deterioration testing requirements for small-volume marinizers. We will assign a de-
terioration factor for use in calculating end-of-life emission factors for certification. The advan-
tage of this approach is to minimize compliance testing.
Streamlined certification. The Panel recommended that EPA propose to specifically
include NTE in a design-based approach. As noted above, we have concerns regarding a design-
based approach. However, we will allow small-volume marinizers to certify to the not-to-exceed
(NTE) requirements using a streamlined approach. We believe small-volume marinizers can
make a satisfactory showing that they meet NTE standards with limited test data. Once these
manufacturers test engines over the five-mode certification duty cycle (E5), they can use those or
other test points to extrapolate the results to the rest of the NTE zone. For example, an
engineering analysis may consider engine timing and fueling rate to determine how much the
engine's emissions may change at points not included in the E5 cycle. For this streamlined NTE
approach, keeping all four test modes of the E5 cycle within the NTE standards will be enough
for small-volume marinizers to certify compliance with NTE requirements, as long as there are
no significant changes in timing or fueling rate between modes.
Hardship provisions. The Panel recommended that EPA propose two types of hardship
programs for marine engine manufacturers, boat builders and fuel tank manufacturers. First, that
EPA should allow small businesses to petition EPA for additional lead time to comply with the
standards. Second, that EPA should allow small businesses to apply for hardship relief if cir-
cumstances outside their control cause the failure to comply (i.e. supply contract broken by parts
supplier) and if the failure to sell the subject fuel tanks or boats would have a major impact on
the company's solvency. The Panel also recommended that EPA work with small manufacturers
to develop these criteria and how they would be used.
We are adopting two hardship provisions for small-volume marinizers, who may apply
8-24
-------
Chapter 8: Small Business Flexibility Analysis
for this relief on an annual basis. These are essentially the same provisions noted in section 8.6.1.
First, small marinizers may petition us for additional time to comply with the standards. The
marinizer must show that it has taken all possible steps to comply but the burden of compliance
costs will have a major impact on the company's solvency. Also, if a certified base engine is
available, the marinizer must generally use this engine. We believe this provision will protect
small-volume marinizers from undue hardship due to certification burden. Also, some emission
reduction can be gained if a certified base engine becomes available.
Second, small-volume marinizers may also apply for hardship relief if circumstances
outside their control caused the failure to comply (such as a supply contract broken by parts
supplier) and if failure to sell the subject engines will have a major impact on the company's
solvency. We consider this relief mechanism to be an option of last resort. We believe this
provision will protect small-volume marinizers from circumstances outside their control. We,
however, intend to not grant hardship relief if contract problems with a specific company prevent
compliance for a second time.
Although the panel did not specify a time limit for these hardship provisions, and we are
not finalizing any such time limits, we envision these hardship provisions as transitional in na-
ture. We would expect their use to be limited to the early years of the program, in a similar time
frame as we are establishing for the recreational vehicle hardship provisions discussed above.
8.6.4.3 Small Recreational Boat Builders
The SBAR Panel Report also recommended approaches for reducing the burden on small-
volume boat builders. The recommendations were based on the concerns that even though boat
builders are not required to certify their own engines to the emission standards, they are required
to use certified engines, and may need to redesign engine compartments on some boats if engine
designs were to change significantly. EPA proposed the flexibilities recommended by the Panel
and are finalizing them as proposed.
We are adopting four options for small-volume vessel manufacturers using recreational
marine diesel engines. These options are intended to reduce the compliance burden on small
companies which are not able to distribute their fixed costs over a large number of vessels. As
proposed, we are therefore defining a small-volume boat builder as one that produces fewer than
100 boats for sale in the U.S. in one year and has fewer than 500 employees. The production
count includes all engine-powered recreational boats. These options may be used at the manu-
facturer's discretion. The options for small-volume boat builders are discussed below.
Percent-of-production delay. Manufacturers with a written request from a small-volume
boat builder and prior approval from us may produce a limited number of uncertified recreational
marine diesel engines. From 2006 through 2010, small-volume boat builders may purchase un-
certified engines to sell in boats in an amount equal to 80 percent of engine sales for one year.
8-25
-------
Draft Regulatory Support Document
For example, if the small boat builder sells 100 engines per year, a total of 80 uncertified engines
may be sold over the five-year period. This will give small boat builders an option to delay using
new engine designs for a portion of business. Engines produced under this flexibility must be
labeled accordingly so that customs inspectors know which uncertified engines can be imported.
We continue to believe this approach is appropriate and are finalizing it as proposed.
Small-volume allowance. This allowance is similar to the percent-of-production
allowance, but is designed for boat builders with very small production volumes. The only
difference with the above allowance is that the 80-percent allowance described above may be
exceeded, as long as sales do not exceed either 10 engines per year or 20 engines over five years
(2006 to 2010). This applies only to engines less than or equal to 2.5 liters per cylinder.
Existing inventory and replacement engine allowance. Small-volume boat builders may
sell their existing inventory after the implementation date of the new standards. However, no
purposeful stockpiling of uncertified engines is permitted. This provision is intended to allow
small boat builders the ability to turn over engine designs.
Hardship relief provision. Small boat builders may apply for hardship relief if circum-
stances outside their control caused the problem (for example, if a supply contract were broken
by the engine supplier) and if failure to sell the subject vessels will have a major impact on the
company's solvency. This relief allows the boat builder to use an uncertified engine and is
considered a mechanism of last resort. These hardship provisions are consistent with those
currently in place for post-manufacture marinizers of commercial marine diesel engines.
8.7 Conclusion
EPA has conducted a substantial outreach program designed to gather information as to
the effect of this final rule on small entities. This process has included two Small Business
Advocacy Review Panels, which sought out small entities that would be affected by the rule-
making and obtained advice and recommendations from them as to ways in which to minimize
the compliance burden placed upon them. We have also published an Advance Notice of Pro-
posed Rulemaking and a Notice of Proposed Rulemaking which requested comments from the
affected entities as well as from other interested parties in the public at large. Further, we have
reopened the comment period to take comments on the permeation issue raised during the initial
comment period, and have included permeation in the analysis of the effects of this rule on small
entities. We have met with a number of stakeholders, including state and environmental organi-
zations, engine manufacturers, and equipment manufacturers. From the information we have
gathered during this process, as well as information provided by contractor studies, we have
found that only 3 small entities are likely to be impacted by more than 3 percent of their sales,
and estimate that the degree of impact is likely to be further reduced by the flexibilities that are
being finalized in this rulemaking. EPA has thus determined that this final rule will not have a
significant economic impact on a substantial number of small entities.
8-26
-------
Chapter 9: Economic Impact Analysis
Chapter 9: Economic Impact Analysis
This chapter presents the economic impacts on the markets of the various vehicle
categories affected by the emissions control program. Each category of vehicles is modeled
separately. However the structure of the economic model used to estimate impacts is essentially
the same. The first section of this chapter provides a summary of the economic impact results for
each of the categories of vehicles affected by the rule. Next, we provide a general description of
the economic theory used to estimate market impacts. We then discuss the concept of fuel
efficiency gains resulting from the emissions control program and how they have been
incorporated into the economic analysis. Also addressed is the potential for product attribute
changes that may result due to the regulation. This is followed by a description of the
methodology used to develop the economic model and the supply and demand elasticity
estimates.
The remainder of the chapter takes each vehicle category in turn and describes the
baseline market characterization, the per vehicle control costs of the regulation, the future years
in which the costs are expected to be incurred, and the economic impact results generated from
the model (excluding fuel efficiency gains). We compare the future year streams of engineering
costs to the estimated economic welfare losses for each vehicle category for which the standards
apply. Economic welfare loss is equal to the sum of the loss in consumer and producer surplus
measures, excluding fuel efficiency gains. Last, we calculate a future year stream of social
costs/gains by adding fuel cost savings to economic welfare losses and compare this stream to the
stream of engineering costs of the rule (including fuel efficiency gains).
For each vehicle market, the economic model relies upon the most current year of data
available (either the year 2000 or 2001) and examines the effect of the emissions control program
as if the standards took effect in this year. The per engine control costs change over time as
different phases of the standard are implemented and the learning curve is applied (see Chapter 5
for details concerning the learning curve). It is important to note that the per engine control costs
reflect the variable cost and annual portion of capital cost associated with the regulations. To
examine the effect of these cost changes, we calculate estimated impacts using baseline year
price and output. This allows us to generate relative changes in prices and market quantities and
compute losses in consumer and producer surplus. Price and quantity data from a baseline year
are used rather than future year projections of prices and quantities because price projections for
the future time stream are not available for the various vehicle markets, though quantity
projections are.
As stated above, a future stream of welfare (or surplus) losses (excluding fuel cost
savings) is calculated by summing of the losses of consumer and producer surplus. This stream
9-1
-------
Draft Regulatory Support Document
of surplus losses, developed from baseline year price and quantity data, is compared to a
hypothetical future stream of engineering costs that are calculated by multiplying the annual
regulatory cost per vehicle in each year by the baseline year quantity. We calculate hypothetical
engineering costs holding quantity constant so that we can make a valid comparison between the
loss in surplus and engineering costs. The purpose of this comparison is to generate a surplus
loss stream that accounts for projected changes in quantity.
Through our comparison, we develop an annual ratio of surplus loss to engineering costs,
which is used to project the annual loss in surplus without fuel efficiency for the future year time
stream (this projection is made by multiplying the annual ratio of surplus loss to engineering
costs by the annual engineering costs shown in Chapter 7 for each vehicle category). The future
stream of surplus losses differs from baseline estimates due to the projected growth in vehicle
sales expected through the year 2030. Last, we calculate the future stream of annual social
costs/gains by adding fuel cost savings to the projected loss in surplus and compare this stream of
social costs/gains to the engineering costs accounting for fuel efficiency.
9.1 Summary of Economic Impact Results
An economic impact analysis of the emissions control program has been carried out to
estimate its effects on the recreational diesel marine vessel, Large SI, snowmobile, ATV, and off-
highway motorcycle markets. A summary of the economic impact results is presented in this
section to show the relative changes in price and quantity and the future year streams of
consumer and producer surplus losses (which exclude fuel cost savings), engineering costs, and
social costs/gains (which include fuel cost savings) in each vehicle market. The net present value
of the stream of surplus loss, fuel savings, and social costs/gains for each vehicle category is also
presented. Discussions of the economic theory, methodology, and full estimation of the
economic impacts are presented in the sections that follow. The results presented here for each
vehicle category summarizes the full results provided in Section 9.6 through 9.10.
As mentioned above, the relative changes in price and quantity have been estimated for
each vehicle category using the per vehicle costs as they change over future years. We calculate
these economic impacts assuming baseline market price and quantity is the same as it was in the
most current year for which data were available (year 2000 or 2001, depending on the vehicle
category).
9.1.1 Summary Results for Marine
The focus of the diesel recreational marine vessel analysis is the market for diesel inboard
cruisers. Based on discussions with industry representatives, inboard cruisers are the main type
of recreational marine vessel equipped with diesel engines. Using a year 2001 baseline average
market price of $341,945 (taken from data provided by the National Marine Manufacturers
Association) and market quantity of 8,435 inboard cruisers (taken from EPA projections based
9-2
-------
Chapter 9: Economic Impact Analysis
on data from the National Marine Manufacturers Association), the future year stream of
economic impacts were estimated for the changes in per marine vessel costs. These results are
presented in Table 9.1-1.
As the table shows, the price and quantity changes are all less than one-quarter of a
percent and by the year 2012, the relative price increase and quantity decrease are less than one-
tenth of a percent. These impacts are considered minimal. Projected surplus losses are equal to
over 99 percent of engineering costs for the diesel inboard cruiser market. The surplus losses are
highest in the year 2010 (approximately $9.6 million), which coincides with the implementation
of the second phase of the emissions control program for two of the three engine power classes
affected by the rule. They fall to their lowest level (approximately $4.9 million) in the year 2014.
They then steadily increase up through the year 2030. This trend of increased surplus losses
occurs because a larger population of engines are projected further out into the future, hence a
larger number of engines need to be controlled. Note that beyond the year 2010, loss in surplus
of the rule for recreational diesel marine vessels are in the $5 to $7 million range. For the
recreational diesel marine engine market, no fuel cost savings are projected. Therefore, the
annual stream of surplus losses equals the social costs of the regulation for this vehicle category.
9-3
-------
Draft Regulatory Support Document
Table 9.1-1
Summary Economic Impact Results for the Diesel Inboard Cruiser Market
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Cost/unit
($)
$808
$844
$844
$905
$905
$478
$464
$464
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
Change in
Price (%)*
0.12%
0.13%
0.13%
0.14%
0.14%
0.07%
0.07%
0.07%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
0.06%
Change in
Quantity (%)*
-0.18%
-0.19%
-0.19%
-0.20%
-0.20%
-0.10%
-0.10%
-0.10%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
-0.09%
Surplus Losses
($103)**
$7,795.3
$8,350.3
$8,558.2
$9,398.8
$9,621.7
$5,203.9
$5,165.6
$5,279.4
$4,952.0
$5,056.6
$5,161.4
$5,266.2
$5,371.2
$5,476.0
$5,580.8
$5,685.5
$5,790.3
$5,895.3
$6,000.1
$6,104.9
$6,209.7
$6,314.3
$6,419.0
$6,523.6
$6,628.4
Engineering
Costs ($103)
$7,806.0
$8,365.3
$8,573.8
$9,413.5
$9,637.0
$5,213.4
$5,176.7
$5,290.8
$4,958.1
$5,062.7
$5,167.7
$5,272.7
$5,377.6
$5,482.6
$5,587.6
$5,692.5
$5,797.5
$5,902.5
$6,007.4
$6,112.4
$6,217.2
$6,322.0
$6,426.9
$6,531.7
$6,636.5
Social Costs
($103)***
$7,795.3
$8,350.3
$8,558.2
$9,398.8
$9,621.7
$5,203.9
$5,165.6
$5,279.4
$4,952.0
$5,056.6
$5,161.4
$5,266.2
$5,371.2
$5,476.0
$5,580.8
$5,685.5
$5,790.3
$5,895.3
$6,000.1
$6,104.9
$6,209.7
$6,314.3
$6,419.0
$6,523.6
$6,628.4
*Percent change in price and quantity are based upon baseline market conditions for 2001
** Surplus Loss is equal to the sum of the loss in consumer surplus and producer surplus. This estimate reflects
projected growth in vehicles occurring subsequent to the baseline year of 2001.
***Social Costs are equal to the surplus losses net fuel cost savings. For this vehicle category, there are no fuel cost
savings; the future stream of surplus losses is therefore equal to the future stream of social costs. Cost estimates are
based on 2001 dollars.
9-4
-------
Chapter 9: Economic Impact Analysis
9.1.2 Summary Results for Large SI
As explained in Section 9.7, we performed an economic impact analysis for only the
forklift segment of the Large SI market. A summary of the estimated changes in price and
quantity, and the sum of consumer and producer surplus losses for forklifts is contained in Table
9.1-2. To estimate the total social costs/gains for Large SI, we use the engineering costs to
approximate the sum of consumer and producer surplus losses for Large SI engines other than
forklifts. This approach slightly overestimates the surplus losses for the category since
engineering costs are higher than surplus losses.
The baseline year for the economic analysis of the forklift market is 2000. In this year,
the forklift price is taken to be $26,380 (the price of a representative Class 5 forklift equipped
with a Large SI engine) and the market output is equal to 65,000 forklifts (taken from the Power
Systems Research (PSR) database). Based on these data, the relative changes in market price and
output are calculated, as are the annual future year streams of surplus losses, engineering costs,
and social costs/gains. Results are presented in Table 9.1-2.
Table 9.1-2
Summary Economic Impact Results for the Forklift Market
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
Cost/unit
($)
$610
$610
$493
$537
$537
$418
$418
$418
$390
$390
$390
$390
$390
$390
$390
$390
$390
$390
Change in
Price (%)*
0.75%
0.75%
0.60%
0.66%
0.66%
0.51%
0.51%
0.51%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
Change in
Quantity (%)*
-1.12%
-1.12%
-0.90%
-0.98%
-0.98%
-0.77%
-0.77%
-0.77%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
Surplus
Losses
($103)**
$43,823.1
$44,996.9
$37,410.6
$41,745.3
$42,780.3
$34,194.5
$35,002.2
$35,809.9
$34,185.7
$34,939.8
$34,693.9
$36,448.0
$37,202.1
$37,956.2
$38,710.3
$39,464.3
$40,218.4
$40,972.5
Engineering
Costs ($103)
$44,403.4
$45,592.7
$37,816.0
$42,246.7
$43,294.1
$34,471.7
$35,286.0
$36,100.3
$34,447,5
$35,207.4
$35,967.3
$36,727.2
$37,487.0
$38,246.9
$39,006.8
$39,766.6
$40,526.5
$41,286.4
Social
Costs/Gains
($103)***
$6,724.8
($29,708.1)
($75,354.6)
($108,221.4)
($143,423.9)
($187,187.5)
($220,411.8)
($248,987.1)
($263,690.9)
($273,632.9)
($282,531.5)
($290,434.8)
($297,344.7)
($303,835.7)
($309,915.5)
($315,594.1)
($320,692.6)
($325,792.0)
9-5
-------
Draft Regulatory Support Document
2022
2023
2024
2025
2026
2027
2028
2029
2030
$390
$390
$390
$390
$390
$390
$390
$390
$390
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
0.48%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
-0.72%
$41,726.6
$42,480.7
$43,234.8
$43,988.9
$44,743.0
$45,497.1
$46,251.2
$47,005.3
$47,759.4
$42,046.3
$42,806.1
$43,566.0
$44,325.9
$45,085.7
$45,845.6
$46,605.5
$47,365.4
$48,125.2
($330,892.1)
($336,421.4)
($342,011.8)
($347,604.0)
($352,536.0)
($357,472.3)
($362,412.8)
($367,356.6)
($372,304.0)
*Percent change in price and quantity are based upon baseline market conditions for 2000
** Surplus Loss is equal to the sum of the loss in consumer surplus and producer surplus. This estimate reflects
projected growth in vehicles occurring subsequent to the baseline year of 2000.
***Social Costs/Gains are equal to the surplus losses net fuel cost savings. () represents a negative cost (social
gain). Cost estimates are based upon 2000$.
The relative changes in price and quantity are slightly larger than they were for the
inboard diesel cruiser market, but they are still considered minimal. The price and quantity
changes resulting from the per forklift costs are less than 1 percent, with the exception of the
quantity change during the two years of the rule's implementation. By the year 2014, the relative
increase in market price is estimated to equal about one-half of one percent and the reduction in
quantity is equal to approximately three-quarters of one percent. As the table shows, the annual
surplus losses are approximately equal to 98 to 99 percent of engineering costs. Over the future
year time stream presented, surplus losses range from a low of $34.2 million in 2009 to a high of
$47.8 million in 2030.
An examination of the social costs/gains shows that the gains continually increase in the
future. This growth in social gains arises from the increasing fuel savings over time. The initial
growth in fuel savings can be attributed to the gradual turnover to new forklifts in the
marketplace. After this turnover, the growth in fuel savings can be credited to an increase in the
sales of forklifts. With a larger population of forklifts projected, the fuel savings are expected to
be larger. Hence the rule, as it affects the forklift market, is expected to result in larger social
gains as new forklifts enter the market and as more forklifts are purchased and operated in the
future. In 2030, the social gains of the rule for this vehicle category are just over $370 million.
Note that the figures discussed here and presented in the above table are not discounted.
Finally, to estimate the social costs/gains for the Large SI category as a whole, we can use
engineering costs as an estimate for the sum of consumer and producer surplus losses. These
estimates are contained in Table 9.1-3.
9-6
-------
Chapter 9: Economic Impact Analysis
Table 9.1-3
Surplus Losses, Fuel Efficiency Gains,
and Social Gains/Costs for Large SI Engines in 2030a
Vehicle Category
Forklifts
Other Large SI
All Large SI
Surplus Losses in
2030 ($106)
$47.8
$48.1
$95.9
Fuel Efficiency Gains in
2030 ($106)
$420.1
$138.4
$558.5
Social Gains/Costs
in2030b($106)
$372.3
$90.3
$462.6
a Figures are in 2000 dollars.
b Figures in this column exclude estimated social benefits.
0 Figure is engineering costs; see text for explanation.
d Net Present Value is calculated over the 2002 to 2030 time frame using a 3 percent discount rate.
9.1.3 Summary Results for Snowmobiles
The baseline year for the economic analysis of the snowmobile market is 2001. In this
year, the average snowmobile price is $6,360 and the market output is 140,629. These data are
provided by the International Snowmobile Manufacturing Association (ISMA).1 Based on these
data, the relative changes in market price and output are calculated, as are the annual future year
streams of surplus losses, engineering costs, and social costs or gains. Results are presented on
Table 9.1-4.
Table 9.1-4
Summary Economic Impact Results for the Snowmobile Market
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Cost/unit ($)
$35
$69
$65
$65
$185
$181
$239
$239
$202
$196
$182
Change in
Price (%)*
0.28%
0.56%
0.52%
0.52%
1.49%
1.46%
1.92%
1.92%
1.63%
1.58%
1.47%
Change in
Quantity
(%)*
-0.56%
-1.11%
-1.05%
-1.05%
-2.98%
-2.92%
-3.85%
-3.85%
-3.25%
-3.16%
-2.93%
Surplus
Losses
($103)**
$6,546.9
$13,397.7
$13,047.2
$13,316.0
$37,787.2
$37,571.1
$49,981.9
$50,697.2
$43,852.8
$43,017.6
$40,648.1
Engineering
Costs ($103)
$6,583.5
$13,546.4
$13,183.5
$13,455.2
$38,933.1
$38,685.1
$51,957.6
$52,701.2
$45.309.0
$44,402.3
$41,860.2
Social
Costs/Gains
($103)***
$6,155.4
$12,172.3
$10,577.4
$9,568.5
$28,241.7
$21,937.4
$24,916.0
$15,841.0
($1,007.1)
($11,957.9)
($24,397.9)
9-7
-------
Draft Regulatory Support Document
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
$180
$180
$180
$180
$180
$180
$180
$180
$180
$180
$180
$180
$180
$180
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
1.45%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
-2.9%
$40,543.0
$41,003.0
$41,455.4
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,903.1
$41,738.4
$42,211.9
$42,677.6
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
$43,138.5
($34,420.2)
($43,542.9)
($52,141.8)
($60,276.2)
($68,292.1)
($74,761.8)
($79,630.7)
($83,278.1)
($85,777.8)
($87,804.8)
($89,549.9)
($91,022.3)
($92,224.9)
($93,165.9)
*Percent change in price and quantity are based upon baseline market conditions for 2001.
** Surplus Loss is equal to the sum of the loss in consumer surplus and producer surplus. This estimate reflects
projected growth in vehicles occurring subsequent to the baseline year of 2001.
***Social Costs/Gains are equal to the surplus losses net fuel cost savings.
() represents a negative cost (social gain). Cost estimates are based upon 2001$
The relative increases in price expected to occur due to the rule range from 0.28 percent
to 1.92 percent and reach a steady state level of 1.45 percent in 2015. The peak occurs in 2012
when the Phase in standards are implemented and the impacts decline with the recognition of
learning curve effects. Estimated quantity changes follow a similar trend ranging from decreases
of 0.56 percent to 3.85 percent in 2010 then reaching a steady state of 2.9 percent in 2017. It is
important to note that these price quantity changes are based upon baseline 2001 snowmobile
market conditions. As the table shows, the annual surplus losses are approximately equal to 96
to 99 percent of engineering costs. Over the future year time stream presented, surplus losses
range from a low of $6.5 million in 2006 to a high of $50.7 million in 2012. These surplus losses
account for projected growth in snowmobiles sales during the period.
An examination of the social costs and gains of the snowmobile regulation shows losses
occur through 2013. Social gains begin in 2014 and continually increase in the future. This
growth in social gains arises from the increasing fuel savings over time. The growth in fuel
savings can be attributed to the gradual turnover of the snowmobile fleet to new fuel efficient
technologies and to projected increases in the sales of snowmobiles. With a larger population of
snowmobiles projected, the fuel savings are expected to be larger. Hence the rule, as it affects
the snowmobile market, is expected to result in larger social gains as new snowmobiles enter the
market and as more snowmobiles are purchased and operated in the future. In 2030, the social
gains of the rule for this vehicle category are anticipated to be just over $93.0 million. Note that
9-8
-------
Chapter 9: Economic Impact Analysis
the figures discussed here and presented in the above table are not discounted and reflect 2001$.
9.1.4 Summary Results for ATVs
The baseline year for the economic analysis of the ATV market is 2001. In this year, the
average ATV price is estimated to be $5,123 and the market output is equal to 880,000, this data
was provided by MIC. Based on these data, the relative changes in market price and output are
calculated, as are the annual future year streams of surplus losses, engineering costs, and social
costs/gains. Results are presented in Table 9.1-5.
Table 9.1-5
Summary Economic Impact Results for the ATV Market
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Cost/unit
($)
$43
$82
$78
$71
$66
$57
$53
$53
$53
$53
$51
$48
$48
$48
$48
$48
$48
$48
$48
$48
$48
$48
$48
$48
$48
Change in
Price (%)*
0.28%
0.53%
0.51%
0.46%
0.43%
0.37%
0.34%
0.34%
0.34%
0.34%
0.33%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
0.31%
Change in
Quantity (%)*
-0.56%
-1.07%
-1.02%
-0.92%
-0.86%
-0.74%
-0.69%
-0.69%
-0.69%
-0.69%
-0.66%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
-0.62%
Surplus
Losses
($103)**
$42,186.6
$80,258.8
$75,611.8
$69,529.4
$64,681.3
$55,891.6
$52,019.5
$52,019.5
$52,019.5
$52,019.5
$49,612.0
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
$47,210.3
Engineering
Costs ($103)
$42,463.9
$80,270.6
$76,518.0
$70,287.0
$65,302.2
$56,379.5
$52,441.5
$52,441.5
$52,441.5
$52,441.5
$49,999.1
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
$47,556.8
Social
Costs/Gains
($103)***
$41,252.7
$76,563.7
$68,657.0
$58,605.5
$49,541.9
$36,400.4
$28,143.4
$23,830.7
$19,705.2
$15,801.2
$9,780.7
$4,086.6
$1,360.2
($456.0)
($1,630.4)
($2,429.8)
($2,924.0)
($3,298.2)
($3,580.7)
($3,790.0)
($3,942.6)
($4,054.2)
($4,132.9)
($4,189.3)
($4,227.9)
9-9
-------
Draft Regulatory Support Document
*Percent change in price and quantity are based upon baseline market conditions for 2001
** Surplus Loss is equal to the sum of the loss in consumer surplus and producer surplus. This estimate reflects
projected growth in vehicles occurring subsequent to the baseline year of 2001.
***Social Costs/Gains are equal to the surplus losses net fuel cost savings. () represents a negative cost (social
gain). Cost estimates are based upon 2001$
The relative changes in price and quantity resulting from the ATV regulations are
considered minimal. The anticipated price change increases resulting from the per ATV costs are
0.53 percent or less. The quantity change decreases resulting from the engine modification costs
are 1 percent or less. As the table shows, the annual surplus losses are approximately equal to 98
to 99 percent of engineering costs. Over the future year time stream presented, surplus losses
range from a low of $42.2 million in 2006 to a high of $80.3 million in 2007 and reach a steady
state of $47.2 million in 2017.
An examination of the social costs/gains shows that the losses decrease beginning in 2008
and become gains in 2019 with gains continually increasing in the future through 2030. This
growth in social gains arises from the increasing fuel savings over time. The initial growth in
fuel savings can be attributed to the gradual conversion of ATVs to new fuel saving technologies
in the marketplace. After this turnover, the growth in fuel savings can be credited to an increase
in the sales of ATVs. With a larger population of ATVs projected, the fuel savings are expected
to be larger. Hence the rule, as it affects the ATV market, is expected to result in larger social
gains as new ATVs enter the market and as more ATVs are purchased and operated in the future.
In 2030, the social gains of the rule for this vehicle category are just over $4.2 million. Note that
the figures discussed here and presented in the above table are not discounted and reflect 2001$.
9.1.5 Summary Results for Off-Highway Motorcycles
The baseline year for the economic analysis of the off-highway motorcycle market is
2001. In this year, the average off-highway motorcycle price is estimated to be $2,253 and the
market sales are equal to!95,250 off-highway motorcycles. These data were provided by MIC.
Based on these data, the relative changes in market price and output are calculated, as are the
annual future year streams of surplus losses, engineering costs, and social costs/gains. Results
are presented in Table 9.1-6.
Table 9.1-6
Summary Economic Impact Results for the Off-Highway Motorcycle Market
Year
2006
2007
Cost/unit
($)
$79
$155
Change in
Price (%)*
1.11%
2.18%
Change in
Quantity (%)*
-2.23%
-4.37%
Surplus
Losses
($103)**
$15,840.8
$30,551.2
Engineering
Costs ($103)
$16,269.1
$32,215.0
Social
Costs/Gains
($103)***
$15,207.4
$28,489.4
9-10
-------
Chapter 9: Economic Impact Analysis
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
$143
$128
$117
$102
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
$99
2.01%
1.80%
1.65%
1.44%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
1.39%
-4.03%
-3.61%
-3.30%
-2.87%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
-2.79%
$28,424.3
$25,970.3
$23,984.8
$21,328.9
$20,895.5
$21,104.4
$21,315.5
$21,528.6
$21,743.9
$21,961.4
$22,181.0
$22,402.8
$22,626.8
$22,853.1
$23,081.6
$23,312.4
$23,545.6
$23,781.6
$24,018.0
$24,259.0
$24,501.6
$24,746.6
$24,994.1
$29,846.5
$27,127.3
$24,957.7
$22,079.4
$21,630.7
$21,847.0
$22,065.4
$22,508.9
$22,734.0
$22,961.4
$22,961.4
$23,191.0
$23,422.9
$23,657.1
$23,893.7
$24,132.6
$24,374.0
$24,617.7
$24,863.9
$25,112.2
$25,363.7
$25,617.3
$25,873.5
$24,658.7
$20,302.3
$16,332.2
$11,658.7
$9,242.8
$7,551.0
$5,910.8
$4,332.7
$2,893.5
$1,757.2
$1,039.5
$609.1
$325.0
$119.2
($35.0)
($133.4)
($195.4)
($240.6)
($256.0)
($252.0)
($244.9)
($214.4)
($170.7)
*Percent change in price and quantity are based upon baseline market conditions for 2001
** Surplus Loss is equal to the sum of the loss in consumer surplus and producer surplus. This estimate reflects
projected growth in vehicles occurring subsequent to the baseline year of 2001.
***Social Costs/Gains are equal to the surplus losses net fuel cost savings. () represents a negative cost (social
gain). Cost estimates are based upon 2001$
The anticipated price change increases resulting from the engine modification costs range
from 1.11 percent to 2.18 percent and reach a steady state of 1.39 percent in 2012. The quantity
change decreases resulting from the per off-highway motorcycle costs range from 2.23 percent to
4.37 percent and reach a steady state of 2.79 percent in 2012. As the table shows, the annual
surplus losses are approximately equal to 98 to 99 percent of engineering costs. Over the future
year time stream presented, surplus losses range from a low of $15.8 million in 2006 to a high of
$30.6 million in 2007.
An examination of the social costs/gains shows that the social costs reach a peak in 2007
and diminish annually through 2021. In 2020, annual social gains occur for this rule and annual
gains occur through 2030. This diminishing social cost and increasing social gain arise from the
9-11
-------
Draft Regulatory Support Document
increasing fuel savings over time. The initial growth in fuel savings can be attributed to the
gradual conversion of off-highway motorcycles new fuel saving technologies in the marketplace.
Hence the rule, as it affects the off-highway motorcycle market, is expected to result in larger
social gains as new off-highway motorcycles enter the market and as more off-highway
motorcycles are purchased and operated in the future. In 2030, the social gains of the rule for
this vehicle category are $170,700. Note that the figures discussed here and presented in the
above table are not discounted and reflect 2001$.
9.1.6 Net Present Value of Surplus Loss, Fuel Cost Savings, and Social Costs/Gains
For each of the vehicle categories, the net present value of the future streams of surplus
losses, fuel savings, and social costs/gains have been calculated. The net present values of these
future streams are calculated using a 3 percent discount rate and are calculated over the 2002 to
2030 time frame. We also show this information using a 7 percent discount rate. Table 9.1-7
presents the net present values and the surplus loss, fuel savings, and social costs/gains for the
year 2030 for each of the vehicle categories.
9-12
-------
s
O)
3
H
So
*>
V5
o
U
s
Cv
Ł
<&l
o
VI
S
~Q.
S
O
V)
4*
_S
"Ł
Ł
PH
•M
z
•a
=
o
fO
o
*l!
Ł
'a"
.2
s
1
o
•^
o
U
is
'w
o
•a
«
13
'o
o
CO
o
^.
^
g
'o
o
CO
o
^
PH
^
^H
'o
0
CO
1
fe
^H
O
>
Sz
NPVof
o
O
PH
f | |
0
fe
^
0
>
g
J2
o {
CO
Q
1
C3
i
w
o
U
CQ"
S
•a
o
to
to
O
O
TO O
O "^
^ o
Ł3
ts
^^
oo
oo
"*"
o
fN
oq
5
sf>
(N
fN
OS
%ft
OO
t~~'
Forklifts
oo
(A
^""^
o
OS
t^
^ — '
^
0
OS
^— ^
oo
0
oo
OS
oo
r~:
OS
Sf>
•^
oo
OS
6ft
^
OO
w
CO
Other Large
0?
fN
(A
^""^
^^
"*
^ — '
-^
m'
OS
^— ^
t^
OS
OS
OS
OS
p
OS
OS
fN
f&
'—\
en
%ft
OS
„_«
s
to
Snowmobile
OS
oo
fN
oo
^^
&
p
>n
~~...
IQff-Highwa:
Motorcycles
00
ts'
OS
ts"
CJQ
o\
oo
OS
oo^
sf$
*— -^
W>
•n
v_>
ts'
OO
^}
o
oo"
o
f-
r-
•T)
OS
oo
^
•
s
m"
^°.
so
w i>
s a a
resent
o o
o o
a
aa
a
s o a
" s -a
aj
W
i—' i—'
1 Cy Cy
| aj
il
W t«
W CL> (L> W t«
aa Ł Ł | aa
-------
Draft Regulatory Support Document
9.2 Economic Theory
Economic theory is based on the examination of choice behavior. As market conditions
change, producers and consumers alter their production and purchasing decisions. In essence,
this approach models the expected reallocation of society's resources in response to a regulation.
The behavioral approach explicitly models the changes in market prices and production. These
changes can be used to compute other impact variables, such as changes in producer and
consumer surplus, changes in employment, and total changes in economic welfare. EPA relies
heavily on this approach to develop impacts for the economic analysis. In order to develop a
methodological approach to examine the economic impacts of the emissions standards applied to
diesel recreational marine vessels, forklifts, and recreational vehicles, certain issues such as the
model scope and length of run for the analysis must be considered. These concepts are discussed
in detail here and can also be found in the OAQPS Economic Analysis Resource Document2.
9.2.1 Partial vs. General Equilibrium Model Scope
A partial equilibrium market model examines the effect of a regulatory action on a single
market, ignoring all other possible market interactions. Such an approach is justified in cases
where a regulation's effect is expected to be concentrated in one market sector (i.e., the effect of
the regulation in indirectly affected markets is relatively small). Other times this approach is
used because of the difficulties of acquiring data for indirectly affected markets.
A general equilibrium market model tracks the effects of a regulation in all sectors of the
economy. In this case, all inter-sectoral linkages are accounted for and examined. It is often
difficult to examine every effect of a regulation on every market. Many market models therefore
examine the most important linkages between sectors of the economy. These are generally
referred to as "general" equilibrium models or multi-market partial equilibrium models.
For the analysis of the recreational vehicles emission standards, we rely upon a partial
equilibrium market model to examine the economic impacts on the markets of each affected
vehicle category. This choice was made because most of the economic impacts are expected to
be incurred in the directly affected market and because of data availability issues.
9.2.2 Length-of-Run Considerations
In developing the partial equilibrium model for this analysis, the choices available to
producers must be considered. The choices are largely dependent upon the time horizon for
which the analysis is performed. Three benchmark time horizons are presented here: the very
short run, the long run, and the intermediate run. For this analysis, we focus on the partial
quilibrium intermediate run analysis. Though these horizons refer to different lengths of time,
they will likely differ depending upon the market in question. What defines these time horizons
is the set of options or degree of flexibility producers have to respond to changing market
9-14
-------
Chapter 9: Economic Impact Analysis
conditions.
In the very short run, all factors of production are assumed to be fixed, thus leaving the
directly affected entity with no means to respond. Within a short time horizon, regulated
producers are unable to adjust inputs or outputs due to contractual, institutional, or other factors.
In this scenario, the impacts of the regulation fall entirely on the regulated entities. Producers in
this case incur the entire regulatory burden as a one-to-one reduction in their profit. This is often
referred to as the "full-cost absorption" scenario.
In the long run, all factors of production are variable and producers can be expected to
adjust their production plans in response to changes in cost resulting from a regulation. Entry
and exit of firms into the industry is feasible. Figure 9.2-1 illustrates one example of a typical, if
somewhat simplified, long-run supply function. In this example, the supply curve is horizontal,
indicating that the marginal and average costs of production are constant with respect to output.
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.
Industry long run supply curves may exhibit constant, increasing, or decreasing returns to scale
even in perfectly competitive markets. In many industries expansion of production in the long
run may bid input prices up leading to increasing returns to scale. Constant returns to scale are
assumed for illustrative purposes.
9-15
-------
Draft Regulatory Support Document
Figure 9.2-1
Full-Cost Pass Through of Regulatory Costs
St: with regulation
Unit cost increase
S0: without regulation
Q
Market demand is represented by the standard downward-sloping curve. A constant cost
industry is assumed; equilibrium is determined by the intersection of the supply and demand
curves. In this case, the upward parallel 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 P0 to PJ). 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., Q0 to Qj). As a result, consumers incur the entire regulatory burden as
represented by the loss in consumer surplus (i.e., the area PgacPj). In the nomenclature of EIAs,
this long-run scenario is typically referred to as "full-cost pass-through."
The "intermediate" run can best be defined by what it is not. It is not the very short run
and it is not the long run. In the intermediate-run, some factors are fixed; some are variable. 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 function (which occupies the same
locus of points as the supply curve) that rises with the output rate, as shown in Figure 9.2-2.
Again, the regulation causes an inward shift in the supply function due to the increase in
production costs. The lack of resource mobility may cause profit (producer surplus) losses for
producers in the face of regulation. However, unlike the full-cost absorption scenario, producers
9-16
-------
Chapter 9: Economic Impact Analysis
are able to pass through 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 P0 to Px) that
is less than the per-unit increase in costs (fb), so that the regulatory burden is shared by producers
(net reduction in profits) and consumers (rise in price). In this case, the change in consumer
surplus is equal to PoCbPj. Producer surplus is equal to an increase in revenues on units it had
previously sold prior to the cost increase (PjCdPo) and a loss due to the costs per unit they now
face (area edba). The producer surplus is therefore equal to area edba - PjCdPg. The combined
consumer and producer surplus loss is equal to PiCdP0 - PiCbP0 - edba. This is represented by
area ecba and is referred to throughout this analysis as the surplus loss.
Figure 9.2-2
Partial-Cost Pass-Through of Regulatory Costs
a
Unit cost increase
j: with regulation
S0: without regulation
Q, Qo
Q
As mentioned earlier, the economic analysis for each vehicle category focuses on an
intermediate run approach. This is justified as the supply curve for each vehicle category shifts
inwards by the total annualized cost per vehicle, not simply variable costs. Though this rule goes
into effect over a number of years, there is a loss in economic welfare that is distributed across
producers and consumers as the rule goes into effect. The analysis presented here chooses to
focus on this loss in surplus and how it affects producers and consumers. Even if we were to
take a long-run approach, the industry supply curve for each vehicle category may not be
horizontal, (and thus represent a constant-cost industry). In fact, in many industries an
increasing-cost industry might be the norm as the prices of factors of production are bid upwards
as these industries expand.
9-17
-------
Draft Regulatory Support Document
9.3 Fuel Efficiency Gains
The main purpose of the emissions control program is to reduce emissions. However the
changes made to the engines in forklifts, snowmobiles, ATVs, and off-highway motorcycles are
also expected to result in fuel cost savings over the lifetime operation of these vehicles. Though
the prices of these vehicles are expected to increase due to the regulatory costs imposed,
consumers will spend less on fuel to operate the vehicles than they would have had the emissions
control program not been implemented. This reduced spending on fuel is a benefit to consumers.
This section qualitatively discusses the market impacts and welfare gains that may result from the
savings in fuel costs.
When recreational vehicle and large SI engine producers are required to meet the
emissions standard, they face an increase in the cost of production. This production cost increase
causes an inward shift of the supply curve equal to the regulatory cost per vehicle, shown in
Figure 9.2-2. As discussed earlier in Section 9.2.2, this leads to a loss in economic welfare equal
to the sum of the loss in producer surplus and consumer surplus. What is not accounted for in
Figure 9.2-2, however, is how fuel cost savings might affect the market equilibrium and what
surplus gain is reaped from the improved fuel efficiency. Consumers may or may not incorporate
the fuel efficiency gains into their valuation of a particular vehicle and the extent to which they
do affects the market equilibrium quantity and price, surplus changes, and social costs.
If consumers value the improvement in fuel efficiency of a particular recreational vehicle,
their demand curve for this product will shift out. The degree to which demand shifts reflects the
magnitude of the potential fuel cost savings, the costs of being informed about the savings, and
consumer time preferences. It may be the case that consumers are unaware of the fuel cost
savings, that they don't perceive them to be as large as they are, or that they heavily discount
their value. In those cases, there may be little or no shift in demand. Larger shifts in demand are
expected if consumers face low information costs and/or have a low discount rate for the future
savings in fuel costs.
For demonstration purposes, we can examine the hypothetical market for snowmobiles
depicted in Figures 9.3-1 through 9.3-3 to see how market equilibrium price and quantity (point
A) may change in response to the emissions control program and the fuel cost savings it
generates. It is important to note that this discussion applies to all vehicle categories affected by
the rule and the snowmobile market is used for explanatory purposes. This entails an
examination of the changes in both supply and demand. Looking at Figure 9.3-1, assume that the
net present value (NPV) of fuel cost savings per vehicle exceeds the regulatory control costs per
snowmobile. As described above, the increase in the costs of producing snowmobiles results in a
parallel shift inward of the supply curve. This leads to a higher price (Pj) and lower quantity (Qj)
sold, resulting in a new equilibrium point B. Now however, snowmobiles can operate using less
fuel due to the technology advancements that are adopted to reduce emissions. This change in
attribute may result in an outwards shift of the demand curve. If consumers fully value the fuel
9-18
-------
Chapter 9: Economic Impact Analysis
cost savings, demand will shift out to Dj^. The new equilibrium price (PJE) and quantity (Qre) is
represented by point C, which exceeds the market equilibrium price (P0) and quantity (Q0) before
the emissions control program was adopted (point A). If producers were certain that consumers
would fully value the fuel efficiency attribute, this change in technology may have occurred
without the implementation of the regulation. If consumers and producers view the world in this
manner, this scenario appears to be a market failure. What appears to be a win-win situation for
consumers and producers does not occur in the market place absent regulation. The risk of
producing new technology engines is borne by the producer as it is the producer that incurs the
increased production costs. In contrast, fuel efficiency gains are experienced by the consumer to
the extent the consumer is willing to pay the higher initial purchase price to gain fuel efficiency
over the useful life of the vehicle. Producers offering the new technologies only gain from the
new technology investment to the extent consumer's demand increases (demand curve shifts
outward) sufficiently to offset the increased cost of production. Thus investment in the new fuel
efficient technologies does represent a business risk for the producer and issues such as risk
aversion may enter into the decision to introduce these newer, cleaner, and fuel efficient
technologies into the marketplace absent regulatory requirements. As is depicted by the next two
scenarios, perfect information does not exist regarding consumers preferences for fuel efficiency.
Thus absent regulation, producers are making expenditures with uncertain potential for returns.
Figure 9.3-1
New Equilibrium with Full Consumer
Valuation of Fuel Efficiency
FE
Pi
PO
S'
D
Qi Qo Q-t
FE
Q
9-19
-------
Draft Regulatory Support Document
If consumers do not fully value the fuel cost savings resulting from the regulation,
demand may not shift out to D^, but instead shift to D'. As Figure 9.3-2 shows, market
equilibrium is now represented by point D where new equilibrium market price (P2) exceeds the
original market price (P0). However, the new equilibrium quantity (Q2) is lower than the original
equilibrium quantity (Q0). In such a scenario, consumers do value the attribute somewhat and are
willing to pay an increased price for the fuel efficient vehicles. However the price consumers are
willing to pay does not fully compensate the producers for the cost of making the vehicle
modification. In this scenario, it is likely that producers will be unwilling to make the engine
technology improvements absent regulation.
Another possibility is that demand may not shift at all if consumers do not perceive the
fuel cost savings associated with the new technology. In this case, Figure 9.3-3 represents the
market outcome. In this final scenario consumers do not value fuel efficiency for these vehicles
and, there is no profit motivation for producer to implement the technology changes absent
regulation.
Figure 9.3-2
New Equilibrium With Partial Consumer
Valuation of Fuel Efficiency
PI
PO
Qi Q2 Qo
Q
9-20
-------
Chapter 9: Economic Impact Analysis
Figure 9.3-3
New Equilibrium with
No Consumer Valuation of Fuel Efficiency
P,=P.
Qo
Q
It is important to recognize that the new price and quantity in the market for snowmobiles
is determined by both a shift in supply as the cost of producing snowmobiles increases and a shift
in demand to account for consumers' valuation of fuel cost savings. The potential gains to
producers from making engine technology changes that increase fuel efficiency are uncertain and
provide an explanation as to why these changes have not occurred in some recreational vehicle
markets absent regulation.
Another effect not depicted in the graphs above occurs in the fuel or gasoline market
where consumers now demand a smaller quantity of fuel to operate the fuel efficient vehicles.
Since consumers will now require less fuel to operate snowmobiles than would be required
absent the regulation, there is an inward shift in demand for gasoline. This shift in demand will
likely be so small as to not affect the price of fuel since consumers of large SI engine equipment
and recreational vehicles are a small segment of the total gasoline market. However, consumers
experience a gain equal to the NPV of the change in the quantity of fuel consumed multiplied by
the price of fuel over the lifetime of the vehicle. This is taken to equal the fuel cost savings for
each vehicle category as calculated and presented in Chapter 7. This gain occurs independently
of consumer preferences for fuel efficient vehicles. Specifically, if a consumer chooses to
purchase a more fuel efficient vehicle, the consumer will experience the gain of increased fuel
cost savings while using the product regardless of his or her preference for the fuel efficient
attributes of the vehicle.
9-21
-------
Draft Regulatory Support Document
For this analysis, we are uncertain of the size of the outward shift in demand. We
therefore do not project the price and quantity changes that occur taking fuel savings into
account. However, we do account for the fuel cost savings by subtracting it from the surplus
losses of the rule for each vehicle category over the future year time stream to generate a more
accurate assessment of the social costs/gains of the regulation. The annual fuel efficiency gains
are projected for each vehicle category in the future as described in Chapter 7 and appropriately
consider the fleet of fuel efficient vehicles operating annually through 2030 and expected vehicle
usage. The fuel efficiency gains represent the fuel cost savings consumers will experience over
the useful life of the more fuel efficient vehicle. We calculate these results for each vehicle
category analyzed. Surplus losses without fuel savings and total social costs/gains with fuel
savings are presented in the following analysis.
9.4 Potential Product Attribute Changes
It is anticipated that the air emission standards for recreational vehicles will be met by
utilizing newer, cleaner, and quieter engine technologies. Anticipated engine technology changes
are perhaps most significant for the snowmobile industry. While the ATV and off-highway
motorcycle industries have utilized 4-stroke engine technology extensively absent regulation, the
snowmobile manufacturers have been slow to introduce this technology. Current models of
ATVs are comprised by approximately 80 percent 4-stroke technologies, while the 4-stroke
technology represents approximately 55 percent of off-highway motorcycles sales. In contrast,
only nine 4-stroke snowmobile models are currently available in the marketplace, and the sales of
these vehicles are estimated to account for a small percentage of annual total snowmobile sales.
An issue has been raised as to whether the technology changes envisioned to meet the emission
standards for recreational vehicles will create attribute changes in vehicles sold. Since the engine
technology changes contemplated may be the most significant for snowmobiles, this issue is
addressed specifically for this industry in the economic analysis. The relevant question to be
addressed from an economic perspective is will snowmobiles post-regulation be perceived from
the consumer's perspective as the same product as snowmobiles pre-regulation? Further, will
any product attribute changes be adversely or positively viewed by consumers impacting
snowmobile demand post-regulation?
Particular product attribute changes alleged to negatively impact snowmobile sales relate
specifically to potential performance changes. Modifications to engines may impact the
versatility, reliability, or compactness of snowmobiles. Assertions have arisen that consumers of
snowmobiles demand high power-to-weight ratio machines and that the new engine technologies
contemplated will impair this product attribute. The issue of whether the increased costs per
engine will make entry level machines too costly for the entry level or marginal consumer have
also been claimed.
Potential product attribute changes are relevant to evaluate the economic impacts of the
rule. The economic analysis conducted for this rule postulates that the post-regulation demand
9-22
-------
Chapter 9: Economic Impact Analysis
for snowmobiles will be identical to the pre-regulation demand for snowmobiles. Consumers
will simply respond to the increased cost of an engine and based upon this increased price will
likely reduce the quantity of snowmobiles purchased (a movement along a demand curve as
opposed to a shift). If however, consumers view these product attribute changes as significant,
demand for the product may increase or decrease (demand shift inward or outward). For positive
attributes demand may increase (demand shifts outward). Under this scenario, consumers will be
willing to pay a higher price for the product because they value the enhanced or new product
attribute. If consumers view the product changes negatively, the opposite reaction occurs and
demand decreases (demand shifts inward). With decreased demand, consumers will pay a lesser
price for the product due to their perceptions that the attribute change negatively affects the value
of the product to them. If consumers view the attribute changes positively, the economic analysis
overstates market impacts. However, if consumers view the attribute changes negatively, the
economic analysis understates the market impacts of the rule. Thus it is important to account for
potential product attribute changes in order to provide a reasonable estimation of the potential
economic consequences of the rule.
The technology changes envisioned for snowmobiles will enhance the fuel efficiency of
snowmobiles. The issue of consumer potential reactions to fuel efficiency gains, a possible
positive product attribute change are discussed in Section 9.3. The 4-stroke and direct fuel
injection (dfi) technologies also offer the positive attribute of "cleaner and quieter" vehicles. The
health and environmental benefits analysis of the rule presented in Chapter 10 assesses the
public's willingness to pay for the human health and environmental benefits of these "cleaner and
quieter" technologies. A separate, but somewhat related question is whether snowmobile
consumers are willing to pay for these product attributes. It is the latter issue that is relevant for
the study of attributes.
The National Park Service (NFS) banned the use of snowmobiles for Yellowstone and
Grand Teton National Parks in January 2001. This ban on snowmobile use was based upon the
belief that snowmobile usage "adversely affects air quality, wildlife, natural soundscapes, and the
enjoyment of other visitors" to the parks.3 Both the "clean and quiet" aspects of snowmobile
attributes are reflected in the NFS ruling. The NFS service is now reviewing their ban and may
reverse the ban and allow snowmobiles in the parks with restrictions. It is possible that these
actions may impact consumer's demand for "clean and quiet" engine technologies versus the
older technologies. The outcome of the NFS activities on sales of snowmobiles and the mix of
technologies consumers will demand is an uncertainty in the economic analysis conducted for
this market and the evaluation of consumer's valuation of product attributes.
The EPA has conducted a product attribute analysis for snowmobiles to address the issue
of potential product attribute changes that may occur as a result of this regulation. Specifically,
the EPA has looked at the products currently available in the marketplace and those attributes
associated with the machines sold. Special emphasis is made to address those attributes that may
change with the regulation.
9-23
-------
Draft Regulatory Support Document
9.4.1 Technology Changes for Snowmobiles
The technology changes anticipated for the snowmobile industry to meet the standards are
addressed in Chapter 4 of this report. These standards do not dictate the use of a particular
technology, but the engineering analysis evaluates currently available technologies that will meet
the emission standards. With the Phase 2 standards for snowmobiles, 50 percent reductions in
HC and CO emissions are mandated. While snowmobile manufacturers may meet these
standards in a variety of ways, the EPA estimates 20 percent of the market will use 4-stroke
technology, 50 percent direct fuel injection technology, 20 percent modified 2-stroke engines
with pulse air, and 10 percent will use unmodified 2-stroke technologies. This technology mix is
used to calculate the engineering costs of the rule. It is relevant to note that the standards allow
for fleet emissions averaging. Thus particular manufacturers may choose the vehicles most
suited to the new technologies to meet the standards. Technologies chosen to meet the standards
are also the choice of the manufacturer. This means a manufacturer fearing the loss of
consumers for entry level machines may opt not to convert those machines to the newer
technologies.
Currently all four manufacturers of snowmobiles produce machines with the 4-stroke
technology. In its 2003 product line, Yamaha has introduced a new 4-stroke high performance
model.4 This machine represents a total redesign for the company's highest performance
machine. The Yamaha RX-1 is reported to have a horsepower rating of 145 making it one of the
most powerful snowmobiles available in the market. The redesigned machine offers a high
power-to-weight ratio that compares favorably to high performance 2 stroke competitor models.
Yamaha has redesigned the chassis and suspension of its 4-stroke model to achieve the goal of
high power to weight performance. Not only is the cleaner and quieter technology compatible
with the high performance and maneuverability, this combination has already been introduced
into the market with positive reviews.5 For several snowmobile manufacturers, the 4-stroke
technology is offered in more moderately priced, low to middle power range vehicles. For
example, the two 4-stroke machines offered for sale by Arctic Cat have estimated horsepower of
approximately 53. Thus, different manufacturers within the market place are introducing the
newer technologies using dissimilar marketing strategies. A relevant issue from the economic
impact perspective is whether snowmobile manufacturers currently in the market are in the same
competitive position to introduce these new technologies. This issue is discussed in Section 9.8
of this report.
9.4.2 Statistical Analysis of Snowmobile Product Attributes
In order to address the issue of potential product attribute changes, a statistical analysis of
product attributes for all snowmobiles in the 2003 model line is conducted. One technique
frequently used to value product attributes is the hedonic model. This model is used extensively
in the economic literature to measure consumer's willingness to pay for particular product
attributes. The hedonic model assumes that there is a continuous function relating the market
9-24
-------
Chapter 9: Economic Impact Analysis
price of a good to its constituent attributes. The assumption is made that snowmobile consumers
select a snowmobile based upon the marginal value they place on individual snowmobile
attributes and the price of those attributes. By analyzing the prices of products currently
available in the market, one may gain knowledge of those product attributes consumers value and
perhaps gain some insight as to consumer's view of potential changes in those product attributes.
An important limitation of the analysis must be addressed. The hedonic model estimated
reflects a market equilibrium relationship between price and product attributes for a single model
year. The equilibrium exists because producers of snowmobiles equate the marginal cost of
producing attributes to consumer's willingness to pay for available attributes. The hedonic model
adjusts until the marginal cost equals the marginal willingness to pay and equilibrium is
achieved. However, the regulations considered will impose a non-marginal change in the
product characteristics; therefore one cannot equate the value to consumers directly from this
model. Thus the statistical hedonic models estimated cannot be used predictively to evaluate
potential market impacts of the regulation (potential shifts in market demand). Additional
modeling is required to conduct this type of estimation. Rather, these statistical models provide
insight into implicit attribute prices for current product attributes. As stated previously in 9.3, the
market model used to assess market impacts for these regulations assumes that no shifts in
demand will occur as a result of this regulation.
9.4.2.1 Relevant Product Attributes
An assumption is made that different snowmobiles model prices may be represented by
accounting for individual product attributes. Thus, the price of a particular snowmobile model is
assumed to be a function of these characteristics. The goal of the hedonic analysis is to determine
those product attributes that account for the product price and to analyze those attributes likely to
change with regulation.
In order to complete the snowmobile hedonic analysis, an accounting of current product
characteristics and those likely to change with regulation is conducted. Product specifications
may be separated into the following categories: engine, chassis, dimensions, features, and other
attributes. Engine specifications likely to contribute positively to the price of a snowmobile
include engine type, engine size (displacement cc), number of cylinders, cooling system, ignition,
transmission, breaking system and carburetion. Chassis characteristics involve elements that
affect the maneuverability and handling of the vehicle such as suspension and shocks. The
length, width, height, weight and fuel capacity are examples of dimension attributes of
snowmobiles. Snowmobiles features include a variety of items such as electric start, reverse,
seating capacity, color and other enhancements to the vehicles. Finally the brand of snowmobile
may have some influence upon product price. Each of the previously listed product attributes
potentially influence the price of a vehicle. Those directly measured in the study are chosen
based upon the availability of data and the ability to measure these attributes. The characteristics
hypothesized to influence price for purpose of this study include engine type, engine
9-25
-------
Draft Regulatory Support Document
displacement cc, the cooling system type, carburetion type, vehicle dimensions (length, width),
fuel capacity (impacts the range a vehicle may travel on a tank of gas), seating, electric start,
reverse, and color. Color is essentially eliminated as an issue relevant for study by using
Manufacturers Suggested Retail Price (MSRP) values for the basic paint vehicles. Other product
attributes not evaluated in the study are either unavailable from publicly available sources
(snowmobile manufacturers websites), available for a subset of the companies, or difficult to
evaluate given the information provided. For example, transmission changes may occur when
using new technologies, but transmission types are difficult to measure in a quantitative or
qualitative manner as all snowmobiles have automatic transmissions.
Of these attributes, engine type, engine displacement, carburetion, cooling system, and
vehicle dimensions (length, width, and fuel tank size) may change with the regulation. Each of
these attributes potentially impact the performance of the vehicle. Engine displacement is a
measure of the power of the vehicle. In general for 2-stroke engines the greater the engine size
the greater the power. In contrast, the relationship between engine displacement and power in
the 4-stroke engine is less direct, and this phenomenon may introduce measurement error when
looking at a data set that combines 2-stroke and 4-stroke vehicles. While horsepower (hp) may
be a better measure of this attribute, hp data are not readily available for all vehicle models.
Ideally weight would be the better measure than vehicle length and width to test power-to-weight
influence upon price. However, weight data are available for only a subset of snowmobiles
offered for sales. Thus width and length proxy for the weight of the vehicle. Consumer's taste
and preferences for engine power appear to be changing over time with the demand for greater
power machines increasing. According to PSR data, the average engine displacement sized
snowmobile produced rose significantly between 1995 and 2000.6
The issue of fuel efficiency and consumers willingness to pay for increased fuel efficiency
is addressed in part with the fuel tank size variable. Gasoline mileage (miles per gallon) and
range (length in hours of a ride with a single tank of gas) information are not available for any
snowmobile models on any of the company websites. The absence of any information
concerning fuel efficiency is somewhat surprising and may perhaps indicate that snowmobile
sellers do not perceive that consumers of snowmobiles have great interest in the relative fuel
efficiency of different products. Thus informational problems exist currently for consumers to be
able to assess the fuel efficiency of products on the market. However, those products with 4-
stroke and dfi technologies are reported to have fuel savings of up to 30% over comparable
vehicles with older technologies.7 Due to the absence of published fuel efficiency data, engine
testing data provided by ISMA and from publications are used to construct a statistical
relationship between mileage and engine size.8 All data in the sample are based upon the 2-
stroke engine technology. Based upon the sample engine test data, the statistical relationship
estimated follows:
Hypothesized relationship: Gallons per hour = f (engine displacement cc)
9-26
-------
Chapter 9: Economic Impact Analysis
Fitted Equation: Gallons per hour = -1.56615 + .00920 engine displacement cc
This equation is used to estimate gallons per mile for each of the vehicles in the data set. The
gallons per hour are then converted to miles per gallon to estimate mileage for each vehicle type.
This information is used along with fuel tank size to estimate the range of each vehicle. The
descriptive statistics for data used in the model, parameter estimates, and relevant statistical
model information are displayed on Table 9.4-1. The fitted model estimates gallons per hour for
2-stroke vehicles only. It is assumed that 4-stroke vehicles and those equipped with dfi have fuel
efficiency gains over comparable 2-stroke vehicles of 25 percent. The mileage and range
estimates constructed appear to systematically underestimate the mileage experienced by the
typical snowmobile and the range for many of the vehicles appears to be understated suggesting
measurement error in these estimates. While these data are used in the analysis, potential
measurement errors in the data exist.
As indicated in the fitted equation, mileage is a function of engine size and as the engine
size increases fuel consumption increases. The implications of this relationship are quite
interesting. If consumers positively value power and power is inversely related to fuel efficiency,
product prices may indicate consumers negatively value fuel efficiency. This is an inaccurate
conclusion. We assume consumers are rational and value fuel efficiency. A more accurate
description of this phenomenon is consumers value power and are willing to pay higher prices for
larger engine sizes with greater power. Fuel efficiency declines within 2-stroke models with
larger engines.
The prices consumers pay for the attributes of power (measured as engine size
displacement) and fuel efficiency (mileage) are jointly determined. The modeling approach
taken evaluates the implicit price of the attribute engine size. It is likely that consumers currently
have a lower implicit price for engine displacement than would occur if this engine displacement
also included greater fuel efficiency. Thus it is important to recognize these attributes are
inextricably linked when consumers make purchase choices. The new technologies of dfi and 4-
stroke engines do, however, represent the potential to gain fuel efficiencies for a given level of
engine power, all other factors held constant.
9-27
-------
Draft Regulatory Support Document
Table 9.4-1
Statistical Model of Snowmobile Gas Mileage
Data Descriptive Statistics9
Sample Size =15
Variable description:
Engine Size
(displacement cc)
Gallons per hour
Mean
540.9
3.41
Standard Deviation
173.2
1.73
Statistical Model Specification:
Gallons per hours = f (engine displacement)
Gallons per hour = Pi + P2 (engine displacement) + e
Model Results:
Gallons per hour = -1.56615 + .00920 engine displacement cc
Statistical Information
Variable:
Intercept
Engine displacement
F-Value
Adjusted R Square
Parameter Estimate
-1.56615
0.00920
73.95
0.839
Standard Errors
0.60571*
0.00107**
Pr>F
< 0.0001
* Statistically significant at the 2% significance level.
** Statistically significant at the 1% significance level
9.4.2.2 Data for Hedonic Analysis
The websites of Polaris, Arctic Cat, Bombardier, and Yamaha include listings of the 2003
models available for sale.10' n'12'13 The specifications for each snowmobile model are listed on
these websites and these data are used as the data set for the study. Data are presented for the one
hundred and forty four models offered for sale in the 2003 product lines of these manufacturers.
Children's snowmobiles are excluded from the study, because the technologies used in this
application differ greatly from the typical snowmobile available for sale.
The price of a snowmobile is the dependent variable in the statistical estimation and price
must be measured to complete the hedonic analysis. MSRP are used to measure the price of
vehicles offered for sale. While the actual price paid for a snowmobile typically is a negotiated
price between the buyer and seller, only MSRP are published and readily available for models
currently offered for sale. Descriptive Statistics for snowmobile prices and product attributes are
shown on Table 9.4-2.
9-28
-------
Chapter 9: Economic Impact Analysis
Table 9.4-2
Snowmobile Price and Product Attribute Descriptive Statistics - All Vehicles 14
(Sample Size = 144)
Product Attributes
Engine Type
Engine Size
Cooling System
Length
Width
Fuel Tank Size
Seating Capacity
Electric Start
Reverse
Electronic Fuel Injection (efi)
Direct Fuel Injection
(dfi)
Brand Name
Measurement
2-stroke versus 4-stroke
cubic centimeters
air cooled or liquid cooled
inches
inches
gallons
1 or 2 person vehicle
standard equipment or optional
standard equipment or optional
Included or not included
Included or not included
Polaris, Arctic Cat,
Bombardier, or Yamaha
Mean Value
Dummy Variable
0 = 2-stroke
1 = 4-stroke
(9 4-stroke)
642
Dummy Variable
0 = air cooled
1 = liquid cooled
(114 liquid cooled)
116.6
46.6
11.3
Dummy Variable
0 = 2 person
1 = 1 person
(106 1 -person)
Dummy Variable
0 = option
1 = standard
(55 standard)
Dummy Variable
0 = optional
1 = standard
(81 standard)
Dummy Variable
0 = no efi
l=efi
(27 efi)
Dummy Variable
0 = no dfi
l=dfi
(6 dfi)
Dummy Variables
1 = particular brand
Standard
Deviation
N/A
144
N/A
6.7
1.9
1
N/A
N/A
N/A
N/A
N/A
N/A
9-29
-------
Draft Regulatory Support Document
Mileage
Range
Dependent Variable:
Snowmobile price
Miles per gallon
Miles traveled on a tank of gas
Manufacturers suggested retail
price
6.2
69.3
$7,291
2.7
26.3
$1,411
Since the 4-stroke engine represents a significant technical departure from the 2-stroke
engines, alternative models are estimated for the 2-stroke and 4-stroke models exclusively. The
descriptive statistics for those variables subject to quantitative estimates for the 4-stroke and 2-
stroke models are shown on Tables 9.4-3 and 9.4-4, respectively. In general, qualitative variables
measured by dummy variables are measured as depicted for all vehicles. Some features that are
measured using dummy variables are not applicable for the 4-stroke technology. For example,
all 4-stroke engines are liquid cooled and have electric start as standard features. Dfi technology
is available exclusively on 2-stroke models. Horsepower data are available for all nine 4-stroke
models.
Table 9.4-3
Snowmobile Price and Product Attribute Descriptive Statistics15
Four-Stroke Models Only (Sample Size =9)
Product Attributes
Engine Size
HP
Length
Width
Fuel Tank Size
Brand Name
Mileage
Range
Dependent Variable:
Snowmobile price
Measurement
cubic centimeters
number
inches
inches
gallons
Polaris, Arctic Cat,
Bombardier, or Yamaha
Miles per gallon
Miles traveled on a tank of gas
Manufacturers suggested retail
price
Mean Value
872
88.6
116.6
47.3
11.1
Dummy Variables
1 = particular brand
4.9
55.4
$8,316
Standard
Deviation
150.7
44
8.5
1.4
1.1
N/A
1.3
20.7
$687
9-30
-------
Chapter 9: Economic Impact Analysis
Table 9.4-4. Snowmobile Price and Product Attribute Descriptive Statistics16
Two-Stroke Models Only (Sample Size = 135)
Product Attributes
Engine Size
Length
Width
Fuel Tank Size
Brand Name
Mileage
Range
Dependent Variable:
Snowmobile price
Measurement
cubic centimeters
inches
inches
gallons
Polaris, Arctic Cat,
Bombardier, or Yamaha
Miles per gallon
Miles traveled on a tank of
gas
Manufacturers suggested
retail price
Mean Value
626.4
116.5
46.6
11.2
Dummy Variables
1 = particular brand
6.3
69.9
$7,213
Standard
Deviation
130.7
6.5
1.9
0.9
N/A
2.8
26.3
$1,423
9.4.2.3 Statistical Model Results
This section presents the results of statistical estimations including results of statistical
tests. The statistical package, SAS 8.2 for Windows was used to generate all statistical results.
Various model specifications were estimated including log-log, log-linear and linear models.
Generally, the log-log model specification provided the best statistical fit. In this model, all
variables are transformed to natural logs except the dummy variables. Numerous model
variations were estimated. In nearly all model specifications, the variables electric start,
electronic fuel injection, brand name, length, fuel tank size, and electric start are consistently not
statistically significant. Since the range and mileage variables are a function of the engine size,
these variables are highly correlated. For this reason, model runs were conducted with engine
size, range or mileage exclusively. The 4-stroke parameter is correlated with engine size
variable. When the model is specified using both of the parameters, the 4-stroke variable appears
to have a negative coefficient and to be statistically significant. When the model is estimated
with the 4-stroke variable and excludes engine size, the parameter estimates are not significantly
different from zero. Thus the fitted model excludes 4-stroke technology from the estimation. It
is possible that a dummy variable is not an adequate method of capturing the attributes associated
with the technology. Given this results a hedonic models of 2-stroke and 4-stoke models only are
estimated. The estimated hedonic function for the full model using engine size follows:
9-31
-------
Draft Regulatory Support Document
log MSRP = 8.2419 + 0.5821 log ( engine displacement cc) + 0.8561 log (width)
+ 0.2397 cooling - 0.0685 seat + 0.0495 reverse + 0.1066 dfi.
All parameter estimates are significant at a 1 percent significance level. Relevant statistical
model results are shown on Table 9.4-5.
Table 9.4-5
Full Model Statistical Results Using Engine Displacement
Variable
Intercept
log (engine displacement cc)
log (width)
cool
seat
reverse
dfi
Parameter Estimate
8.2419
0.5821
0.8561
0.2397
-0.0685
0.0495
0.1066
Standard Error*
0.6987
0.0362
0.1713
0.0223
0.0159
0.0143
0.0343
F Value 157.28*
Adjusted R-Square 0.8677
: All parameter estimates are statistically significant at a 1% significance level.
The model is re-estimated using the same specifications and variables shown in Table 9.4-5, but
replacing engine size with a mileage variable and in a subsequent run with the range variable.
The models and parameter estimates remain statistically significant. The mileage variable and
range variable have negative signs as previously postulated and are statistically significant in
each of the runs.
Based upon the statistical results, one may conclude that the relative prices (as measured
by MSRP) are higher for vehicles with larger engine sizes, greater width, liquid cooling systems,
reverse, and dfi. Alternatively, one-seating capacity machines are priced generally lower than
two-seat machines. In the alternative model specifications, the mileage and range variables have
negative signs and are statistically significant. This result may be interpreted to mean that
consumers value power even when greater power translates into less fuel efficiency.
The full data set is split into a 4-stroke data set and a 2-stroke data set to assess the model
differences with these two technologies. The model estimation results for the 2-stroke
technology are as follows:
9-32
-------
Chapter 9: Economic Impact Analysis
Log (MSRP) = 7.5689 + 0.6461 log (engine displacement cc) + 0.7847 log (width)
+ 0.2260 cool + 0.0626 reverse -0.0722 reverse + 0.0906 dfi
Statistical results are shown in Table 9.4-6. In general, the results of this run differ little from the
full model. This is not surprising since 135 observations of the full data set are represented in the
2-stroke model specification. Thus the conclusions for the full model apply to the two-stroke
technology.
Table 9.4-6
Two-Stroke Model Statistical Results Using Engine Displacement
Variable
Intercept
log (engine displacement cc)
log (width)
cool
reverse
seat
dfi
Parameter Estimate
7.5689
0.6461
0.7847
0.226
0.0626
-0.0722
0.0906
Standard Error*
0.6984
0.0386
0.1683
0.0218
0.0143
0.0143
0.0333
F Value 165.49*
Adjusted R-Square 0.8805
* All parameter estimates are statistically significant at a 1% significance level.
Only nine 4-stroke models are currently available for sale. Thus the sample size is quite small.
In general, only engine size or horsepower are statistically significant. Horsepower provides a
stronger statistical relationship to MSRP and the model results are shown below:
log (MSRP) = 8.3330 + 0.1577 log (hp)
Model results are shown in Table 9.4-7.
9-33
-------
Draft Regulatory Support Document
Table 9.4-7
Four-Stroke Model Statistical Results Using Engine Horsepower
Variable
Intercept
log (horsepower)
Parameter Estimate
8.333
0.1577
Standard Error *
0.1064
0.0242
F Value 42.53*
Adjusted R-Square 0.8941
* All parameter estimates are statistically significant at a 1% significance level.
The model results tend to provide confirmation that higher powered (greater hp) four-
stroke machines are higher priced that than lower powered 4-stroke machines.
In general, the statistical results from all model runs tend to indicate that higher MSRP
exist in the current snowmobile market for power (larger engine size or hp), wider machines,
liquid cooling, reverse, and dfi product attributes. One-seat machines, all other factors held
constant, are lower priced than two-seat machines. The statistical results also indicate prices are
higher for vehicles equipped with the dfi technology.
The statistical results indicate that fuel efficiency is inversely related with engine size.
Since prices are relatively higher for more powerful machines, this translates to lower fuel
efficiency. This phenomenon is related to the two-stroke technology. This does not likely reflect
a negative view of fuel efficiency so much as a positive view of greater power. While consumers
of 4-stroke models also are willing to pay higher prices for greater power, greater fuel efficiency
is an intrinsic attribute of the 4-stroke technology. The model results are not satisfying with
regard to the 4-stroke technology. This is likely due to the fact that the dummy variable does not
adequately capture the attributes associated with the 4-stroke technology and may also be due to
the relatively small number of models with this technology.
9.4.3 Anecdotal Pricing Information For Snowmobiles
The statistical analysis is unsuccessful at identifying product price differentials for the 4-
stroke technology versus 2-stroke. For this reason, a model by model comparison is conducted
of the 4-stroke snowmobile models that are similar except for engine type. The MSRP
differential typically ranges from $500 to $600 for the 4-stroke model when compared to the 2-
stroke comparable model.17 The prices consumers actually pay for these comparison vehicles are
ultimately dependent upon a negotiated price rather than MSRP.
9-34
-------
Chapter 9: Economic Impact Analysis
9.4.4 Uncertainties and Limitations of the Attribute Study
The statistical uncertainties of the attribute study are presented in the discussions of the
models estimated. In additional to the statistical uncertainties, other uncertainties exist. The
outcome of NFS issues with snowmobile usage in national parks is an uncertainty that cannot be
adequately addressed in the analysis. To the extent that NFS actions, spur demand for "cleaner
and quieter" snowmobiles, demand for the new technologies may increase. However, the overall
impact of a ban on snowmobile usage in the parks is a recognized uncertainty of the economic
impact analysis conducted for this rule.
The hedonic model estimated reflects a market equilibrium relationship between price
and attributes for a single model year. The equilibrium exists because producers of snowmobiles
equate the marginal cost of producing attributes to consumer's willingness to pay for available
attributes. The hedonic model adjusts until the marginal cost equals the marginal willingness to
pay and equilibrium is achieved. However, the regulations considered will impose a non-
marginal change in the product characteristics; therefore one cannot equate the value to
consumers directly from this model. Additional modeling is required to conduct this type of
estimation.
9.4.5 Conclusions
Two questions are posed at the beginning of this analysis regarding potential product
attribute changes. Those questions are: will snowmobiles post-regulation be perceived from the
consumer's perspective as the same product as snowmobiles pre-regulation and will product
attribute changes be adversely or positively viewed by consumers impacting snowmobile demand
post-regulation? The answer to the first question is that the technology changes envisioned by
the rule do alter the attributes of snowmobiles such that the typical consumers of snowmobiles
post-regulation will view these products as different from the pre-regulation snowmobile. Two
qualifiers to this conclusion exists. The first is that these technologies are already available in the
market place. The regulation will simply encourage the proliferation of these new technologies
throughout the snowmobile market. The second is a mix of technologies will exist that include
older technologies. Thus consumers of the older technology machines will not likely perceive
product changes post regulation.
With regard to the second question, consumer demand may change as a result of these
altered product attributes. However, quantification of any demand changes is not possible with
the data evaluated. The negative aspects of product changes alleged by some involve potential
degradation of the power-to-weight ratio for high performance machines. Yamaha's introduction
of its new high performance 4-stroke machine is evidence that the "clean and quiet" technologies
can coexist with high power-to-weight ratios. Thus consumers will be able to obtain "clean and
quiet" high powered snowmobiles. The question then becomes are consumers willing to pay
higher prices for the new attributes of cleaner, quieter, greater fuel efficiency, and other
9-35
-------
Draft Regulatory Support Document
performance attributes of snowmobiles equipped with dfi or 4-stroke engines. The statistical
analysis provides evidence that MSRP is higher for vehicles equipped with dfi, all other factors
held constant. A comparison of the suggested MSRP of comparable 4-stroke and 2-stroke
vehicles reflects higher prices for the 4-stroke engine vehicles currently offered in the market of
approximately $500 to $600. Thus snowmobile manufacturer's recommend higher prices for the
newer technologies. This recommendation reflects the belief that certain consumers will value
the bundle of product attributes of the new cleaner quieter machines and be willing to pay a
premium for these attributes. The actual price differences paid for new versus old technology
vehicles is determined by those prices negotiated in the market. Further, the increased price may
reflect an increased cost of production and not necessarily translate into additional profits for the
manufacturer.
With regard to the issue of whether entry level consumers will leave the market, fleet
emissions averaging will allow producers to use older less costly technologies on entry level
machines to avoid sales losses for this segment of the market.
9.5 Methodology
For the economic impact analysis of the effects of the emissions control program, we rely
upon a national-level partial equilibrium market model. Inputs to this model include baseline
market price, market output (domestic and imported quantities), and estimates of price elasticity
of supply and demand. Price elasticities measure the responsiveness of quantity demanded and
supplied to changes in price. This section describes the conceptual model used to generate the
economic impacts and it provides the methodology and data inputs used to develop estimates of
supply and demand price elasticities for each vehicle category.
9.5.1 Conceptual Model
The regulatory compliance costs provide an exogenous shock to the model with the per
unit total compliance costs (c) resulting in a shift of the domestic supply curve (S0 to St in Figure
9.2-2 above). This shift, expressed as the cost increase per vehicle, is based on the cost
information presented in Chapter 5 (generally, the regulatory cost per engine is taken to equal the
cost per vehicle). The model equations that respond to this exogenous shock are described
below.
The change in domestic supply (dq°) due to the imposition of the regulation will depend
upon the typical supply response to a price increase and the change in the "net" price of a given
vehicle (i.e., dP - c) so that
(Eq.9-1)
9-36
-------
Chapter 9: Economic Impact Analysis
where ^D is the domestic supply elasticity. Supply elasticities have been estimated for each of
the vehicle categories affected by the emissions standards and a description of the estimation
procedure used is provided below.
International trade is included through the specification of an equation to characterize
imports to the U.S. Thus, the change in imports from these foreign countries is included through
the following equation:
p
(dP - c) (Eq. 9-2)
where c^1 is the import supply elasticity. Data to estimate import supply elasticities for the
various vehicle categories were not available. For the economic impact analysis, the value of the
import supply elasticity is assumed to equal the value of the domestic supply elasticity.
Next, the change in market supply must equal the change in the quantity of individual
suppliers both domestic and foreign, i.e.,
dQ = dq D + dql (Eq. 9-3)
where dcf is the change in domestic supply and dcf is the change in imports.
Lastly, the market demand condition must hold, i.e.,
dP (Eq. 9-4)
where 77 is the market demand elasticity. The economic model relies upon demand elasticities
that have been estimated or found in the economics literature for the various vehicle categories.
Estimation procedures for demand elasticity are discussed below.
Equations 9-1 through 9-4 form four linear equations with four unknowns (dcf, dcf, dQ,
and dP) that can be solved using linear algebra, i.e.,
b = A V
where b is the vector containing the four unknowns (dcf, dcf', dQ, and dP), A"1 is the inverse of
A, a 4x4 matrix, and c is the vector (c, c, 0, 0). Using this model, we develop our national-level
9-37
-------
Draft Regulatory Support Document
economic impacts resulting from the rule. The full system of equations (Ab = c) is as follows:
- -V Kr 0
-1
- ^r - 0
-1
-1 71^
P.
dqd
dq'
dQ
[dP
=
c
c
0
0
(Eq. 9-5)
9.5.2 Price Elasticity Estimation
As discussed above, demand and supply elasticities are crucial components of the partial
equilibrium model used to quantify the economic impacts of the emission standards. The price
elasticity of demand is a measure of the sensitivity of buyers of a product to a change in price of
the product. The price elasticity of demand represents the percentage change in the quantity
demanded resulting from each 1 percent change in the price of the product. The price elasticity
of supply is a measure of the responsiveness of producers to changes in the price of a product.
The price elasticity of supply indicates the percentage change in the quantity supplied of a
product resulting from each 1 percent change in the price of the product.
This section presents the analytical approach employed to estimate the demand and
supply price elasticities used in the partial equilibrium analysis for each vehicle category. As
discussed below, demand and supply elasticity estimates used in the market model are either
estimated, assumed, or retrieved from previous studies that have carried out these estimations. In
the case of recreational diesel marine vessels, a demand elasticity measure was available from a
previous study, but the supply elasticity was estimated. For forklifts, both supply and demand
elasticities were estimated. Because of data limitations, EPA's estimates of demand elasticity for
the forklift model are not considered robust. Two estimates were generated; one was not
significant while the other was significant but not of reasonable size. The economic impact
analysis therefore relies upon an assumed price elasticity of demand for forklifts based on the
results generated for this vehicle category. A sensitivity analysis is included in an appendix to
show the economic impacts of the rule on the forklift market when the large estimate of demand
elasticity is used. For the snowmobile, ATV, and OHM markets, attempts were made at
9-38
-------
Chapter 9: Economic Impact Analysis
econometric estimation of the price elasticity of demand. These attempts were unsuccessful as
was a search to find these data in the literature. In lieu of estimates specific to the snowmobile,
ATV and the OHM markets, an estimate of the price elasticity of demand for recreational boats
obtained from a study are used to estimated market impacts. This value is assumed to be a
reasonable estimate of the price elasticity of demand for the snowmobile, ATV and OHM
markets. The uncertainties involved in this estimate are acknowledged. A sensitivity analysis is
included in the Appendix to Chapter 9 to recognize the uncertainties associated with this
estimate. The price elasticity of supply is estimated for the snowmobile and OHM markets.
Attempts to estimate this value for the ATV market were unsuccessful. The price elasticity of
supply estimate generated for the OHM market is assumed to be a reasonable estimate of this
value for the ATV market. Sensitivity analyses are presented in the appendix to this chapter to
evaluate the uncertainties involved in these estimates. A summary of the price elasticity of
demand and supply used in the study for each vehicle type are summarized in Table 9-5.0 shown
below.
Table 9-5.0 Summary of Price Elasticity of Demand and Supply
Used in the Market Analyses
Market
Inboard Cruisers
Forklifts
Snowmobiles
ATVs
Off-highway motorcycles
Price Elasticity of Demand
-1.41
-1.52
-2.03
-2.03
-2.03
Price Elasticity of Supply
1.62
0.72
2.12
1.04
0.92
1 Raboy, David. G. 1987. Results of an Economic Analysis of Proposed Excise Taxes on Boats.
Washington, D.C: Fatten, Boggs, and Blow. Prepared for the National Marine Manufacturing Association. Docket
A-2000-01, Document IV-A-129.
2 Assumed value.
3 Econometrically estimated.
4 Assumed value based upon the price elasticity of demand estimate for recreational boats in the Raboy study listed
above.
5 Assumed value based upon the price elasticity of supply estimate for off-highway motorcycles.
9.5.2.1 Price Elasticity Estimation for Marine
Demand Elasticity
The economic model developed for the CI recreational marine vessel market concentrates
solely on the inboard cruiser market. This is the segment of the recreational marine vessel
market which relies upon diesel engines more than any other. Fortunately, a previously estimated
9-39
-------
Draft Regulatory Support Document
price elasticity of demand for the inboard cruiser market is available18. For this reason, demand
elasticity was not estimated. The previously estimated value that is used in the economic model
is-1.44.
Supply Elasticity
Published sources of the price elasticity of inboard marine cruisers were not readily
available. Therefore, an econometric analysis of the price elasticity of supply for boat
manufacturing was conducted, assuming that this estimate is representative of the supply
elasticity for the inboard cruiser market. The approach used to estimate the supply elasticity
makes use of the production function. The methodology of deriving a supply elasticity from an
estimated production function will be briefly discussed with the industry production function
defined as follows:
Qs = f(L,K,M,t] (Eq.9-6)
where:
Qs = output or production
L = the labor input, or number of labor hours,
K = real capital stock,
M = the material inputs, and
t = a time variable to reflect technology changes.
In a 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 for a boat producer is as follows:
TC =h(C,K,t,Qs) (Eq. 9-7)
where:
TC = the total cost of production, and
C = the cost of production (including cost of materials and labor).
All other variables have been previously defined.
This methodology assumes that capital stock is fixed, or a sunk cost of production. The
assumption of a fixed capital stock may be viewed as a short-run modeling assumption. This
assumption is consistent with the objective of modeling the adjustment of supply to price
changes after implementation of controls. Firms will make economic decisions that consider
those costs of production that are discretionary or avoidable. These avoidable costs include
production costs, such as the costs associated with labor and materials. In contrast, costs
associated with existing capital are not avoidable or discretionary. Differentiating the total cost
9-40
-------
Chapter 9: Economic Impact Analysis
function with respect to Qs derives the following marginal cost function:
MC =h'(C,K,t,Qs) (Eq. 9-8)
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
equates market price, P, to the marginal cost of production. Setting the price equal to the
preceding marginal cost function and solving for Qs yields the following implied supply function:
Qs =(P,PL,PM,K,t) (Eq.9-9)
where:
P = the price of recreational marine vessels,
PL = the price of labor, and
PM = the price of materials input.
All other variables have been previously defined.
An explicit functional form of the production function may be assumed to facilitate
estimation of the model. For this analysis, the Cobb-Douglas, or multiplicative form, of the
production function is postulated. The Cobb-Douglas production function has the convenient
property of yielding constant elasticity measures. The functional form of the production function
becomes:
Qt = AK"f-t^L^M"M (Eq. 9-10)
where:
Qt = output or production in year t,
Kt = the real capital stock in year t,
Lt = the quantity of labor hours used in year t,
Mt = the material inputs in year t, and
A, aK, aL, aM, A, = parameters to be estimated by the model.
This equation can be written in linear form by taking the natural logarithms of both sides
of the equation. Linear regression techniques may then be applied. Using the approach
described, the implied supply function may be derived as:
Ing = j80 + /In P + /?, In Ł + /32 In Pl + /33 In PM + /34 In/ (Eq. 9-11)
where:
9-41
-------
Draft Regulatory Support Document
PL = the factor price of the labor input,
PM = the factor price of the material input, and
K = fixed real capital.
The /3t and y coefficients are functions of the «;., the coefficients of the production function. The
supply elasticity, y, is equal to the following:
= al+aM
1-«Z-«M
It is necessary to place some restrictions on the estimated coefficients of the production
function in order to have well-defined supply function coefficients. The sum of the coefficients
for labor and materials should be less than one. Coefficient values for aL and aM that equal to
one result in a price elasticity of supply that is undefined, and values greater than one result in
negative supply elasticity measures. For these reasons, the production function is estimated with
the restriction that the sum of the coefficients for the inputs equal one. This is analogous to
assuming that the boat manufacturing industry exhibits constant returns to scale, or is a long-run
constant cost industry. This assumption seems reasonable on an a priori basis and is not
inconsistent with the data.
The estimated model reflects the production function for boats, using annual time series
data for the years from 1958 through 1999. The following model was estimated econometrically,
using real values of capital stock, production wages, and material inputs:
Ing, = In A + aK lnŁ, + Alnf + aL Ini, + aM In Mt (Eq. 9-13)
where each of the variables and coefficients have been previously defined.
The data inputs used to estimate the supply elasticity are enumerated in Table 9.5-1. This
table contains a list of the variables included in the model and the units of measure. The data for
the price elasticity of supply estimation model includes: the value of domestic shipments in
millions of dollars; the price index for the value of domestic shipments (the value of domestic
shipments deflated by the price index represents the quantity variable which is the dependent
variable in the analysis); a technology time variable; production wages in millions of dollars; the
implicit GDP deflator (used to deflate production wages), the material inputs in millions of
dollars; the price index for value of materials; investment in millions of dollars; the price index
for investment; and real net capital stock in millions of dollars.
9-42
-------
Chapter 9: Economic Impact Analysis
Table 9.5-1
Data Inputs for the Estimation of
Supply Elasticity for the Boat Building Industry19 2021222324
Variable Unit of Measure
1. Value of Shipments for the Boat Building Industry (SIC 3732) millions of $
2. Price Index of Shipments for the Boat Building Industry (SIC 3732) index
3. Time trend
4. Production Worker Wages millions of $
5. Implicit GDP Deflator index
6. Cost of Material Inputs millions of $
7. Price Index of Material Inputs index
8. Investment millions of $
9. Price Index of Investment index
10. Real Capital Stock millions of 1987$
Data to estimate the production function exclusively for inboard cruisers were largely
unavailable; therefore, data for SIC code 3732 (Boat Building) is utilized for each of the
variables previously enumerated with the exception of the time variable. All data for the supply
elasticity estimation were retrieved from the National Bureau of Economic Research-Center for
Economic Studies (NBER-CES) Productivity Database and the U.S. Census Bureau's Annual
Survey of Manufactures (ASM), with the exception of the technology time trend, the implicit
GDP deflator, the price index for investment for SIC 3732 for the years 1997 through 1999, the
price indices of shipments and material inputs for SIC 3732 for the years 1998 and 1999, and real
capital stock for the years 1998 and 1999 (these data for real capital stock were not available).
These variables (except the time trend and real capital stock for 1998 and 1999), were retrieved
from the Bureau of Economic Analysis (BEA).
More specifically, the price index of shipments for 1998 and 1999 was retrieved from the
BEA's Shipments of Manufacturing Industries. Note that since a price index of material inputs
for SIC 3732 was not available beyond 1997, we relied upon a general price index for
intermediate materials from BEA's Survey of Current Business. A price index for investment for
SIC 3732 was also not available beyond 1996, so a general price index for capital equipment was
used for the years 1997 - 1999 from the same source. Last, real capital stock for the years 1998
and 1999 was calculated using the following formula:
real cap stockj = real cap stock^ + real investment; - depreciation rate*real cap stock^ (Eq. 9-14)
where i = 1998, 1999. The depreciation rate for capital for SIC 3732 was taken as the average
depreciation rate over the last 10 years for which investment and capital stock data were
available (1987 - 1996).
9-43
-------
Draft Regulatory Support Document
The capital stock variable was the most difficult variable to quantify for use in the
econometric model. Ideally, this variable should represent the economic value of the capital
stock actually used by each facility to produce boats for each year of the study. The most
reasonable data for this variable would be the number of machine hours actually used to produce
boats each year. These data are unavailable. In lieu of machine hours data, the dollar value of
net capital stock in constant 1987 prices, or real net capital stock, is used as a proxy for this
variable. However, these data are imperfect because they represent accounting valuations of
capital stock rather than economic valuations. This aberration is not easily remedied, but is
generally considered unavoidable in most studies of this kind.
SAS Release 8.2 for Windows was used to develop econometric estimates of the price
elasticity of supply for the boat manufacturing industry. A restricted least squares estimator was
used to estimate the coefficients of the production function model. A log-linear specification
was estimated with the sum of the at restricted to unity. This procedure is consistent with the
assumption of constant returns to scale. The model was further adjusted to correct for first-order
serial correlation using the Yule-Walker estimation method. The results of the estimated model
are presented in Table 9.5-2 with p-values listed in parentheses below each coefficient estimate.
Table 9.5-2
Estimated Supply Model Coefficients for the Boat Building Industry
Variables
In(Time) (t)
ln(Real Capital Stock) (K)
ln(Real Production Wages) (L)
ln(Real Material Inputs) (M)
Estimated Coefficients
0.3445*
(<0001)
0.3888*
(<0001)
0.7604*
(<0001)
-0.1492*
(<0001)
* statistically significant
The coefficients for real capital and real production wages have the anticipated signs and
are significant at a high level of confidence. The real material inputs coefficient does not have
the anticipated sign but does test significantly different from zero. Using the estimated
coefficients and the formula for supply elasticity shown above, the price elasticity of supply for
boat manufacturing is derived to be 1.57. The calculation of statistical significance for this
elasticity measure is not a straightforward calculation since the estimated function is non-linear.
No attempt has been made to assess the statistical significance of the estimated elasticity. The
corrections for serial correlation and the restricted model results yield inaccurate standard
measures of goodness of fit (R2). However, the model that is unrestricted and unadjusted for
serial correlation has an R2 of 0.99.
9-44
-------
Chapter 9: Economic Impact Analysis
The estimated price elasticity of supply for the boat manufacturing industry reflects that
the industry in the United States will increase production of boats by 1.57 percent for every 1.0
percent increase in the price of this product. The preceding methodology does not directly
estimate the supply elasticity of inboard cruisers due to a lack of necessary data. The assumption
implicit in the use of this estimate of price elasticity of supply is that the supply elasticity of
inboard cruisers will not differ significantly from the price elasticity of supply for all products
classified under SIC code 3732.
9.5.2.2 Price Elasticity Estimation for Forklifts
Demand Elasticity
Forklifts are used as intermediate products to produce final goods. The demand for large
SI engine forklifts is therefore derived from the demand for these final products. Information is
provided in Section 2.2 concerning the end uses of forklifts. According to this information,
forklifts are used primarily as an input in the manufacturing and wholesale trade sectors. One
primary use for forklifts is to lift and transport materials and merchandise in warehouse or retail
trade settings. Forklifts are therefore used in the production of a wide variety of goods
manufactured by these sectors of the economy.
The assumption was made that firms using forklifts as inputs into their productive
processes seek to maximize profits. The profit function for these firms may be written as
follows:
MAX n=PPPxf(Q,I)-(PxQ)- (POI x /) (Eq. 9-15)
where:
Tl
PPP
= profit,
= the price of the final product or end-use product,
f(Q> I) = the production function of the firm producing the final product,
P = the price of the forklifts,
Q = the quantity input use of forklifts
POI = a vector of prices of other inputs used to produce the final product,
and
/ = a vector of other inputs used to produce the final product.
The solution to the profit function maximization results in a system of derived demand
equations for forklifts. The derived demand equations are of the following form:
Q9g(P,PPP,POI) (Eq.9-16)
9-45
-------
Draft Regulatory Support Document
A multiplicative functional form of the derived demand equations are assumed because of the
useful properties associated with this functional form. The functional form of the derived
demand function is expressed in the following formula:
PP'P (Eq.9-17)
where:
A = a constant
P = the price elasticity of demand for forklifts, and
PpP = the final product price elasticity with respect to the use of forklifts.
All other variables have been previously defined and P, ppp, and A are parameters to be estimated
by the model. In the above equation, p represents the own-price elasticity of demand. The price
of other inputs (represented by POI) has been omitted from the estimated model, because data
relevant to these inputs were unavailable. The implication of this omission is that the use of
forklifts in production is fixed by technology.
The market price and quantity sold of forklifts are simultaneously determined by the
demand and supply equations. For this reason, it is advantageous to apply a systems estimator to
obtain unbiased and consistent estimates of the coefficients for the demand equations.25 Two-
stage least squares (2SLS) is the estimation procedure used in this analysis to estimate the
demand equation for forklifts. Two-stage least squares uses the information available from the
specification of an equation system to obtain a unique estimate for each structural parameter.
The first stage of the 2SLS procedure involves regressing the observed price of forklifts against
the supply and demand "shifter" variables that are exogenous to the system. These are referred
to as instruments. This first stage produces fitted (or predicted) values for the forklift price
variable that are, by definition, uncorrelated with the error term by construction and thus do not
incur endogeneity bias. These fitted values for price are then used in the second stage equation
(see Eq. 9-17). By converting the above equation to natural logarithms, the coefficient on the
forklift price variable (/?) yields an estimate of constant elasticity of supply.
The exogenous supply-side variables used to estimate the demand function include: the
real capital stock variable for SIC code 3537 (the industry that manufactures forklifts), a
technology time trend (i), and the price indices for the cost of labor and the cost of materials for
SIC code 3537. A price index for the cost of labor was generated by dividing real production
worker wages (derived by dividing nominal production worker wages by the implicit GDP
deflator) by production worker hours. The demand-side variables include: real GDP and the
price indices of manufacturing and wholesale trade. Generally, the price of final products are
used as demand-side variables, but because forklifts are used as an input to the production of a
wide variety of goods, we rely upon price indices of the manufacturing and wholesale trade
sectors.
Data relevant to the econometric modeling of the price elasticity of demand for forklifts
9-46
-------
Chapter 9: Economic Impact Analysis
are listed in Table 9.5-3. Consistent time series data for the period 1970 through 1999 were
obtained. The annual domestic quantity of forklift shipments was retrieved from the Industrial
Truck Association Membership Handbook. Price data for forklifts over this time period were not
available, so the price index of shipments for SIC code 3537 was retrieved from both the NBER-
CES Productivity Database and BEA's Shipments of Manufacturing Industries instead. The
following variables were also retrieved from the NBER-CES Productivity Database and the
Census Bureau's ASM: production worker wages, production worker hours, real capital stock
(except for the years 1998 and 1999), investment, the price index of investment (except for the
years 1997 through 1999), and the price indices of shipments and material inputs (except for the
years 1998 and 1999).
Other variables, including the price indices for the manufacturing and wholesale trade
industries, the implicit GDP deflator, real GDP, the price index of investment for SIC code 3537
for the years 1997 to 1999, and the price indices of shipments and material inputs for the years
1998 and 1999 were retrieved from the Bureau of Economic Analysis. Note that since a price
index of material inputs for SIC 3537 was not available beyond 1997, we relied upon a general
price index for intermediate materials from BEA's Survey of Current Business. A price index
for investment for SIC 3537 was also not available beyond 1996, so a general price index for
capital equipment was used for the years 1997 - 1999 from the same source. Real capital stock
for the years 1998 and 1999 was derived for SIC 3537 (see Equation 9-13 for the equation used
to calculate real capital stock for these years).
Table 9.5-3
Data Inputs for the Estimation of
Demand Equations for the Forklift industry26272829303132
Variable Unit of Measure
1. Time Trend
2. Price Index of Shipments for the Industrial Truck, Tractor, Trailer, and index
Stacker Mainery Industry (SIC 3537)
3. Quantity of Forklift Shipments units
4. Price Index for the Manufacturing Industry index
5. Price Index for the Wholesale Trade Industry index
6. Price Index of Material Inputs index
7. Production Worker Wages millions of $
8. Implicit GDP Deflator index
9. Production Worker Hours thousands of worker hours
10. Investment millions of $
11. Price Index of Investment index
12. Real Capital Stock millions of $ 1987
13. Real Gross Domestic Product billions of $ 1987
9-47
-------
Draft Regulatory Support Document
SAS Release 8.2 for Windows was used to econometrically estimate the price elasticity of
demand. Two-stage least squares econometric models were estimated for the forklift industry
using the price indices of manufacturing and wholesale trade as the end-use products,
respectively. Relying on price indices for entire sectors of the economy to represent specific end-
use products is not ideal, but price data on specific products that forklifts are used to manufacture
are not readily available. Additionally, forklifts are used in the production of a large variety of
goods and it would therefore be difficult to determine which products to focus on for the
estimation of demand elasticity. The data limitations are recognized and the demand elasticity
estimates generated here are therefore, interpreted with caution.
Overall, the models using price indices for these end products were not successful. This
may be due in part to the fact that price indices for entire sectors of the economy are not reliable
instruments for the prices of the final products that forklifts are used to produce. The coefficient
for the price index of shipments for SIC 3537 was not statistically different from zero in the
model which included manufacturing. In the second model, which used the price index of
wholesale trade in lieu of price index of manufacturing, the coefficient on the price index of
shipments for SIC 3537 was significantly different than zero, but was equal to -5.8, an extremely
large estimate of demand elasticity. The model results using the price indices of manufacturing
and wholesale trade as the final product prices are reported in Table 9.5-4. with p-values listed
below each coefficient estimate. Each of the coefficients reported has the anticipated sign,
however not all of the estimates are significantly different from zero.
The price elasticity of demand estimate reflects an elastic demand for forklifts.
Regulatory control costs are less likely to be paid by consumers of products with elastic demand
when compared to products with inelastic demand, all other things held constant. Price increases
for products with elastic price elasticity of demand lead to decreases in revenues for producers,
however it does say anything with regard to producer profits.
A degree of uncertainty is associated with this method of demand estimation. The
estimation is not robust since the model results vary depending upon the instruments used in the
estimation process. For this reason, the above results are used as an indication that the elasticity
of demand is elastic and we instead rely upon an assumed measure of-1.5 for the own-price
elasticity of demand for forklifts.
9-48
-------
Chapter 9: Economic Impact Analysis
Table 9.5-4
Derived Demand Coefficients Equations for the Forklift Industry
Variables
Own Price ft
ln(PI of Shipments for SIC 3537)
End-Use ppp
ln(PI of Manufacturing)
End-Use ppp
ln(PI of Wholesale Trade)
ln(Real GDP)
F value
Adjusted R-Square
Estimation 1
-3.03
(0.1113)
0.17
(0.9203)
3.44*
(<.0001)
24.25*
(<.0001)
0.76
Estimation 2
-5.76*
(<.0001)
3.11*
(0.0142)
4.23*
(<.0001)
32.96*
(<.0001)
0.813
statistically significant.
Supply Elasticity
Published sources of the price elasticity of forklift supply were not readily available. For
this reason, an econometric analysis of the price elasticity of supply for forklifts was conducted
using the same approach as the one used to estimate the supply elasticity for boat manufacturing
described above.
The estimated model reflects the production function for forklifts, using annual time
series data for the years from 1958 through 1999. The data used to estimate supply elasticity are
enumerated in Table 9.5-5. The data for the price elasticity of supply estimation model includes:
the value of domestic shipments of SIC 3537 in millions of dollars; the price index for value of
domestic shipments (the value of domestic shipments deflated by the price index represents the
quantity variable which is the dependent variable in the analysis); a technology time variable;
production wages in millions of dollars; the implicit GDP deflator (used to deflate production
wages), the material inputs in millions of dollars; the price index for value of materials;
investment in millions of dollars; the price index of investment; and real net capital stock in
millions of dollars.
Data to estimate the production function for the forklifts exclusively were largely
unavailable; therefore, data for SIC code 3537 is utilized for each of the variables previously
enumerated with the exception of the time variable. All data for the supply elasticity estimation
were retrieved from the National Bureau of Economic Research-Center for Economic Studies
(NBER-CES) Productivity Database and the U.S. Census Bureau's Annual Survey of
Manufactures (ASM), with the exception of the technology time trend, the implicit GDP deflator,
the price index for investment for SIC 3537 for the years 1997 through 1999, the price indices of
9-49
-------
Draft Regulatory Support Document
shipments and material inputs for SIC 3537 for the years 1998 and 1999, and real capital stock
for the years 1998 and 1999 (these data for real capital stock were not available). These variables
(except the time trend and real capital stock for 1998 and 1999), were retrieved from the Bureau
of Economic Analysis (BEA).
More specifically, the price index of shipments for SIC 3537 for the years 1998 and 1999
was retrieved from the BEA's Shipments of Manufacturing Industries. Similar to the boat
manufacturing industry, a price index of material inputs for SIC 3537 was not available beyond
1997. We therefore relied upon a general price index for intermediate materials from BEA's
Survey of Current Business. A price index for investment for SIC 3537 was also not available
beyond 1996, so a general price index for capital equipment was used for the years 1997 - 1999
from the same source. Real capital stock for the years 1998 and 1999 was derived for SIC 3537
(see Equation 9-13 for the equation used to calculate real capital stock for these years).
Again, the capital stock variable was the most difficult variable to quantify for use in the
econometric model. Ideally, this variable should represent the economic value of the capital
stock actually used by each facility to produce forklifts for each year of the study. The most
reasonable data for this variable would be the number of machine hours actually used to produce
forklifts each year, but we do not possess this information. In lieu of machine hours data, the
dollar value of net capital stock in constant 1987 prices, or real net capital stock, is used as a
proxy for this variable.
Table 9.5-5
Data Inputs for the Estimation of Supply Elasticity for the Forklift Industry33'3435'36-37'38
Variable Unit of Measure
1. Value of Shipments for the Industrial Truck, Tractor, Trailer, and millions of $
Stacker Machinery Industry (SIC 3537)
2. Price Index of Shipments for the Industrial Truck, Tractor, Trailer, and index
Stacker Machinery Industry (SIC 3537)
3. Time trend
4. Production Worker Wages millions of $
5. Implicit GDP Deflator index
6. Cost of Material Inputs millions of $
7. Price Index of Material Inputs index
8. Investment millions of $
9. Price Index of Investment index
8. Real Capital Stock millions of 1987$
SAS Release 8.2 for Windows was used to estimate econometric estimates of the price
elasticity of supply for the forklift manufacturing industry. A restricted least squares estimator
9-50
-------
Chapter 9: Economic Impact Analysis
was used to estimate the coefficients of the production function model. A log-linear
specification was estimated with the sum of the at restricted to unity. This procedure is
consistent with the assumption of constant returns to scale. The model was further adjusted to
correct for first-order serial correlation using the Yule-Walker estimation method. The results of
the estimated model are presented in Table 9.5-6 with p-values listed in parentheses below each
coefficient estimate.
Table 9.5-6
Estimated Supply Model Coefficients for the Forklift Industry
Variables
In(Time) (/)
ln(Real Capital Stock) (K)
ln(Real Production Wages) (Lt)
ln(Real Material Inputs) (Mt)
Estimated Coefficients
0.1676
(.2066)
0.5833*
(0.0070)
1.1632*
(0.0001)
-0.7466*
(0.0002)
: statistically significant
The coefficients for real capital and real production wages have the anticipated signs and
are significant at a high level of confidence. The real material inputs coefficient does not have
the anticipated sign and also tests significantly different from zero. Using the estimated
coefficients and the formula for supply elasticity shown above, the price elasticity of supply for
forklift manufacturing is derived to be 0.714. The calculation of statistical significance for this
elasticity measure is not a straightforward calculation since the estimated function is non-linear.
No attempt has been made to assess the statistical significance of the estimated elasticity. The
corrections for serial correlation and the restricted model results yield inaccurate standard
measures of goodness of fit (R2). However, the model that is unrestricted and unadjusted for
serial correlation has an R2 of 0.99.
The estimated price elasticity of supply for the forklift manufacturing industry reflects
that the industry in the United States will increase production of forklifts by 0.714 percent for
every 1.0 percent increase in the price of this product. The preceding methodology does not
directly estimate the supply elasticities for forklifts due to a lack of necessary data. The
assumption implicit in the use of this price elasticity of supply estimate is that the supply
elasticity of forklifts will not differ significantly from the price elasticity of supply for all
products classified under SIC code 3537.
9-51
-------
Draft Regulatory Support Document
9.5.2.3 Price Elasticity Estimation for Snowmobiles
Demand Elasticity
The price elasticity of demand is an important input into the market model, and this
information is required to characterize the demand for snowmobiles. Econometric estimation of
the price elasticity of demand for snowmobiles was unsuccessful despite numerous model
specifications and varied statistical techniques evaluated. A search of the literature did not
provide snowmobile price elasticity of demand estimates. A study was conducted for the
recreational boat industry in 1987.39 This study estimates the price elasticity of demand for boats
to be -1.78. The price elasticity of demand for a variety of pleasure boat categories were
estimated. These estimates range from -1.4 to -2.17. For purposes of this analysis a price
elasticity of demand for snowmobiles of -2 is postulated. Since this estimate does not relate
specifically to the snowmobile market but to another category of recreational vehicles, and there
are uncertainties associated with elasticity estimates, a sensitivity analysis of the impact of this
estimate on model results is shown in the Appendix to Chapter 9 of this report.
Supply Elasticity
The price elasticity of supply for snowmobiles is a necessary input into the market model.
A literature search did not provide any estimates of this required input. An econometric analysis
is conducted and a value for this parameter is estimated. Several approaches were considered
including a simultaneous equation approach, a production function approach and a simple supply
function specification. Econometric results from the latter approach are presented. With this
approach, the quantity of snowmobiles produced is hypothesized to be a function of the price of
the product and the price of factors of production including the materials, labor, and capital as
follows:
Where Qt is the quantity of snowmobiles produced and sold in period t and PMt , PLt PKt are the
factor prices for inputs of production (materials, labor and capital, respectively) in period t. The
data used to estimate the elasticity are enumerated in Table 9.5-7. Consistent time series data for
the years 1986 through 2000 are used in the analysis. All price data have been restated into real
values using the implicit GDP deflator. Snowmobile price and quantity data are provided by
ISMA. The quantity of snowmobiles sold are restated to be values sold on a per household basis.
Cost of production data for the snowmobile industry are largely unavailable. In lieu of the cost
production data specific to snowmobile production, cost of production data for SIC 3799/NAICS
code 336999 Other Transportation Equipment (includes snowmobiles as a product category) are
used in the analysis as a proxy for the cost of production data for snowmobiles. The data used for
the analysis are listed in Table 9.5-7.
9-52
-------
Chapter 9: Economic Impact Analysis
Table 9.5-7
Data Inputs for the Estimation of
Supply Elasticity for the Snowmobile industry40 414243444S 4647
Variable
Unit of Measure
1. Quantity of Snowmobiles Sold
2. US Households
3. Average price of snowmobiles sold
4. Price Index - Materials (SIC 3799 /NAICS
336999)
5. Price Index - Investment (SIC 3799 /NAICS
336999)
6. Wages per employee (SIC 3799 /NAICS
336999)
7. Real Implicit Gross Domestic Product Deflator
units
number of households
dollars
price index
price index
dollars
price index
SAS Release 8.2 for Windows was used to develop econometric estimates of the price elasticity
of supply for the snowmobile industry. A log-log specification of the model was estimated. The
price of capital was omitted from the model specification due to high correlation with the
snowmobile price data. The model was further adjusted to correct for serial correlation using the
Yule-Walker estimation method. Alternative lag periods were considered. The results of the
estimated model are presented in Table 9.5-8 with related standard errors. Based upon this
analysis the price elasticity of supply for the snowmobile industry is estimated to be 2.10.
Table 9.5-8
Estimated Supply Model Coefficients for the Snowmobile Industry
Variables
Intercept
log (real price of snowmobiles)
log (real wages per employee)
log (real price of materials)(PM/f)
Total R-Square
Durbin-Watson Statistic
Estimated Coefficient
-16.4236
2.1043
-0.2858
0.1617
0.9771
1.9728
Standard Errors
1.9094*
0.2441*
0.5479
0.1322
Statistically significant at the 1% significance level.
The estimated model is statistically significant. The coefficient for real wages per
employee has the anticipated signs but is not statistically significant. The coefficient for the
9-53
-------
Draft Regulatory Support Document
materials variable does not have the anticipated sign and is not statistically significant. The
coefficient for the price variable has the expected sign and is statistically significant. This value
provides an estimate for the price elasticity of supply for snowmobiles. The estimated model is
statistically significant. This value of 2.10 represents the price elasticity of supply used in the
study. The uncertainty associated with this estimate is acknowledged. A sensitivity analysis of
this model input is conducted in the appendix to this chapter.
9.5.2.4 Price Elasticity Estimation for All-Terrain Vehicles
Demand Elasticity
The price elasticity of demand is an important input to the market model, and this
information is required to characterize the demand for ATVs. Econometric estimation of the
price elasticity of demand for this market was unsuccessful despite numerous model
specifications and varied statistical techniques evaluated. A search of the literature did not
provide ATV price elasticity of demand estimates. A study was conducted for the recreational
boat industry in 1987.48 This study estimates the price elasticity of demand for boats to be -1.78.
The price elasticity of demand for a variety of pleasure boat categories were estimated. These
estimates range from -1.4 to -2.17. For purposes of this analysis, a price elasticity of demand for
ATVs of -2 is postulated. Since this estimate does not relate specifically to the ATV market but
another category of recreational vehicles and there are uncertainties associated with elasticity
estimates in general, a sensitivity analysis of the impact of this estimate on model results is
shown in the Appendix to Chapter 9 of this report.
Supply Elasticity
The price elasticity of supply is a necessary input in the market model. This estimate is
required to characterize the way producers of ATVs respond to a change in the price of the
product. A search of the economic literature was conducted without success. Econometric
estimation of this variable were undertaken also without success. Numerous model specification
and variable combinations were investigated, but the results were not satisfactory from a
statistical perspective. The price elasticity of supply for off-highway motorcycles was estimated
to be -0.93. Since the productive processes are similar for ATVs and off-highway motorcycles
and many of the producers of ATVs also produce off-highway motorcycles, the supply elasticity
for off-highway motorcycles appears to be a reasonable proxy for the supply elasticity for ATVs.
A discussion of the techniques and data used to econometrically estimate this value follows in
Section 9.5.2.5.
9-54
-------
Chapter 9: Economic Impact Analysis
9.5.2.5 Price Elasticity Estimation for Off-Highway Motorcycles
Demand Elasticity
The price elasticity of demand is an important component of the market model and this
information is required to characterize the demand for off-highway motorcycles. Econometric
estimation of the price elasticity of demand for this market was unsuccessful despite numerous
model specifications and varied statistical techniques evaluated. A search of the literature did
not provide off-highway motorcycle price elasticity of demand estimates. A study was conducted
for the recreational boat industry in 1987.49 This study estimates the price elasticity of demand
for boats to be -1.78. The price elasticity of demand for a variety of pleasure boat categories
were estimated. These estimates range from -1.4 to-2.17. For purposes of this analysis a price
elasticity of demand for off-highway motorcycles of -2 is postulated. Since this estimate does
not relate specifically to the off-highway motorcycle market but another category of recreational
vehicles and there are uncertainties associated with elasticity estimates in general, a sensitivity
analysis of the impact of this estimate on model results is shown in the Appendix to Chapter 9 of
this report.
Supply Elasticity
The price elasticity of supply for off-highway motorcycles is econometrically estimated.
Data for the study is provided by the MIC and collected from publicly available sources. A
description of the data used in the study, the modeling techniques used, and the model results are
presented.
Methodology
A partial equilibrium market demand/supply 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 variable in other equations, the
error terms are correlated with the endogenous variables (price and output). In this case, single-
equation ordinary least squares (OLS) estimation of individual equations will lead to biased and
inconsistent parameter estimates. Thus, simultaneous estimation of this system to obtain
elasticity estimates requires that each equation be identified through the inclusion of exogenous
variable to control for shifts in the supply and demand curves over time.
The supply/demand system for OHM over time (t) is defined as follows:
Qtd = f(Pt,Zt) + ut
9-55
-------
Draft Regulatory Support Document
Qts=(P,,Wt) + vt
Q,d = Q,s
The first equation above shows quantity demanded in year t as a function of price, Pt and an array
of demand factors (e.g., measures of economic activity and substitute prices), and an error term,
ut. The second equation characterizes supply for the OHM market. The quantity supplied, Qts in
year t is a function of price and other supply factors, Wt (e.g., input prices) and an error term, vt.
The third equation specifies the equilibrium condition that quantity supplied equals quantity
demanded in year t creating a system of three equations in three variables . The interaction of the
specified market forces solves this system generating equilibrium values for the variables Pt* and
Q,' = Q/ = Q/
Since the objective is to generate estimates of the supply equation for use in the economic
model, the EPA employed the two-stage least squares (2SLS) regression procedure to estimate
only the parameters of the supply equation. Similar techniques for the demand equation were
unsuccessful. EPA specified the logarithm of the quantity supplied as a linear function of the
logarithm of the price so that the coefficient on the price variable yields the estimate of the
constant elasticity of supply for OHM. All prices employed in the estimation process were
deflated by the gross domestic product (GDP) implicit price deflator to reflect real rather than
nominal prices. The first stage produces fitted (or predicted) values for the price variables that
are, by definition, highly correlated with the error term. In the second stage, these fitted values
are then employed as observations of the right hand side price variable in the supply function.
This fitted value is uncorrelated with the error term by construction and thus does not incur the
endogeneity bias.
Data
Price and quantity data were provided by MIC for the period 1990 through 2000. Thus the
study uses annual data for the period 1990 through 2000. For the supply equation estimated,
supply is postulated to be a function of price, a trend variable to recognize technology changes
over time, and the price of inputs of production. A number of factor prices were considered
including the price of materials, labor, and capital. Unfortunately these inputs price are some
cases highly correlated. For this reason, the price of materials is used in estimation. A listing of
the data used in the analysis and the source of the data are shown in Table 9.5-9. All data used in
the analysis are deflated to real values using the real gross domestic product implicit price
deflator. Sales quantities and income values are restated to per US household values. All values
are restated to natural logs.
9-56
-------
Chapter 9: Economic Impact Analysis
Table 9.5-9
Data Inputs for Off-Highway Motorcycle Supply Estimation50 5152535455 5657
Variable
Unit of Measure
1. Quantity of OHM sold
2. US households
3. Average price OHM
4. Time trend
5. Price index for materials used in production
6. Price of a substitute product (SIC 3799/NAICS 336999)
7. Disposable household income
8. Real implicit GDP deflator
units
number
dollars
N/A
price index
price index
dollars
price index
Results
The results of the supply estimation are shown in Table 9.5-10
Table 9.5-10
Estimated Supply Model for the Off-Highway Motorcycle Industry
Parameter
Intercept
log (Trend Variable)
log (Real Price)
log (Price of materials used in
production of OHM)
Parameter Estimate
-10.7632*
-0.03399*
0.93323*
-0.36977
Standard Error
0.179407
0.005626
0.017468
0.294203
Adjusted R Square 0 .9996
F-Value 8867.69*
Durbin Watson 1.65
* Statistically significant at the 1% significance level.
The estimated equation and coefficients have the expected sign and are statistically significant at
a 1% significance level with the exception of the cost of materials variable. While the coefficient
for the price of materials variable has the expected sign, it is not statistically significant. The
coefficient for the natural log of the real price variable of 0.93 is the estimate of the price
elasticity of supply for the off-highway motorcycle market. The uncertainty surrounding this
9-57
-------
Draft Regulatory Support Document
estimate is recognized and a sensitivity analysis of this model input is conducted in the appendix
to this chapter.
9.6 Marine
The following section describes the baseline characterization of the market in the year
2001, the per unit regulatory control costs incurred by producers of recreational diesel marine
vessels, and the economic impacts that would have resulted had the emissions control program
been implemented in the baseline year. We also examine the economic impacts on the diesel
inboard cruiser market using baseline year data for each change in the per unit control costs that
occurs. This section concludes with a comparison of the stream of engineering costs and
estimated welfare losses (excluding fuel efficiency gains) projected to occur after the regulation's
implementation. No fuel efficiency gains are projected to occur from the standard affecting
diesel recreational marine vessels, therefore the social costs (surplus losses net fuel cost savings)
are equal to the surplus losses projected from the model.
9.6.1 Marine Baseline Market Characterization
Inputs to the economic analysis are a year 2001 baseline characterization of the diesel
inboard cruiser market that includes the domestic quantity produced, quantity of imports,
baseline market price, demand elasticity, and domestic and foreign supply elasticity measures.
Table 9.6-1 provides the baseline data on the U.S. diesel inboard cruiser market used in this
analysis.
Table 9.6-1
Baseline Characterization of the U.S. Diesel Inboard Cruiser Market: 2001
Inputs Baseline Observation
Market price ($/boat) $341,945.00
Market output (boats) 843 5
Domestic 8098
Foreign 337
Elasticities
Domestic supply (estimated) 1.57
Foreign supply (assumed) 1.57
Demand (previously estimated) -1.44
The total market output of diesel inboard cruiser marine vessels was derived from data
taken from publications of the National Marine Manufacturers Association58'59. EPA projected
9-58
-------
Chapter 9: Economic Impact Analysis
the quantity of CI marine engines for the years 1998 through 2030 based upon NMMA's
historical data on the quantity of inboard cruisers sold in the U.S. For the year 2001, EPA's
projection shows that 16,068 engines were sold domestically. This total includes those engines
sold in the U.S. whether they were produced domestically or abroad. A simplifying assumption
has been made that all of these engines are used in inboard cruisers, though we acknowledge that
there is an extremely small fraction of these engines that are used in inboard runabouts
(approximately 2 percent) and an even smaller fraction used in marine vessels with outboard
engine configurations.60 A majority (95 percent) of inboard cruisers contain two engines.61
Using this information, we find that the 16,068 recreational diesel marine engines sold in 2001
would yield 8,435 diesel inboard cruisers.
Market output is not partitioned into domestically produced and imported quantities of
recreational diesel marine engines. In order to determine the share of imported boats, historical
import quantities of inboard cruisers were compared with the domestically produced quantities
reported in Table 2.1-7 for the years 1992 to 200062. On average, imported inboard cruisers were
equal to about 4 percent of the inboard cruisers produced and sold in the U.S. This information
was used to partition the total quantity of diesel inboard boats for the year 2001.
The price of diesel inboard cruisers was taken to be equal to the average retail price of all
inboard cruisers sold in the year 2001. NMMA quotes this price at $341,945.63 The estimates of
demand and supply elasticity have been discussed in detail in Section 9.5.2.1. A separate
estimate of foreign supply elasticity has not been carried out. For modeling purposes, we assume
that the foreign supply elasticity is equal to the domestic supply elasticity.
9.6.2 Marine Control Costs
In order to determine a per diesel inboard cruiser cost over the years 2006 to 2030 for use
in the economic analysis, the future stream of engineering costs (without fuel savings) provided
in Chapter 7 is divided by the number of boats EPA projected from the NMMA data. This yields
a stream of average cost per diesel inboard cruiser. As stated in the section above, the EPA
projected the quantity of recreational diesel marine engines sold in the U.S. for the years 1998
through 2030. Using these engine quantities and the fact that approximately 95 percent of
inboard cruisers contain two engines, we developed a projected stream of domestic diesel
inboard cruiser sales. The total stream of engineering costs from Chapter 7, the projected
number of diesel inboard cruisers, and the average regulatory cost per boat are provided in Table
9.6-2. During the initial years of implementation, the per unit costs change but by 2014, they are
projected to remain the same.
Table 9.6-2
Projected Future Stream of Engineering Costs ($103), Quantity of
Diesel Inboard Cruisers, and Per Diesel Inboard Cruiser Regulatory Costs
9-59
-------
Draft Regulatory Support Document
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated
Engineering Costs
$7,806.0
$8,365.3
$8,573.8
$9,413.5
$9,637.0
$5,213.4
$5,176.7
$5,290.8
$4,958.1
$5,062.7
$5,167.7
$5,272.7
$5,377.6
$5,482.6
$5,587.6
$5,692.5
$5,797.5
$5,902.5
$6,007.4
$6,112.4
$6,217.2
$6,322.0
$6,426.9
$6,531.7
$6,636.5
Projected Quantity of Diesel
Inboard Cruisers
9665
9913
10159
10407
10653
10899
11145
11390
11636
11882
12128
12374
12621
12867
13113
13360
13606
13853
14099
14345
14591
14837
15083
15329
15575
Cost Per Diesel
Inboard Cruiser
$808
$844
$844
$905
$905
$478
$464
$464
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
$426
9.6.3 Marine Economic Impact Results
The economic impacts of the emissions control program for recreational diesel marine
vessels are estimated for each year in which the per vessel regulatory costs change, assuming the
baseline year 2001 price and quantity. Though we possess projected quantities of diesel inboard
cruiser marine vessels through the year 2030, we do not have future year prices. We are
therefore unable to estimate the economic impacts of the future costs assuming future year
quantities and prices. For this reason, we rely upon the most current year of data to inform the
model when we impose the future costs per vessel on producers. Using baseline year data allows
9-60
-------
Chapter 9: Economic Impact Analysis
us to estimate relative changes in price and quantity as opposed to absolute changes. The
estimated percent changes in price and quantity, the changes consumer and producer surplus, and
the total loss in surplus are presented for various years in Tables 9.6-3 and 9.6-4.
Table 9.6-3
Price and Quantity Changes for the Diesel Inboard Cruiser Market*
Impact Measure
Cost Per Unit
Change in Market Price
Change in Market Output
Domestic
Foreign
2006
$808
0.12%
-0.18%
-0.18%
-0.18%
2007/8
$844
0.13%
-0.19%
-0.19%
-0.19%
2009/10
$905
0.14%
-0.20%
-0.20%
-0.20%
2011
$478
0.07%
-0.10%
-0.10%
-0.10%
2012/13
$464
0.07%
-0.10%
-0.10%
-0.10%
2014+
$426
0.06%
-0.09%
-0.09%
-0.09%
*Results are the same for the years 2007 and 2008, 2009 and 2010, and for the years 2012 and 2013. They are also
the same for the years 2014 and beyond. These results are not reported in separate columns to avoid repetition.
Results are based on baseline year 2001 market conditions and fuel cost savings are not included.
Table 9.6-4
Annual Losses in Consumer and
Producer Surplus and for the Diesel Inboard Cruiser Market*
Impact Measure
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
2006
$3,551.8
$3,251.9
$3,122.0
$129.9
$6,083.7
2007/8
$3,709.9
$3,396.4
$3,260.7
$135.7
$7,106.3
2009/10
$3,977.7
$3,641.1
$3,495.6
$145.5
$7,618.8
2011
$2,101.9
$1,925.8
$1,848.9
$76.9
$4,027.7
2012/13
$2,040.4
$1,869.5
$1,794.8
$74.7
$3,909.9
2014+
$1,873.4
$1,716.6
$1,648.0
$68.6
$3,590.0
*Results are the same for the years 2007 and 2008, 2009 and 2010, and for the years 2012 and 2013. They are also
the same for the years 2014 and beyond. These results are not reported in separate columns to avoid repetition.
Results are based on baseline year 2001 market conditions and fuel cost savings are not included.
** CS refers to consumer surplus and is rounded to the nearest hundredths. For a description of the change in
consumer surplus, see Section 9.2.2
*** PS refers to producer surplus and is rounded to the nearest hundredths. For a description of the change in
producer surplus, see Section 9.2.2.
As Table 9.6-3 shows, the relative increases in price due to the regulatory costs are less
than two-tenths of a percent while the reductions in output are less than one-quarter of a percent.
These impacts are considered minimal. Also notable is that the percent changes in price and
quantity peak in the years 2009 and 2010 but then are smaller further out into the future. The
9-61
-------
Draft Regulatory Support Document
percent reduction in quantity is the same for both domestic and foreign output because it has
been assumed that domestic and foreign supply have the same price elasticity.
Table 9.6-4 presents the loss in consumer surplus, the loss in producer surplus, and the
loss in surplus (equal to the sum of the changes in consumer and producer surplus). These results
show that the losses in consumer and producer surplus are approximately equal in size, though
the loss in producer surplus is slightly less than the loss in consumer surplus. Consumer surplus
losses range from a high of just under $4 million to a low of $1.9 million, while the losses in
producer surplus vary from $3.6 million to $1.7 million. Like the price and quantity changes,
these measures are largest in the years 2009 and 2010. They then decline to their lowest value in
2014 and beyond.
9.6.4 Marine Engineering Cost and Surplus Loss Comparison
Table 9.6-5 presents the future stream of estimated engineering costs holding quantity
constant to the baseline year quantity and the loss in surplus that has been estimated from the
economic impact model. Because economic modeling takes into account consumer and producer
behavior, the estimated surplus losses are less than the engineering costs under a perfectly
competitive market setting. In this case, surplus losses are, on average equal to over 99 percent
of the calculated engineering costs. Note that the costs provided in this table are not discounted.
Based upon the annual ratio of surplus losses to engineering costs holding quantity
constant to baseline year quantity, a projection of surplus losses over the future year stream is
calculated from the future stream of engineering costs that appear in Chapter 7. The projected
future stream of surplus loss is calculated by multiplying the annual ratio by the future stream of
engineering costs and is presented in Table 9.6-6. Again, these costs are not discounted.
9.6.5 Marine Economic Impact Results with Fuel Cost Savings
No fuel savings are projected for the recreational diesel marine engine category, therefore
there are no alternative results to present for this vehicle category. The stream of social costs for
this vehicle category are equal to the stream of estimated surplus losses shown in Table 9.6-6.
Table 9.6-5
Interim Engineering Cost and Surplus Loss Comparison for the Recreational Diesel
Marine Vessel Market Based on Year 2001 Quantity (Q =8,435 inboard cruisers)
Year
2006
2007
2008
2009
Estimated Engineering Costs
$6,812,980
$7,119,006
$7,119,006
$7,630,744
Estimated Surplus Loss
$6,803,645
$7,106,227
$7,106,227
$7,618,828
9-62
-------
Chapter 9: Economic Impact Analysis
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
$7,630,982
$4,035,120
$3,918,352
$3,918,326
$3,594,386
$3,594,365
$3,594,403
$3,594,441
$3,594,328
$3,594,365
$3,594,401
$3,594,436
$3,594,470
$3,594,365
$3,594,399
$3,549,432
$3,594,373
$3,594,444
$3,594,388
$3,594,456
$3,594,401
$7,618,828
$4,027,788
$3,909,937
$3,909,937
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
$3,590,020
Table 9.6-6
Engineering Costs and Surplus Loss Comparison for
the Recreational Diesel Marine Vessel Market
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Estimated Engineering Costs
$7,806,010
$8,365,319
$8,573,839
$9,413,530
$9,637,035
$5,213,411
$5,176,672
$5,290,764
$4,958,052
$5,062,713
$5,167,682
Estimated Surplus Loss
$7,795,314
$8,350,303
$8,558,165
$9,398,831
$9,621,686
$5,203,938
$5,165,555
$5,279,437
$4,952,029
$5,056,593
$5,161,380
9-63
-------
Draft Regulatory Support Document
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
$5,272,652
$5,377,623
$5,482,592
$5,587,562
$5,692,532
$5,797,503
$5,902,472
$6,007,442
$6,112,413
$6,217,227
$6,322,042
$6,426,858
$6,531,673
$6,636,488
$5,266,167
$5,371,178
$5,475,965
$5,580,752
$5,685,539
$5,790,326
$5,895,337
$6,000,124
$6,104,911
$6,209,698
$6,314,262
$6,419,049
$6,523,512
$6,628,400
9.7 Large SI Engines
As described in Chapter 2 and illustrated in Table 6.2.2-1, Large SI engines are used in
nearly 50 different applications ranging from fairly small, low horsepower equipment used in
lawncare applications to agricultural and construction equipment exceeding 100 horsepower.
Forklifts are clearly the dominant application in this category, accounting for about 52 percent of
the 2000 populations of Large SI engines. The next largest applications are generators,
accounting for about 15 percent, and commercial turf applications, accounting for about 6
percent. Forklifts are also used more than other applications, for about 15,000 hours over the
average operating life of the equipment, compared to about 6,000 hours for the next most-used
applications (e.g., aerial lifts, refrigeration/AC, cranes). Similarly, forklifts accounted for nearly
81 percent of the NOx, 64 percent of the HC, 54 percent of the CO, and 76 percent of the PM
emissions from Large SI engines in 2000. Because of their dominant position in this category,
the following economic impact analysis focuses on the forklift segment. Specifically, we
estimate the change in price and quantity, and the sum of consumer and producer surplus losses
only for forklifts. To estimate the total social costs/gains for Large SI, we use the engineering
costs to approximate the sum of consumer and producer surplus losses for Large SI engines other
than forklifts. This approach slightly overestimates the surplus losses for the category since
engineering costs are higher than surplus losses.
While it would be possible to perform a market analysis for each of the Large SI
applications, we chose not to. Annual sales in some of these categories are so small that the
results of separate analysis would not be meaningful and would imply a degree of precision that
would not be reflected in the data inputs. Grouping the applications by horsepower, load factor,
9-64
-------
Chapter 9: Economic Impact Analysis
or usage rates would not necessarily reduce the complexity of the analysis because equipment
that use similar size engines are often not used with the same intensity. In addition, their markets
may not necessarily share the same demand and supply characteristics.
The results of our economic impact analysis for forklifts with regard to price and quantity
changes is not meant to be interpreted as representing the estimated impacts for all Large SI
engines. Changes in price and quantity are likely to be different for applications other than
forklifts due to differences in their market characteristics.
The remainder of this section describes the baseline characterization of the forklift market
in the year 2000, the regulatory control costs incurred by producers of forklifts, and the
economic impacts that would have resulted had the emissions control program been imposed in
the baseline year. We examine the economic impacts on the forklift market using the baseline
year data for each change in the per unit control costs that occurs. A comparison is then made
between the engineering cost and surplus loss streams projected to occur after the regulation's
implementation. This initial comparison of the cost streams assumes no fuel cost savings. A
comparison is then made between engineering costs and social costs/gains accounting for fuel
cost savings of the emissions control program. Finally, an estimate of the social costs/gains for
Large SI engines other than forklifts is presented, using engineering costs as a substitute for
consumer and producer surplus losses.
9.7.1 Forklift Baseline Market Characterization
Inputs to the economic analysis are a year 2000 baseline characterization of the forklift
market that includes the domestic quantity produced, quantity of imports, baseline market price,
demand elasticity, and domestic and foreign supply elasticity measures. Table 9.7-1 provides the
baseline data on the U.S. forklift market used in this analysis.
Table 9.7-1
Baseline Characterization of the U.S. Forklift Market: 2000
Inputs Baseline Observation
Market price ($/forklift) $26,380.00
Market output (forklifts) 65000
Domestic 48750
Foreign 16250
Elasticities
Domestic supply (estimated) 0.714
Foreign supply (assumed) 0.714
Demand (assumed) -1.5
9-65
-------
Draft Regulatory Support Document
The total quantity of Large SI engines sold in the U.S. was retrieved from the PSR
database, which contains projections of U.S. sales of Large SI engines for the year 2000 and the
years 2004 through 2030. Though we possess year 2000 quantity of imports and domestic
shipments of forklifts from the International Trade Commission and the Industrial Truck
Association, respectively, we have chosen to rely on PSR's database to maintain consistency with
the projections of forklift engines used in other sections of this rule's analysis. Based on the PSR
database, we have determined that approximately 50 percent of the population of Large SI
engines are used in the production of forklifts. This quantity of engines is taken as a measure of
the quantity of forklifts sold, based on the assumption that each forklift contains one engine.
The PSR database does not separate the quantity of forklift engines that are produced and
used in the U.S. from those that are imported. In order to determine the share of imported
forklifts of this total, historical import quantities of forklifts were compared with domestically
produced quantities. On average, imported forklifts were equal to about 25 percent of forklifts
produced in the U.S. in the past 10 years. This information was used to partition the total
quantity of forklifts listed in the PSR database into the share of domestically produced forklifts
and the share of imports for the year 2000.
The price of forklifts used in the model is taken as the year 2000 price of a representative
model of Class 5 forklift. The year 2000 price of Nissan's JC50 pneumatic tire 1C engine forklift
was $26,380 and it is used as the nationwide market price of forklifts. It is acknowledged that
there are a variety of Class 4, 5, and 6 forklifts with varying prices. The range of prices of these
forklifts are discussed in Chapter 2. However, we require a single price to operationalize the
perfectly competitive national-level market model used to examine the economic impacts of this
rule on the U.S. forklift market.
The estimates of demand and supply elasticity have been discussed in detail in Section
9.5.2.2. A separate estimate of foreign supply elasticity has not been carried out. For modeling
purposes, we assume that the foreign supply elasticity is equal to the domestic supply elasticity.
9.7.2 Forklift Control Costs
The emissions control costs used in the economic analysis are developed and reported in
Chapter 5. In this section, we briefly recount the estimated regulatory cost per forklift that are
used to in the model. The regulatory cost per unit faced by forklift producers leads to a parallel
shift inward of the market supply curve. As stated earlier, the compliance costs per forklift are
projected to change in future years as different phases of the emissions control program are
implemented and as the learning curve is applied (see Chapter 5 for a discussion of the learning
curve). The regulatory cost per forklift are presented in Table 9.7-2 for the years in which they
change.
9-66
-------
Chapter 9: Economic Impact Analysis
Table 9.7-2
Regulatory Costs Per Forklift
Year
2004/5
2006
2007/8
2009/10/11
2012-2030
Cost Per Forklift
$610
$493
$537
$418
$390
Cost Description
Phase I/year 1 costs
Phase I/year 3 costs
Phase I/year 3 costs
Phase I/year 6 costs
Phase I/year 6 costs
+ Phase 2/year 1 costs
+ Phase 2/year 3 costs
+ Phase 2/year 6 costs
Economic impacts are estimated based upon these costs. In the model, the baseline year
quantity and price of forklifts are used and the per unit costs are imposed on the model to
determine price, quantity, and consumer and producer surplus changes.
9.7.3 Forklift Economic Impact Results
The economic impacts of the regulation on the forklift market are estimated for each year
in which the per engine regulatory costs change, assuming the baseline year 2000 price and
quantity. We possess projected quantities of forklifts through the year 2030, however we do not
have projected future year prices. Without this information, we cannot estimate the economic
impacts of the future costs assuming future year quantities and prices. We instead rely upon the
most current year of data to inform the model when we impose the future costs per forklift on
producers. Using baseline year data allows us to estimate relative changes in price and quantity
as opposed to absolute changes. The estimated percent changes in price and quantity, the losses
in consumer and producer surplus, and total surplus loss are presented for various years in Tables
9.7-3 and 9.7-4. These results do not account for fuel cost savings that may arise from this
emissions control program.
Table 9.7-3
Price and Quantity Changes for the Forklift Market*
Impact Measure
Cost Per Unit
Change in Market Price
Change in Market Output
Domestic
Foreign
2004/5
$610
0.75%
-1.12%
-1.12%
-1.12%
2006
$493
0.60%
-0.90%
-0.90%
-0.90%
2007/8
$537
0.66%
-0.98%
-0.98%
-0.98%
2009
$418
0.51%
-0.77%
-0.77%
-0.77%
2012
$390
0.48%
-0.72%
-0.72%
-0.72%
*Results are the same for the years 2004 and 2005, 2007 and 2008, and the years 2009, 2010, and 2011. They are
also the same for the years 2012 and beyond. These results are not reported in separate columns to avoid repetition.
Results are based on baseline year 2000 market conditions and fuel cost savings are not included.
9-67
-------
Draft Regulatory Support Document
Table 9.7-4
Annual Losses in Consumer and Producer Surplus for the Forklift Market*
Impact Measure
LossinCS**($103)
Loss in PS***
Domestic
Foreign
Loss in Surplus
($103)
($103)
2004/5
$12,715
$26,412
$19,809
$6,603
$39,127
3
4
3
1
7
2006
$10,287.6
$21,416
$16,062
$5,354
$31,703
3
2
1
9
2007/8
2009
$11,201.2
$23,299
$17,474
$5,824
$34,500
1
3
8
3
$8
$18
$13
$4
$26
,728.6
,196.2
,647.2
,549.0
,924.8
2012
$8
$16
$12
$4
$25
,146.0
,990.5
,742.9
,247.6
,136.5
*Results are the same for the years 2004 and 2005, 2007 and 2008, and the years 2009, 2010, and 2011. They are also
the same for the years 2012 and beyond. These results are not reported in separate columns to avoid repetition. Results
are based on baseline year 2000 market conditions and fuel cost savings are not included.
** CS refers to consumer surplus and is rounded to the nearest hundredths. For a description of the change in consumer
surplus, see Section 9.2.2
*** PS refers to producer surplus and is rounded to the nearest hundredths. For a description of the change in producer
surplus, see Section 9.2.2.
For the per forklift engine costs resulting from the implementation of the emissions
control program, the relative increases in price over the future time period examined are three-
quarters of one percent or less. By the year 2014, the relative price increase falls to
approximately one-half of one percent. The percent reductions in the market quantity of forklifts
are initially projected to be slightly greater than one percent, but by 2006, the relative reduction
in market quantity falls below one percent. Though these impacts are larger than those in the
inboard diesel cruiser market, they are still considered minimal. Note that the percent reduction
in quantity is the same for both domestic and foreign output because it has been assumed that
domestic and foreign supply have the same price elasticity.
Table 9.7-4 above presents the loss in consumer surplus, the loss in producer surplus, and
the total loss in surplus (equal to the sum of the changes in consumer and producer surplus)
without fuel cost savings. As the table shows, the consumer surplus loss is approximately half
the size of the loss in producer surplus. Consumer surplus losses range from $12.7 million in
year 2004 when the rule is first implemented to $8.1 million in 2012 and the years beyond
through 2030. The losses in producer surplus are at their largest at $26.4 million in the first year
of implementation and they reach their lowest value in 2012 and the years beyond at just below
$17 million. Note that the annual surplus loss associated with the forklift market declines as the
per forklift engine costs fall. Loss in surplus is equal to $39.1 million in 2004 and it falls to
$25.1 million by 2012.
9-68
-------
Chapter 9: Economic Impact Analysis
9.7.4 Forklift Engineering Cost and Surplus Loss Comparison
This section presents a comparison of the future stream of engineering costs (excluding
fuel cost savings) and surplus losses for the forklift market. In Table 9.7-5, we first present an
interim comparison of the estimated engineering costs, holding quantity constant to the baseline
year quantity, with the surplus losses that were estimated from the economic impact model.
Because economic modeling takes into account consumer and producer behavior, the estimated
loss in surplus is less than the engineering costs under a perfectly competitive market setting. In
this case, the annual surplus losses are, on average, equal to 98 to 99 percent of the calculated
engineering costs. The cost numbers in this table and Table 9.7-6 are not discounted.
Based upon a ratio of the loss in surplus to engineering costs, holding baseline quantity
constant, a projection of the surplus loss over the future year stream is calculated from the future
stream of engineering costs that appear in Chapter 7. This projection of the future stream of
surplus losses is compared to the future stream of engineering costs in Table 9.7-6. Note that
these results are not discounted nor do they account for fuel cost savings.
9.7.5 Forklift Economic Impact Results with Fuel Cost Savings
In Table 9.7-7, the social costs/gains are calculated by adding the annual savings in fuel
costs (presented initially in Chapter 7) to the projected annual surplus loss. These social gains
are compared to the engineering costs with fuel efficiency gains. As you can see from this table,
the emissions control program is expected to yield social gains rather than losses beyond the
initial year of implementation. Only the initial year of implementation results in a social loss
from this regulation for the forklift market.
9-69
-------
Draft Regulatory Support Document
Interim Engineering
Forklift Market Based
Table 9.7-5
Cost and Surplus Loss Comparison
on Year 2000 Quantity (Q = 65,000
for the
forklifts)
Year Estimated Engineering Costs Estimated Surplus Loss
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
$39,645,853
$39,645,853
$32,047.483
$34,914,619
$34,914,619
$27,143,050
$27,143,050
$27,143,050
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$25,329,069
$39,127,756
$39,127,756
$31,703,880
$34,500,273
$34,500,273
$26,924,774
$26,924,774
$26,924,774
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
$25,136,527
9-70
-------
Chapter 9: Economic Impact Analysis
Engineering Cost
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Table 9.7-6
and Surplus Loss Comparison for
without Fuel Cost Savings
Estimated Engineering Costs
$44,403,355
$45,592,731
$37,816,030
$42,246,689
$43,294,128
$34,471,674
$35,285,965
$36,100,257
$345447,534
$35,207,406
$35,967,278
$36,727,150
$37,487,022
$38,246,894
$39,006,766
$39,766,638
$40,526,510
$41,286,382
$42,046,254
$42,806,126
$43,565,998
$44,325,871
$45,085,743
$45,845,615
$46,605,487
$47,365,359
$48,125,231
the Forklift Market
Estimated Surplus Loss
$43,823,087
$44,996,919
$37,410,578
$41,745,330
$42,780,339
$34,194,463
$35,002,206
$35,809,949
$34,185,677
$34,939,773
$34,693,868
$36,447,964
$37,202,060
$37,956,156
$38,710,252
$39,464,347
$40,218,443
$40,972,539
$41,726,635
$42,480,731
$43,234,826
$43,988,922
$44,743,018
$45,497,114
$46,251,210
$47,005,305
$47,759,401
9-71
-------
Draft Regulatory Support Document
Table 9.7-7
Engineering and Social Cost Comparison
for the Forklift Market with Fuel Cost Savings
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Engineering Costs
with Fuel Cost Savings
$7,305,024
($29,112.307)
($74,949,193)
($107,719,996)
($142,910,106)
($186,910,292)
($220,128,020)
($248,696,789)
($263,429,050)
($273,365,256)
($282,258,050)
($290,155,574)
($297,059,701)
($303,544,978)
($309,618,970)
($315,291,768)
($320,384,517)
($325,478,111)
($330,572,494)
($336,095,973)
($341,680,638)
($347,267,003)
($352,193,263)
($357,123,770)
($362,058,551)
($366,996,593)
($371,938,165)
Estimated Social Costs/Gains
(Surplus Loss - Fuel Savings)*
$6,724,756
($29,708,119)
($75,354,645)
($108,221,355)
($143,423,895)
($187,187,502)
($220,411,779)
($248,987,097)
($263,690,906)
($273,632,888)
($282,531,460)
($290,434,760)
($297,344,663)
($303,835,716)
($309,915,484)
($315,594,059)
($320,692,585)
($325,791,955)
($330,892,113)
($336,421,369)
($342,011,810)
($347,603,952)
($352,535,988)
($357,472,271)
($362,412,827)
($367,356,646)
($372,303,995)
( ) represents a negative cost (social gain). Cost estimates are based upon 2000$.
9-72
-------
Chapter 9: Economic Impact Analysis
9.7.6 Economic Impacts - Other Large SI Engines
To complete the analysis of the economic impacts of this rulemaking on Large SI engines,
we used engineering costs as a surrogate for consumer and producer surplus losses. As noted
above, this approach slightly overestimates the surplus losses, suggesting that the standards will
have a slightly larger total impact on consumers and producers. This approach does not allow
disaggregating to determine the portion of the costs borne by consumers and the portion borne by
producers. The estimated fuel cost savings for Large SI engines other than forklifts are based on
the methodology used for forklifts. The results of this analysis are contained in Table 9.7-8.
According to this analysis, the emissions control program is expected to yield social gains rather
than losses beyond the first two years of implementation.
9-73
-------
Draft Regulatory Support Document
Table 9.7-8
Engineering Cost and Surplus Loss Comparison
Large SI Engines Other Than Forklifts
for
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Surplus
Loss (Engineering
Costs)
$44,403,355
$45,592,731
$37,816,030
$42,246,689
$43,294,128
$34,471,674
$35,285,965
$36,100,257
$34,447,534
$35,207,406
$35,967,278
$36,727,150
$37,487,022
$38,246,894
$39,006,766
$39,766,638
$40,526,510
$41,286,382
$42,046,254
$42,806,126
$43,565,998
$44,325,871
$45,085,743
$45,845,615
$46,605,487
$47,365,359
$48,125,231
Estimated Fuel
Savings
($15,627,144)
($28,275,848)
($40,160,970)
($48,976,681)
($56,624,806)
($63,712,068)
($70,327,718)
($76,172,728)
($81,521,871)
($86,460,491)
($90,759,859)
($94,347,999)
($97,888,686)
($101,329,714)
($104,666,222)
($107,916,691)
($111,080,698)
($114,155,459)
($117,123,427)
($117,123,427)
($122,621,375)
($125,268,725)
($128,102,036)
($130,896,877)
($133,533,546)
($135,988,425)
($138,409,359)
Estimated Social
Costs/Gains
(Surplus Loss - Fuel
Savings)*
$28,776,211
$17,316,883
($2,344,940)
($6,729,992)
($13,330,678)
($29,240,394)
($35,041,753)
($40,072,471)
($47,074,337)
($51,253,085)
($54,792,581)
($57,620,849)
($60,401,664)
($63,082,820)
($65,659,456)
($68,150,053)
($70,554,188)
($72,869,077)
($75,077,173)
($74,317,301)
($79,055,377)
($80,942,854)
($83,016,293)
($85,051,262)
($86,928,059)
($88,623,066)
($90,284,128)
9-74
-------
Chapter 9: Economic Impact Analysis
9.8 Snowmobiles
The following section describes the baseline characterization of the snowmobile market
in the year 2001, the regulatory control costs incurred by producers of snowmobiles, and the
economic impacts that would have resulted had the emissions control program been imposed in
the baseline year. We examine the economic impacts on the snowmobile market using the
baseline year data for each change in the per unit control costs that occurs. A comparison is then
made between the engineering cost and surplus loss streams projected to occur after the
regulation's implementation. This initial comparison of the cost streams assumes no fuel cost
savings. A comparison is then made between engineering costs and social costs/gains accounting
for fuel cost savings of the emissions control program.
9.8.1 Snowmobile Baseline Market Characterization
Inputs to the economic analysis are provide a baseline characterization for the
snowmobile market for the year 2001. Baseline market data include the domestic quantity
produced, quantity of imports, baseline market price, demand elasticity, and domestic and foreign
supply elasticity measures. Table 9.8-1 provides the baseline data for the U.S. snowmobile
market used in this analysis.
Table 9.8-1
Baseline Characterization of the U.S. Snowmobile Market: 20016465
Inputs Baseline Observation
Market price ($/snowmobile) $6,360.00
Market output (snowmobiles) 140,629
Domestic 80,015
Foreign 60,614
Elasticities
Domestic supply (estimated) 2.1
Foreign supply (assumed) 2.1
Demand (assumed) -2
The market sales and quantity data are available from the ISMA website. Import and
export estimates are based upon data from the PSR. PSR lists vehicles that are imports. For the
year 2000, approximately 60 percent of snowmobiles produced by the 4 largest producers were
produced domestically by Polaris and Arctic Cat. It is assumed that the production relationship
between imports and exports is mirrored in sales for 2001. Based upon this import ratio, we
estimate that approximately 61 thousand of the snowmobiles sold in the US in 2001 were
9-75
-------
Draft Regulatory Support Document
imported.
The estimates of demand and supply elasticity have been discussed in detail in Section
9.5.2.3. A separate estimate of foreign supply elasticity has not been carried out. For modeling
purposes, we assume that the foreign supply elasticity is equal to the domestic supply elasticity.
It is important to note that imports and domestically produced vehicles must meet the US
emission standards in order to be sold in this country.
9.8.2 Snowmobile Control Costs
The emissions control costs used in the economic analysis are developed and reported in
Chapter 5. In this section, we briefly recount the estimated regulatory cost per snowmobile that
are used in the model. The regulatory cost per unit faced by snowmobile producers leads to a
parallel shift inward of the market supply curve. As stated earlier, the compliance costs per
snowmobile are projected to change in future years as different phases of the emissions control
program are implemented and as the learning curve is applied (see Chapter 5 for a discussion of
the learning curve). The regulatory cost per snowmobile are presented in Table 9.8-2 for the
years in which they change.
Table 9.8-2
Regulatory Costs Per Snowmobile
Year
2006
2007
2008-2009
2010
2011
2012
2013
2014
2015
2016
2017-2030
Cost Per
Snowmobile
$35
$69
$65
$185
$181
$239
$239
$202
$196
$182
$180
Cost Description
Phase I/year 1 costs
Phase I/year 2 costs
Phase I/year 3 and 4 costs
Phase 2/year 1 costs
Phase 2 /year 2 costs
Phase 3 /year 1 costs
Phase 3/year 2 costs
Phase 3/year 3 costs
Phase 3/year 4 costs
Phase 3/year 5 costs
Phase 3/year 6 and years thereafter costs
Economic impacts are estimated based upon these costs. In the model, the baseline year
quantity and price of snowmobiles are used and the per unit costs are imposed on the model to
determine price, quantity, and consumer and producer surplus changes.
9-76
-------
Chapter 9: Economic Impact Analysis
9.8.3 Snowmobile Economic Impact Results
The economic impacts of the regulation on the snowmobile market are estimated for each
year in which the per engine regulatory costs change, assuming the baseline year 2001 price and
quantity. We possess projected quantities of snowmobiles through the year 2030, however we do
not have projected future year prices. Without this information, we cannot estimate the
economic impacts of the future costs assuming future year quantities and prices. We instead rely
upon the most current year of data to inform the model when we impose the future costs per
snowmobile on producers. Using baseline year data allows us to estimate relative changes in
price and quantity as opposed to absolute changes. The estimated percent changes in price and
quantity, the losses in consumer and producer surplus, and total surplus loss are presented for
various years in Tables 9.8-3 and 9.8-4. These results do not account for fuel cost savings that
may arise from this emissions control program.
Table 9.8-3
Price and Quantity Changes for the Snowmobile Market*
Impact
Measure
Cost Per Unit
Change
in Price
Change in
Output:
2006
$35
0.28%
-0.56%
2007
$69
0.56%
-1.11%
2008-
2009
$65
0.52%
-1.05%
2010
$185
1.49%
-2.98%
2011
$181
1.46%
-2.92%
2012-
2013
$239
1.92%
-3.85%
2014
$202
1.63%
-3.25%
2015
$196
1.58%
-3.16%
2016
$182
1.47%
-2.93%
2017-
2030
$180
1.450/
-2.90/
*Based upon 2001baseline market conditions and impacts estimated to occur from the regulation. Assumes 2001$.
9-77
-------
Draft Regulatory Support Document
Table 9.8-4
Annual Losses in Consumer and Producer Surplus for the Snowmobile Market*
Impact Measure
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
LossmCS**($103)
LossmPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
Year
2006
$2,513.9
$2,380.7
$1,354.6
$1,026.1
$4,894.6
2011
$12,847.3
$11,873.5
$6,755.8
$5,117.7
$24,720.8
2016
$12,917.2
$11,936.1
$6,791.4
$5,144.7
$24,853.3
2007
$4,942.4
$4,654.5
$2,648.3
$2,006.2
$9,596.9
2012-2013
$16,883.7
$15,448.6
$8,798.9
$6,658.7
$32,332.3
2017-2030
$12,777.4
$11,810.9
$6,720.2
$5,090.8
$24,588.3
2008-2009
$4,657.4
$4,338.9
$2,497.2
$1,891.7
$9,049.4
2014
$14,313.3
$13,180.8
$7,499.6
$5,681.2
$27,494.1
2010
$13,126.9
$12,123.7
$6,898.1
$5,225.6
$25,250.6
2015
$13,894.9
$12,808.8
$7,287.9
$5,520.9
$26,703.7
* Based upon 2001 baseline market conditions and the impact of the regulations on those market conditions. Assumes
2001$.
** CS refers to consumer surplus and is rounded to the nearest hundredths. For a description of the change in consumer
surplus, see Section 9.2.2
*** PS refers to producer surplus and is rounded to the nearest hundredths. For a description of the change in producer
surplus, see Section 9.2.2.
For the per snowmobile engine costs resulting from the implementation of the emissions
control program, the relative increases in price over the future time period examined ranges from
0.28% to approximately 1.92% and achieve a steady state in 2017 of approximately 1.45%. The
percent reductions in the market quantity of snowmobiles are initially projected to be 0.28% but
increase to around 3.85% in 2012, the first year of the Phase 3 regulations. The steady state
quantity reductions begin in 2017 and are approximately 2.9%. The percentage change in
9-78
-------
Chapter 9: Economic Impact Analysis
domestic and foreign production are the same. This is based upon the assumption that the
foreign price elasticity of demand is equivalent to the domestic price elasticity of demand, and
the fact that both foreign and domestic snowmobiles are subject to the emission standards. All
price quantity change estimates are based upon 2001 baseline market conditions and the impact
of the regulation on those baseline market conditions.
Table 9.8-4 above presents the loss in consumer surplus, the loss in producer surplus, and
the total loss in surplus (equal to the sum of the changes in consumer and producer surplus)
without fuel cost savings. As the table shows, the consumer surplus loss is approximately half
the size of the loss in producer surplus. Producer surplus losses range from $2.4 million to
$15.4 million in 2012 and reach a steady state value of $11.8 million in 2017 and beyond. The
losses in consumer surplus range from $2.5 to $16.9 million and reach a steady state of $12.8 in
2017. Note that the annual surplus loss associated with the snowmobile market increases as the
per snowmobile engine costs increase and declines as the per snowmobile engine costs fall.
Annual loss in surplus ranges from $4.9 million to $32.3 million in 2010 and decrease to a steady
state level in 2017 of $24.6 million. It is important to note that these estimates are based upon
2001 baseline conditions and the impact of the regulation on those market conditions.
9.8.4 Snowmobile Engineering Cost and Surplus Loss Comparison
This section presents a comparison of the future stream of engineering costs (excluding
fuel cost savings) and surplus losses for the snowmobile market. In Table 9.8-5, we first present
an interim comparison of the estimated engineering costs, holding quantity constant to the
baseline year quantity. The surplus losses are estimated from the economic impact model.
Because economic modeling takes into account consumer and producer behavior, the estimated
loss in surplus is less than the engineering costs under a perfectly competitive market setting. In
this case, the annual surplus losses are, on average, equal to 96 to 99 percent of the calculated
engineering costs. It is important to note that the relationship between engineering and economic
costs are based upon this comparison. It is the relationship between these costs that are assumed
to actually occur in the market in future years. The cost numbers in Table 9.8-5 and 9.8-6 are not
discounted.
Based upon a ratio of the loss in surplus to engineering costs, holding baseline quantity
constant, a projection of the surplus loss over the future year stream is calculated from the future
stream of engineering costs that appear in Chapter 7. This projection of future stream of
engineering costs is based upon projected snowmobiles sales provided by ISMA and estimated
per unit engineering engine modification costs. This projection of the future stream of surplus
losses is compared to the future stream of engineering costs in Table 9.8-6. Note that these
results are not discounted nor do they account for fuel cost savings. The relationship between
engineering costs and surplus losses are determined using the market model are assumed to occur
in future years. Thus the engineering costs and surplus losses shown in Table 9.8-6 are based
upon forecasted sales volumes in the future, the engineering cost estimate for those sales.
9-79
-------
Draft Regulatory Support Document
Surplus losses represent the estimated value of those losses as informed by the market model, but
accounting for projected sales growth in the future.
Table 9.8-5
Interim Engineering Cost and Surplus Loss Comparison for the
Snowmobile Market Based on Year 2001 Baseline Market Conditions
(millions of 2001 $)
Year
2006
2007
2008 - 2009 (annually)
2010
2011
20 12 -20 13 (annually)
2014
2015
2016
20 17 -2030 (annually)
Estimated Engineering Costs
$4.9
$9.7
$9.1
$26.0
$25.5
$33.6
$28.4
$27.6
$25.6
$25.3
Estimated Surplus Loss
$4.9
$9.6
$9.0
$25.2
$24.7
$32.3
$27.5
$26.7
$24.9
$24.6
9.8.5 Snowmobile Economic Impact Results with Fuel Cost Savings
In Table 9.8-7, the social costs/gains are calculated by adding the annual savings in fuel
costs (presented initially in Chapter 7) to the projected annual surplus loss. These social gains
are compared to the engineering costs with fuel efficiency gains. As you can see from this table,
the emissions control program is expected to yield social gains rather than losses beyond the year
2014.
9-80
-------
Chapter 9: Economic Impact Analysis
Engineering Cost
without Fuel
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020-2030
Table 9.8-6
and Surplus Loss Comparison for the Snowmobile Market
Cost Savings Assumes Sales Growth in Future Years*
(millions of 2001 $)
Estimated Engineering Costs
$6.6
$13.5
$13.2
$13.5
$38.9
$38.7
$52.0
$52.7
$45.3
$44.4
$41.9
$41.7
$42.2
$42.7
$43.1
Estimated Surplus Loss
$6.5
$13.4
$13.0
$13.3
$37.8
$37.6
$50.0
$50.7
$43.9
$43.0
$40.6
$40.5
$41.0
$41.5
$41.9
Snowmobile sales growth provided by ISMA. Sales are not projected to grow after 2020.
9-81
-------
Draft Regulatory Support Document
Table 9.8-7
Engineering and Social Cost Comparison for the Snowmobile Market
with Fuel Cost Savings - Assumes Sales Growth In Future Years*
(millions of 2001$)
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Engineering Costs
with Fuel Cost Savings
$6.2
$12.3
$10.7
$9.7
$29.4
$23.1
$26.9
$17.8
$0.4
($10.5)
($23.2)
($33.2)
($42.3)
($50.9)
($59.0)
($67.0)
($73.5)
($78.4)
($82.0)
($84.5)
($86.5)
($88.3)
($89.8)
($90.9)
($91.8)
Estimated Social Costs/Gains
(Surplus Loss - Fuel Savings)*
$6.2
$12.1
$10.6
$9.6
$28.2
$21.9
$24.9
$15.8
($1.0)
($12.0)
($24.4)
($34.4)
($43.5)
($52.1)
($60.3)
($68.3)
($74.8)
($79.6)
($83.3)
($85.8)
($87.8)
($89.5)
($91.0)
($92.2)
($93.2)
: () represents a negative cost (social gain). Cost estimates are based upon 2001$
9.8.6 Economic Impacts on Individual Engine Manufacturers, Snowmobile Retailers and
Snowmobile Rental Firms
Insufficient data were obtained to conduct an analysis of the impact of the regulation on
individual producers in the market. Thus, this analysis does not address individual producer
9-82
-------
Chapter 9: Economic Impact Analysis
impacts. Each snowmobile manufacturer must meet the emission standards for vehicles sold
domestically. Since Yamaha and Bombardier produce their own engines, it is possible that these
firms may be at a competitive advantage relative to Arctic Cat and Polaris who purchase engines
from other firms. No analysis has been conducted to determine the impact of the difference in
cost of production or cost of compliance for the individual firms within the industry. The EPA
sought information concerning individual firm's cost of producing snowmobiles, but was unable
to obtain sufficient data to conduct an analysis.
With regard to snowmobile retail and rental firms. To the extent that the price of
snowmobiles increases, these firms will be impacted by the regulation The increase in market
price estimated for the steady state of 1.45% does not appear sufficient to create significant
impacts for these firms. In addition, most retail firms sell a variety of products, and snowmobiles
are only one product in their product line. This will tend to mitigate the impact for these firms.
9.9 All-Terrain Vehicles (ATVs)
9.9.1 ATV Baseline Market Characterization
Inputs to the economic analysis are for the year 2001. Baseline characterization of the
ATV market includes the domestic quantity of ATVs produced, quantity of imports, baseline
market price, demand elasticity, and domestic and foreign supply elasticity measures. Table 9.9-
1 provides the baseline data on the U.S. ATV market used in this analysis.
Table 9.9-1
Baseline Characterization of the U.S. ATV Market: 2001
Inputs Baseline Observation
Market price ($/ATV) $5,123.00
Market output (ATV) 880000
Domestic 874746
Foreign 5254
Elasticities
Domestic supply (assumed) 1
Foreign supply (assumed) 1
Demand (assumed) -2
The total quantity of ATVs sold in the U.S. was retrieved from the MIC. Trade data
specific to the ATV market were unavailable. However, the International Trade Commission
publishes international trade data for NAICS code 336999 - Other Transportation Equipment.
9-83
-------
Draft Regulatory Support Document
According to ITC data, imports for NAICS code 336999 account for less than 1 percent of
domestic sales. The import ratio for Other Transportation Equipment is assumed to be a
reasonable proxy for imports for the ATV market.
The price of ATVs used in the model is the average ATV price in 2001 provided by MIC.
An average ATV market price is required to operationalize the perfectly competitive national-
level market model used to examine the economic impacts of this rule on the U.S. ATV market.
The estimates of demand and supply elasticity have been discussed in detail in Section
9.5.2.4. A separate estimate of foreign supply elasticity has not been carried out. For modeling
purposes, we assume that the foreign supply elasticity is equal to the domestic supply elasticity.
9.9.2 ATV Control Costs
The emission control costs used in the economic analysis are developed and reported in
Chapter 5. In this section, we briefly recount the estimated regulatory cost per ATV that are used
in the model. The regulatory cost per unit faced by ATV producers leads to a parallel shift
inward of the market supply curve. As stated earlier, the compliance costs per ATV are projected
to change in future years as different phases of the emissions control program are implemented
and as the learning curve is applied (see Chapter 5 for a discussion of the learning curve). The
regulatory cost per ATV are presented in Table 9.9-2 for the years in which these costs change.
Table 9.9-2
Regulatory Costs Per ATV
Year
2006
2007
2008
2009
2010
2011
2012-2015
2016
2017-2030
Cost Per ATV
$43
$82
$78
$71
$66
$57
$53
$51
$48
Cost Description
Phase I/year 1 costs
Phase I/year 2 costs
Phase I/year 3 costs
Phase I/year 4 costs
Phase I/year 5 costs
Phase I/year 6 costs
Phase I/year 7-10 costs
Phase I/year 1 1 costs
Phase I/year 12-25 costs
Economic impacts are estimated based upon these costs. In the model, the baseline year
quantity and price of ATVs are used and the per unit costs are imposed on the model to
determine price, quantity, and consumer and producer surplus changes.
9-84
-------
Chapter 9: Economic Impact Analysis
9.9.3 ATV Economic Impact Results
The economic impacts of the regulation on the ATV market are estimated for each year in
which the per engine regulatory costs change, assuming the baseline year 2001 price and
quantity. Estimated projected quantities of ATVs sales through the year 2030 are available,
however we do not have projected future year prices. Any price projections would be subject to
significant uncertainties. Without this information, we cannot estimate the economic impacts of
the future costs assuming future year quantities and prices. We instead rely upon the most
current year of data to inform the model when we impose the future costs per ATV on producers.
Assuming annual sales and average prices are increasing for ATVs, this model approach tends to
overstate potential price and quantity impacts. Using baseline year data allows us to estimate
relative changes in price and quantity as opposed to absolute changes. The estimated percent
changes in price and quantity, the losses in consumer and producer surplus, and total surplus loss
are presented for various years in Tables 9.9-3 and 9.9-4. These results do not account for fuel
cost savings that may arise from this emissions control program.
Table 9.9-3
Price and Quantity Changes for the ATV Market*
Impact Measure
Cost Per Unit
Change in Market Price
Change in Market Output
Domestic
Foreign
Cost Per Unit
Change in Market Price
Change in Market Output
Domestic
Foreign
Year
2006
$43
0.28%
-.56%
-.56%
-.56%
2011
$57
0.37%
-.74%
-.74%
-.74%
2007
$82
0.53%
-1.07%
-1.07%
-1.07%
2012/2015
$53
0.34%
-.69%
-.69%
-.69%
2008
$78
0.51%
-1.02%
-1.02%
-1.02%
2016
$51
0.33%
-0.66%
-0.66%
-0.66%
2009
$71
0.46%
-.92%
-.92%
-.92%
2017/2030
$48
0.31%
-0.62%
-0.62%
-0.62%
2010
$66
0.43%
-.86%
-.86%
-.86%
*Results are the same for the years 2012 through 2015 and for 2017 through 2030. These results are not reported in
separate columns to avoid repetition. Results are based on baseline year 2001 market conditions and fuel cost savings are
not included.
9-85
-------
Draft Regulatory Support Document
Table 9.9-4
Annual Losses in Consumer and Producer Surplus for the ATV Market*
Impact Measure
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
Year
2006
$12,578.0
$25,015.0
$24,865.6
$149.4
$37,593.0
2011
$16,658.0
$33,068.0
$32,870.5
$197.4
$49,726.0
2007
$23,925.0
$47,336.7
$47,054.0
$282.6
$71,261.7
2012-2015
$15,493.0
$30,771.7
$30,587.9
$183.7
$46,264.7
2008
$22,763.9
$45,063.3
$44,794.2
$269.1
$67,827.2
2016
$14,910.4
$29,622.1
$29,445.3
$176.9
$44,532.5
2009
$20,730.5
$41,076.0
$40,830.8
$245.2
$61,806.5
2017-2030
$14,036.0
$27,896.2
$27,729.6
$166.6
$41,932.2
2010
$19,276.9
$38,221.2
$37,993.0
$228.2
$57,498.0
*Results are based on baseline year 2001 market conditions and fuel cost savings are not included.
** CS refers to consumer surplus and is rounded to the nearest hundred. For a description of the change in consumer
surplus, see Section 9.2.2
*** PS refers to producer surplus and is rounded to the nearest hundred. For a description of the change in producer
surplus, see Section 9.2.2.
For the per ATV engine costs resulting from the implementation of the emissions control
program, the relative increases in price over the future time period examined are one-half of one
percent or less. The market quantity reductions are estimated to be approximately one percent or
less and reach a steady state decrease of 0.62 percent in 2017. Note that the percent reduction in
quantity is the same for both domestic and foreign output because it has been assumed that
domestic and foreign supply have the same price elasticity.
Table 9.9-4 above presents the loss in consumer surplus, the loss in producer surplus, and
the total loss in surplus (equal to the sum of the changes in consumer and producer surplus)
without fuel cost savings. As the tables show, the consumer surplus loss is approximately half
the size of the loss in producer surplus. Consumer surplus losses range from nearly $12.6 million
in year 2006 when the rule is first implemented, it rises to $23.9 million in 2007 and falls to $14
million in 2017 and the years beyond. The losses in producer surplus range from $25 million in
the first year of implementation, rising to $47.3 million in 2007 and falls to $27.9 million in 2012
and the years beyond. Note that the annual surplus loss associated with the ATV market declines
9-86
-------
Chapter 9: Economic Impact Analysis
as the per ATV engine costs fall starting in 2008. Loss in surplus is equal to $37.6 million in
2006, rises to 71.3 in 2007 and it falls to $42 million by 2017. The surplus estimate presented in
Table 9.9-4 is based upon 2001 baseline market conditions and do not consider fuel cost savings.
9.9.4 ATV Engineering Cost and Surplus Loss Comparison
This section presents a comparison of the future stream of engineering costs (excluding
fuel cost savings) and surplus losses for the ATV market. In Table 9.9-5, we first present an
interim comparison of the estimated engineering costs, holding quantity constant to the baseline
year quantity, with the surplus losses that were estimated from the economic impact model.
Because economic modeling takes into account consumer and producer behavior, the estimated
loss in surplus is less than the engineering costs under a perfectly competitive market setting. In
this case, the annual surplus losses are, on average, equal to 98 to 99 percent of the calculated
engineering costs. The cost numbers in Table 9.9-5 are not discounted.
Based upon a ratio of the loss in surplus to engineering costs, holding baseline quantity
constant, a projection of the surplus loss over the future year stream is calculated from the future
stream of engineering costs that appear in Chapter 7. This projection of the future stream of
surplus losses is compared to the future stream of engineering costs in Table 9.9-6. Note that
these results are not discounted nor do they account for fuel cost savings.
Table 9.9-5
Interim Engineering Cost and Surplus Loss Comparison for the
ATV Based on Year 2001 Quantity (Q = 880,000 ATV)*
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017-2030
Estimated Engineering Costs
$37,840.0
$72,160.0
$68,640.0
$62,480.0
$58,080.0
$50,160.0
$46,640.0
$46,640.0
$46,640.0
$46,640.0
$44,880.0
$42,240.0
Estimated Surplus Loss
$37,593.0
$71,261.7
$67,827.2
$61,806.5
$57,498.0
$49,726.0
$46,264.7
$46,264.7
$46,264.7
$46,264.7
$44,532.5
$41,932.2
*Estimates are based on baseline year of 2001 and reflect 2001 dollars.
9-87
-------
Draft Regulatory Support Document
Table 9.9-6
Engineering Cost and Surplus Loss Comparison for the ATV Market
without Fuel Cost Savings (Q = ATV projected sales for 2006 through 2030)*
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017-2030
Estimated Engineering Costs
$42,463.9
$81,270.6
$76,518.0
$70,287.0
$65,302.2
$56,379.5
$52,441.5
$52,441.5
$52,441.5
$52,441.5
$50,000.0
$47,556.8
Estimated Surplus Loss
$42,186.6
$80,258.8
$75,611.8
$69,529.4
$64.681.3
$55,891.6
$52,019.5
$52,019.5
$52,019.5
$52,019.5
$49,612.0
$47,210.3
*Estimates reflect growth in sales projected in the future and are based on 2001 dollars.
9.7.5 ATV Economic Impact Results with Fuel Cost Savings
In Table 9.9-7, the social costs/gains are calculated by adding the annual savings in fuel
costs (presented initially in Chapter 7) to the projected annual surplus loss. These social gains
are compared to the engineering costs with fuel efficiency gains. As you can see from this table,
the emissions control program is expected to yield social gains rather than losses beginning in
2019.
-------
Chapter 9: Economic Impact Analysis
Table 9.9-7
Engineering and Social Cost Comparison for the ATV Market
with Fuel Cost Savings (Q = ATV projected sales for 2006 through 2030)
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Engineering Costs
with Fuel Cost Savings
$41,529.9
$77,878.5
$69,563.1
$59,363.1
$50,192.8
$36,888.3
$28,565.3
$24,252.7
$20,127.2
$16,223.2
$10,167.9
$4,433.1
$1,706.8
($109.4)
($1,283.9)
($2,083.2)
($2,577.5)
($2,951.6)
($3,234.2)
($3,443.4)
($3,596.0)
($3,707.7)
($3.786.4)
($3,842.7)
($3,881.4)
Estimated Social Costs/Gains
(Surplus Loss - Fuel Savings)*
$41,252.7
$76,563.7
$68,657.0
$58,605.5
$49,541.9
$36,400.4
$28,143.4
$23,830.7
$19,705.2
$15,801.2
$9,780.7
$4,086.6
$1,360.2
($456.0)
($1,630.4)
($2,429.8)
($2,924.0)
($3,298.2)
($3,580.7)
($3,790.0)
($3,942.6)
($4,054.2)
($4,132.9)
($4,189.3)
($4,227.9)
* () represents a negative cost (social gain). Cost estimates are based upon 2001$
9-89
-------
Draft Regulatory Support Document
9.10 Off-Highway Motorcycles
9.10.1 Off-Highway Motorcycle Baseline Market Characterization
Inputs to the economic analysis are for the year 2001. Baseline characterization of the
off-highway motorcycle market includes the domestic quantity of off-highway motorcycles
produced, quantity of imports, baseline market price, demand elasticity, and domestic and foreign
supply elasticity measures. Table 9.10-1 provides the baseline data on the U.S. off-highway
motorcycle market used in this analysis.
Table 9.10-1
Baseline Characterization of the U.S. Off-Highway Motorcycle Market: 2001
Inputs Baseline Observation
Market price ($/off-highway motorcycle) $2,253.00
Market output (off-highway motorcycle) 195250
Domestic 82463
Foreign 112787
Elasticities
Domestic supply (estimated) 0.93
Foreign supply (assumed) 0.93
Demand (assumed) -2
The total quantity of off-highway motorcycle sold in the U.S. was obtained from the MIC
The quantity of imports of off-highway motorcycle from the International Trade Commission.
According to ITC data, imports for NAICS code 336991 account for nearly 58 percent of
domestic sales.
The price of off-highway motorcycles used is the average off-highway motorcycle price
in 2001 provide by MIC. An average off-highway motorcycle market price is required to
operationalize the perfectly competitive national-level market model used to examine the
economic impacts of this rule on the U.S. off-highway motorcycle market. The import ratios for
Motorcycles, Bicycles, and Parts Manufactures are assumed to be a reasonable proxy for off-
highway motorcycle imports.
The estimates of demand and supply elasticity have been discussed in detail in Section
9.5.2.5. A separate estimate of foreign supply elasticity has not been carried out. For modeling
purposes, we assume that the foreign supply elasticity is equal to the domestic supply elasticity.
9-90
-------
Chapter 9: Economic Impact Analysis
9.10.2 Off- Highway Motorcycle Control Costs
The emissions control costs used in the economic analysis are developed and reported in
Chapter 5. In this section, we briefly recount the estimated regulatory cost per off-highway
motorcycle that are used to in the model. The regulatory cost per unit faced by off-highway
motorcycle producers leads to a decrease in the market supply curve. As stated earlier, the
compliance costs per off-highway motorcycle are projected to change in future years as different
phases of the emissions control program are implemented and as the learning curve is applied
(see Chapter 5 for a discussion of the learning curve). The regulatory cost per off-highway
motorcycles are presented in Table 9.10-2 for the years in which they change.
Table 9.10-2
Regulatory Costs Per Off-Highway Motorcycle
Year
2006
2007
2008
2009
2010
2011
2012-2030
Cost Per Off-Highway
Motorcycle
$79
$155
$143
$128
$117
$102
$99
Cost Description
Phase I/year 1 costs
Phase I/year 2 costs
Phase I/year 3 costs
Phase I/year 4 costs
Phase I/year 5 costs
Phase I/year 6 costs
Phase I/year 7 costs
Economic impacts are estimated based upon these costs. In the model, the baseline year
quantity and price of off-highway motorcycle are used and the per unit costs are imposed on the
model to determine price, quantity, and consumer and producer surplus changes.
9.10.3 Off-Highway Motorcycles Economic Impact Results
The economic impacts of the regulation on the off-highway motorcycle market are
estimated for each year in which the per engine regulatory costs change, assuming the baseline
year 2001 price and quantity. Estimated projected quantities of off-highway motorcycle sales
through the year 2030 are available, however we do not have projected future year prices.
Without this information, we cannot estimate the economic impacts of the future costs assuming
future year quantities and prices. Any price projections would be subject to significant
uncertainties. We instead rely upon the most current year of data to inform the model when we
impose the future costs per off-highway motorcycle on producers. Assuming annual sales and
average prices are increasing for off-highway motorcycles, this model approach tends to overstate
the potential price and quantity impacts. Using baseline year data allows us to estimate relative
9-91
-------
Draft Regulatory Support Document
changes in price and quantity as opposed to absolute changes. The estimated percent changes in
price and quantity, the losses in consumer and producer surplus, and total surplus loss are
presented for various years in Tables 9.10-3. These results do not account for fuel cost savings
that may arise from this emissions control program.
Table 9.10-3
Price and Quantity Changes for the Off-Highway Motorcycle Market*
Impact Measure
Cost Per Unit
Change in Market Price
Change in Market Output
Domestic
Foreign
2006
$79
1.11%
-2.23%
-2.23%
-2.23%
2007
$155
2.18%
-4.37%
-4.37%
-4.37%
2008
$143
2.01%
-4.03%
-4.03%
-4.03%
2009
$128
1.80%
-3.61%
-3.61%
-3.61%
2010
$117
1.65%
-3.30%
-3.30%
-3.30%
2011
$102
1.44%
-2.87%
-2.87%
-2.87%
2012-
2030
$99
1.39%
-2.79%
-2.79%
-2.79%
*Results are the same for the years 2012 through 2030. These results are not reported in separate columns to avoid
repetition. Results are based on baseline year 2001 market conditions and fuel cost savings are not included.
For the per off-highway motorcycle engine costs resulting from the implementation of the
emissions control program, the relative increases in price over the future time period examined
are 2.18 percent or less. By the year 2012, the relative price increase falls to approximately 1.4
percent. The percent reductions in the market quantity of off-highway motorcycles ranges from
2.23 percent to 4.37 percent, reaching a steady state of 2.79 percent in 2012. Note that the
percent reduction in quantity is the same for both domestic and foreign output because it has
been assumed that domestic and foreign supply have the same price elasticity.
Table 9.10-4 presents the loss in consumer surplus, the loss in producer surplus, and the
total loss in surplus (equal to the sum of the changes in consumer and producer surplus) without
fuel cost savings. As the table shows, the consumer surplus loss is approximately half the size of
the loss in producer surplus. Consumer surplus losses range from nearly $5 million in year 2006
when the rule is first implemented, it rises to $9 million in 2007 and falls to $ 6 million in 2012
and the years beyond. The losses in producer surplus range from $10 million in the first year of
implementation, rising to $19 million in 2007 and falls to $12.7 million in 2012 and the years
beyond. Note that the annual surplus loss associated with the off-highway motorcycle market
declines as the per off-highway motorcycle engine costs fall starting in 2008. Loss in surplus is
equal to $15 million in 2006, rises to 28.7 in 2007 and it falls to $18.7 million by 2012. The
surplus estimate presented in Table 9.10-4 is based upon 2001 baseline market conditions and do
not consider fuel cost savings.
9-92
-------
Chapter 9: Economic Impact Analysis
Table 9.10-4
Annual Losses in Consumer and
Producer Surplus for the Off-Highway Motorcycle Market*
Impact Measure
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
LossinCS**($103)
LossinPS***($103)
Domestic
Foreign
Loss in Surplus ($103)
Year
2006
$4,841.4
$10,177.3
$ 4,298.3
$ 5,879.0
$15,018.7
2010
$7,131.4
$14,822.3
$ 6,260.2
$8,562.1
$21,953.7
2007
$9,369.1
$19,304.6
$8,153.2
$11,151.4
$28,700.7
2011
$ 6,230.5
$13,008.2
$ 5,493.9
$7,514.2
$19,238.6
2008
$ 8,683.7
$17,906.7
$ 7,562.8
$10,343.9
$26,590.3
2012-2030
$ 6,049.8
$12,642.3
$5,339.4
$ 7,302.9
$18,692.1
2009
$ 7,789.6
$16,136.5
$6,815.2
$9,321.3
$23,926.1
*Results are based on baseline year 2001 market conditions and fuel cost savings are not included.
** CS refers to consumer surplus and is rounded to the nearest hundredths. For a description of the change in consumer
surplus, see Section 9.2.2
*** PS refers to producer surplus and is rounded to the nearest hundredths. For a description of the change in producer
surplus, see Section 9.2.2.
9.10.4 Off-Highway Motorcycle Engineering Cost and Surplus Loss Comparison
This section presents a comparison of the future stream of engineering costs (excluding
fuel cost savings) and surplus losses for the off-highway motorcycle market. In Table 9.10-5, we
first present an interim comparison of the estimated engineering costs, holding quantity constant
to the baseline year quantity, with the surplus losses that were estimated from the economic
impact model. Because economic modeling takes into account consumer and producer behavior,
the estimated loss in surplus is less than the engineering costs under a perfectly competitive
market setting. In this case, the annual surplus losses are, on average, equal to 98 to 99 percent
of the calculated engineering costs. The cost numbers in this table and Table 9.10-6 are not
discounted.
Based upon a ratio of the loss in surplus to engineering costs, holding baseline quantity
constant, a projection of the surplus loss over the future year stream is calculated from the future
9-93
-------
Draft Regulatory Support Document
stream of engineering costs that appear in Chapter 7. This projection of the future stream of
surplus losses is compared to the future stream of engineering costs in Table 9.10-6. Note that
these results are not discounted nor do they account for fuel cost savings.
9.10.5 Off-Highway Motorcycle Economic Impact Results with Fuel Cost Savings
In Table 9.10-7, the social costs/gains are calculated by adding the annual savings in fuel
costs (presented initially in Chapter 7) to the projected annual surplus loss. These social gains
are compared to the engineering costs with fuel efficiency gains. As you can see from this table,
the emissions control program is expected to yield social gains rather than losses beyond the
initial year of implementation. Only the initial year of implementation results in a social loss
from this regulation for the off-highway motorcycle market.
Table 9.10-5
Interim Engineering Cost and Surplus Loss Comparison for the
Off-Highway Motorcycle Market Based on Year 2001 Quantity
(Q = 195,250 off-highway motorcycle)
Year
2006
2007
2008
2009
2010
2011-2030
Estimated Engineering Costs
$15,424.8
$30,263.8
$27,920.8
$24,992.0
$22,844.3
$19,915.5
Estimated Surplus Loss
$15,018.7
$28,700.7
$26,590.3
$23,926.1
$21,953.7
$19,238.6
9-94
-------
Chapter 9: Economic Impact Analysis
Table 9.10-6
Engineering Cost and Surplus Loss Comparison for the
Off-Highway Motorcycle Market without Fuel Cost Savings
(Q = Off-Highway Motorcycle projected sales for 2006 through 2030)
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Engineering Costs
$16,269.1
$32,215.0
$29,846.5
$27,127.3
$24,957.7
$22,079.4
$21,630.7
$21,847.0
$22,065.4
$22,286.1
$22,508.9
$22,734.0
$22,961.4
$23,191.0
$23,422.9
$23,657.1
$23,893.7
$24,132.6
$24,374.0
$24,617.7
$24,863.9
$25,112.5
$25,363.6
$25,617.3
$25,873.5
Estimated Surplus Loss
$15,840.8
$30,551.2
$28,424.3
$25,970.3
$23,984.8
$21,328.9
$20,895.5
$21,104.4
$21,315.5
$21,528.6
$21,743.9
$21,961.4
$22.181.0
$22,402.8
$22.626.8
$22,853.1
$23,081.6
$23,312.4
$23,545.6
$23,781.0
$24,018.8
$24,259.0
$24,501.6
$24,746.6
$24,994.1
9-95
-------
Draft Regulatory Support Document
Table 9.10-7
Engineering and Social Cost Comparison for the
Off-Highway Motorcycle Market with Fuel Cost Savings
(Q = Off-Highway Motorcycle projected sales for 2006 through 2030)
Year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Estimated Engineering Costs
with Fuel Cost Savings
$15,635.6
$30,153.2
$26,080.9
$21,459.3
$17,305.2
$12,409.1
$9,978.0
$8,293.5
$6,660.8
$5,090.2
$3,658.5
$2,529.9
$1,818.9
$1,397.3
$1,121.1
$923.2
$777.1
$686.8
$633.0
$596.1
$589.0
$601.6
$617.6
$656.3
$708.7
Estimated Social Costs/Gains
(Surplus Loss - Fuel Savings)*
$15,207.4
$28,489.4
$24,658.7
$20,302.3
$16,332.2
$11,658.7
$9,242.8
$7,551.0
$5,910.8
$4,332.7
$2,893.5
$1,757.2
$1,039.5
$609.1
$325.0
$119.2
($35.0)
($133.4)
($195.4)
($240.6)
($256.0)
($252.0)
($244.9)
($214.4)
($170.7)
: () represents a negative cost (social gain). Cost estimates are based upon 2001$
9-96
-------
Chapter 9: Economic Impact Analysis
Appendix to Chapter 9: Sensitivity Analyses
This appendix presents the results from a series of sensitivity analyses completed for the
recreational vehicles emissions standard. The sensitivity analyses examine how the market
impacts for each vehicle category would be affected if different measures of supply and demand
elasticities were used. For each vehicle category, changes in market price, quantity, and loss of
consumer and producer surplus are calculated by first varying the elasticity of supply, holding the
elasticity of demand fixed at the original value and then varying the elasticity of demand, holding
supply elasticity fixed at its original value. The sensitivity analyses are conducted using the
highest per vehicle costs over the future time stream of the regulation. We use the highest annual
per vehicle costs to ensure that our sensitivity analysis examines a worst-case scenario. Analysis
results are presented in comparison tables.
In order to estimate the economic impacts of the regulation on the each of the vehicle
markets, we rely upon the most current year of data (either 2000 or 2001, depending on the
vehicle category) to inform the model when we impose the regulatory costs per vessel on
producers. Using baseline year data allows us to estimate relative changes in price and quantity
as opposed to absolute changes. The results presented in these sensitivity analyses do not
account for fuel cost savings that may arise from this emissions control program.
Some general observations can be made about the market impacts resulting from a
regulation that affects production costs when different measures of supply and demand elasticity
are used and when demand and supply are assumed to be linear. The changes in market price and
quantity are smaller for an inward shift in the supply curve the more inelastic is the supply curve.
The more inelastic is the demand curve, the larger is the equilibrium change in market price and
the smaller is the change in market quantity from an inward shift in the supply curve.
9A.1 Sensitivity Analyses for Marine
The original estimates of supply and demand elasticity for the diesel inboard cruiser
market are e = 1.57 (for domestic and foreign supply) and r| = -1.44, both of which are elastic.
Using the highest per vessel costs of $905 which first occur in the year 2009, the market impacts
on price, quantity, and surplus losses are calculated first by varying measures of supply elasticity
holding demand elasticity constant and then by varying measures of demand elasticity holding
supply elasticity constant. These results are presented in Tables 9A. 1-1 and 9A. 1-2.
In the first column of Table 9A. 1-1, we reproduce the original market impacts for the year
2009 that were originally presented in Section 9.6 and compare them to the market impacts
calculated when supply elasticity is assumed to be equal to e = 1.00 (supply is unit elastic) and e
9-97
-------
Draft Regulatory Support Document
= 0.50 (supply is inelastic). Demand elasticity is assumed to equal -1.44 for each of these cases.
As the results show, the relative increase in market price and decrease in market output are
smaller as supply becomes more inelastic. Additionally, the more inelastic is supply, the smaller
is the loss in consumer surplus and larger is the loss in producer surplus. Consumer surplus loss
falls to just below $2 million from approximately $4 million while producer surplus losses
increases to $5.7 million from $3.6 million. While there is a change in the distribution of surplus
loss across consumers and producers, there is almost no change in the overall loss in surplus with
more inelastic supply. The overall surplus loss increases only by $5.6 thousand.
Table 9A.1-1
Supply Elasticity Sensitivity Analysis: Market Impacts
for the Diesel Inboard Cruiser Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
| = 1.57, 77 = -1.44
0.14%
-0.20%
$3,977.7
$3,641.1
$7,618.8
Unit Elastic
Supply
| = 1.00, 77 = -1.44
0.11%
-0.16%
$3,126.1
$4,494.6
$7.620.7
Inelastic
Supply
| = 0.50, 77 = -144
0.07%
-0.10%
$1,966.5
$5,657.9
$7,624.4
*Results are calculated using the highest per vehicle regulatory costs, which are equal to $905 and are projected to
occur in the year 2009/10. Results are based on baseline year 2001 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
Table 9A. 1-2 presents a comparison of the market impacts when demand elasticity is
varied while holding supply elasticity constant at 1.57. We calculate the changes in market price,
quantity, and surplus losses assuming r\ = -1.00 (demand is unit elastic) and r| = -0.50 (demand is
inelastic) and compare these results to the original results first presented in Section 9.6. As we
assume a more inelastic demand curve, the change in market price increases while the change in
quantity decreases. However, even when we assume inelastic demand, the change in market
price for diesel inboard cruisers is still under one-quarter of one percent. We also can examine
the change in consumer and producer surplus. In this case, consumer surplus loss increases and
producer surplus loss decreases as demand becomes more inelastic. The loss in consumer
surplus rises from $3.9 million to $5.9 million while producer surplus loss decreases from $3.6
million to $1.8 million. Overall surplus loss rises by approximately $9.2 thousand as demand
becomes more inelastic, again a minuscule amount.
9-98
-------
Chapter 9: Economic Impact Analysis
Table 9A.1-2
Demand Elasticity Sensitivity Analysis: Market Impacts
for the Diesel Inboard Cruiser Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
| = 1.57, 77 = -1.44
0.14%
-0.20%
$3,977.7
$3,641.1
$7,618.8
Unit Elastic
Demand
| = 1.57, 77 = -1.00
0.16%
-0.16%
$4,659.6
$2,963.1
$7,622.7
Inelastic
Demand
0.20%
-0.10%
$5,786.9
$1,841.1
$7,628.0
*Results are calculated using the highest per vehicle regulatory costs, which are equal to $1,552 and are projected to
occur in the year 2009/10. Results are based on baseline year 2001 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9A.2 Sensitivity Analyses for Forklifts
For the forklift market, the original economic impact analysis used an inelastic estimate
of supply, equal to e = 0.714 (for domestic and foreign supply), and an elastic estimate of
demand, equal to r\ = -1.5. The highest per vehicle costs for the forklift market, $610, are
incurred during 2004, which is the first year the regulation is implemented. Tables 9A.2-1 and
9A.2-2 present the sensitivity analyses assuming varying supply elasticities and varying demand
elasticities, respectively. The results include the changes in market price, quantity, and losses in
consumer and producer surplus.
Table 9A.2-1 presents the original results for the year 2004 from Section 9.7 of the
analysis and then presents the market impacts assuming e = 1.00 (supply is unit elastic) and e =
1.50 (supply in elastic). According to these results, we find that as the supply curve becomes
more elastic, the changes in both market price and quantity are larger. Assuming elastic supply,
we find that the increase in market price is equal to 1.16 percent and the decrease in market
quantity is equal to -1.73 percent. These market impacts, though larger than those we find when
supply is assumed to be inelastic, are not significant. We also examine the changes in consumer
and producer surplus to find that as supply becomes more elastic, the loss in consumer surplus
increases from $12.7 million to $19.7 million and the loss in producer surplus falls from $26.4
million to $19.3 million. Along with this redistribution of surplus loss is a reduction in the
overall loss in surplus as supply is assumed to be elastic. The overall loss in surplus originally
9-99
-------
Draft Regulatory Support Document
was equal to $39.1 million but falls to just under $39 million when e = 1.50.
Table 9A.2-1
Supply Elasticity Sensitivity Analysis: Market Impacts
for the Forklift Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
Ł = 0.714, 77 = -1.50
0.75%
-1.12%
$12,715.3
$26,412.4
$39,127.7
Unit Elastic
Supply
Ł=1.00,77 = -1.50
0.92%
-1.39%
$15,750.0
$23,294.9
$29,044.9
Elastic
Supply
1.16%
-1.73%
$19,653.1
$19,309.3
$38,962.4
*Results are calculated using the highest per vehicle regulatory costs, which are projected to occur in the year 2004
and are equal to $610 per forklift. Results are based on baseline year 2000 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
In the next table, demand elasticity is varied holding supply elasticity constant. The
original results were generated assuming e = 0.714 and r\ = -1.5. To conduct the sensitivity
analysis, we estimated the market impacts when demand elasticity was equal to -1 (unit elastic)
and also when it was equal to -0.5 (inelastic). The results in Table 9A.2-2 show that as demand
becomes more inelastic, the change in market price increases while the change in quantity
decreases. The largest change in market price is approximately 1.4 percent, which is still small
in scale. An examination of the surplus measures shows that the loss in consumer surplus
increases and the loss in producer surplus decreases as demand is more inelastic. Originally,
consumer surplus loss was equal to $12.7 million and producer surplus was equal to $26.4
million. For the inelastic demand case, consumer surplus loss increases to $23.4 million while
the loss in producer surplus falls to $16.2 million. Like the diesel marine vessel case, the overall
change in the total loss in surplus is negligible, approximately $3 thousand.
A sensitivity analysis for forklifts was also conducted using the estimated elasticity of
demand discussed in Section 9.5 of Chapter 9. The demand elasticity estimated is equal to -5.76,
a rather large estimate. Table 9A.2-3 presents a comparison of the original market impacts
originally presented in Chapter 9 with the market impacts when e = 0.714 and r| = -5.76. From
this sensitivity analysis, EPA finds that the relative increase in market price is one-quarter of one
percent while the decrease in market output is approximately one and one-half percent. The price
9-100
-------
Chapter 9: Economic Impact Analysis
increase is smaller relative to the original results because of the extremely elastic demand
measure. Overall, these market impacts are not very different from the original results.
What does differ a great deal is the distribution of the loss in welfare. Originally, the loss
in producer surplus was approximately two times the size of the loss in consumer surplus. When
the elasticity of demand is equal to -5.76, however, virtually all of the loss in economic welfare is
incurred by producers. Almost 90 percent of the loss in welfare is borne by producers while 10
percent is borne by consumers.
Table 9A.2-2
Demand Elasticity Sensitivity Analysis: Market Impacts
for the Forklift Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
1 = 0.714, 77 = -1.50
0.75%
-1.12%
$12,715.3
$26,412.4
$39,127.7
Unit Elastic
Demand
Ł = 0.714,77 = -1.00
0.96%
-0.96%
$16,437.4
$22,798.8
$39,236.2
Inelastic
Demand
Ł = 0.714, 77 = -0.50
1.36%
-0.68%
$23,240.4
$16,163.7
$39,404.1
*Results are calculated using the highest per vehicle regulatory costs, which are projected to occur in the year 2004
and are equal to $610 per forklift. Results are based on baseline year 2000 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9-101
-------
Draft Regulatory Support Document
Table 9A.2-3
Alternative Demand Elasticity Sensitivity Analysis: Market Impacts
for the Forklift Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
1 = 0.714, 77 = -1.50
0.75%
-1.12%
$12,715.3
$26,412.4
$39,127.7
Alternative Elastic
Demand
Ł = 0.7 14, 77 = -5.76
0.25%
-1.47%
$4,340.8
$34,499.8
$38,840.6
*Results are calculated using the highest per vehicle regulatory costs, which are projected to occur in the year 2004
and are equal to $610 per forklift. Results are based on baseline year 2000 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9A.3 Sensitivity Analyses for Snowmobiles
For the snowmobile market, the original economic impact analysis used an elastic
estimate of supply, equal to e = 2.1 (for domestic and foreign supply), and an elastic estimate of
demand, equal to r\ = -2.0. The steady state per vehicle engine modification costs resulting from
the regulation for the snowmobiles market of $180, are incurred during 2017 through 2030. This
per unit vehicle cost of emission controls is based upon 2001 price levels, Phase 3 regulatory
requirements, and incorporates the impact of the learning curve for the engine modification costs.
The EPA contends these per unit costs represent those the snowmobile manufacturers will
experience on an ongoing basis due to this regulation. Tables 9A.3-1 and 9A.3-2 present the
sensitivity analyses assuming varying supply elasticities and varying demand elasticities,
respectively. The results include the changes in market price, quantity, and losses in consumer
and producer surplus. All estimates are based upon the 2001 baseline market conditions.
Table 9A.3-1 presents the original results for the year 2017-2030 from Section 9.8 of the
analysis and then presents the market impacts assuming e = 2.6 (supply is more elastic) and e =
1.60 (supply is less elastic). According to these results, we find that as the supply curve becomes
more elastic, the changes in both market price and quantity are somewhat larger. These market
impacts, though larger than those we find when supply is assumed to be 2.1, are not significantly
different. We also examine the changes in consumer and producer surplus to find that as supply
becomes more elastic, the loss in consumer surplus increases from $12.8 million to $14.1 million
9-102
-------
Chapter 9: Economic Impact Analysis
and the loss in producer surplus falls from $11.8 million to $10.5 million. Along with this
redistribution of surplus loss is a reduction in the overall loss in surplus as supply is assumed to
be more elastic. When supply is assumed to be less elastic, price and quantity impacts decrease.
With less elastic supply producers bear more of the cost of the regulation. As illustrated by this
sensitivity analysis, price and quantity market impacts do not change substantially with
reasonable changes in the supply elasticity measures. As supply become less elastic producers
bear more of the cost of the regulation.
Table 9A.3-1
Supply Elasticity Sensitivity Analysis: Market Impacts
for the Snowmobile Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
Ł = 2.1,77= -2.0
1.45%
-2.90%
$12,777.4
$11,810.9
$24,588.3
More Elastic
Supply
Ł=2.6, 77= -2.0
1.60%
-3.20%
$14,078.6
$10,447.6
$24,556.2
Less Elastic
Supply
Ł=1.60,77= -.2.0
1.26%
-2.52%
$11,108.8
$13,532.7
$24,641.0
*Results are calculated using the steady-state per vehicle regulatory costs, which are projected to occur in the year
2015 through 2030 and are equal to $178 per snowmobile. Results are based on baseline year 2001 market
conditions.
** CS refers to consumer surplus and is rounded to the nearest hundred.
*** PS refers to producer surplus and is rounded to the nearest hundred.
In the next table, demand elasticity is varied holding supply elasticity constant. The
original results were generated assuming e = 2.1 and T| = -2.0. To conduct the sensitivity
analysis, we estimated the market impacts when demand elasticity was equal to -2.5 (more
elastic) and also when it was equal to -1.5 (less elastic). The results in Table 9A.3-2 show that
as demand becomes more elastic, the change in market price decreases while the change in
quantity increases. With more elastic demand, producers bear more of the burden of the
regulation, while consumers bear less. The overall surplus loss declines slightly. With less
elastic demand, the price change increases and quantity change decreases somewhat. Consumers
pay a larger share of the cost of the regulation with less elastic demand and producers a smaller
share. The surplus losses associated with the regulation increase slightly.
On August 2, 2002, National Economic Research Associates (NERA) provided the EPA
9-103
-------
Draft Regulatory Support Document
with the document Economic Assessments of Alternative Emission Standards for Snowmobile
Engines on behalf of ISMA. In this report, an estimate of the price elasticity of demand for
snowmobiles is presented. The EPA does not accept the validity of this elasticity estimate for a
number of reasons (see September 11, 2002 memorandum from Chris Lieske and Linda Chappell
to Docket A-2000-01, Document IV-B-45). In an effort to provide additional information to
quantify the market impacts of a more elastic price elasticity of demand, market impacts for a
price elasticity of demand estimate of-4.63 are presented in
Table 9 A. 3-2. As shown in the third column of this table, x - 21 ft = -453
projected price increases are smaller and market quantity
decreases are somewhat larger assuming a price elasticity of demand estimate of-4.63. In
addition, producers bear a greater portion of the burden of the regulation assuming the more
elastic price elasticity of demand.
Table 9A.3-2
Demand Elasticity Sensitivity Analysis: Market Impacts
for the Snowmobile Market*
Impact Measures
Change in Market Price
Change in Market
Output
Loss in CS** ($103)
LossinPS***($103)
Loss in Surplus ($103)
Original
Results
x = 2.1, h = -2.0
1.45%
-2.90%
$12,777.4
$11,810.9
$24,588.3
More Elastic
Demand
1=2.1,77= -2.5
1.29%
-3.23%
$11,369.4
$13,090.6
$24,460.0
More Elastic
Demand
0.88%
-1.09%
$7,737.1
$16,364.5
$24,083.6
Less Elastic
Demand
|=2.1, 7] = -1.5
1.65%
-2.48%
$14,583.2
$10,155.4
$24,738.6
*Results are calculated using the steady-state per vehicle regulatory costs, which are projected to occur in the year
2015 through 2030 and are equal to $$178 per snowmobile. Results are based on baseline year 2001 market
conditions.
** CS refers to consumer surplus and is rounded to the nearest hundred.
*** PS refers to producer surplus and is rounded to the nearest hundred.
In general, the sensitivity analysis indicates that market impacts are not particularly
sensitive to reasonable changes in the price elasticity of supply and demand. However, this
sensitivity analysis does indicate that the surplus losses borne by consumers and producers are
impacted by these estimates. Less elastic supply leads to the producer bearing a greater
percentage of the losses due to the regulation. Less elastic demand leads to consumers bearing
more of the cost of the regulation.
9-104
-------
Chapter 9: Economic Impact Analysis
9A.4 Sensitivity Analyses for ATV
For the ATV market, the original economic impact analysis used an original estimate of
supply, equal to e = 1.0 (for domestic and foreign supply), and an elastic estimate of demand,
equal to r| = -2.0. The steady state per vehicle costs for the ATV market, $48, are incurred
during 2012 through 2030. Tables 9A.4-1 and 9A.4-2 present the sensitivity analyses assuming
varying supply elasticities and varying demand elasticities, respectively. The results include the
changes in market price, quantity, and losses in consumer and producer surplus.
Table 9A.4-1 presents the original results for the year 2012 from Section 9.9 of the
analysis and then presents the market impacts assuming e = 1.50 (supply is more elastic) and e =
.50 (supply in elastic). Assuming the more elastic supply of e = 1.50, we find that the increase in
market price is equal to 0.40 percent and the decrease in market quantity is equal to -0.80
percent. Assuming the in elastic supply of e = 0.50, we find that the increase in market price is
equal to 0.19 percent and the decrease in market quantity is equal to -0.37 percent. We also
examine the changes in consumer and producer surplus to find that as supply becomes more
elastic, the loss in consumer surplus increases were $18.0 million and $8.4 million and the loss in
producer surplus are $23.8 million and $33.4 million, respectively . The overall loss in surplus
originally was equal to $41.9 million and $42.0 million, respectively.
Table 9A.4-1
Supply Elasticity Sensitivity Analysis: Market Impacts
for the ATV Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
Ł= 1.0,77= -2.0
0.31%
-0.62%
$14,036.0
$27,896.2
$41,932.2
More Elastic
Supply
Ł= 1.5,77= -2.0
0.40%
-0.80%
$18,030.2
$23,846.4
$41,876.5
InElastic
Supply
0.19%
-0.37%
$8,432.2
$33,401.4
$42,034.2
*Results are calculated using the steady state per vehicle regulatory costs, which are projected to occur in the year
2012 through 2030 and are equal to $48 per ATV. Results are based on baseline year 2001 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9-105
-------
Draft Regulatory Support Document
In the next table, demand elasticity is varied holding supply elasticity constant. The
original results were generated assuming e = 1.0 and r| = -2.0. To conduct the sensitivity
analysis, we estimated the market impacts when demand elasticity was equal to -2.5 (more
elastic) and also when it was equal to -1.5 (less elastic). The results in Table 9A.4-2 show that as
demand becomes more inelastic, the change in market price increases while the change in
quantity decreases. An examination of the surplus measures shows that the loss in consumer
surplus increases and the loss in producer surplus decreases as demand is more inelastic.
Originally, consumer surplus loss was equal to $14.0 million and producer surplus was equal to
$27.9 million. For the more elastic demand case, consumer surplus loss falls to $12.0 million
while the loss in producer surplus increase to $29.9 million. The overall change in the total loss
in surplus is negligible, approximately $20.
Table 9A.4-2
Demand Elasticity Sensitivity Analysis: Market Impacts
for the ATV Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
0.31%
-0.62%
$14,036.0
$27,896.2
$41,932.2
More Elastic
Demand
0.27%
-0.67%
$12,028.2
$29,868.6
$41,876.7
Inelastic
Demand
Y - 1 n h - 1 R
0.37%
-0.56%
$16,848.5
$25,130.3
$41,978.8
*Results are calculated using the steady state per vehicle regulatory costs, which are projected to occur in the year
2012 through 2030 and are equal to $48 per ATV. Results are based on baseline year 2001 market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9A.5 Sensitivity Analyses for Off-Highway Motorcycle
For the off-highway motorcycle market, the original economic impact analysis used an
original estimate of supply, equal to e = 0.93 (for domestic and foreign supply), and an elastic
estimate of demand, equal to T| = -2.0. The steady state per vehicle costs for the off-highway
motorcycle market, $99, are incurred during 2012 through 2030. Tables 9A.5-1 and 9A.5-2
present the sensitivity analyses assuming varying supply elasticities and varying demand
elasticities, respectively. The results include the changes in market price, quantity, and losses in
consumer and producer surplus.
9-106
-------
Chapter 9: Economic Impact Analysis
Table 9A.5-1 presents the original results for the year 2012 from Section 9.10 of the
analysis and then presents the market impacts assuming e = 1.50 (supply is more elastic) and e =
.50 (supply in elastic). Assuming the more elastic supply of e = 1.50, we find that the increase in
market price is equal to 1.88 percent and the decrease in market quantity is equal to -3.77
percent. Assuming the in elastic supply of e = 0.50, we find that the increase in market price is
equal to 0.88 percent and the decrease in market quantity is equal to -1.76 percent. We also
examine the changes in consumer and producer surplus to find that as supply becomes more
elastic, the loss in consumer surplus increases were $8.1 million and $3.8 million and the loss in
producer surplus are $10.4 million and $15.1 million, respectively . The overall loss in surplus
originally was equal to $18.6 million and $18.9 million, respectively.
Table 9A.5-1
Supply Elasticity Sensitivity Analysis: Market Impacts
for the Off-highway Motorcycle Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
x = 0.93, h = -2.0
1.39%
-2.79%
$6,049.8
$12,642.3
$5,339.42
More Elastic
Supply
x = 1.5, h = -2.0
1.88%
-3.77%
$8,128.2
$10,421.5
$18,549.7
InElastic
Supply
x =.50, h = -2.0
.88%
-1.76%
$3,832.0
$15,056.1
$18,888.1
*Results are calculated using the steady state per vehicle regulatory costs, which are projected to occur in the year
2012 through 2030 and are equal to $99 per off-highway motorcycle. Results are based on baseline year 2001
market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
In the next table, demand elasticity is varied holding supply elasticity constant. The
original results were generated assuming e = 0.93 and r| = -2.0. To conduct the sensitivity
analysis, we estimated the market impacts when demand elasticity was equal to -2.5 (more
elastic) and also when it was equal to -1.5 (less elastic). The results in Table 9A.2-5 show that as
demand becomes more inelastic, the change in market price increases while the change in
quantity decreases. An examination of the surplus measures shows that the loss in consumer
surplus increases and the loss in producer surplus decreases as demand is more inelastic.
Originally, consumer surplus loss was equal to $6.1 million and producer surplus was equal to
9-107
-------
Draft Regulatory Support Document
$12.7 million. For the more elastic demand case, consumer surplus loss falls to $5.6 million
while the loss in producer surplus increase to $13.5 million. The overall change in the total loss
in surplus is negligible, approximately $10.
Table 9A.5-2
Demand Elasticity Sensitivity Analysis: Market Impacts
for the Off-highway Motorcycle Market*
Impact Measures
Change in Market Price
Change in Market Output
LossinCS**($103)
Loss in PS*** ($103)
Loss in Surplus ($103)
Original
Results
Ł=0.93, 77= -2.0
1.39%
-2.79%
$6,049.8
$12,649.3
$18,692.1
More Elastic
Demand
Ł=0.93,77 = -2.5
1.19%
-2.98%
$5,163.0
$13,459.3
$18,622.2
Inelastic
Demand
Ł=0.93,77= -1-5
1.68%
-2.52%
$7,304.5
$11,480.5
$18,785.0
*Results are calculated using the steady state per vehicle regulatory costs, which are projected to occur in the year
2012 through 2030 and are equal to $99 per off-highway motorcycle. Results are based on baseline year 2001
market conditions.
** CS refers to consumer surplus and is rounded to the nearest hundredths.
*** PS refers to producer surplus and is rounded to the nearest hundredths.
9-108
-------
Chapter 9: Economic Impact Analysis
Chapter 9 References
1. International Snowmobile Manufacturing Association, www.snowmobile.org. 2002. Docket
A-2000-01, Document IV-A-140.
2. U.S. EPA. 1999. OAQPS Economic Analysis Resource Document. Office of Air Quality
Planning and Standards, Air Quality Strategies and Standards Division, Innovative Strategies and
Economics Group. April. Docket A-2000-01, Document IV-A-71.
3. National Park Service News Release. Final Snowmobile Regulations For Yellowstone and
Grand Teton National Parks Published in Federal Register. January 23, 2001. Docket A-2000-
01, Document No. IV-A-119.
4. See Letter from Line Wehrly to Docket, titled "News Articles for Yamaha 4-Stroke
Snowmobile." Docket A-2000-01. Docket A-2000-01, Document IV-B-25.
5. Internet Search, www.yamaha-motors.com. August 2002. Docket A-2000-01, Document
Nos. IV-A-111, IV-A-121, andIV-A-139.
6. Production data were taken from OELINK Database owned by Power Systems Research.
7. Internet Search, www.yamaha-motors.com. June 2002. Docket A-2000-01, Document Nos.
IV-A-111, IV-A-121, and IV-A-139.
8. Estimated based upon submission of data by ISMA manufacturers and Southwest Research
Institute. Carroll (1999), White et.al. (1997), Hare and Springer 1974, White and Wright (1998).
These documents can be found in Docket A-2000-01, Documents U-B-19, U-D-15, IV-A-142, II-
A-50, andII-D-05.
9. Estimated based upon submission of data by ISMA manufacturers and Southwest Research
Institute. Carroll (1999), White et.al. (1997), Hare and Springer 1974, White and Wright (1998).
These documents can be found in Docket A-2000-01, Documents U-B-19, U-D-15, IV-A-142, II-
A-50, andII-D-05.
10. Internet Search, www.polarisindustries.com/product/default.asp7contentid=sno June 2002.
Docket A-2000-01, Document Nos. IV-A-111, IV-A-120, and IV-A-139.
11. Internet Search, www.yamaha-motors.com. June 2002 Docket A-2000-01, Document Nos.
IV-A-111, IV-A-120, and IV-A-139.
12. Internet Search, www.yamaha-motors.com. June 2002. Docket A-2000-01, Document Nos.
IV-A-111, IV-A-120, and IV-A-139.
9-109
-------
Draft Regulatory Support Document
13. Internet Search, http://www.arctic-cat.com/snowmobiles/index.asp June 2002. Docket A-
2000-01, Document Nos. IV-A-111, IV-A-120, and IV-A-139.
14. Internet Search. www.polarisindustries.com/product/default.asp?contentid=sno.
www.yamaha-motors.com.. www.yamaha-motors.com..
http://www.arctic-cat.com/snowmobiles/index.asp June 2002. Docket A-2000-01, Document
Nos. IV-A-111, IV-A-120, IV-A-121, and IV-A-139.
15. Internet Search. www.polarisindustries.com/product/default.asp?contentid=sno.
www.yamaha-motors.com.. www.yamaha-motors.com..
http://www.arctic-cat.com/snowmobiles/index.asp June 2002. Docket A-2000-01, Document
Nos. IV-A-111, IV-A-120, IV-A-121, and IV-A-139
16. Internet Search. www.polarisindustries.com/product/default.asp?contentid=sno.
www.yamaha-motors.com.. www.yamaha-motors.com..
http://www.arctic-cat.com/snowmobiles/index.asp June 2002. Docket A-2000-01, Document
Nos. IV-A-111, IV-A-120, IV-A-121, and IV-A-139.
17. Internet Search. www.polarisindustries.com/product/default.asp?contentid=sno.
www.yamaha-motors.com.. www.yamaha-motors.com..
http://www.arctic-cat.com/snowmobiles/index.asp June 2002. Docket A-2000-01, Document
Nos. IV-A-111, IV-A-120, IV-A-121, and IV-A-139.
18. Raboy, David G. 1987. Results of an Economic Analysis of Proposed Excise Taxes on
Boats mimeo. Washington DC: Patton, Boggs, and Blow. Prepared for the National Marine
Manufacturing Association. Docket A-2000-01, Document IV-A-129.
19. 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 Docket A-2000-01, Document IV-A-48 and IV-A-
190.
20. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-5 8
21. Bureau of Economic Analysis. 2002. Survey of Current Business - D. Domestic
Perspectives. January. http://www.bea.doc.gov/bea/ARTICLES/2002/01January/D-
Pages/0102DpgD.pdf Docket A-2000-01, Document IV-A-55.
22. Bureau of Economic Analysis. 1999. Survey of Current Business - D. Domestic Perspectives.
December.
http://www.bea.doc.gov/bea/ARTICLES/NATIONAL/NIPAREL/1999/1299dpgd.pdf. Docket
9-110
-------
Chapter 9: Economic Impact Analysis
A-2000-01, Document IV-A-56.
23. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://www.bea.doc.gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-
191.
24. Bureau of Economic Analysis. National Income and Product Accounts Tables. Table 7.1-
Quantity and Price Indices for GDP.
http://www.bea.gov^ea/dn/nipaweb/SelectTable.asp?Selected=Y#s7 Docket A-2000-01,
Document IV-A-192.
25. Pindyck, Robert S. and Daniel L. Rubinfeld. Econometric Models and Economic Forecasts,
2nd edition. McGraw Hill Publishing. 1981. Pages 174-201.
26. Industrial Truck Association. 2002. Membership Handbook, http://www.indtrk.com/
Copies of relevant pages can be found in Docket A-2000-01, Document IV-A-188.
27. 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 Docket A-2000-01, Document IV-A-48 and IV-A-
190.
28. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-58.
29. Bureau of Economic Analysis. 2002. Survey of Current Business - D. Domestic
Perspectives. January. http://www.bea.doc.gov/bea/ARTICLES/2002/01January/D-
Pages/0102DpgD.pdf Docket A-2000-01, Document IV-A-55
30. Bureau of Economic Analysis. 1999. Survey of Current Business - D. Domestic
Perspectives. December.
http://www.bea.doc.gov/bea/ARTICLES/NATIONAL/NIPAREL/1999/1299dpgd.pdf Docket A-
2000-01, Document IV-A-56.
31. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://www.bea.doc.gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-
191.
32. Bureau of Economic Analysis. National Income and Product Accounts Tables. Table 7.1-
Quantity and Price Indices for GDP.
http://www.bea.gov/bea/dn/nipaweb/SelectTable.asp? Selected=Y#s7 Docket A-2000-01,
Document IV-A-192.
9-111
-------
Draft Regulatory Support Document
33. 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 Docket A-2000-01, Document IV-A-48 and IV-A-
190.
34. U.S. Department of Commerce. 2001. 1999 Annual Survey of Manufactures - Statistics for
Industry Groups and Industries. Washington, DC: Government Printing Office. Docket A-2000-
01, Document IV-A-58.
35. Bureau of Economic Analysis. 2002. Survey of Current Business - D. Domestic
Perspectives. January. http://www.bea.doc.gov/bea/ARTICLES/2002/01January/D-
Pages/0102DpgD.pdf Docket A-2000-01, Document IV-A-55.
36. Bureau of Economic Analysis. 1999. Survey of Current Business - D. Domestic
Perspectives. December.
http://www.bea.doc.gov/bea/ARTICLES/NATIONAL/NIPAREL/1999/1299dpgd.pdf Docket A-
2000-01, Document IV-A-56.
37. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://www.bea.doc.gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-
191.
38. Bureau of Economic Analysis. National Income and Product Accounts Tables. Table 7.1-
Quantity and Price Indices for GDP.
http://www.bea.gov/bea/dn/nipaweb/SelectTable.asp? Selected=Y#s7 Docket A-2000-01,
Document IV-A-192.
39. Raboy, David G. 1987. Results of an Economic Analysis of Proposed Excise Taxes on
Boats mimeo. Washington DC: Patton, Boggs, and Blow. Prepared for the National Marine
Manufacturing Association. Docket A-2000-01, Document IV-A-129.
40. National Bureau of Economic Research and U.S. Census Bureau, Center for Economic
Research. 2002 NBER-CES Manufacturing Industry Database, 1958 - 1996. Docket A-2000-01,
Document IV-A-48 and IV-A-190.
41. U. S. Department of Commerce. 2001. 1999 Annual Survey of Manufacturers - Statistics for
Industry Groups and Industries. Washington, DC Government Printing Office. Docket A-2000-
01, Document IV-A-58.
42. U. S. Department of Commerce. 2001. 2000 Annual Survey of Manufacturers - Statistics for
Industry Groups and Industries. Washington, DC Government Printing Office.
9-112
-------
Chapter 9: Economic Impact Analysis
43. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://bea. gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-191.
44. Bureau of Economic Analysis. 2002. Survey of Current Business -D Domestic Perspectives.
January, http://www.bea.doc/bea/ARTICLES/2002/01 January/D-Pages/0102Dpg.pdf Docket A-
2000-01, Document IV-A-55.
45. Bureau of Economic Analysis. 1999. Survey of Current Business -D Domestic Perspectives.
December. http://www.bea.doc/bea/ARTICLES/NATIONAL/NIPAREL/19991299dpgd.pdf
Docket A-2000-01, Document IV-A-56.
46. International Snowmobile Manufacturers Association, www.snowmobile.org. June 2002.
47. US Census Bureau, Historical Income Tables - Households, Table H-6, Regions -
Households (All Races) by Median and Mean Income, 1975-2000. A copy of this information
can be found in Docket A-2000-01, Document VI-A-203.
48. Raboy, David G. 1987. Results of an Economic Analysis of Proposed Excise Taxes on
Boats mimeo. Washington DC: Patton, Boggs, and Blow. Prepared for the National Marine
Manufacturing Association. Docket A-2000-01, Document IV-A-129.
49. Raboy, David G. 1987. Results of an Economic Analysis of Proposed Excise Taxes on
Boats mimeo. Washington DC: Patton, Boggs, and Blow. Prepared for the National Marine
Manufacturing Association. Docket A-2000-01, Document IV-A-129.
50. National Bureau of Economic Research and U.S. Census Bureau, Center for Economic
Research. 2002 NBER-CES Manufacturing Industry Database, 1958 - 1996. Docket A-2000-01,
Document IV-A-48 and IV-A-190.
51. U. S. Department of Commerce. 2001. 1999 Annual Survey of Manufacturers - Statistics for
Industry Groups and Industries. Washington, DC Government Printing Office. Docket A-2000-
01, Document IV-A-5 8.
52. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://bea. gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-191.
53. Bureau of Economic Analysis. 2002. Survey of Current Business -D Domestic Perspectives.
January, http://www.bea.doc/bea/ARTICLES/2002/01 January/D-Pages/0102Dpg.pdf Docket A-
2000-01, Document IV-A-55.
54. Bureau of Economic Analysis. 2002. Shipments of Manufacturing Industries.
http://bea. gov/bea/dn2/gpo.htm Docket A-2000-01, Document IV-A-44 and IV-A-191.
9-113
-------
Draft Regulatory Support Document
55. Bureau of Economic Analysis. 1999. Survey of Current Business -D Domestic Perspectives.
December. http://www.bea.doc/bea/ARTICLES/NATIONAL/NIPAREL/19991299dpgd.pdf
Docket A-2000-01, Document IV-A-56.
56. Motorcycle Industry Council, Inc. 2002. Docket A-2000-01, Document No. IV-A-138.
57. US Census Bureau, Historical Income Tables - Households, Table H-6, Regions -
Households (All Races) by Median and Mean Income, 1975-2000. A copy of this information
can be found in Docket A-2000-01, Document VI-A-203.
58. National Marine Manufacturing Association. U.S. Recreational Boating Domestic Shipment
Statistics, 1970 - 1998. http://63.236.237.146/facts/historv.pdf Docket A-2000-01, Document
IV-A-49.
59. National Marine Manufacturing Association. Boating 2001 - Facts and Figures at a Glance.
http://www.nmma.org/facts/boatingstats/2001/ Docket A-2000-01, Document IV-A-50.
60. Teleconference with John McKnight of the National Marine Manufacturers Association on
April 15,2002.
61. Boating Industry. 1996. The Boating Business, 1995 Annual Industry Review. January.
Docket A-2000-01, Document IV-A-43.
62. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Table 4 - Product Analysis: Inboard Boats,
Imports. Docket A-2000-01, Document IV-A-53.
63. National Marine Manufacturing Association. 2002. Economic Impact Analysis of the
Diesel Engine Rule on the U.S. Boat Market. Docket A-2000-01, Document IV-A-53.
64. ISMA Snowmobile Statistics www.snowmobile.org. June 2002.
65. Production data were taken from OELINK Database owned by Power Systems Research.
9-114
-------
Chapter 10: Benefit-Cost Analysis
Chapter 10: Benefit-Cost Analysis
10.1 Introduction
This chapter contains EPA's analysis of the economic benefits of the Large
Si/Recreational Vehicle rule. The analysis presented here attempts to answer three questions
• What are the physical health and welfare effects of changes in ambient air quality
resulting from reductions in nitrogen oxides (NOx), hydrocarbons (HC) (including
air toxics), carbon monoxide (CO), and particulate matter (PM) emissions?
• What is the value placed on these emission reductions by U.S. citizens as a
whole?
• How do these estimated benefits compare to the estimated costs associated with
this rule?
In the benefits analysis, we calculate a limited set of PM-related health benefits (our base-
case estimate). In this part of the analysis, we estimate nationwide PM health effects benefits
associated with reduction of Nox and direct PM emissions from Large SI only. Reductions
related to ATVs, OHMs, snowmobiles and recreational marine diesel are not quantified. This
analysis is based on estimated reductions in NOx and PM emissions and uses a benefits transfer
technique to determine the changes in human health and welfare, both in terms of physical effects
and monetary value
These analyses yield a stream of monetized benefits which we compare to the costs of
the standards. It is important to note that there are significant categories of benefits associated
with the control program which cannot be monetized (or in many cases even quantified),
including visibility, ozone health benefits, ecological effects, most species of air toxics' health
and ecological effects. We identify these benefits in the discussion below and carry them
through our estimates as nonmonetized health benefits.
10.2 General Methodology
10.2.1 PM Methodology - Benefits Transfer
In performing the analysis for the PM benefits, we relied on the results of a similar
analysis performed for our emission controls for on-highway heavy-duty engines (called the
10-1
-------
Draft Regulatory Support Document
HD07 rule.11 see 99 FR 5002, January 18, 2001). This approach was necessary due to time and
resource constraints. To apply that analysis to this control program, we used a benefits transfer
technique, described below. Benefits transfer is the science and art of adapting primary benefits
research from similar contexts to obtain the most accurate measure of benefits for the
environmental quality change under analysis. Where appropriate, adjustments are made for the
level of environmental quality change, the sociodemographic and economic characteristics of the
affected population, and other factors in order to improve the accuracy and robustness of benefits
estimates. Additional information on the technique used can be found in Hubbell 2002
memorandum to the Docket (Docket A-2000-01, Document IV-A-146).
The HD07 analysis followed the same general methodology used in the benefits analysis
for the passenger vehicle Tier 2/Gasoline Sulfur final rulemm and other EPA air benefits reports,
with routine updates in response to public comment and to reflect advances in modeling and the
literature for economics and health effects. This analysis also reflects the advice of its
independent Science Advisory Board (SAB) in determining the health and welfare effects
considered in the benefits analysis and in establishing the most scientifically valid measurement
and valuation techniques.
10.2.2 CO and Air Toxics Methodology : WTP
In this component of the analysis, we discuss the benefits of reducing air toxics pollution
from vehicles subject to the rule. The only segment for which willingness to pay for reductions
in pollution were reported in the literature was for use-values for snowmobiles; however, the
estimates pertained only to use value and were not judged to be reliable. There were no studies
estimating the changes in consumer surplus to other non-snowmobilers such as cross-country
skiers, nature enthusiasts, and residents near where snowmobiles are operated. We are not able
to estimate the value of changes in air toxics or CO from other engines subject to this rule.
"Additional information about the Regulatory Model System for Aerosols and Deposition
(REMSAD) and our modeling protocols can be found in our Regulatory Impact Analysis: Heavy-
Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements,
document EPA420-R-00-026, December 2000. Docket No. A-2000-01, Document No. A-II-13.
This document is also available at http://www.epa.gov/otaq/disel.htm#documents. Information
can also be found in the docket for the HD07 rulemaking: A-99-06.
mm US EPA. Regulatory Impact Analysis: Control of Air Pollution from New Motor
Vehicles: Tier 2 Emission Standards. Report No EPA420-R-99-023. December 1999. A copy
of this document can be found in Docket A-99-06, Document IV-A-09.
10-2
-------
Chapter 10: Benefit-Cost Analysis
10.2.3 Benefits Quantification
We use the term benefits to refer to any and all positive effects of emissions changes on
social welfare that we expect to result from the final rule. We use the term environmental costs
(also commonly referred to as "disbenefits") to refer to any and all negative effects of emissions
changes on social welfare that result from the final rule. We include both benefits and
environmental costs in this analysis. Where it is possible to quantify benefits and environmental
costs, our measures are those associated with economic surplus in accepted applications of
welfare economics. They measure the value of changes in air quality by estimating (primarily
through benefits transfer) the willingness of the affected population to pay for changes in
environmental quality and associated health and welfare effects.
Not all the benefits of the rule can be estimated with sufficient reliability to be
quantified and included in monetary terms. The omission of these items from the total of
monetary benefits reflects our inability to measure them. It does not indicate their lack of
importance in the consideration of the benefits of this rulemaking.
This analysis presents estimates of the potential benefits from the Large Si/Recreational
Vehicle rule expected to occur in 2030 as well as a stream of benefits and net present value from
2002 to 2030. The predicted emissions reductions that will result from the rule have yet to occur,
and therefore the actual changes in human health and welfare outcomes to which economic
values are ascribed are predictions. These predictions are based on the best available scientific
evidence and judgment, but there is unavoidable uncertainty associated with each step in the
complex process between regulation and specific health and welfare outcomes.
Changes in ambient concentrations will lead to new levels of environmental quality in the
U.S., reflected both in human health and in non-health welfare effects. Thus, the predicted
changes in ambient air quality serve as inputs into functions that predict changes in health and
welfare outcomes. We use the term "endpoints" to refer to specific effects that can be associated
with changes in air quality. Table 10.2-1 lists the human health and welfare effects identified for
changes in air quality as they related to ozone, PM, CO, and HC.m This list includes both those
effects quantified (and/or monetized) in this analysis and those for which we are unable to
provide quantified estimates.
For changes in risks to human health from changes in PM, quantified endpoints include
changes in mortality and in a number of pollution-related non-fatal health effects. Only the
benefits related to changes in NOx-related PM and directly emitted PM were estimated for Large
SI. HC-related PM and any PM-related benefits for recreational marine, ATVs, OHMs, and
m The HC listed in Table 10.2-1 are also listed as hazardous air pollutants in the Clean
Air Act. We are not able to quantify their direct effects. To the extent that they are precursors to
ozone or PM, they are included in our quantitative results.
10-3
-------
Draft Regulatory Support Document
snowmobiles were not estimated because of uncertainties with the benefits transfer to those
categories and due to lack of information about HC-related PM from the original data set.
The benefits related to changes in CO and HC are not directly quantified for our primary
analysis due to a lack of direct estimates of willingness to pay or appropriate exposure and air
quality models for these pollutants.
Table 10.2-1
Human Health and Welfare Effects of Pollutants
Affected by the Large Si/Recreational Vehicle Rule
Pollutant/Effect
Primary Quantified and Monetized
EffectsA
Unqualified Effects
Ozone/Health
Not quantified in this analysis
Minor restricted activity days
Hospital admissions - respiratory and cardiovascular
Emergency room visits for asthma
Non-asthma respiratory emergency room visits
Asthma symptoms
Chronic asthma0
Premature mortality13
Increased airway responsiveness to stimuli
Inflammation in the lung
Chronic respiratory damage
Premature aging of the lungs
Acute inflammation and respiratory cell damage
Increased susceptibility to respiratory infection
Ozone/Welfare
Not quantified in this analysis
Decreased worker productivity
Decreased yields for commercial crops
Decreased commercial forest productivity
Decreased yields for fruits and vegetables
Decreased yields for other commercial and
non-commercial crops
Damage to urban ornamental plants
Impacts on recreational demand from damaged
forest aesthetics
Damage to ecosystem functions
10-4
-------
Chapter 10: Benefit-Cost Analysis
Pollutant/Effect
Primary Quantified and Monetized
EffectsA
Unqualified Effects
PM/Health
Premature mortality
Bronchitis - chronic and acute
Hospital admissions - respiratory and
cardiovascular8
Emergency room visits for asthma
Asthma attacks
Lower and upper respiratory illness
Minor restricted activity days
Work loss days
Infant mortality
Low birth weight
Changes in pulmonary function
Chronic respiratory diseases other than chronic
bronchitis
Morphological changes
Altered host defense mechanisms
Cancer
Non-asthma respiratory emergency room visits
PMAVelfare
Not quantified in this analysis
Visibility in areas where people live, work and
recreate
Visibility in Class I national parks and forest areas
Household soiling
Materials damage
Nitrogen and
Sulfate
Deposition/
Welfare
Not quantified in this analysis
Impacts of acidic sulfate and nitrate deposition on
commercial forests
Impacts of acidic deposition on commercial
freshwater fishing
Impacts of acidic deposition on recreation in
terrestrial ecosystems
Impacts of nitrogen deposition on commercial
fishing, agriculture, and forests
Impacts of nitrogen deposition on recreation in
estuarine ecosystems
Costs of nitrogen controls to reduce eutrophication in
estuaries
Reduced existence values for currently healthy
ecosystems
NOx/Health
Not quantified in this analysis
Lung irritation
Lowered resistance to respiratory infection
Hospital Admissions for respiratory and cardiac
diseases
CO/Health
Not quantified in this analysis
As a supplemental calculation, some
behavior effects (choice-reaction
time) are quantified for one category
for which an exposure model was
available
Premature mortality8
Behavioral effects
Hospital admissions - respiratory, cardiovascular,
and other
Other cardiovascular effects
Developmental effects
Decreased time to onset of angina
Non-asthma respiratory ER visits
10-5
-------
Draft Regulatory Support Document
Pollutant/Effect
HCsE
Health
HCs E Welfare
Primary Quantified and Monetized
EffectsA
Not quantified in this analysis
As a supplemental calculation, some
behavior effects (choice-reaction time
and toluene) are quantified for one
category for which an exposure model
was available
Not quantified in this analysis
Unqualified Effects
Cancer (diesel PM, benzene, 1,3-butadiene,
formaldehyde, acetaldehyde)
Anemia (benzene)
Disruption of production of blood components
(benzene)
Reduction in the number of blood platelets
(benzene)
Excessive bone marrow formation (benzene)
Depression of lymphocyte counts (benzene)
Reproductive and developmental effects
(1,3-butadiene)
Irritation of eyes and mucous membranes
(formaldehyde)
Respiratory and respiratory tract
Asthma attacks in asthmatics (formaldehyde)
Asthma-like symptoms in non-asthmatics
(formaldehyde)
Irritation of the eyes, skin, and respiratory tract
(acetaldehyde)
Upper respiratory tract irritation & congestion
(acrolein)
Direct toxic effects to animals
Bioaccumulation in the food chain
A Primary quantified and monetized effects are those included when determining the base-case estimate of total monetized
benefits of the Large Si/Recreational Vehicle rule.
B Our examination of the original studies used in this analysis finds that the health endpoints that are potentially
affected by the GAM issues include: reduced hospital admissions and reduced lower respiratory symptoms. While
resolution of these issues is likely to take some time, the preliminary results from ongoing reanalyses of some of the
studies suggest a more modest effect of the S-plus error than reported for the NMMAPS PM10 mortality study.
While we wait for further clarification from the scientific community, we have chosen not to remove these results
from the benefits estimates, nor have we elected to apply any interim adjustment factor based on the preliminary
reanalyses. EPA will continue to monitor the progress of this concern, and make appropriate adjustments as further
information is made available.
c While no causal mechanism has been identified linking new incidences of chronic asthma to ozone exposure, an
epidemiological study shows a statistical association between long-term exposure to ozone and incidences of chronic asthma in
some non-smoking men (McDonnell, et al., 1999).
D Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the American Cancer
Society (ACS)/ Krewski, et al., 2000 C-R function we use for premature mortality captures both PM mortality benefits and any
mortality benefits associated with other air pollutants (ACS/ Krewski, et al., 2000).
E Many of the hydrocarbons (HCs) listed in the table are also hazardous air pollutants listed in the Clean Air Act.
This remainder of this chapter proceeds as follows: in Sections 10.3, we describe the
10-6
-------
Chapter 10: Benefit-Cost Analysis
categories of benefits that are estimated, present the techniques and inputs that are used, and
provide a discussion of how we incorporate uncertainty into our analysis. In Section 10.4, we
briefly discuss the CO and air toxics benefits in a qualitative manner. In Section 10.5, we report
our estimates of total monetized benefits.
10.3 PM-Related Health Benefits Estimation
10.3.1 Emissions Inventory Implications
The national inventories for NOx, HC, CO and PM have already been presented and
discussed in Chapters 1 and 6 and in the supporting documents referenced in those chapters.
Interested readers desiring more information about the inventory methodologies or results should
consult that chapter for details. This section explains the specific inventories that were used in
our quantitative estimates of benefits and the implications of those inventories related to
interpreting results.
As noted in the previous section, this analysis focuses on the PM-related health benefits
from emission reductions from Large SI engines only. To quantify these PM-related health
benefits, we used NOx and direct PM emission changes (both reductions and increases, where
applicable) for the categories Large SI. Our underlying air quality modeling which forms the
basis for the transfer technique considers NOx as a precursor for both PM and ozone; thus,
oxidant chemistry in the model would not lead to over-estimation of secondary PM formation.
We did not include HC-related PM because we do not currently have an appropriate transfer
technique.
We did not quantify the NOx, direct PM, or HC-related PM benefits for ATVs, OHMs,
recreational marine diesels or snowmobiles because in our judgement there are substantial
uncertainties in making the transfer from the on-highway vehicle modeling to these categories.
This is because their operating characteristics and the locations in which these nonroad engines
are used can be very different from on-highway vehicles. We had more reason to believe that the
distribution of vehicles with respect to human populations was more similar for Large SI.
However, in the analyses of alternatives, we present a sensitivity calculation for ATVs, noting the
large uncertainties inherent in that application of this technique.
As described in the previous chapters of this Regulatory Support Document, the emission
controls for Large SI engines and recreational vehicles begin at various times and in some cases
phase in over time. This means that during the early years of the program there would not be a
consistent match between cost and benefits. This is especially true for the vehicle control
portions and initial fuel changes required by the program, where the full vehicle cost would be
incurred at the time of vehicle purchase, while the fuel cost along with the emission reductions
and benefits resulting from all these costs would occur throughout the lifetime of the vehicle.
Because of this inconsistency and our desire to more appropriately match the costs and emission
10-7
-------
Draft Regulatory Support Document
reductions of our program, our analysis uses a future year when the fleet is nearly fully turned
over (2030). Consequently, we developed emission inventories through 2030 for both baseline
conditions and a control scenario. We present both the benefits as a snapshot in 2030 and as a
stream of benefits in the years leading up to 2030. However, our discussion of this analysis
focuses on 2030 because the benefits transfer technique applied to these inventories relies on air
quality modeling conducted for the year 2030.
10.3.2 Benefits Transfer Methodology
This section summarizes the benefits transfer methodology used in this analysis. This
method provides a relatively simple analysis of the health costs of NOX, and direct PM emissions
from Large SI engines. It is important to distinguish these estimates from an analysis that
employs full-scale air quality modeling and benefits modeling. The transfer technique used here
produces reasonable approximations. Nevertheless, the method also adds uncertainty to the
analysis and the results may under or overstate actual benefits of the control program.
Our approach is to develop estimates of health costs expressed in per ton terms. From
the Regulatory Model System for Aerosols and Deposition (REMSAD) air quality modeling used
for the HD07 rule benefits analysis, we estimated environmental and health costs per ton of NOx
and PM. Aggregate environmental and health cost estimates at the national level are scaled to
account for human population changes between years of analysis. Complete details of the
emissions, air quality, and benefits modeling conducted for the HD07 rule can be found at
http://www.epa.gov/otaq/diesel.htm and http://www.epa.gov/ttn/ecas/regdata/tsdhddv8.pdf.
Further details of the transfer technique calculations and inputs can be found in the supporting
memorandum to the docket (Hubbell 2002a). An alternative approach is presented to provide
some insight into the potential of importance of key elements underlying estimates of benefits
(Hubbell 2002b).
We examined the impacts of NOx, and direct PM emissions. NOx emissions are
associated with both ambient ozone and particulate matter (PM) levels. Due to data limitations,
we are providing estimates only for PM related health impacts. The underlying REMSAD
modeling partitions the NOx into formation of both ozone and PM in 2030, oxidant chemistry in
the model would not lead to over-estimation of secondary PM formation.00 Note that we do not
attempt to quantify ozone-related benefits. Because the vast majority of the benefits we are able
00Additional information about the Regulatory Model System for Aerosols and
Deposition (REMSAD) and our modeling protocols can be found in our Regulatory Impact
Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements, document EPA420-R-00-026, December 2000. Docket No. A-2000-01,
Document No. A-II-13. This document is also available at
http://www.epa.gov/otaq/disel.htm#documents. Information can also be found in the docket for
the HD07 rulemaking: A-99-06.
10-8
-------
Chapter 10: Benefit-Cost Analysis
to measure and place a monetary value on are PM related, these estimates will capture most of
the benefits we are able to monetize associated with the NOX, and direct PM emission control.
However, one important limitation is that benefits from ozone reductions, air toxics reductions,
visibility improvement, and other unquantifiable health and welfare endpoints are not captured in
these estimates. The results of this original analysis are summarized in Table 10.3-1.
The cost-per-ton estimate presented in Table 10.3-1 is for estimating tons reduced in 2001
based on a U.S. population of 277 million people. To apply this figure to future years, it is
necessary to adjust for increases in population (e.g., in 2030, the U.S. population is estimated to
be 345 million) and for growth in real income (see Hubbell 2002a and Equation 1 below).
10-9
-------
Draft Regulatory Support Document
Table 10.3-1
Summary of Health Effects and Economic Cost Estimates for Transfer
Health Effect3
All-cause Premature Mortality from Long-
term Exposure
Chronic Bronchitis
Hospital Admissions - COPD
Hospital Admissions - Pneumonia
Hospital Admissions - Asthma
Hospital Admissions - Total Cardiovascular
Asthma-Related ER Visits
Asthma Attacks
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Work Loss Days
Minor Restricted Activity Days (minus
asthma attacks)
Totals
Incidence/ton in 2001 based
on U. S. population of 277
million
NOX
0.0016
0.0010
0.0002
0.0002
0.0002
0.0005
0.0004
0.0324
0.0034
0.0368
0.0373
0.2849
1.3875
PM
0.0221
0.0143
0.0024
0.0030
0.0023
0.0072
0.0053
0.4566
0.0479
0.5188
0.5270
4.0180
20.9184
Estimated $/ton economic costs
in 2001 based on U.S.
population of 277 million
(1999$)
NOX
$9,726
$350
$2
$3
$1
$10
$0
$1
<$1
$1
$1
$30
$68
$10,193
PM
$136,164
$5,012
$30
$44
$15
$132
$2
$19
$3
$13
$8
$402
$1,023
$142,867
Note that the wide discrepancy between the per ton values of NO,, and direct PM is due to differences in their relative
contributions to ambient concentrations of PM25. The underlying REMSAD modeling partitions NOx between ozone and
secondary PM formation.The HD07 analysis examined the impacts in 2030 of reducing SO2 emissions by 141,000 tons and NOX
emissions by 2,570,000 tons, as well as a 109,000 ton reduction in direct PM emissions.
a Our examination of the original studies used in this analysis finds that the health endpoints that are potentially
affected by the GAM issues include: reduced hospital admissions and reduced lower respiratory symptoms. While
resolution of these issues is likely to take some time, the preliminary results from ongoing reanalyses of some of the
studies suggest a more modest effect of the S-plus error than reported for the NMMAPS PM10 mortality study.
While we wait for further clarification from the scientific community, we have chosen not to remove these results
from the benefits estimates, nor have we elected to apply any interim adjustment factor based on the preliminary
reanalyses. EPA will continue to monitor the progress of this concern, and make appropriate adjustments as further
information is made available.
10-10
-------
Chapter 10: Benefit-Cost Analysis
10.3.3 Overview of Heavy Duty Engine/Diesel Fuel Benefits Analysis and Development of
Benefits Transfer Technique
This section provides an overview of the original Heavy Duty Engine/Diesel Fuel 2007
rule (HD07) benefits analysis as it relates to the development of a benefits transfer technique.
The HD07 analysis examined the impacts in 2030 of reducing SO2 emissions by 141,000 tons
and NOX emissions by 2,570,000 tons, as well as a 109,000 ton reduction in direct PM emissions.
Table 10.3-2 summarizes the NOx and direct PM results in aggregate and on a per ton basis.
10-11
-------
Draft Regulatory Support Document
Table 10.3-2
Summary of Results from 2030 HP Engine/Diesel Fuel Health Benefits Analysis
Health Outcome
Premature Mortality
All-cause premature mortality from
long-term exposure
NOX
Avoided Incidences
Total
5,027
Per Ton
0.00196
PM
Avoided Incidences
Total
3,007
Per Ton
0.02759
Chronic Illness
Chronic Bronchitis
(pooled estimate)
3,243
0.00126
1,941
0.01781
Hospital Admissions
COPD
Pneumonia
Asthma
Total Cardiovascular
Asthma-Related ER Visits
554
676
523
1,635
1,209
0.00022
0.00026
0.00002
0.00064
0.00047
331
404
313
978
723
0.00304
0.00371
0.00289
0.00897
0.00663
Other Effects
Asthma Attacks
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Work Loss Days
Minor Restricted Activity Days (minus
asthma attacks)
103,905
10,874
118,063
119,760
914,055
4,763,239
0.04043
0.00423
0.04594
0.04660
0.35566
1.85300
62,135
6,515
70,601
71,711
546,744
2,846,434
0.57005
0.05977
0.64771
0.65790
5.01600
26.11407
In the original HD07 analyses, we used the air quality model, REMSAD, which is a three-
dimensional grid-based Eulerian air quality model designed to estimate annual particulate
concentrations and deposition over large spatial scales (e.g., over the contiguous U.S.) as
summarized in Chapter 1 above. The HD07 RIA benefits analysis applies the modeling system
to the entire U.S. for two future-year scenarios: a 2030 base case and a 2030 HD Engine/Diesel
Fuel control scenario. The PM species modeled by REMSAD include a primary fine fraction
(corresponding to particulates less than 2.5 microns in diameter) and several secondary particles
(e.g., sulfates, nitrates, and organics). PM25 is calculated as the sum of the primary fine fraction
10-12
-------
Chapter 10: Benefit-Cost Analysis
and all of the secondary particles.
For the purposes of this analysis, we separated the predicted 2030 change in the primary
and secondarily-formed components of PM25 (i.e., sulfates and nitrates) to provide attributable
health effects for SO2 and NOX. We did this by separating these chemically speciated fractions of
PM (e.g., particulate elemental carbon, and total organic aerosols, sulfate, and particulate nitrate
(PNO3)). It is reasonable to separate these predicted concentrations because of the limited
interactions of secondary sulfate and nitrates within the modeling system and the limited
contribution of secondary organic aerosols (SOA) to TOA (i.e., since there little or no change in
HCs in the original HD07 scenario). Because the original HD07 modeling did not examine the
type of HC reductions that are present in this rulemaking, we are not able to create a transfer
technique for the HC that would contribute to PM formation. Thus, we limit our consideration of
secondary formation of PM to the NOX emissions in this analysis.
To develop the NOx transfer values, we estimated the incidences of the health endpoints
we are able to quantify using the population weighted change in nitrate of -0.388 micrograms per
cubic meter into each of the concentration-response functions used in the HD07 benefits analysis.
This yields estimates of the health effects associated with the NOX emission reductions. Based on
2030 populations, this change leads to the estimated reductions in health effects listed in the
second column of Table 10.3-2. Note that for concentration response (C-R) functions that use
daily average PM2 5 or PM10 levels, use of the annual mean as a proxy for daily averages will over
or underestimate the annual incidence by a small amount (less than five percent). We then
divided the attributable incidences by NOX tons reduced in the HD07 analysis, resulting in
incidences per ton of NOX reduced in 2030 as listed in the third column of Table 10.3-2. We
then scaled the incidences per ton by the ratio of population in the year of analysis to population
in 2030 to obtain incidences per ton for each year (Hubbell 2002).
We conducted a similar operation to develop coefficients for direct PM. In this instance,
we started with the population-weighted change in primary PM of -0.232 micrograms per cubic
meter in the HD07 analysis.
[1]
l = I IP, Ł X Trearl, p X RatlOPOpYearl X ValUBTearl, E
Where
BenefitsYearI = Monetized Benefits in Year /, pollutant P
IP E = Avoided Incidence per ton pollutant P for endpoint E
T year i, p = Tons pollutant P in Year /
RatioPopYearI = Population ratio between year of analysis and 2030
ValueYearI, E = Monetary value per avoided incidence of endpoint E in Year /
10-13
-------
Draft Regulatory Support Document
10.3.4. Quantifying and Valuing Individual Health Endpoints
This section summarizes the studies used to calculate the health incidences and valuation
of those incidences both in the original HD07 benefits analysis and relied on here. Quantifiable
health benefits of the final Large Si/Recreational Vehicle rule may be related to PM only, or both
PM and ozone. We are not estimating any ozone-related benefits, so this analysis is only a
partial quantification of the benefits associated with the emission controls for these categories.
PM-only health effects include premature mortality, chronic bronchitis, acute bronchitis, upper
and lower respiratory symptoms, and work loss days.pp Health effects related to both PM and
ozone include hospital admissions, asthma attacks, and minor restricted activity days.
For this analysis, we rely on concentration response (C-R) functions estimated in
published epidemiological studies relating serious health effects to ambient air quality. The
specific studies from which C-R functions are drawn are included in Table 10.3-3. A complete
discussion of the C-R functions used for this analysis and information about each endpoint are
contained in the HD07 RIA and supporting documents. It is important to note that although there
may be biologically relevant differences between direct PM from diesels and from gasoline
engines, the primary health studies on which the HD07 benefits assessment is based relied on
ambient measurements of PM, not diesel-specific exposure information. Thus, we avoid an
uncertainty of transferring a diesel-PM health estimate to gasoline-PM situation.
While a broad range of serious health effects have been associated with exposure to
elevated PM levels (as noted for example in Table 10.2-1 and described more fully in the ozone
and PM Criteria Documents (US EPA, 1996a, 1996b), we include only a subset of health effects
in this quantified benefit analysis. Health effects are excluded from this analysis for four
reasons:
(i) lack of an adequate benefits transfer technique;
(ii) the possibility of double counting (such as hospital admissions for specific
respiratory diseases);
pp Some evidence has been found linking both PM and ozone exposures with premature
mortality. The SAB has raised concerns that mortality-related benefits of air pollution reductions
may be overstated if separate pollutant-specific estimates, some of which may have been
obtained from models excluding the other pollutants, are aggregated. In addition, there may be
important interactions between pollutants and their effect on mortality (EPA-SAB-Council-
ADV-99-012, 1999; a copy of this document is available in Docket A-99-06, Document IV-A-
20). Because of concern about overstating of benefits and because the evidence associating
mortality with exposure to PM is currently stronger than for ozone, only the benefits related to
the long-term exposure study (ACS/Krewkski, et al, 2000) of mortality are included in the total
primary benefits estimate. A copy of Krewski, et al., can be found in Docket A-99-06, Document
No. IV-G-75.
10-14
-------
Chapter 10: Benefit-Cost Analysis
(iii) uncertainties in applying effect relationships based on clinical studies to the
affected population; and
(iv) a lack of an established C-R relationship.
Table 10.3-3
Endpoints and Studies Included in the Primary Analysis
Endpoint
Premature Mortality
Long-term exposure
Chronic Illness
Chronic Bronchitis (pooled estimate)
Hospital Admissions
COPD
Pneumonia
Asthma
Total Cardiovascular
Asthma-Related ER Visits
Other Illness
Asthma Attacks
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Work Loss Days
Minor Restricted Activity Days (minus asthma
attacks)
Study
Krewski, et al. (2000 )A
Abbey, etal. (1995)
Schwartz, etal. (1993)
Samet, et al. (2000)
Samet, et al. (2000)
Sheppard, etal. (1999)
Samet, et al. (2000)
Schwartz, etal. (1993)
Whittemore and Kom (1 980)
Dockery etal. (1996)
Pope etal. (1991)
Schwartz et al. (1994)
Ostro(1987)
Ostro and Rothschild (1989)
Study Population
Adults, 30 and older
> 26 years
> 29 years
> 64 years
> 64 years
< 65 years
> 64 years
All ages
Asthmatics, all ages
Children, 8-12 years
Asthmatic children, 9-11
Children, 7-14 years
Adults, 18-65 years
Adults, 18-65 years
A Estimate derived from Table 31, PM2.5(DC), All Causes Model (Relative Risk =1.12 for a 24.5 |o.g/m3 increase in mean PM25).
Recently, the Health Effects Institute (HEI) reported findings by investigators at Johns
Hopkins University and others that have raised concerns about aspects of the statistical
methodology used in a number of recent time-series studies of short-term exposures to air
pollution and health effects (Greenbaum, 2002). Some of the concentration-response functions
used in this benefits analysis were derived from such short-term studies. The estimates derived
from the long-term mortality studies, which account for a major share of the benefits in the Base
10-15
-------
Draft Regulatory Support Document
Estimate, are not affected. As discussed in HEI materials provided to sponsors and to the Clean
Air Scientific Advisory Committee (Greenbaum, 2002) these investigators found problems in the
default "convergence criteria" used in Generalized Additive Models (GAM) and a separate issue
first identified by Canadian investigators about the potential to underestimate standard errors in
the same statistical package.qq These and other investigators have begun to reanalyze the results
of several important time series studies with alternative approaches that address these issues and
have found a downward revision of some results. For example, the mortality risk estimates for
short-term exposure to PM10 from The National Morbidity, Mortality and Air Pollution Study
(NMMAPS) were overestimated (this study was not used in this benefits analysis of fine particle
effects)." However, both the relative magnitude and the direction of bias introduced by the
convergence issue is case-specific. In most cases, the concentration-response relationship may
be overestimated; in other cases, it may be underestimated. The preliminary reanalyses of the
mortality and morbidity components of NMMAPS suggest that analyses reporting the lowest
relative risks appear to be affected more greatly by this error than studies reporting higher
relative risks (Dominici et al., 2002; Schwartz and Zanobetti, 2002).
Our examination of the original studies used in this analysis finds that the health
endpoints that are potentially affected by the GAM issues include: reduced hospital admissions
and reduced lower respiratory symptoms in the both the Base and Alternative Estimates; and
reduced premature mortality due to short-term PM exposures in the Alternative Estimate.
While resolution of these issues is likely to take some time, the preliminary results from ongoing
reanalyses of some of the studies used in our analyses (Dominici et al, 2002; Schwartz and
Zanobetti, 2002; Schwartz, personal communication 2002) suggest a more modest effect of the
S-plus error than reported for the NMMAPS PM10 mortality study. While we wait for further
clarification from the scientific community, we have chosen not to remove these results from the
estimated benefits, nor have we elected to apply any interim adjustment factor based on the
preliminary reanalyses. EPA will continue to monitor the progress of this concern, and make
appropriate adjustments as further information is made available.
In Table 10.3-4, we present how we have valued the estimated changes in health effects
and the value functions selected from the peer reviewed literature to provide monetized
estimates. One of the most important effects is premature mortality. While the base value for a
mortality incidence is $6.1 million (1999$), this number is always adjusted downward to reflect
the impact of discounting over the assumed 5 year lag period between reductions in PM
concentrations and full realization of reduced mortality. The lag-adjusted base VSL is $5.8
qqMost of the studies used a statistical package known as "S-plus." For further details, see
http://www.healtheffects.org/Pubs/NMMAPSletter.pdf.
"HEI sponsored the multi-city the National Morbidity, Mortality, and Air Pollution Study
(NMMAPS). See http://biosun01 .biostat.jhsph.edu/~fdominic/NMMAPS/nmmaps-revised.pdf
for revised mortality results. A copy of this document can be found in Docket A-2000-01,
Document IV-A-201.
10-16
-------
Chapter 10: Benefit-Cost Analysis
million (1999$) when a 3% discount rate is assumed. Thus the attached table reflects income
adjustments applied to these lag adjusted base values.
10-17
-------
Draft Regulatory Support Document
Table 10.3-4
Unit Values Used for Economic Valuation of Health Endpoints
Health or Welfare Endpoint
Estimated Value
per Incidence
(1999$)
Central Estimate
Derivation of Estimates
Respiratory Ailments Not Requiring Hospitalization
Premature Mortality
Chronic Bronchitis (CB)
$6 million per
statistical life
$331,000
Value is the mean of value-of-statistical-life estimates from 26
studies (5 contingent valuation and 21 labor market studies)
reviewed for the Section 812 Costs and Benefits of the Clean
Air Act, 1990-2010 (US EPA, 1999).
Value is the mean of a generated distribution of WTP to avoid
a case of pollution-related CB. WTP to avoid a case of
pollution-related CB is derived by adjusting WTP (as
described in Viscusi et al., 1991) to avoid a severe case of CB
for the difference in severity and taking into account the
elasticity of WTP with respect to severity of CB.
Hospital Admissions
Chronic Obstructive
Pulmonary Disease (COPD)
(ICD codes 490-492, 494-496)
Pneumonia
(ICD codes 480-487)
Asthma admissions
All Cardiovascular
(ICD codes 390-429)
Emergency room visits for
asthma
$12,378
$14,693
$6,634
$18,387
$299
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total COPD category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total pneumonia category illnesses)
reported in Elixhauser (1993).
The COI estimates are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total asthma category illnesses) reported in
Elixhauser (1993).
The COI estimates are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital
stay, and weighted share of total cardiovascular illnesses)
reported in Elixhauser (1993).
COI estimate based on data reported by Smith, et al. (1997).
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
$24
Combinations of the 3 symptoms for which WTP estimates are
available that closely match those listed by Pope, et al. result in
10-18
-------
Chapter 10: Benefit-Cost Analysis
Lower Respiratory Symptoms
(LRS)
Acute Bronchitis
$15
$57
Combinations of the 4 symptoms for which WTP estimates are
available that closely match those listed by Schwartz, et al.
result in 1 1 different "symptom clusters," each describing a
"type" of LRS. A dollar value was derived for each type of
LRS, using mid-range estimates of WTP (lEc, 1994) to avoid
each symptom in the cluster and assuming additivity of WTPs.
The dollar value for LRS is the average of the dollar values for
the 11 different types of LRS.
Average of low and high values recommended for use in
Section 812 analysis (Neumann, et al. 1994)
Restricted Activity and Work Loss Days
Work Loss Days (WLDs)
Minor Restricted Activity
Days (MRADs)
Variable
$48
Regionally adjusted median weekly wage for 1990 divided by
5 (adjusted to 1999$) (US Bureau of the Census, 1992).
Median WTP estimate to avoid one MRAD from Tolley, et al.
(1986) .
10.3.5. Estimating Monetized Benefits Anticipated in Each Year
We applied these estimates of the value per incidence to calculate a stream of benefits in
future years. We scaled the benefits to the appropriate future year national populations to reflect
growth in population. Our projections reflect the U.S. Bureau of the Census predictions.
Our analysis accounts for expected growth in real income over time. Economic theory
argues that willingness to pay (WTP) for most goods (such as environmental protection) will
increase if real incomes increase. There is substantial empirical evidence that the income
elasticity88 of WTP for health risk reductions is positive, although there is uncertainty about its
exact value. Thus, as real income increases the WTP for environmental improvements also
increases. While many analyses assume that the income elasticity of WTP is unit elastic (i.e., ten
percent higher real income level implies a ten percent higher WTP to reduce risk changes),
empirical evidence suggests that income elasticity is substantially less than one and thus
relatively inelastic. As real income rises, the WTP value also rises but at a slower rate than real
income.
The effects of real income changes on WTP estimates can influence benefit estimates in
two different ways: (1) through real income growth between the year a WTP study was
conducted and the year for which benefits are estimated, and (2) through differences in income
between study populations and the affected populations at a particular time. Empirical evidence
88Income elasticity is a common economic measure equal to the percentage change in
WTP for a one percent change in income.
10-19
-------
Draft Regulatory Support Document
of the effect of real income on WTP gathered to date is based on studies examining the former.
The Environmental Economics Advisory Committee (EEAC) of the SAB advised EPA to adjust
WTP for increases in real income over time, but not to adjust WTP to account for cross-sectional
income differences "because of the sensitivity of making such distinctions, and because of
insufficient evidence available at present" (EPA-SAB-EEAC-00-013).
Based on a review of the available income elasticity literature, we adjust the valuation of
human health benefits upward to account for projected growth in real U.S. income. Faced with a
dearth of estimates of income elasticities derived from time-series studies, we applied estimates
derived from cross-sectional studies in our analysis. Details of the procedure can be found in
Kleckner and Neumann (1999). An abbreviated description of the procedure we used to account
for WTP for real income growth between 1990 and 2030 is presented in the HD07 TSD.
Incidences in future years will have different values based on adjustments to WTP for
growth in income over time. (The schedule of adjustment factors and adjusted WTP values to be
applied for each year is listed in attachment 2 of the Hubbell 2002, Docket A-2000-01,
Document number IV-A-146.) Adjustment factors should not be applied to the values for
avoided hospital admissions, as these are cost-of-illness estimates and not WTP estimates.
Likewise, adjustment factors should not be applied to the value of work loss days, as this is a
wage-based estimate, not WTP.
10.3.6. Methods for Describing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty." This analysis is no exception. As outlined
both in this and preceding chapters, there are many inputs used to derive the final estimate of
benefits, including emission inventories, air quality models (with their associated parameters and
inputs), epidemiological estimates of C-R functions, estimates of values (both from WTP and
cost-of-illness studies), population estimates, income estimates, and estimates of the future state
of the world (i.e., regulations, technology, and human behavior). Each of these inputs may be
uncertain, and depending on their location in the benefits analysis, may have a disproportionately
large impact on final estimates of total benefits. For example, emissions estimates are a
foundation of the analysis. As such, any uncertainty in emissions estimates will be propagated
through the entire analysis. When compounded with uncertainty in later stages, small
tt It should be recognized that in addition to uncertainty, the annual benefit estimates for
the final Large Si/Recreational Vehicle rule presented in this analysis are also inherently variable,
due to the truly random processes that govern pollutant emissions and ambient air quality in a
given year. Factors such as weather display constant variability regardless of our ability to
accurately measure them. As such, the estimates of annual benefits should be viewed as
representative of the types of benefits that will be realized, rather than the actual benefits that
would occur every year.
10-20
-------
Chapter 10: Benefit-Cost Analysis
uncertainties in emission levels can lead to much larger impacts on total benefits. A more
thorough discussion of uncertainty can be found in the HD07 benefits TSD (Abt Associates,
2000).
Some key sources of uncertainty in each stage of the benefits analysis are:
Gaps in scientific data and inquiry;
• Uncertainties in the benefit transfer process from the HD07 case to the vehicles
covered in this rulemaking;
• Variability in estimated relationships, such as C-R functions, introduced through
differences in study design and statistical modeling;
• Errors in measurement and projection for variables such as population growth
rates;
• Errors due to misspecification of model structures, including the use of surrogate
variables, such as using PM10 when PM2 5 is not available, excluded variables, and
simplification of complex functions; and
Biases due to omissions or other research limitations.
Some of the key uncertainties in the benefits analysis are presented in Table 10.3-5.
There are a wide variety of sources for uncertainty and the potentially large degree of uncertainty
in our estimate. In the original HD07 benefits assessment, sensitivity analyses were performed
including qualitative discussions, probabilistic assessments, alternative calculations, and
bounding exercises. For some parameters or inputs it may be possible to provide a statistical
representation of the underlying uncertainty distribution. For other parameters or inputs, the
information necessary to estimate an uncertainty distribution is not available. Even for
individual endpoints, there is usually more than one source of uncertainty. This makes it difficult
to provide a quantified uncertainty estimate. For example, the C-R function used to estimate
avoided premature mortality has an associated standard error which represents the sampling error
around the pollution coefficient in the estimated C-R function. It would be possible to report a
confidence interval around the estimated incidences of avoided premature mortality based on this
standard error. However, this would omit the contribution of air quality changes, baseline
population incidences, projected populations exposed, and transferability of the C-R function to
diverse locations to uncertainty about premature mortality. Thus, a confidence interval based on
the standard error would provide a misleading picture about the overall uncertainty in the
estimates. Information on the uncertainty surrounding particular C-R and valuation functions is
provided in the HD07 benefits TSD (Abt Associates, 2000). But, this information should be
interpreted within the context of the larger uncertainty surrounding the entire analysis.
Many benefits categories, while known to exist, do not have enough information
available to provide a quantified or monetized estimate. One significant limitation of both the
health and welfare benefits analyses is the inability to quantify many of the serious effects listed
in Table 10.2-1. The uncertainty regarding these endpoints is such that we could determine
10-21
-------
Draft Regulatory Support Document
neither a primary estimate nor a plausible range of values. The net effect of excluding benefit
and disbenefit categories from the estimate of total benefits depends on the relative magnitude of
the effects.
Our estimate of total benefits should be viewed as an approximate result because of the
sources of uncertainty discussed above (see Table 10.3-5). The total benefits estimate may
understate or overstate actual benefits of the rule. In considering the monetized benefits
estimates, the reader should remain aware of the many limitations of conducting these analyses
mentioned throughout this chapter.
Table 10.3-5
Primary Sources of Uncertainty in the Benefit Analysis
1. Uncertainties Associated With Concentration-Response Functions
The value of the PM-coefficient in each C-R function.
Application of a single C-R function to pollutant changes and populations in all locations.
Similarity of future year C-R relationships to current C-R relationships.
Correct functional form of each C-R relationship.
Extrapolation of C-R relationships beyond the range of PM concentrations observed in the study.
Application of C-R relationships only to those subpopulations matching the original study population.
2. Uncertainties Associated With Original Modeled Ambient PM Concentrations
Responsiveness of the models to changes in precursor emissions resulting from the control policy.
Projections of future levels of precursor emissions, especially ammonia and crustal materials.
Model chemistry for the formation of ambient nitrate concentrations.
Comparison of model predictions of particulate nitrate with observed rural monitored nitrate levels indicates
that REMSAD overpredicts nitrate in some parts of the Eastern US and underpredicts nitrate in parts of
the Western US.
3. Uncertainties Associated with PM Mortality Risk
No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
Direct causal agents within the complex mixture of PM have not been identified.
The extent to which adverse health effects are associated with low level exposures that occur many times in
the year versus peak exposures.
The extent to which effects reported in the long-term exposure studies are associated with historically higher
levels of PM rather than the levels occurring during the period of study.
Reliability of the limited ambient PM2 5 monitoring data in reflecting actual PM2 5 exposures.
4. Uncertainties Associated With Possible Lagged Effects
The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM
levels would occur in a single year is uncertain as well as the portion that might occur in subsequent years.
5. Uncertainties Associated With Baseline Incidence Rates
10-22
-------
Chapter 10: Benefit-Cost Analysis
- Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
accurately represent the actual location-specific rates.
Current baseline incidence rates may not approximate well baseline incidence rates in 2030.
— Projected population and demographics may not represent well future-year population and demographics.
6.
Uncertainties Associated With Economic Valuation
Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and
have uncertainty surrounding them.
— Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due
differences in income or other factors.
7.
Uncertainties Associated With Aggregation of Monetized Benefits
therefore
to
- Health and welfare benefits estimates are limited to the available C-R functions. Thus, unquantified or
unmonetized benefits are not included.
8.
Uncertainties introduced by Transferring Benefits from a Previous Mobile Source Benefits Analysis
— The reasonableness of the benefits transfer depends on the similarity of the original analysis and the emission
reductions analyzed with respect to the relationship between emissions and human populations.
10.3.7. Estimated Reductions in Incidences of Health Endpoints and Associated Monetary
Values
Applying the techniques (including the C-R and valuation functions described above) to
the estimated changes in NOx and direct PM emissions yields estimates of the number of
avoided incidences (i.e. premature mortalities, cases, admissions, etc.) and the associated
monetary values for those avoided incidences. These estimates are presented in Table 10.3-6 for
2030. All of the monetary benefits are in constant 2002 dollars.
Not all known PM- and ozone-related health effects could be quantified or monetized.
These unmonetized benefits are indicated by place holders, labeled Bx and B2. In addition,
unmonetized benefits associated with ozone, CO and HC reductions are indicated by the
placeholders B2 B3 and B4. Unquantified physical effects are indicated by Uj through U4. The
estimate of total monetized health benefits is thus equal to the subset of monetized PM-related
health benefits plus BH, the sum of the unmonetized health benefits.
The largest monetized health benefit is associated with reductions in the risk of premature
mortality, which accounts for over $7.5 billion, which is over 95 percent of total monetized
health benefits."" The next largest benefit is for chronic bronchitis reductions, although this value
""Alternative calculations for premature mortality incidences and valuation are presented
in the HD07 RIA in Tables VII-24 and "VTi-25, respectively. An alternative calculation is also
provided in Table "VTi-25 for chronic bronchitis incidences and for chronic asthma incidences.
The HD07 RIA can be found in Docket A-2000-01, Document II-A-13.
10-23
-------
Draft Regulatory Support Document
is more than an order of magnitude lower than for premature mortality. Minor restricted activity
days, work loss days, and worker productivity account for the majority of the remaining benefits.
The remaining categories account for less than $10 million each; however, they represent a large
number of avoided incidences affecting many individuals.
10-24
-------
Chapter 10: Benefit-Cost Analysis
Table 10.3-6
Base-Case Estimate of Annual Health Benefits Associated With Air Quality
Changes Resulting from the Large SI Requirements Only in 2030
Endpoint
PM-related Endpointsc
Premature mortality13 (adults, 30 and over)
Chronic bronchitis (adults, 26 and over)
Hospital Admissions - Pneumonia (adults, over 64)
Hospital Admissions - COPD (adults, 64 and over)
Hospital Admissions - Asthma (65 and younger)
Hospital Admissions - Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)E
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Other PM-related health effects13
Ozone-related Endpoints
CO and HC-related health effects13
Monetized Total Health-related Benefits0
Avoided
IncidenceA
(cases/year)
1,000
640
100
100
100
300
300
20,600
2,200
23,700
23,400
181,300
944,400
u,
U2
U3+U4
—
Monetary Benefits8
(millions 2002$, adjusted
for growth in real
income)
$7,510
$280
<$5
<$5
<$1
<$10
<$1
<$1
<$1
<$1
<$1
$20
$50
B,
B2
B3+B4
$7,880+BH
A Incidences are rounded to the nearest 100.
B Dollar values are rounded to the nearest $10 million.
c PM-related benefits are based on the assumption that Eastern U.S. nitrate reductions are equal to one-fifth the nitrate reductions predicted by
REMSAD (see HD07 RIA Chapter II for a discussion of REMSAD and model performance).
D Premature mortality associated with ozone is not separately included in this analysis (also note that the estimated value for PM-related
premature mortality assumes the 5 year distributed lag structure). Further, PM-related reductions are not quantified for ATVs, OHMs,
snowmobiles and recreational marine diesel.
E A detailed listing of unquantified PM, ozone, CO, and HC related health effects is provided in Table 10.2-1.
F Based upon recent preliminary findings by the Health Effects Institute, the concentration-response functions used to estimate reductions in
hospital admissions may over- or under-estimate the true concentration-response relationship. Our examination of the original studies used in
10-25
-------
Draft Regulatory Support Document
this analysis finds that the health endpoints that are potentially affected by the GAM issues include: reduced hospital admissions and reduced
lower respiratory symptoms. While resolution of these issues is likely to take some time, the preliminary results from ongoing reanalyses of
some of the studies suggest a more modest effect of the S-plus error than reported for the NMMAPS PM10 mortality study. While we wait for
further clarification from the scientific community, we have chosen not to remove these results from the benefits estimates, nor have we elected
to apply any interim adjustment factor based on the preliminary reanalyses. EPA will continue to monitor the progress of this concern, and
make appropriate adjustments as further information is made available.
G Bg is equal to the sum of all unmonetized categories, i.e. Ba+B1+B2+B3+B4.
In Table 10.3-7, we present the benefits over time as the regulations phase in over time
and a net present value, assuming a 3 percent social discount rate.
10-26
-------
Chapter 10: Benefit-Cost Analysis
Table 10.3-7
Monetized Benefits for Large SI Category OnlyA
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Net Present
Nox
Reductions
(tons)
40117
74541
108754
152431
193218
233094
271554
306016
328022
347920
365688
378511
389820
400470
410477
419931
428805
437527
446085
454549
462994
471382
479206
486998
494665
502188
509684
PM
Reductions
(tons)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
Value 2002 -2030
Total
Large SI
Benefits
(thousands $)
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
420,000
800,000
1,180,000
1,670,000
2,150,000
2,630,000
3,110,000
3,820,000
4,160,000
4,480,000
4,790,000
5,030,000
5,270,000
5,490,000
5,710,000
5,900,000
6,130,000
6,320,000
6,540,000
6,750,000
6,950,000
7,120,000
7,280,000
7,440,000
7,600,000
7,740,000
7,880,000
77,180,000
A This analysis excludes the health effects we are not able to quantify for PM, ozone, CO, and HC. A detailed list is
provided in Table 10.2-1. Only NOx and PM reductions from Large SI are quantified. The sizable PM and Nox
reductions from ATVs, OHMs, snowmobiles, and recreational marine diesel are not quantified.
B Dollar values are rounded to the nearest $10 million.
0 A social discount rate of 3 percent is used to calculate the net present value. If a discount rate of 7 percent is used,
the net present value (2002 - 2030) is $40.07 billion.
10-27
-------
Draft Regulatory Support Document
10.3.8 Alternative Calculations of Estimated Reductions in Incidences of Health Endpoints
and Associated Monetary Values
We have also evaluated an alternative, more conservative estimate, that can provide
useful insight into the potential impacts of the key elements underlying estimates of the benefits
of reducing NOx, and PM emissions from this rule through calculated alternative benefits for
mortality and chronic bronchitis. The alternative estimate of mortality reduction relies on certain
recent available scientific studies. These studies found an association between increased
mortality and short-term exposure to PM over days to weeks. The alternative approach uses
different data on valuation and makes adjustments relating to the health status and potential
longevity of the populations most likely affected by PM (for more details see Hubbell 2002b).
We are continuing to examine the merits of applying this alternative approach to the calculation
of benefits. Some of the issues that warrant further investigation are described below.
10.3.9 Alternative Calculations of PM Mortality Risk Estimates and Associated Monetary
Values
The Alternative Estimate addresses uncertainty about the relationship between premature
mortality and long-term exposures to ambient levels of fine particles by assuming that there is no
mortality effect of chronic exposures to fine particles. Instead, it assumes that the full impact of
fine particles on premature mortality can be captured using a concentration-response function
relating daily mortality to short-term fine particle levels. Specifically, a concentration-response
function based on Schwartz et al. (1996) is employed, with an adjustment to account for recent
evidence that daily mortality is associated with particle levels from a number of previous days
(Schwartz, 2000). Previous daily mortality studies (Schwartz et al., 1996) examined the impact
of PM25 on mortality on a single day or over the average of two or more days. Recent analyses
have found that impacts of elevated PM2 5 on a given day can elevate mortality on a number of
following days (Schwartz, 2000; Samet et al., 2000). Multi-day models are often referred to as
"distributed lag" models because they assume that mortality following a PM event will be
distributed over a number of days following or "lagging" the PM event.w
There are no PM25 daily mortality studies which report numeric estimates of relative risks
from distributed lag models; only PM10 studies are available. Daily mortality C-R functions for
PM10 are consistently lower in magnitude than PM2 5-mortality C-R functions, because fine
particles are believed to be more closely associated with mortality than the coarse fraction of PM.
Given that the emissions reductions from heavy duty vehicles result primarily in reduced ambient
concentrations of PM25, use of a PM10 based C-R function results in a significant downward bias
in the estimated reductions in mortality. To account for the full potential multi-day mortality
w It is of note that, based on recent preliminary findings from the Health Effects Institute (http://www.healtheffects.org), the
magnitude of mortality from short-tern exposure may be under or overestimated.
10-28
-------
Chapter 10: Benefit-Cost Analysis
impact of acute PM2 5 events, we use the distributed lag model for PM10 reported in Schwartz
(2000) to develop an adjustment factor which we then apply to the PM2 5 based C-R function
reported in Schwartz et al. (1996). If most of the increase in mortality is expected to be
associated with the fine fraction of PM10, then it is reasonable to assume that the same
proportional increase in risk would be observed if a distributed lag model were applied to the
PM25 data. There are two relevant coefficients from the Schwartz et al. (1996) study, one
corresponding to all-cause mortality, and one corresponding to chronic obstructive pulmonary
disease (COPD) mortality (separation by cause is necessary to implement the life years lost
approach detailed below).
These estimates, while approximating the full impact of daily pollution levels on daily
death counts, do not capture any impacts of long-term exposure to air pollution. EPA's Science
Advisory Board, while acknowledging the uncertainties in estimation of a PM-mortality
relationship, has recommended the use of a study that does reflect the impacts of long-term
exposure. The omission of long-term impacts accounts for an approximately 40 percent
reduction in the estimate of avoided premature mortality in the alternative estimates relative to
the primary estimates.
Furthermore, the alternative estimates reflect the impact of changes to key assumptions
associated with the valuation of mortality. These include: 1) the impact of using wage-risk and
contingent valuation-based value of statistical life estimates in valuing risk reductions from air
pollution as opposed to contingent valuation-based estimates alone, 2) the relationship between
age and willingness-to-pay for fatal risk reductions, and 3) the degree of prematurity in
mortalities from air pollution.
The alternative estimates address this issue by using an estimate of the value of statistical
life that is based only on the set of five contingent valuation studies included in the larger set of
26 studies recommended by Viscusi (1992) as applicable to policy analysis. The mean of the five
contingent valuation based VSL estimates is $3.7 million (1999$), which is approximately 60
percent of the mean value of the full set of 26 studies.
The second issue is addressed by assuming that the relationship between age and
willingness-to-pay for fatal risk reductions can be approximated using an adjustment factor
derived from Jones-Lee (1989). The SAB has advised the EPA that the appropriate way to
account for age differences is to obtain the values for risk reductions from the age groups
affected by the risk reduction.
To show the maximum impact of the age adjustment, the Alternative Estimate is based on
the Jones-Lee (1989) adjustment factor of 0.63, which yields a VSL of $2.3 million for
populations over the age of 70. Deaths of individuals under the age of 70 are valued using the
unadjusted mean VSL value of $3.7 million (1999$). Since these are acute mortalities, it is
assumed that there is no lag between reduced exposure and reduced risk of mortality.
10-29
-------
Draft Regulatory Support Document
A simpler and potentially less biased approach is to simply apply a single age adjustment
based on whether the individual was over or under 65 years of age at the time of death. This is
consistent with the range of observed ages in the Jones-Lee studies and also agrees with the
findings of more recent studies by Krupnick et al. (2000) that the only significant difference in
WTP is between the over 70 and under 70 age groups. To correct for the potential extrapolation
error for ages beyond 70, the adjustment factor is selected as the ratio of a 70 year old
individual's WTP to a 40 year old individual's WTP, which is 0.63, based on the Jones-Lee
(1989) results and 0.92 based on the Jones-Lee (1993) results.
The third issue is addressed in the Alternative Estimate by assuming that deaths from
chronic obstructive pulmonary disease (COPD) are advanced by 6 months, and deaths from all
other causes are advanced by 5 years. These reductions in life years lost are applied regardless of
the age at death. Actuarial evidence suggests that individuals with serious preexisting
cardiovascular conditions have a remaining life expectancy of around 5 years. While many
deaths from daily exposure to PM may occur in individuals with cardiovascular disease, studies
have shown relationships between all cause mortality and PM, and between PM and mortality
from pneumonia (Schwartz, 2000). In addition, recent studies have shown a relationship
between PM and non-fatal heart attacks, which suggests that some of the deaths due to PM may
be due to fatal heart attacks (Peters et al., 2001). And, a recent meta-analysis has shown little
effect of age on the relative risk from PM exposure (Stieb et al. 2002), which suggests that the
number of deaths in non-elderly populations (and thus the potential for greater loss of life years)
may be significant. Indeed, this analysis estimates that 21 percent of non-COPD premature
deaths avoided are in populations under 65. Thus, while the assumption of 5 years of life lost
may be appropriate for a subset of total avoided premature mortalities, it may over or
underestimate the degree of life shortening attributable to PM for the remaining deaths.
In order to value the expected life years lost for COPD and non-COPD deaths, we need to
construct estimates of the value of a statistical life year. The value of a life year varies based on
the age at death, due to the differences in the base VSL between the 65 and older population and
the under 65 population. The valuation approach used is a value of statistical life years (VSLY)
approach, based on amortizing the base VSL for each age cohort. Previous applications have
arrived at a single value per life year based on the discounted stream of values that correspond to
the VSL for a 40 year old worker (U.S. EPA, 1999a). This assumes 35 years of life lost is the
base value associated with the mean VSL value of $3.7 million (1999$). The VSLY associated
with the $3.7 million VSL is $163,000, annualized assuming EPA's guideline value of a 3
percent discount rate, or $270,000, annualized assuming OMB's guideline value of a 7 percent
discount rate.
The VSL applied in this analysis is then built up from that VSLY by taking the present
value of the stream of life years, again assuming a 3% discount rate. Thus, if you assume that a
40 year-old dying from pneumonia would lose 5 years of life, the VSL applied to that death
would be $0.79 million. For populations over age 65, we then develop a VSLY from the age-
10-30
-------
Chapter 10: Benefit-Cost Analysis
adjusted base VSL of $2.3 million. Given an assumed remaining life expectancy of 10 years, this
gives a VSLY of $258,000, assuming a 3 percent discount rate. Again, the VSL is built based on
the present value of 5 years of lost life, so in this case, we have a 70 year old individual dying
from pneumonia losing 5 years of life, implying an estimated VSL of $1.25 million. COPD
deaths for populations aged 65 and older are valued at $0.13 million per incidence. Finally,
COPD deaths for populations aged 64 and younger are valued at $0.09 million per incidence.
The implied VSL for younger populations is less than that for older populations because the
value per life year is higher for older populations. Since we assume that there is a 5 year loss in
life years for a PM related mortality, regardless of the age of person dying, this necessarily leads
to a lower VSL for younger populations. As a final step, these estimated VSL values are
multiplied by the appropriate adjustment factors to account for changes in WTP over time.
10.3.9.1 Alternative Calulations of Chronic Bronchitis Monetary Values
For the alternative estimate, a cost-of illness value is used in place of willingness-to-pay
to reflect uncertainty about the value of reductions in incidences of chronic bronchitis. In the
primary estimate, the willingness-to-pay estimate was derived from two contingent valuation
studies (Viscusi et al., 1991; Krupnick and Cropper, 1992). These studies were experimental
studies intended to examine new methodologies for eliciting values for morbidity endpoints.
Although these studies were not specifically designed for policy analysis, the SAB (EPA-SAB-
COUNCIL-ADV-00-002, 1999) has indicated that the severity-adjusted values from this study
provide reasonable estimates of the WTP for avoidance of chronic bronchitis. As with other
contingent valuation studies, the reliability of the WTP estimates depends on the methods used to
obtain the WTP values. In order to investigate the impact of using the CV based WTP estimates,
the alternative estimates rely on a value for incidence of chronic bronchitis using a cost-of-illness
estimate based Cropper and Krupnick (1990) which calculates the present value of the lifetime
expected costs associated with the illness. The current cost-of-illness (COI) estimate for chronic
bronchitis is around $107,000 per case, compared with the current WTP estimate of $330,000.
Because the alternative estimate is based on cost-of-illness, no income adjustments are applied
when applying the estimate in future year analyses.
10.3.9.2 Alternative Calulations Results
Applying the techniques (including the C-R and valuation alternatives described above)
to the estimated changes in NOx and direct PM emissions for Large SI engines from this rule
yields estimates of the number of avoided incidences of premature mortalities and chronic
bronchitis cases and the associated monetary values for those avoided incidences. These
estimates are presented in Table 10.3-8 for 2030. All of the monetary benefits are in constant
2002 dollars.
10-31
-------
Draft Regulatory Support Document
Table 10.3-8.
Alternative Benefits in 2030 from PM-related Reductions from the Large SI Categories
Short-term exposure
mortality
Chronic bronchitis
Alternative
Estimate IncidenceA
600
640
Alternative
Estimation Valuation8
(million $)
$810
$90
A Incidences are rounded to the nearest 10.
B Dollar values are rounded to the nearest $10 million.
In Table 10.3-9, we present the benefits over time as the regulations phase in over time and a net
present value, assuming a 3 percent social discount rate.
10-32
-------
Chapter 10: Benefit-Cost Analysis
Table 10.3-9
Alternative Monetized Benefits Mortality and Chronic Bronchitis
for Large SI Category Only'
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Net Present
Nox
Reductions
40,117
74,541
108,754
152,431
193,218
233,094
271,554
306,016
328,022
347,920
365,688
378,511
389,820
400,470
410,477
419,931
428,805
437,527
446,085
454,549
462,994
471,382
479,206
486,998
494,665
502,188
509,684
Value 2002 to 2030
PM
Reductions
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
Total
Benefits
(thousands)
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
$
50,000
90,000
130,000
190,000
250,000
300,000
350,000
440,000
470,000
510,000
550,000
570,000
600,000
620,000
650,000
670,000
700,000
720,000
750,000
770,000
790,000
810,000
830,000
850,000
870,000
880,000
900,000
$8,800 million
A This alternative analysis excludes the health effects we are not able to quantify for PM, ozone, CO, and HC as well
as excluding benefits from long-term exposure mortality, hospital admissions, emergency department visits, upper
and lower respiratory symptoms, asthma attacks, acute bronchitis, work loss days and minor restricted activity days.
A detailed list is provided in Table 10.2-1. Only NOx and PM reductions from Large SI are quantified. The sizable
PM and Nox reductions from ATVs, OHMs, snowmobiles, and recreational marine diesel are not quantified.
B Dollar values are rounded to the nearest $10 million.
0 A social discount rate of 3 percent is used to calculate the net present value. If a discount rate of 7 percent is used,
the net present value (2002 - 2030) is $4.57 billion.
10-33
-------
Draft Regulatory Support Document
10.4 CO and Air Toxics Health Benefits Estimation
Although we achieve substantial reductions in CO and HC (many of which are hazardous
air pollutants), we are unable to quantify benefits for these reductions. We present two
techniques for estimating the economic benefits of changes in emissions from snowmobiles that
are possible areas for further reserach.
10.4.1 Direct Valuation of "Clean" Snowmobiles
In general, economists tend to view an individual's willingness-to-pay (WTP) for a
improvement in environmental quality as the appropriate measure of the value of a risk
reduction. An individual's willingness-to-accept (WTA) compensation for not receiving the
improvement is also a valid measure. However, WTP is generally considered to be a more readily
available and conservative measure of benefits. Adoption of WTP as the measure of value
implies that the value of environmental quality improvements is dependent on the individual
preferences of the affected population and that the existing distribution of income (ability to pay)
is appropriate.
For many goods, WTP can be observed by examining actual market transactions. For
example, if a gallon of bottled drinking water sells for one dollar, it can be observed that at least
some persons are willing to pay one dollar for such water. For goods not exchanged in the
market, such as most environmental "goods," valuation is not as straightforward. Nevertheless, a
value may be inferred from observed behavior, such as sales and prices of products that result in
similar effects or risk reductions, (e.g., non-toxic cleaners or safety devices). Alternatively,
surveys may be used in an attempt to directly elicit WTP for an environmental improvement.
One distinction in environmental benefits estimation is between use values and non-use
values. Although no general agreement exists among economists on a precise distinction
between the two (see Freeman, 1993), the general nature of the difference is clear. Use values
are those aspects of environmental quality that affect an individual's welfare more or less
directly. These effects include changes in product prices, quality, and availability, changes in the
quality of outdoor recreation and outdoor aesthetics, changes in health or life expectancy, and the
costs of actions taken to avoid negative effects of environmental quality changes.
Non-use values are those for which an individual is willing to pay for reasons that do not
relate to the direct use or enjoyment of any environmental benefit, but might relate to existence
values and bequest values. Non-use values are not traded, directly or indirectly, in markets. For
this reason, the measurement of non-use values has proved to be significantly more difficult than
the measurement of use values. The air quality changes produced by the final Large
Si/Recreational Vehicle rule cause changes in both use and non-use values, but the monetary
10-34
-------
Chapter 10: Benefit-Cost Analysis
benefit estimates are almost exclusively for use values.
The most direct way to measure the economic value of air quality changes is in cases
where the endpoints have market prices. More frequently than not, the economic benefits from
environmental quality changes are not traded in markets, so direct measurement techniques can
not be used.
Estimating benefits for public land activities or its existence value is a more difficult and
less precise exercise because the endpoints are not directly or indirectly valued in markets. For
example, the loss of a species of animal or plant from a particular habitat does not have a well-
defined price, neither does a crisp winter day of quietude. The contingent valuation (CV) method
has been employed in the economics literature to value endpoint changes for both visibility and
ecosystem functions (Chestnut and Dennis, 1997). There is an extensive scientific literature and
body of practice on both the theory and technique of CV. EPA believes that well-designed and
well-executed CV studies are valid for estimating the benefits of air quality regulation."™
The contingent valuation (CV) method uses survey techniques to estimate values
individuals place on goods and services for which no market exists. Contingent valuation has
been widely applied (Mitchell and Carson 1989, and Walsh, Johnson, and McKean 1992), and
the U.S. Water Resources Council recognizes this as an appropriate method. The U.S.
Department of Interior's federal guidelines have designated CV as the best available procedure
for valuing damages arising in Superfund natural resource damage cases (U.S. DOT 1986, 1991).
The CV method values endpoints by using carefully structured surveys to ask a sample of
people what amount of compensation is equivalent to a given change in environmental quality.
In a CV survey, individuals are asked about their willingness to pay for a given service or
commodity contingent on their acceptance of a hypothetical but plausible and realistic market
situation. Thus, there are three main elements in the approach: 1) a description of the commodity
to be valued; 2) the payment vehicle (i.e., how the individual will pay for the good or service);
and 3) the form of the question (e.g., open-ended or dichotomous choice questions). A study that
""Concerns about the reliability of value estimates from CV studies arose because
research has shown that bias can be introduced easily into these studies if they are not carefully
conducted. Accurately measuring WTP for avoided health and welfare losses depends on the
reliability and validity of the data collected. There are several issues to consider when evaluating
study quality, including but not limited to 1) whether the sample estimates of WTP are
representative of the population WTP; 2) whether the good to be valued is comprehended and
accepted by the respondent; 3) whether the WTP elicitation format is designed to minimize
strategic responses; 4) whether WTP is sensitive to respondent familiarity with the good, to the
size of the change in the good, and to income; 5) whether the estimates of WTP are broadly
consistent with other estimates of WTP for similar goods; and 6) the extent to which WTP
responses are consistent with established economic principles.
10-35
-------
Draft Regulatory Support Document
contained information about use value for "clean, quiet" snowmobiles was recently conducted
(Duffield and Neher 2000).xx However, the study was judged to have limitations in its
application here. The National Park Service is endeavoring to conduct a new study that may
address the short-comings of this study.
10.4.2 Overview of Benefits Estimation for CO and Air Toxics from the Final Rule
A large variety of substances is emitted from tail pipes of snowmobiles powered by two-
stroke engines.1 Some of these substances may be acutely neurotoxic at sufficiently high
concentration, including volatile hydrocarbons (HC) and carbon monoxide (CO). The acute
neurotoxicity of only two of the identified exhaust components have been studied extensively on
an individual basis (toluene and CO), but the combined toxicity of the mixture of toluene and CO
has not been evaluated.2 Toluene comprises about 20 percent of the total amount of
hydrocarbons in the exhaust of snowmobiles.3 As discussed above, up to a third of the fuel and
lubricating oil mixture delivered to the 2-stroke snowmobile engine is emitted directly without
being burned.
Ideally, we would have quantified the economic benefit of reductions in all of these
pollutants from vehicles subject to our final rule. In developing a method to quantify economic
benefits for the reduction of these toxic pollutants, however, we were limited by the available
exposure literature to modeling a specific common exposure scenario for snowmobiles. After
detailed subsequent investigation of the limited exposure information, we judge the study to
contain too many unresolved uncertainties to be used in this analysis. Further, we are not able to
quantify exposures related to other high-emitting 2-stroke engines in ATVs or OHMCs.
Furthermore, there are substantial uncertainties in the analysis and gaps in our underlying
knowledge. More research is needed, especially regarding exposure to neurotoxicants emitted
from these and other categories of 2-stroke engines to facilitate benefits calculations.
If after further study, we learn that off-road vehicle operators are exposed to combined
levels of neurotoxi cants at levels that impair skills related to driving ability,4 then reductions in
these exposures could result in fewer accidents and avoided medical and property damage costs.
However, we were limited by gaps in knowledge about exposure estimates and health effects
related to most neurotoxic compounds. For air toxics and CO, it can be important to consider
both momentary blood dose as well as longer term exposures in evaluating the health effects and
monetary benefits.
xxDuffield, JW and CJ Neher. Winter 1998-99 Visitor Survey: Yellowstone National
Park, Grand Teton National Park, and Greater Yellowstone Area. May 2000. Docket A-2000-01,
Document IV-A-113. The survey instrument and the report were independently peer-reviewed.
10-36
-------
Chapter 10: Benefit-Cost Analysis
10.5 Total Benefits
We provide our base-case estimate of benefits for each health and welfare endpoint as
well as the resulting base-case estimate of total benefits. To obtain this estimate, we aggregate
dollar benefits associated with each of the effects examined, such as hospital admissions, into a
total benefits estimate assuming that none of the included health and welfare effects overlap.
The base-case estimate of the total benefits associated with the health and welfare effects is the
sum of the separate effects estimates. Total monetized benefits associated with the final Large
Si/Recreational Vehicle rule are listed in Table 10.5-1, along with a breakdown of benefits for
the Large SI category only by endpoint. Note that the value of endpoints known to be affected by
ozone and/or PM that we are not able to monetize are assigned a placeholder value (e.g., Bl3 B2,
etc.). Unquantified physical effects are indicated by a U. The estimate of total benefits is thus
the sum of the monetized benefits and a constant, B, equal to the sum of the unmonetized
benefits, B1+B2+...+Bn.
A comparison of the incidence column to the monetary benefits column reveals that there
is not always a close correspondence between the number of incidences avoided for a given
endpoint and the monetary value associated with that endpoint. For example, there many times
more asthma attacks than premature mortalities, yet these asthma attacks account for only a very
small fraction of total monetized benefits. This reflects the fact that many of the less severe
health effects, while more common, are valued at a lower level than the more severe health
effects. Also, some effects, such as asthma attacks, are valued using a proxy measure of WTP.
As such the true value of these effects may be higher than that reported in Table 10.5-1.
10-37
-------
Draft Regulatory Support Document
Table 10.5-1
Base-Case Estimate of Annual Health Benefits Associated With
Air Quality Changes Resulting from the Large Si/Recreational Vehicle Rule in 2030
Endpoint
PM-related Endpoints0
Premature mortality13 (adults, 30 and over)
Chronic bronchitis (adults, 26 and over)
Hospital Admissions - Pneumonia (adults, over 64)F
Hospital Admissions - COPD (adults, 64 and over)
Hospital Admissions - Asthma (65 and younger)
Hospital Admissions - Cardiovascular (adults, over 64)
Emergency Room Visits for Asthma (65 and younger)
Asthma Attacks (asthmatics, all ages)E
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Other PM-related health effects13
Ozone-related Endpoints
Quantified HC-related WTP
CO and HC-related health effects13
Monetized Total Health-related Benefits0
Avoided
Incidence*
(cases/year)
1,000
640
100
100
100
300
300
20,600
2,200
23,700
23,400
181,300
944,400
u,
U2
-
U4+U5
—
Monetary Benefits8
(millions 2002$, adjusted
for growth in real
income)
$7,510
$280
<$5
<$5
<$1
<$10
<$1
<$1
<$1
<$1
<$1
$20
$50
B,
B2
U3
B3
$7,880 +BH
A Incidences are rounded to the nearest 100. Nox and PM-related reductions are not quantified for ATVs, OHMs, snowmobiles and recreational
marine diesel.
B Dollar values are rounded to the nearest $10 million.
0 PM-related benefits are based on the assumption that Eastern U.S. nitrate reductions are equal to one-fifth the nitrate reductions predicted by
REMSAD (see HD07 RIA Chapter II for a discussion of REMSAD and model performance).
D Premature mortality associated with ozone is not separately included in this analysis (also note that the estimated value for PM-related
premature mortality assumes the 5 year distributed lag structure).
E A detailed listing of unquantified PM, ozone, CO, and HC related health effects is provided in Table 10.2-1.
Based upon recent preliminary findings by the Health Effects Institute, the concentration-response functions used to estimate reductions in hospital admissions may over- or
under-estimate the true concentration-response relationship. Our examination of the original studies used in this analysis finds that the health endpoints that
are potentially affected by the GAM issues include: reduced hospital admissions and reduced lower respiratory symptoms. While resolution of
these issues is likely to take some time, the preliminary results from ongoing reanalyses of some of the studies suggest a more modest effect of
the S-plus error than reported for the NMMAPS PM10 mortality study. While we wait for further clarification from the scientific community, we
have chosen not to remove these results from the benefits estimates, nor have we elected to apply any interim adjustment factor based on the
preliminary reanalyses. EPA will continue to monitor the progress of this concern, and make appropriate adjustments as further information is
made available.
G Bg is equal to the sum of all unmonetized categories, i.e. Bj+Bj
-------
Chapter 10: Benefit-Cost Analysis
10.6 Comparison of Costs to Benefits
Benefit-cost analysis provides a valuable framework for organizing and evaluating
information on the effects of environmental programs. When used properly, benefit-cost analysis
helps illuminate important potential effects of alternative policies and helps set priorities for
closing information gaps and reducing uncertainty. According to economic theory, the efficient
policy alternative maximizes net benefits to society (i.e., social benefits minus social costs).
However, not all relevant costs and benefits can be captured in any analysis. Executive Order
12866 clearly indicates that unquantifiable or nonmonetizable categories of both costs and
benefits should not be ignored. There are many important unquantified and unmonetized costs
and benefits associated with reductions in emissions, including many health and welfare effects.
Potential benefit categories that have not been quantified and monetized are listed in Table 10.2-
1 of this chapter.
The estimated social cost (measured as changes in consumer and producer surplus) in
2030 to implement the final Large Si/Recreational Vehicle program from Chapter 9 is $216
million (2001$). The net social gain, considering fuel efficiency, is $553 million. The monetized
benefits are approximately $7.8 billion, and EPA believes there is considerable value to the
public of the benefits it could not monetize. The net benefit that can be monetized is $8.4 billion.
Therefore, implementation of the Large Si/Recreational Vehicle program is expected to provide
society with a net gain in social welfare based on economic efficiency criteria. Table 10.6-1
summarizes the costs, benefits, and net benefits.
10-39
-------
Draft Regulatory Support Document
Table 10.6-1
Social Gains
Monetized PM-related benefits'"'0
Monetized Ozone-related benefitsM
HC-related benefits
CO-related benefits
Total annual benefits
Monetized net benefits6
Millions of 2001$"
$550
$7,880 + BPM
not monetized ( B0zone)
not monetized ( BHC )
not monetized (Bco)
$7,880 +BPM + B0zolle + BHC + Bco
$8,430 + B
a For this section, all costs and benefits are rounded to the nearest 10 million. Thus, figures presented in this chapter may
not exactly equal benefit and cost numbers presented in earlier sections of the chapter.
b Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that
have not been quantified and monetized are listed in Table IX-E.2. Unmonetized PM- and ozone-related benefits are
indicated by BPM. And B0zQne, respectively.
0 Based upon recent preliminary findings by the Health Effects Institute, the concentration-response functions used to
estimate reductions in hospital admissions may over- or under-estimate the true concentration-response relationship.
dThere are substantial uncertainties associated with the benefit estimates presented here, as compared to other EPA
analyses that are supported by specific modeling. This analysis used a benefits transfer technique described in the RSD.
e B is equal to the sum of all unmonetized benefits, including those associated with PM, ozone, CO, and HC.
The net present value of the future benefits has also been calculated, using a 3 percent
discount rate over the 2002 to 2030 time frame. The net present value of the social gains, from
Table 9.1-7 of Chapter 9, is $4,930 million. The net present value of the total annual benefits,
from Tables 10.3-7 and 10.4-3, is $77,177 million + B. Consequently, the net present value of
the monetized net benefits of this program is $82,107 million.
For each of the vehicle categories, the net present value of the future streams of surplus
losses, fuel savings, social costs/gains, health and environmental benefits and net cost/benefits
have been calculated. The net present values of these future streams are calculated using a 3
percent discount rate (in Chapters 9, 10, and 11) and are calculated over the 2002 to 2030 time
frame.
These net present value estimates are sensitive to the discount rate. Table 10.6-2 presents
an alternative net present value calculation of the surplus loss, fuel savings, social costs/gains,
10-40
-------
Chapter 10: Benefit-Cost Analysis
health and environmental benefits, and net cost or benefits for the control programs being
adopted in this rulemaking, for each vehicle category, for the period 2002 to 2030, assuming an
alternative discount rate of 7%.
Table 10.6-2
Net Present Values*, Fuel Cost Savings, and Social Costs/Gains
(millions of 2001$)**
Vehicle Category
CI Marine
Forklifts
Other Large Si""
Snowmobiles
ATVs
Off-Highway Motorcycles
Total
NPV of Surplus
Loss
$59.0
$415.8
$419.7
$296.9
$491.9
$206.2
$1,889.5
NPV of Fuel Cost
Savings
$0.0
$2,644.2
$804.8
$459.7
$253.0
$120.6
$4,282.3
NPV of Social
Costs/Gains
***
$59.0
($2,228.4)
($385.1)
($162.8)
$238.9
$85.6
($2,392.8)
* Net Present Values are calculated using a discount rate of 7 percent over the 2002 - 2030 time period.
** Figures are in year 2000 and 2001 dollars, depending on the vehicle category; () represents a negative cost
(social gain).
***Figures in this column exclude estimated health and environmental benefits.
****Figures in this row are engineering cost estimates. See Section 9.7.6 of Chapter 9.
The net present value of the future benefits has also been calculated, using a 7 percent
discount rate over the 2002 to 2030 time frame. The net present value of the social gains from
above, is $2,393 million. The net present value of the total annual health and environmental
benefits that we were able to quantify using a 7 percent discount rate is $40,070 million + B.
Consequently, the net present value of the monetized net benefits of this program using a 7
percent discount rate is $42,477 + B million.
10-41
-------
Draft Regulatory Support Document
Chapter 10 Bibliography
Abbey, D.E., B.L. Hwang, RJ. Burchette, T. Vancuren, and P.K. Mills. 1995a. "Estimated
Long-Term Ambient Concentrations of PM(10) and Development of Respiratory Symptoms in a
Nonsmoking Population." Archives of Environmental Health 50(2): 139-152.
Abbey, D.E., F. Petersen, P. K. Mills, and W. L. Beeson. 1993. "Long-Term Ambient
Concentrations of Total Suspended Particulates, Ozone, and Sulfur Dioxide and Respiratory
Symptoms in a Nonsmoking Population." Archives of Environmental Health 48(1): 33-46.
Abt Associates Inc., 2000.. Heavy Duty Engine/Diesel Fuel Rule: Air Quality Estimation,
Selected Health and Welfare Benefits Methods, and Benefit Analysis Results. Prepared for US
EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC. November. A
copy of this document can be found in Docket A-2000-01, Document IV-A-219.
Abt Associates, Inc., 1999a. Tier2 Final Rule: Air Quality Estimation, Selected Health
and Welfare Benefits Methods, and Benefit Analysis Results, Prepared for the US Environmental
Protection Agency, Office of Air Quality Planning and Standards. Research Triangle Park, NC,
November.
Abt Associates Inc., 1999b. Memorandum to Lisa Conner, US EPA, Office of Air Quality
Planning and Standards. "Derivation of 2030 Population Estimates for the Tier II Analysis."
September.
Abt Associates Inc., 1998. Air Quality Estimation for the NOx SIP Call RIA. Prepared for US
EPA, Office of Air Quality Planning and Standards, under contract no. 68-D-98-001. Research
Triangle Park, NC. September. See EPA Air Docket A-96-56, Document No. VI-B-09.
American Lung Association, 1999. Chronic Bronchitis. Web site available at:
http://www.lungusa.org/diseases/lungchronic.html.
Blumenschein, K. and M. Johannesson. 1998. "Relationship Between Quality of Life
Instruments, Health State Utilities, and Willingness to Pay in Patients with Asthma." Annals of
Allergy, Asthma, and Immunology 80:189-194.
Benignus, V, W Boyes, and P. Bushnell, US Environmental Protection Agency. National Health
and Environmental Effects Research Laboratory. Memorandum to the Docket. Acute
Behavioral Effects of Exposure to Toluene and Carbon Monoxide from Snowmobile Exhaust.
September 2002. A copy of this document can be found in Docket A-2000-01, Document IV-
A-143.
Chestnut, L.G. 1997. Draft Memorandum: Methodology for Estimating Values for Changes in
10-42
-------
Chapter 10: Benefit-Cost Analysis
Visibility at National Parks. April 15.
Chestnut, L.G. and R.L. Dennis. 1997. Economic Benefits of Improvements in Visibility: Acid
Rain Provisions of the 1990 Clean Air Act Amendments. Journal of Air and Waste Management
Association 47:395-402.
Chestnut, L.G., and R.D. Rowe. 1990b. A New National Park Visibility Value Estimates. In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C.V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
Chestnut, L.G. and R.D. Rowe. 1990a. Preservation Values for Visibility Protection at the
National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards,
US Environmental Protection Agency, Research Triangle Park, NC and Air Quality Management
Division, National Park Service, Denver, CO.
Crocker T. D. and R. L. Horst, Jr. 1981. "Hours of Work, Labor Productivity, and Environmental
Conditions: a Case Study." The Review of Economics and Statistics 63:361-368.
Delfmo, R.J., R.S. Zeiger, J.M. Seltzer, and D.H. Street. 1998. Symptoms in Pediatric
Asthmatics and Air Pollution: Differences in Effects by Symptom Severity, Anti-inflammatory
Medication Use and Particulate Averaging Time. Environmental Health Perspectives., 106(11):
751-761.
Delfmo, R.J., R.S. Zeiger, J.M. Seltzer, and D.H. Street, 1998. Symptoms in Pediatric
Asthmatics and Air Pollution: Differences in Effects by Symptom Severity. Anti-inflammatory
Medication Use and Particulate Averaging Time. Environmental Health Perspectives, 106(11):
751-761.
Dennis, R.L., 1997. "Using the Regional Acid Deposition Model to Determine the Nitrogen
Deposition Airshed of the Chesapeake Bay Watershed." In: Atmospheric Deposition to the Great
Lakes and Coastal Waters. Baker, J.E., ed. Society of Environmental Toxicology and
Chemistry, Pensacola, FL. pp. 393-413. See EPA Air Docket A-96-56, Document No. VI-B-09.
Desvouges, W.H., F. R. Johnson, andH.S. Banzhaf. 1998. Environmental Policy Analysis with
Limited Information: Principles and Applications of the Transfer Method. Edward Elgar,
Northhampton, MA.
Dickie, M. et al. 1991. Reconciling Averting Behavior and Contingent Valuation Benefit
Estimates of Reducing Symptoms of Ozone Exposure (draft), as cited in Neumann, J.E., Dickie,
M.T., and R.E. Unsworth. 1994. Industrial Economics, Incorporated. Memorandum to Jim
DeMocker, US EPA, Office of Air and Radiation. March 31.
10-43
-------
Draft Regulatory Support Document
Dillman, D. 1992. Mail and Telephone Surveys. John Wiley Publishing. New York.
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, E.G. Ferris and F.E.
Speizer. 1993. "An association between air pollution and mortality in six U.S. cities." New
England Journal of Medicine 329(24): 1753-1759.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H.
Ware, M. Raizenne and F.E. Speizer. 1996. "Health Effects of Acid Aerosols On North
American Children-Respiratory Symptoms." Environmental Health Perspectives. 104(5): 500-
505.
Douglas, S.G. and R.K. Iwamiya, 1999. Estimating the Effects of the Tier 2 Motor-Vehicle
Standards on Air Quality: Ozone. Prepared for Abt Associates, Inc. Prepared by Systems
Applications International, Inc. SYSAPP-99-98/50. January.
Duffield, J.W., Neher, C. J. Winter 1998-99 Visitor Survey Yellowstone N.P., Grand Teton
N.P., and the Greater Yellowstone Area Final Report. Report prepared for National Park
Service, Denver, CO. May, 2000. This document is available in Docket A-2000-01, Document
IV-A-114.
E.H. Pechan & Associates, 1996. Regional Particulate Control Strategies Phase II. Prepared for
the US Environmental Protection Agency, Office of Policy, Planning, and Evaluation.
Washington, DC. September.
E.H. Pechan & Associates, 1999. Emissions and Air Quality Impacts of Proposed Motor Vehicle
Tier 2 and Fuel Sulfur Standards. Prepared by The Pechan-Avanti Group under EPA Contract
No. 68-D9-8052. for US EPA, Office of Air Quality Planning and Standards, Innovative
Strategies and Economics Group. Research Triangle Park, NC. January.
Elixhauser, A., R.M. Andrews, and S. Fox. 1993. Clinical Classifications for Health Policy
Research: Discharge Statistics by Principal Diagnosis and Procedure. Agency for Health Care
Policy and Research (AHCPR), Center for General Health Services Intramural Research, US
Department of Health and Human Services.
Empire State Electric Energy Research Corporation (ESEERCO). 1994. New York State
Environmental Externalities Cost Study. Report 2: Methodology. Prepared by: RCG/Hagler,
Bailly, Inc., November.
EPA-SAB, Health and Ecological Effects Subcommittee of the Advisory Council on Clean Air
Act Compliance Analysis. 1999. Summary Minutes of Public Meeting, April 21-22, 1999.
EPA-SAB-COUNCIL-ADV-99-05, 1999. An SAB Advisory on the Health and Ecological
10-44
-------
Chapter 10: Benefit-Cost Analysis
Effects Initial Studies of the Section 812 Prospective Study: Report to Congress: Advisory by the
Health and Ecological Effects Subcommittee, February.
EPA-SAB-COUNCIL-ADV-98-003, 1998. Advisory Council on Clean Air Compliance Analysis
Advisory on the Clean Air Act Amendments (CAAA) of 1990 Section 812 Prospective Study:
Overview of Air Quality and Emissions Estimates: Modeling, Health and Ecological Valuation
Issues Initial Studies.
EPA-SAB-COUNCIL-ADV-99-012, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects: Part 1. July. A
copy of this document can be found in Docket A-2000-01, Document IV-A-204.
EPA-SAB-COUNCIL-ADV-00-001, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects: Part 2. October,
1999.
EPA-SAB-COUNCIL-ADV-00-002, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Advisory Council on Clean
Air Compliance Analysis: Costs and Benefits of the CAAA. Effects Subcommittee on Initial
Assessments of Health and Ecological Effects: Part 2. October, 1999.
EPA-SAB-EEAC-00-013, 2000. An SAB Report on EPA's White Paper Valuing the Benefits of
Fatal Cancer Risk Reduction .US Environmental Protection Agency, Science Advisory Board,
Environmental Economics Advisory Committee, July 27, 2000. A copy of this document can be
found in Docket A-99-06, Document IV-A-19.
Greenbaum, D., President, Health Effects Institute. "Dear Colleague" letter dated May 30, 2002.
A copy of this letter can be found in Docket A-2000-01, Document IV-A-145, as an attachment
to a letter from Lester D. Grant, PhD, to Dr. Philip K. Hopke, dated August 8, 2002.
Grosclaude, P. and N.C. Soguel. 1994. "Valuing Damage to Historic Buildings Using a
Contingent Market: A Case Study of Road Traffic Externalities." Journal of Environmental
Planning and Management 37: 279-287.
Holland, M., D. Forster, and M. Wenborn. 1999. Economic Evaluation of Proposals Under the
UNECE Multi-effects and Multi-pollutant Protocol. Prepared for: European Commission, DGXI,
Brussels and Luxembourg. January. Report no. AEAT-4587.
Hoskins, G., C. McCowan, R.G. Neville, G.E. Thomas, B. Smith, and S. Silverman. 2000. "Risk
Factors and Costs Associated with and Asthma Attack." Thorax. 55: 19-24.
10-45
-------
Draft Regulatory Support Document
Hoskins, G., C. McCowan, R.G. Neville, G.E. Thomas, B. Smith, S. Silverman. 2000. Risk
Factors and Costs Associated with an Asthma Attack. Thorax, 55: 19-24.
Hubbell, Bryan. 2002a. Senior Economist, OAQPS. Memorandum to Docket A-2000-01.
Estimated NOx, SO2, and PM Emissions Health Damages for Heavy Duty Vehicle Emissions.
April 22, 2002. A copy of this document can be found in Docket A-2000-01, Document IV-A-
146.
Hubbell, Bryan. 2002b. Senior Economist, OAQPS. Memorandum to Trish Koman, OTAQ.
Appendix to April 22 Memo "Estimated NOx, SO2, and PM Emissions Health Damages for
Heavy Duty Vehicle Emissions": Alternative Estimates of Mortality and Chronic Bronchitis
Impacts. A copy of this document can be found in Docket A-2000-01, Document IV-A-202.
Industrial Economics, Incorporated (ffic). 1994. Memorandum to Jim DeMocker, Office of Air
and Radiation, Office of Policy Analysis and Review, US Environmental Protection Agency,
March 31.
Jack, E., L.P. Boss, and W.M. Millington. 2000. Asthma - A Speaker's Kit for Public Health
Professionals. Centers for Disease Control and Prevention, National Center for Health Statistics,
Division of Environmental Hazards and Health Effects, Air Pollution and Respiratory Health
Branch. Available at: http://www.cdc.gov/nceh/asthma/speakit/cover.htm.
Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk. Oxford: Basil Blackwell.
Jones-Lee, M.W., G. Loonies, D. O'Reilly, and P.R. Phillips. 1993. The Value of Preventing
Non-fatal Road Injuries: Findings of a Willingness-to-pay National Sample Survey. TRY
Working Paper, WP SRC2.
Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M,
White WH. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer Society
Study of Particulate Air Pollution and Mortality. Special Report to the Health Effects Institute,
Cambridge MA, July 2000. A copy of this document can be found in Docket A-99-06,
Document IV-G-75.
Krupnick, A.J., W. Harrington and B. Ostro. 1990. "Ambient Ozone and Acute Health Effects -
Evidence From Daily Data." Journal of Environmental Economics and Management 18(1): 1-18.
Krupnick, A.J. and M.L. Cropper. 1992. "The Effect of Information on Health Risk
Valuations." Journal of Risk and Uncertainty 5(2): 29-48.
Krupnick, A.J., A. Alberini, M. Cropper, N. Simon, B. O'Brien, R. Goeree, and M. Heintzelman.
2000. Age, Health, and the Willingness to Pay for Mortality Risk Reductions: A Contingent
10-46
-------
Chapter 10: Benefit-Cost Analysis
Valuation Survey of Ontario Residents. Resources for the Future Discussion Paper 00-37.
Krupnick, AJ. 2000. Personal Communication, December 5.
Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Kerry, F. Horak Jr., V.
Puybonnieux-Texier, P. Quenel, J. Schneider, R. Seethaler, J-C Vergnaud, and H. Sommer.
2000. Public-health Impact of Outdoor and Traffic-related Air Pollution: A European
Assessment. The Lancet, 356: 795-801.
Lang, C., G. Yarwood, F. Lalonde, and R. Bloxam. 1995. Environmental and Health Benefits of
Cleaner Vehicles and Fuels. Prepared for: Canadian Council of Ministers of the Environment
Task Force on Cleaner Vehicles and Fuels, Winnipeg, Manitoba. October.
Laurence, J.A., W.A. Retzlaff, J.S. Kern, E.H.Lee, W.E. Hogsett, and D.A. Weinstein. 2000.
Predicting the regional impact of ozone and precipitation on the growth of loblolly pine and
yellow-poplar using linked TREGRO and ZELIG models. For Ecological Management. In
press.
Loehman, E.T., S.V. Berg, A.A. Arroyo, R.A. Hedinger, J.M. Schwartz, M.E. Shaw, R.W.
Fahien, V.H. De, R.P. Fishe, D.E. Rio, W.F. Rossley, and A.E.S. Green. 1979. "Distributional
Analysis of Regional Benefits and Cost of Air Quality Control." Journal of Environmental
Economics and Management 6: 222-243.
Manuel, E.H., R.L. Horst, K.M. Brennan, W.N. Lanen, M.C. Duff and J.K. Tapiero. 1982.
Benefits Analysis of Alternative Secondary National Ambient Air Quality Standards for Sulfur
Dioxide and Total Suspended Particulates, Volumes I-IV. Prepared for US Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck, and M.
Thayer. 1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
Prepared for Office of Policy, Planning and Evaluation, US Environmental Protection Agency.
September.
McConnell, R., K. Berhane, F. Gilliland, S.J. London, H. Vora, E. Avol, WJ. Gauderman, H.G.
Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. Air Pollution and Bronchitic
Symptoms in Southern California Children with Asthma. Environmental Health Perspectives,
107(9): 757-760.
McConnell, R., K. Berhane, F. Gilliland, S.J. London, H. Vora, E. Avol, W.J. Gauderman, H.G.
Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. Air Pollution and Bronchitic
Symptoms in Southern California Children with Asthma. Environmental Health Perspectives.
107(9): 757-760.
10-47
-------
Draft Regulatory Support Document
McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz. 1999. "Long-term ambient ozone
concentration and the incidence of asthma in nonsmoking adults: the ahsmog study."
Environmental Research. 80(2 Pt 1): 110-21.
Mitchell, RC and RT Carson, 1989. Using Surveys to Value Public Goods: the Contingent
Valuation Method. Resources for the Future. Washing ton DC.
NCLAN. 1988. Assessment of Crop Loss from Air Pollutants. (Eds. Walter W. Heck, O. Clifton
Taylor and David T. Tingey) Elsevier Science Publishing Co.: New York, Pp. 1-5. (ERL,GB
639).
O'Conor, Richard M. and Glenn C. Blomquist, 1997. "Measurement of Consumer-Patient
Preferences Using a Hybrid Contingent Valuation Method." Journal of Health Economics
16:667-683.
Ostro, B.D., MJ. Lipsett, J.K. Mann, H. Braxton-Owens, M.C. White, 1995. Air Pollution and
Asthma Exacerbations Among African-American Children in Los Angeles. Inhalation
Toxicology, 7:711-722.
Ostro, B.D. 1987. Air Pollution and Morbidity Revisited: a Specification Test. Journal of
Environmental Economics Management. 14: 87-98.
Ostro B.D. and S. Rothschild. 1989. "Air Pollution and Acute Respiratory Morbidity: An
Observational Study of Multiple Pollutants." Environmental Research 50:238-247.
Ostro, B.D., MJ. Lipsett, J.K. Mann, H. Braxton-Owens, M.C. White. 1995. Air Pollution and
Asthma Exacerbations Among African-American Children in Los Angeles. Inhalation
Toxicology, 7:711-722.
Pew Environmental Health Commission. 2000. Attack Asthma: Why America Needs A Public
Health Defense System to Battle Environmental Threats. Johns Hopkins School of Public Health,
Baltimore, MD.
Pew Environmental Health Commission. 2000. Attack Asthma: Why America Needs A Public
Health Defense System to Battle Environmental Threats. John Hopkins School of Public Health,
Baltimore, MD.
Pope, C.A., JJI, MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and
C.W. Heath, Jr. 1995. "Particulate Air Pollution as a Predictor of Mortality in a Prospective
Study of U.S. Adults." American Journal of Respiratory Critical Care Medicine 151: 669-674.
Pope, C.A., HI, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory Health
10-48
-------
Chapter 10: Benefit-Cost Analysis
and PM10 Pollution: a Daily Time Series Analysis" American Review of Respiratory Diseases
144: 668-674.
Rowe, R.D. and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis-Executive Summary." Prepared by Energy and Resource Consultants, Inc. Report to
the US EPA, Office of Policy Analysis. EPA-230-09-86-018. Washington, DC March.
Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A.
2000. The National Morbidity, Mortality and Air Pollution Study: Part II: Morbidity, Mortality
and Air Pollution in the United States. Research Report No. 94, Part II. Health Effects Institute,
Cambridge MA, June 2000. Docket A-2000-01, Documents IV-A-205 and 206.
Schwartz, J., Dockery, D.W., Neas, L.M., Wypij, D., Ware, J.H., Spengler, J.D., Koutrakis, P.,
Speizer, F.E., and Ferris, Jr., E.G. 1994. "Acute Effects of Summer Air Pollution on
Respiratory Symptom Reporting in Children" American Journal of Respiratory Critical Care
Medicine 150: 1234-1242.
Schwartz, J. 1993. "Particulate Air Pollution and Chronic Respiratory Disease" Environmental
Research 62: 7-13.
Sheppard, L., D. Levy, G. Norris, T.V. Larson and J.Q. Koenig. 1999. Effects of ambient air
pollution on nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994.
Epidemiology. Vol. 10(1): 23-30.
Sisler, J.F. 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition of
the Haze in the United States: An Analysis of Data from the IMPROVE Network. Cooperative
Institute for Research in the Atmosphere, Colorado State University; Fort Collins, CO July.
Smith, D.H., D.C. Malone, K.A. Lawson, L. J. Okamoto, C. Battista, and W.B. Saunders, 1997.
"A National Estimate of the Economic Costs of Asthma." American Journal of Respiratory
Critical Care Medicine 156: 787-793.
Smith, V.K., G.Van Houtven, and S. Pattanayak. 1999. "Benefit Transfer as Preference
Calibration" Resources for the Future Discussion Paper 99-36.
Taylor, C.R., K.H. Reichelderfer, and S.R. Johnson. 1993. Agricultural Sector Models for the
United States: Descriptions and Selected Policy Applications. Iowa State University Press:
Ames, IA.
Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. Summertime Haze Air
Pollution and Children with Asthma. American Journal of Respiratory Critical Care Medicine.,
155: 654-660.
10-49
-------
Draft Regulatory Support Document
Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. Summertime Haze Air
Pollution and Children with Asthma. American Journal of Respiratory Critical Care Medicine,
155: 654-660.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks.
University of Chicago. Final Report for the US Environmental Protection Agency. January.
US Department of Commerce, Economics and Statistics Administration. 1992. Statistical
Abstract of the United States, 1992: The National Data Book. 112th Edition, Washington, DC.
US Department of Commerce, Bureau of Economic Analysis. 1995. BEA Regional Projections
to 2045: Vol. 1, States. Washington, DC US Govt. Printing Office, July 1995.
US Department of Health and Human Services, Centers for Disease Control and Prevention.
1999. National Center for Environmental Health Web Site. Available at
http://www.cdc. gov/nceh/asthma/factsheets/asthma. htm
US Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Health Statistics. 1994. Vital Statistics of the United States, 1990. Volume
II-Mortality. Hyattsville, MD.
US Department of Health and Human Services, Centers for Disease Control and Prevention.
1999. "Forecasted State-Specific Estimates of Self-Reported Asthma Prevalence — United States,
1998." CDC-Morbidity and Mortality Weekly Report, 47: 1022-1025.
US Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Health Statistics. 1999. National Vital Statistics Reports, 47(19).
US Department of Interior, 1986. Natural Resources Damage Assessment Rules and
Regulations. (51 FR 148: 27725-27753. August 1, 1986)
US Department of Interior, 1991. Natural Resource Damage Assessment: Notice of Proposed
Rulemaking. 56 FR 82: 19752-19753 April 29, 1991).
US Environmental Protection Agency, 1996b. Review of the National Ambient Air Quality
Standards for Paniculate Matter: Assessment of Scientific and Technical Information. Office of
Air Quality Planning and Standards, Research Triangle Park, NC EPA report no. EPA/4521R-96-
013. A copy of this document can be found in Docket A-99-06, Document II-A-23.
US Environmental Protection Agency, 1997e. Regulatory Impact Analyses for the Paniculate
Matter and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule.
US EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC. July. See
10-50
-------
Chapter 10: Benefit-Cost Analysis
EPA Air Docket A-96-56, Document No. VI-B-09.
US Environmental Protection Agency, 1998a. Regulatory Impact Analysis for the NOx SIP
Call, FIP and Section 126 Petitions, Volume 2: Health and Welfare Benefits. Prepared by:
Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards,
Research Triangle Park, NC December.
US Environmental Protection Agency, 1997a. Regulatory Impact Analysis for Paniculate
Matter and Ozone National Ambient Air Quality Standards and Proposed RH Rule. Prepared by:
Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. July.
US Environmental Protection Agency, 1998a. Regulatory Impact Analysis for the NOx SIP
Call, FIP, and Section 126 Petitions. US EPA, Office of Air and Radiation. Washington, DC.
EPA-452/R-98-003. December. See EPA Air Docket A-96-56, VI-B-09.
US Environmental Protection Agency, 1998b. The Regional NOx SIP Call & Reduced
Atmospheric Deposition of Nitrogen: Benefits to Selected Estuaries, September.
US Environmental Protection Agency, 1999a. The Benefits and Costs of the Clean Air Act,
1990-2010. Prepared for US Congress by US EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC, November; EPA report no. EPA-410-R-99-001.
US Environmental Protection Agency, 1999b. Regulatory Impact Analysis: Control of Air
Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and Gasoline
Sulfur Control Requirements. Prepared by: Office of Mobile Sources, Office of Air and
Radiation, December.
US Environmental Protection Agency, 1999c. Regulatory Impact Analysis for the Final
Regional Haze Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
April 22, 1999.
US Environmental Protection Agency, 1997f "Response to Comments Made by AISI on EPA
Methodology for Predicting PM2.5 from PM10." Memorandum to the docket from Terence Fitz-
Simons (Office of Air Quality Planning and Standards, Air Quality Trends Analysis Group).
February 6. See EPA Air Docket A-96-56, Document No. VI-B-09.
US Environmental Protection Agency, 1996a. Proposed Methodology for Predicting PM2.5
from PM10 Values to Assess the Impact of Alternative Forms and Levels of the PMNAAQS.
Prepared by Terence Fitz-Simons, David Mintz and Miki Wayland (US Environmental
Protection Agency, Office of Air Quality Planning and Standard, Air Quality Trends Analysis
Group). June 26. See EPA Air Docket A-96-56, Document No. VI-B-09.
10-51
-------
Draft Regulatory Support Document
US Environmental Protection Agency, 1996b. Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information. Office of Air Quality
Planning and Standards, Research Triangle Park, NC EPA report no. EPA/4521R-96-007.
Docket A-99-06, Document II-A-22.
US Environmental Protection Agency, 1994. Documentation for Oz-One Computer Model
(Version 2.0). Prepared by Mathtech, Inc., under Contract No. 68D30030, WA 1-29. Prepared
for US EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC.
August
US Environmental Protection Agency, 1993. External Draft, Air Quality Criteria for Ozone and
Related Photochemical Oxidants. Volume n. US EPA, Office of Health and Environmental
Assessment. Research Triangle Park, NC, EPA/600/AP-93/004b.3v.
US Environmental Protection Agency, 1999b. Regulatory Impact Analysis for the Final
Regional Haze Rule. Prepared by: Office of Air Quality Planning and Standards, Office of Air
and Radiation, April.
US Environmental Protection Agency, 2000a. Technical Support Document for the Heavy-Duty
Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements- Air
Quality Modeling Analyses. December 2000.
US Environmental Protection Agency, 2000b. Valuing Fatal Cancer Risk Reductions. White
Paper for Review by the EPA Science Advisory Board.
US Environmental Protection Agency,1997c. "Methodology Used to Create PM10 and PM2.5
Air Quality Databases for RIA Work." Memorandum from David Mintz, Air Quality Trends
Analysis Group, Office of Air Quality Planning and Standards to Allyson Siwik, Innovative
Strategies and Economics Group, Office of Air Quality Planning and Standards. July 15. See
EPA Air Docket A-96-56, Document No. VI-B-09.
US Environmental Protection Agency, 1997b. The Benefits and Costs of the Clean Air Act,
1970 to 1990. Prepared for US Congress by US EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC
U. S. EPA Guidelines for the Health Risk Assessment of Chemical Mixtures. EPA Report No.
630/R-98/002. September 24, 1986. Federal Register 51(185) : 34014 - 34025.
http ://cfpub. epa. gov/ncea/
Valigura, Richard A., and W.T. Luke, R.S. Artz, B.B. Hicks. 1996. Atmospheric input to
coastal areas: reducing the uncertainties. NOAA Coastal Ocean Program, Decision Analysis
Series No. 9. NOAA Atmospheric Resources Laboratory, Silver Spring, MD.
10-52
-------
Chapter 10: Benefit-Cost Analysis
Valiela, G. Collins, and J. Kremer, K. Lajtha, M. Geist, B. Seely, J. Brawley, C.H. Sham. 1997.
"Nitrogen loading from coastal watersheds to receiving estuaries: new method and application."
Ecological Applications 7(2), pp. 358-380.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. (New York:
Oxford University Press).
Viscusi, W.K., W.A. Magat, and J. Huber. 1991. "Pricing Environmental Health Risks: Survey
Assessments of Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis" Journal of
Environmental Economics and Management, 21: 32-51.
Walsh, RG, DM Johnson and JR McKean, 1992. Benefits Transfer of Outdoor Recreation
Demand Studies, 1968 - 1988. Water Resource Research. 28: 703-713.
White, JJ and Carroll, JN. (1998) Emissions from snowmobile engines using bio-based fuels and
lubricants. Final Report. Prepared for State of Montana Department of Environmental Quality.
Southwest Research Institute report #7383.
Whittemore, A.S. and E.L. Korn. 1980. Asthma and Air Pollution in the Los Angeles Area. Am
J Public Health. Vol. 70: 687-696.
Yu, O., L. Sheppard, T. Lumley, J. Q. Koenig, and G.G. Shapiro. 2000. Effects of Ambient Air
Pollution on Symptoms of Asthma in Seattle-Area Children Enrolled in the CAMP Study.
Environmental Health Perspectives, 108(12): 1209-1214.
10-53
-------
Draft Regulatory Support Document
10-54
-------
Chapter 11: Regulatory Alternatives
Chapter 11: Regulatory Alternatives
Adopting standards to reduce emissions requires consideration of a variety of alternative
approaches. This rulemaking development effort includes consideration of the timing of
emission standards, the level of stringency, the appropriate test procedures, among other things.
In this chapter, we present a variety of alternatives that we considered in preparing this
rulemaking. While these alternatives were not adopted as part of the final rule, they are
discussed here with an analysis of the associated costs and emission reductions involved and our
rationale for not adopting them.
11.1 Recreational Marine Diesel Engines
While developing the CI recreational marine engine standards we analyzed two
alternative approaches. The first approach was to apply the draft European Commission
recreational marine emission standards to CI recreational marine engines used in the United
States. Another approach we considered was to implement the CI recreational marine engine
standards on the same schedule as for commercial marine engines. These two alternative
approaches are discussed below.
11.1.1 Harmonization with Draft EC Standards
Several manufacturers commented that we should finalize the emission standards
proposed by the European Commission (EC) for CI recreational marine engines for our national
standards. These emission levels are presented in Table 11.1-1. This table also presents the U.S.
standards finalized today and average baseline emissions based on data presented earlier in
Chapter 4 on engines for which we had data on both HC+NOx and PM."7 Based on this data, we
believe that the proposed European emissions standards for recreational marine diesel engines
may not result in a decrease in emissions, and may even allow an increase in emissions from
engines operated in the U.S. because current engines are already performing better than the
proposed EC limits. Also, because the Clean Air Act directs us to set standards that "achieve the
greatest degree of emission reduction achievable" given appropriate considerations, we do not
believe it would be appropriate to finalize emission standards at the levels proposed by the
European Commission.
5/7 If we include HC+NOx data from engine tests that did not include PM measurement,
the HC+NOx average decreases to 8.6 g/kW-hr.
11-1
-------
Draft Regulatory Support Document
Table 11.1-1
EPA and Proposed European Standards Compared to
Average Baseline Levels for CI Recreational Marine Emissions
Pollutant
HC+NOx
PM
CO
EPA Standards
g/kW-hr
7.2-7.5
0.2-0.4
5.0
Proposed EC Standards
g/kW-hr
9.8 NOx, 1.5 HC*
1.4
5.0
Baseline Emissions g/kW-hr
9.2
0.2
1.3
* HC increases slightly with increasing power rating.
We are not presenting an analysis of the cost per ton of emission reduction for this
approach because we do not believe that it would result in emission reductions. However, the
engine manufacturers would still need to incur the certification and compliance costs presented in
Chapter 5. Therefore, setting a standard equal to the draft EC standards would likely result in
costs with few or no benefits.
11.1.2 Earlier Implementation Dates Consistent with Commercial Marine
We believe that the emission-reduction strategies expected for land-based nonroad diesel
engines and commercial marine diesel engines will also be applied to recreational marine diesel
engines. Marine diesel engines are generally derivatives of land-based nonroad and highway
diesel engines. Marine engine manufacturers and marinizers make modifications to the engine to
make it ready for use in a vessel. These modifications can range from basic engine mounting and
cooling changes to a restructuring of the power assembly and fuel management system. Because
we anticipate that the same or similar technology will be used to meet the recreational and
commercial marine standards, we considered including recreational marine engines in the
commercial marine program with the same implementation dates.
Engine manufacturers commented that recreational marine engines need at least two years
of lead time after the commercial marine standards to transfer technology from commercial
marine engines to recreational marine engines and to stagger the need for manufacturers'
research and development costs. We agree that this is necessary. In current production practices,
the recreational marine engines are designed to operate at a higher power to weight ratio than
commercial engines which requires development efforts specific to these engines. Although we
believe that the same technology can be applied to recreational and commercial marine engines
to reduce emissions, we recognize that individual development efforts will be required. In
current practices, manufacturers stagger their development schedules to effectively use resources
which include engineering hours and test cell time. If we were to require that recreational marine
engines meet the new standards in the same year as commercial marine engines, manufacturers
11-2
-------
Chapter 11: Regulatory Alternatives
would likely need to double their research and development resources. We do not consider it
practical for a manufacturer to do this in time for earlier standards, especially if the resources are
only needed for two years. By allowing an additional two years of lead time, manufacturers are
better able to stagger their development efforts.
The advantage of the earlier implementation dates would be to achieve emission
reductions two years earlier. This would not likely affect the hardware costs discussed in
Chapter 5, but would significantly increase the research and development costs if new people had
to be hired and new facilities constructed. In fact, manufacturers would not likely have enough
time to increase their research and development resources in time to meet earlier implementation
dates. Therefore we are giving two years of additional lead time for recreational marine engines
beyond the commercial marine implementation dates.
11.2 Large Industrial Spark-Ignition Engines
Of the several possibilities for Large SI engines, we are choosing one alternative over
several others. For example, we are not analyzing the alternative of adopting only 2004
standards. Given the California certification data showing that some manufacturers are already
achieving 2007 emission levels (with steady-state testing). This alternative would therefore
clearly not meet the Clean Air Act direction to adopt the most stringent standards achievable.
Second, we are not analyzing a scenario of more stringent emission standards. The 2007
standards follow directly from available emission test data showing what level of emission
control is achievable in that time frame. Any significant emission reductions beyond the 2007
standards would be appropriate to consider for a third tier of emission standards. Once
manufacturers gain experience with the new emission-control technologies and the measurement
procedures, additional information will be available to help us evaluate the relative costs and
benefits of more stringent standards. Such information is not available today.
Third, we are not considering the approach of requiring forklifts to convert to battery
power. We don't believe this would be an appropriate policy under Clean Air Act section 213, as
described in the Summary and Analysis of Comments. An analysis comparing the life-cycle
costs and benefits of the two alternative power sources for forklifts would provide useful
information to consumers interested in evaluating their available choices. However, such an
analysis is outside the scope of this rulemaking.
The alternative we have chosen to analyze captures a common input from those
commenting on the proposal. Manufacturers generally questioned the need, value, or cost-
effectiveness of adopting emission procedures requiring transient engine operation. To evaluate
this more carefully, we analyzed the scenario of adopting the 2007 standards based only on
steady-state emission measurement. To assess this alternative, we have calculated the costs and
emission reductions associated with adding the transient controls to an engine already meeting
11-3
-------
Draft Regulatory Support Document
the 2007 standards with steady-state testing.
Estimating the costs of controlling transient emissions is straightforward, with two
simplifying assumptions. First, we need to assume that the technology and costs associated with
the 2004 standards presented in Chapter 5 are sufficient to achieve the 2007 standards with
steady-state testing. The existing California certification data support this. Second, even though
the 2007 cost estimates include an allowance for meeting diagnostic requirements and field-
testing standards, in this analysis we assign the full estimated cost of meeting the 2007 standards
to upgrading for transient control. The resulting estimated first-year cost of $27 per engine
therefore somewhat overestimates the actual cost. This includes engineering time to improve
calibrations with the existing hardware, so there are no variable costs under this scenario.
To estimate the emission reductions associated with the transient test procedure, we rely
primarily on the transient adjustment factors described in Chapter 6. Applying the transient
adjustment factor leads to increased emissions of about 0.77 g/hp-hr HC+NOx and 3 g/hp-hr CO.
Factoring in the lifetime operating parameters from the NONROAD model leads to a discounted
lifetime emission reduction per engine of 0.22 tons for HC+NOx and 0.76 tons for CO.
Comparing costs and emission reductions yields an estimated cost of about $200 per ton
HC+NOx. Estimated nationwide emission reductions after fully phasing in the emission
standards are 17,000 tons HC, 36,000 tons NOx, and 188,000 tons CO. These figures represent
the incremental benefit of adding transient test procedures for the Tier 2 standards.
This analysis supports the decision to adopt emission standards requiring control of
emissions during transient operation.
11.3 Recreational Vehicle Exhaust Emission Standards
11.3.1 Off-highway Motorcycles
We are presenting an analysis of two alternatives to the 2.0 g/km HC+NOx standard
contained in the Final Rule, a less stringent and a more stringent alternative. The less stringent
alternative we are presenting is a 4.0 g/km HC+NOx standard in the same time frame as the 2.0
g/km standard (50 and 100% phase-in for 2006 and 2007). We are finalizing this standard as an
option to the 2.0 g/km standard with the provision that a manufacturer must certify all of their
products, including machines that may otherwise meet the exemption for vehicles used solely for
competition, to the 4.0 g/km standard. This alternative is numerically less stringent than the 2.0
g/km standard, but may actually result in more significant emission reductions than the final
program since machines that may otherwise be exempt in the final program are included in the
optional 4.0 g/km standard. Most competition off-highway motorcycles that could meet the
competition exemption use high performance two-stroke engines that have HC levels
significantly higher than the standard.
11-4
-------
The second alternative we are presenting is the 2.0 g/km standard with an additional more
stringent Phase 2 standard of 1.0 g/km phased in at 50 and 100% in 2009 and 2010. We
proposed this alternative for ATVs, but not for off-highway motorcycles. It is clear from our
analysis of technology, the current off-highway motorcycle market, and the comments received
from manufacturers that four-stroke engines are technologically within reach for all off-highway
motorcycle applications. While it is less clear, based on our analysis of technology and
comments received from manufacturers and user groups it appears that direct fuel injection for
two-stroke engines may also be within reach for some off-highway motorcycle applications. An
analysis of the costs, emission reductions, costs per ton, and economic impacts of the alternatives
are presented here. The methodology used for these analyses are the same as those described for
the final program in the previous chapters.
11.3.1.1 Per Unit Costs
We have analyzed a less stringent standard of 4.0 g/km HC+NOx phased in at 50 and
100% in 2006 and 2007. The per unit average cost for this alternative is presented in Table
11.3.1-1 below. The average costs are based on a technology mix that includes the use of four-
stroke engines and direct fuel injection for two-stroke engines. Because off-highway
motorcycles have been using four-stroke engines for a many years and there is a significant
number of these engines sold, the cost of using a four-stroke engine is less than the cost of using
a direct fuel injection system with a two-stroke engine. Since we do not anticipate that any direct
fuel injection two-stroke engines will be capable of meeting the final standard of 2.0 g/km
HC+NOx, the resulting average cost for this alternative is somewhat higher than that of the final
program, which we estimated at $158 per unit (see Chapter 5).
-------
Draft Regulatory Support Document
Table 11.3.1-1
Estimated Average Costs For Off-Highway Motorcycle Alternative 1 (4.0 g/km)
< 125 cc
(31%)
125<250cc
(27%)
> 250 cc
(42%)
4-stroke engine
Direct injection
compliance
total
4-stroke engine
Direct injection
compliance
total
4-stroke engine
Direct injection
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$219
$375
$7
-
$286
$375
$7
-
$353
$375
$7
-
-
Lifetime
Fuel
Savings
(NPV)
($140)
($140)
-
-
($140)
($140)
-
-
($140)
($140)
-
-
-
Baseline
55%
0%
0%
-
29%
0%
0%
-
29%
0%
0%
-
-
Control
85%
15%
100%
-
85%
15%
100%
-
85%
15%
100%
-
-
Incrementa
ICost
$66
$56
$7
$129
$160
$56
$7
$223
$198
$56
$7
$71
$210
$127
Incremental
Fuel Savings
(NPV)
$42
$21
-
$63
$78
$21
-
$99
$78
$21
-
$99
$88
$88
We have also analyzed an alternative that would include our final standard of 2.0 g/km
plus a Phase 2 standard of 1.0 g/km that would be phased in at 50 and 100% in 2009 and 2010.
This additional level of control would require R&D beyond that projected for the final 2.0 g/km
standard and the incorporation of additional controls for four-stroke engines. We are projecting
that at least half of off-highway motorcycle models would be equipped with catalysts in order to
meet this level of stringency. The estimated average per unit costs for Phase 2 incremental to
Phase 1 are provided in Table 11.3.1-2. We estimate that Phase 2 would cost about $70
incremental to Phase 1.
11-6
-------
Table 11.3.1-2
Estimated Average Costs For Phase 2 Off-highway Motorcycles (Phase 2 = 1.0 g/km)
(Non-competition models only)
< 125 cc
(37%)
125<250cc
(21%)
> 250 cc
(42%)
4-stroke engine
pulse air
R&D including
recalibration
Catalyst
compliance
total
4-stroke engine
pulse air
R&D including
recalibration
Catalyst
compliance
total
4-stroke engine
pulse air
R&D including
recalibration
Catalyst
compliance
total
Near Term Composite Incremental
Cost
Long Term Composite Incremental
Cost
Cost
$219
$39
$15
$68
$1
-
$286
$39
$15
$68
$1
-
$353
$39
$15
$70
$1
-
-
Lifetime
Fuel
Savings
(NPV)
($140)
$0
$0
$0
-
-
($140)
$0
$0
$0
-
-
($140)
$0
$0
$0
-
-
-
Baseline
100%
25%
0%
0%
0%
-
100%
0%
0%
0%
0%
-
100%
0%
0%
0%
0%
-
-
Contr
ol
100%
75%
100%
50%
100%
-
100%
25%
100%
50%
100%
-
100%
25%
100%
50%
100%
-
-
Incremental
Cost
$0
$19
$15
$34
$1
$70
$0
$19
$15
$34
$1
$70
$0
$19
$15
$35
$1
$71
$70
$28
Incremental
Fuel Savings
(NPV)
$0
$0
$0
$0
-
$0
$0
$0
$0
$0
-
$0
$0
$0
$0
$0
-
$0
$0
$0
11.3.1.2 Aggregate Cost Estimates
Based on the above per unit costs, we have estimated the aggregate costs for the two
alternatives. The aggregate costs for Alternative 2 includes the costs for both phases of
standards. The aggregate costs for the alternatives are provided in Table 11.3.1-3, along with the
-------
Draft Regulatory Support Document
aggregate cost estimates for the final off-highway motorcycle program, which are estimated in
Chapter 5. The fuel savings for both alternatives result from the switching of two-stroke to four-
stroke engines. Alternative 1 also experiences fuel savings by the incorporation of competition
machines into the program. Competition machines would either switch from two-stroke to four-
stroke engines or use direct fuel injection with two-stroke engines. Direct fuel injection with
two-stroke technology can result in similar fuel savings as converting from two-stroke to four-
stroke engines.
Table 11.3.1-3
Summary of Annual Aggregate Costs and Fuel Savings (millions of dollars)
OHMC Final Program
Alternative 1
Alternative 2
Fuel Savings (Alt 1)
Fuel Savings (Alt 2)
2006
$16.27
$30.68
$16.27
$1.32
$0.63
2010
$24.24
$46.56
$34.25
$14.13
$7.23
2015
$21.53
$42.90
$28.53
$30.62
$16.19
2020
$22.63
$45.09
$29.99
$39.05
$21.03
2025
$23.79
$47.39
$31.52
$41.98
$22.65
11.3.1.3 Emissions Reductions
In Chapter 6, we estimated the emissions reductions for the final program. We have
estimated the emissions reductions from both alternatives using the same methodology. We
would expect NOx and CO to be similar under the various alternatives. The results for HC are
shown in Table 11.3.1-4 and in the Figure 11.3.1-1. The majority of the HC emissions
reductions occur due to switching those remaining two-stroke off-highway motorcycles over to
four-stroke technology. We expect this to occur in each of the alternatives we have analyzed.
Alternative 1 has significantly greater reductions than alternative 2 or the final program, even
though the numerical standard is less stringent. This is due to the fact that alternative 1 includes
all off-highway motorcycles. Machines that may otherwise qualify for the competition
exemption make up 29-percent of off-highway motorcycle sales, and they tend to use high-
performance two-stroke engines that emit very high levels of HC emissions. Controlling HC
emissions from these machines to the alternative 1 standard of 4.0 g/km would result in
significant reductions.
11-8
-------
Chapter 11: Regulatory Alternatives
Table 11.3.1-4
Summary of HC Reductions (thousands of tons)
OHMC Final Program
Alternative 1
Alternative 2
2006
3.1
5.7
3.1
2010
36.3
63.4
36.8
2015
84.1
142.6
86.6
2020
111.1
184.9
115.4
2025
120.0
199.2
124.8
Figure 11.3.1-1
Off-Highway Motorcycle HC Emissions Inventory
300
250
| 200
co
§ 150
o
.c
H 100
50
0
20
0
Nationwide HC Emissions (tons)
xxo********^
/sOO^00
^ ^> <> ^
O °
[t "A!*"**.*^^^*^****
A A
"AAAAAAAAAAAAZ
05 2010 2015 2020 2025 2030
Year
Baseline A Final Program A Alternative 1 T Alternative 2
11.3.1.4 Cost Per Ton
Chapter 7 provides the cost per ton estimate for the final program. Using the same
methodology, we have estimated the cost per ton of HC+NOx reduced for the two alternatives.
The results are provided in Table 11.3.1-5. The results of Alternative 2 Phase 2 are based on the
incremental change from 2.0 g/km to 1.0 g/km.
11-9
-------
Draft Regulatory Support Document
Table 11.3.1-5
Estimated Off-Highway Motorcycle Average
Cost Per Ton of HC + NOx Reduced (7 percent discount rate)
Final Program
Alternative 1
Alternative 2 Phase 1
Alternative 2 Phase 2*
Lifetime Reductions
per Vehicle
(NPV tons)
0.38
0.50
0.38
0.02
Discounted Per Vehicle
Costs Per Ton without Fuel
Savings ($/ton)
$410
$420
$410
$3,590
Discounted Per Vehicle Costs
Per Ton with Fuel Savings
($/ton)
$280
$210
$280
$3,590
* Phase 2 standards incremental to Phase 1
11.3.1.5 Economic Impacts Analysis
The human health and environmental benefits and economic costs of the regulatory
alternatives for off-highway motorcycles are presented. The methodologies used to estimate the
economic costs of these alternatives are discussed extensively in Chapter 9. We are presenting
two alternatives to the 2.0 g/km HC+NOx standard contained in the Final Rule, a less stringent
and a more stringent alternative.
Table 11.3.1-6
Economic Costs of Alternative
Off-Highway Motorcycle Standards—Values in 2030 ( millions of 2001$)
Standard
(HC/CO Reductions)
OHM Final Program
Alternative 1
Alternative 2
Engineering Costs
$25.9
$33.1
$49.8
Economic Costs
(Surplus Losses)
$25.0
$31.7
$46.6
Fuel Efficiency
Cost Savings
$25.2
$46.4
$25.2
Economic Gains
or Costs1
$0.2
$14.7
($21.5)
1 Economic costs or net economic costs shown in parenthesis. Additional important considerations, such as
potential safety impacts discussed below, are not reflected in these cost estimates.
11-10
-------
Chapter 11: Regulatory Alternatives
Table 11.3.1-7a
Economic Costs of Alternative Off-Highway Motorcycle Standards—Net Present Value
2002 through 2030 (millions of 2001$, using 3 percent discount rate)
Standard
(HC/CO
Reductions)
OHM Final
Program
Alternative 1
Alternative 2
Engineering Costs
$372.6
$461.4
$712.0
Economic Costs
(Surplus Losses)
$358.9
$441.1
$663.1
Fuel Efficiency
Cost Savings
$242.4
$467.8
$242.4
Economic Gains or
Costs1
($116.5)
26.7
($420.7)
1 Economic costs or net economic costs shown in parenthesis.
Table 11.3.1-7b
Economic Costs of Alternative Off-Highway Motorcycle Standards—Net Present Value
2002 through 2030 (millions of 2001$, using 7 percent discount rate)
Standard
(HC/CO
Reductions)
OHM Final
Program
Alternative 1
Alternative 2
Engineering Costs
$214.3
$261.6
$408.6
Economic Costs
(Surplus Losses)
$206.3
$249.9
$379.9
Fuel Efficiency
Cost Savings
$120.6
$232.5
$120.6
Economic Gains or
Costs1
($85.6)
($17.4)
($259.3)
1 Economic costs or net economic costs shown in parenthesis.
11.3.1.6 Discussion
Although alternative 1 is numerically less stringent than the final standard of 2.0 g/km
HC+NOx, it would result in significant additional emissions reductions from the final program.
These reductions are gained by the inclusion of machines that could otherwise qualify as vehicles
used solely for competition into the program. The CAA requires that competition vehicles be
exempt from emission regulations. Moreover, the 4.0 g/km standard would not otherwise meet
the CAA requirements that standards achieve the greatest degree of emissions reduction
achievable through use of available technology, taking cost, noise, energy, and safety into
11-11
-------
Draft Regulatory Support Document
account. Therefore, this alternative cannot be considered as a replacement to the final program.
However, the potential for significant emission reductions resulting from the control of
competition machines is very desirable. That is why we are finalizing alternative 1 as an option
to the 2.0 g/km HC+NOx standard in the final program. This option would result in the use of
four-stroke engines and two-stroke engines equipped with direct fuel injection.
Alternative 2 would require manufacturers to achieve reductions beyond those required
by the California off-highway motorcycle program. We believe that manufacturers would be
required to use high levels of pulse air and would also need to use catalysts on some models. As
discussed in Chapter 4, there are still concerns over the safety, durability and feasibility of the
widespread use of catalysts on off-highway motorcycles. We are concerned that catalysts could
pose safety threats from burns to individual riders as well as the potential for setting fires in the
riding environment, which is frequently forests and grassy fields. There are also concerns over
the ability of a catalyst to be able to physically survive in the very harsh environment that off-
highway motorcycles frequently operate in. In general, we have concerns about the feasibility of
many advanced emission control technologies with off-highway motorcycle applications. Off-
highway motorcycles are exposed to dirt, dust, mud, water, rocks, etc. All of which make the use
of relatively fragile technology such as electronic fuel injection and secondary air injection
questionable. This alternative is based on the standards we proposed for ATVs but are not
finalizing. As discussed in detail in the preamble for the Final Rule, we are not finalizing this
level of control for ATVs due to concerns about the ability of manufacturers to meet the
standards within the time frame proposed. These same concerns apply to off-highway
motorcycles. We believe additional testing and analysis is needed before we can affirm the
feasibility of Phase 2 standards.
11.3.2 All-terrain Vehicles
We are presenting an analysis of two alternatives to the 1.5 g/km HC+NOx standard
contained in the Final Rule, a less stringent and a more stringent alternative. The less stringent
alternative we are presenting is a 2.0 g/km HC+NOx standard in the same time frame as the 1.5
g/km standard (50 and 100 % phase-in for 2006 and 2007). The second alternative we are
presenting is the 2.0 g/km alternative with an additional more stringent Phase 2 standard of 1.0
g/km phased in at 50/100% in 2009/2010. We proposed but did not finalize two phases of
standards for ATVs and the second alternative analyzed below is based on the proposed
standards. It is clear from our analysis of technology, the current ATV market, and the
comments received from manufacturers that 4-stroke engines are technologically within reach for
all ATV applications. Therefore, the focus of the alternatives analysis is on what level of control
to require from 4-stroke ATVs. An analysis of the costs, emissions reductions, costs per ton, and
economic impacts of the alternatives are presented here. The methodology used for these
analyses are the same as those described for the final program in the previous chapters. Also, the
costs for the various technologies is presented in Chapter 5. Finally, a discussion of why these
alternatives were not chosen for the Final Rule is provided in Section 11.3.2.6.
11-12
-------
Chapter 11: Regulatory Alternatives
11.3.2.1 Per unit Costs
We have analyzed a less stringent standard of 2.0 g/km HC+NOx phased in at 50 and
100% in 2006 and 2007. The per unit average cost for this alternative is presented in Table
11.3.2-1 below. The average costs are based on a technology mix similar to that of the final 1.5
g/km standard, but with less reliance on reducing emissions from the 4-stroke engines through
the use of recalibration and secondary air. This results in an average cost that is somewhat lower
than that of the final program, which we estimated would cost $87 per unit (see Chapter 5).
Alternative 2 would require manufacturers to achieve reductions beyond those required
by the California off-highway motorcycle program. We believe that manufacturers would be
required to use high levels of pulse air and would also need to use catalysts on some models. As
discussed in Chapter 4, there are still concerns over the safety, durability and feasibility of the
widespread use of catalysts on off-highway motorcycles. We are concerned that catalysts could
pose safety threats from burns to individual riders as well as the potential for setting fires in the
riding environment, which is frequently forests and grassy fields. There are also concerns over
the ability of a catalyst to be able to physically survive in the very harsh environment that off-
highway motorcycles frequently operate in. In general, we have concerns about the feasibility of
many advanced emission control technologies with off-highway motorcycle applications. Off-
highway motorcycles are exposed to dirt, dust, mud, water, rocks, etc. All of which make the use
of relatively fragile technology such as electronic fuel injection and secondary air injection
questionable. This alternative is based on the standards we proposed for ATVs but are not
finalizing. As discussed in detail in the preamble for the Final Rule, we are not finalizing this
level of control for ATVs due to concerns about the ability of manufacturers to meet the
standards within the time frame proposed. These same concerns apply to off-highway
motorcycles. We believe additional testing and analysis is needed before we can affirm the
feasibility of Phase 2 standards.
11-13
-------
Table 11.3.2-1
Estimated Average Costs For a ATV Alternative 1 (2.0 g/km)
< 200 cc
(15%)
> 200 cc
(85%)
4-stroke engine
pulse air
R&D for
exhaust
including
recalibration
permeation
control
compliance
total
4-stroke engine
pulse air
R&D for
exhaust
including
recalibration
permeation
control
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$219
$33
$16
$3
$13
--
$349
$27
$5
$3
$12
--
--
--
Lifetime
Fuel
Savings
(NPV)
($124)
$0
$0
($5)
--
--
($124)
$0
$0
($5)
--
--
--
--
% of use
Baseline
8%
0%
0%
0%
0%
--
93%
0%
0%
0%
0%
--
--
--
% of use
Control
100%
25%
50%
100%
100%
--
100%
25%
50%
100%
100%
--
--
--
Incrementa
ICost
$202
$8
$8
$3
$13
$234
$24
$7
$2
$3
$12
$49
$76
$36
Incremental
Fuel Savings
(NPV)
($114)
$0
$0
($5)
--
($119)
($9)
$0
$0
($5)
--
($13)
($29)
($29)
-------
Table 11.3.2-2
Estimated Average Costs For ATV Alternative 2 (Phase 2 =1.0 g/km)
< 200 cc
(15%)
> 200 cc
(85%)
4-stroke engine
pulse air
R&D for
exhaust
including
recalibration for
Phase 2
Catalyst
compliance
total
4-stroke engine
pulse air
R&D for
exhaust
including
recalibration for
Phase 2
Catalyst
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$219
$33
$16
$68
$2
--
$349
$27
$5
$70
$2
--
--
--
Lifetime
Fuel
Savings
(NPV)
($124)
$0
$0
$0
--
--
($124)
$0
$0
$0
--
--
--
--
% of use,
Phase 1 =
2.0 g/km
100%
0%
0%
50%
0%
--
100%
0%
0%
50%
0%
--
--
--
% of use,
Phase 2 =
1.0 g/km
100%
50%
100%
100%
100%
--
100%
50%
100%
100%
100%
--
--
--
Incrementa
ICost
$0
$16
$16
$34
$2
$68
$0
$14
$5
$35
$2
$54
$56
$30
Incremental
Fuel Savings
(NPV)
$0
$0
$0
$0
--
$0
$0
$0
$0
$0
--
$0
$0
$0
-------
Draft Regulatory Support Document
11.3.2.2 Aggregate Cost Estimates
Based on the above per unit costs, we have estimated the aggregate costs for the two
alternatives. The aggregate costs for Alternative 2 includes the costs for both phases of
standards. The aggregate costs for the alternatives are provided in Table 11.3.2-3, along with the
aggregate cost estimates for the final ATV program, which are estimated in Chapter 5. The fuel
savings result from switching from 2-stroke to 4-stroke engines and are the same for each
alternative.
Table 11.3.2-3
Summary of Annual Aggregate Costs and Fuel Savings (millions of dollars)
ATV Final Program
Alternative 1
Alternative 2
Fuel Savings
2006
$42.46
$37.43
$37.43
$0.93
2010
$65.30
$57.11
$102.58
$15.14
2015
$52.44
$48.18
$77.28
$36.22
2020
$47.56
$43.29
$72.39
$48.84
2025
$47.56
$43.29
$72.39
$51.00
11.3.2.3 Emissions Reductions
In Chapter 6, we estimated the emissions reductions for the final program. We have
estimated the emissions reductions for both alternatives using the same methodology. We would
expect NOx and CO to be similar under the various alternatives. The results for HC are shown in
Table 11.3.2-4 and in the following figure. The majority of the HC emissions reductions occur
due to switching those remaining 2-stroke ATVs over to 4-stroke technology. The base emission
factor is about 34 g/km for that 20 percent of the ATV fleet which is two-stroke and 1.8 g/km for
the remaining 80 percent which are four stroke. Thus, even though eliminating the four strokes
is significant the reductions from the four strokes is large as well. We expect this to occur in
each of the alternatives we have analyzed.
Table 11.3.2-4
Summary of HC Reductions (thousands of tons)
ATV Final Program
Alternative 1
Alternative 2
2006
6.2
5.9
5.9
2010
92.4
88.0
91.1
2015
225.0
214.9
230.4
2020
304.1
291.0
317.0
2025
315.5
302.0
331.0
11-16
-------
Chapter 11: Regulatory Alternatives
Figure 11.3.2-1: ATV HC Emissions Inventory
/inn
4UU
350
•jnn
oUU
C/5
CO
o 1c-n
.n 1 OU
1-
1 nn
I UU
50
Nationwide HC Emissions (tons)
j^fT^^
L^^^L
Z^^^i
s^N-w
*>^_-
-AAxHHHHt
2005 2010 2015 2020 2025 2030
Year
"•" Baseline ~*~ ATV Alternative 1
0 Final Program (1.5 g/km) A ATV Alternative 2
11.3.2.4 Cost Per Ton
Chapter 1 provides the cost per ton estimates for the final program. Using the same
methodology, we have estimated the cost per ton of HC+NOx reduced for the two alternatives.
The results are provided in table 11.3.2-5. The results for Alternative 2 Phase 2 are based on the
incremental change from 2.0 g/km to 1.0 g/km.
11-17
-------
Draft Regulatory Support Document
Table 11.3.2-5
Estimated ATV Average
Cost Per Ton of HC + NOx Reduced (7 percent discount rate)
Final Program
Alternative 1
Alternative 2 Phase 1
Alternative 2 Phase 2*
Lifetime Reductions
per Vehicle
(NPV tons)
0.21
0.20
0.20
0.02
Discounted Per Vehicle Cost
Per Ton without Fuel
Savings ($/ton)
$400
$370
$370
$2,700
Discounted Per Vehicle Cost
Per Ton with Fuel Savings
($/ton)
$290
$250
$250
$2,700
* Phase 2 standards incremental to Phase 1
11.3.2.5 Economic Impacts Analysis
The economic costs of the regulatory alternatives for ATVs are presented. The
methodologies used to estimate economic costs of these alternatives are discussed extensively in
Chapter 9. We are presenting two alternatives to the 1.5 g/km HC+NOx standard contained in the
Final Rule, a less stringent and a more stringent alternative.
Table 11.3.2-6
Economic Costs of Alternative ATV Standards—Values in 2030 ( millions of 2001$)
Standard
(HC/CO
Reductions)
ATV Final Program
Alternative 1
Alternative 2
Engineering Costs
$496.3
$445.2
$662.0
Economic Costs
(Surplus Losses)
$491.9
$441.7
$654.1
Fuel Efficiency
Cost Savings
$253.0
$253.0
$253.0
Economic Gains or
Costs1
($238.9)
($188.6)
($401.0)
1 Economic costs or net economic costs shown in parenthesis.
11-18
-------
Chapter 11: Regulatory Alternatives
Table 11.3.2-7a
Economic Costs of Alternative ATV Standards
Net Present Value 2002 through 2030
(millions of 2001$, using 3 percent discount rate)
Standard
(HC/CO
Reductions)
ATV Final Program
Alternative 1
Alternative 2
Engineering Costs
$836.3
$752.9
$1,154.1
Economic Costs
(Surplus Losses)
$829.2
$747.0
$1,140.5
Fuel Efficiency
Cost Savings
$510.5
$510.5
$510.5
Economic Gains or
Costs1
($318.7)
($236.5)
($630.0)
1 Economic costs or net economic costs shown in parenthesis.
Table 11.3.2-7b
Economic Costs of Alternative ATV Standards
Net Present Value 2002 through 2030
(millions of 2001$, using 7 percent discount rate)
Standard
(HC/CO
Reductions)
ATV Final Program
Alternative 1
Alternative 2
Engineering Costs
$836.3
$752.9
$1,154.1
Economic Costs
(Surplus Losses)
$829.2
$747.0
$1,140.5
Fuel Efficiency
Cost Savings
$510.5
$510.5
$510.5
Economic Gains or
Costs1
($318.7)
($236.5)
($630.0)
1 Economic costs or net economic costs shown in parenthesis.
11.3.2.6 Discussion
Alternative 1 would require only modest additional emissions reductions from 4-strokes,
in general, and many models would meet the standard in their base configuration. In addition,
this alternative is less stringent than the current California standard for ATVs. Most, if not all 4-
stroke ATV models are certified to the California requirements. We received support for
harmonizing standards with California and this level of control is feasible for 4-stroke equipped
ATVs. Therefore, we do not believe that a standard less stringent than that contained in the
California program would meet the basic criteria of the Clean Air Act which requires us to set a
standard based on the greatest degree of emission reduction achievable. Our consideration of
11-19
-------
Draft Regulatory Support Document
costs and economic impacts did not change our view that a 1.5 g/km standard was appropriate for
ATVs.
Alternative 2 would require manufacturers to achieve reductions beyond those required in
by the California program. We believe that manufacturers would be required to use a high level
of pulse air and would also need to use catalyst on some ATV models. For our cost analysis
above, we projected that catalysts would be used on half of all ATV models. This alternative is
based on the standards we proposed for ATVs but are not finalizing. As discussed in detail in the
preamble for the Final Rule, we are not finalizing this level of control due to concerns about the
ability of manufacturers to meet the standards within time frame proposed. We believe
additional testing and analysis is needed before we can affirm the feasibility of the Phase 2
standards.
11.3.3 Snowmobiles
While developing the final snowmobile emissions standards we analyzed four alternative
sets of emissions standards, including options both less stringent and more stringent than the
final standards. These alternatives are as follows:
Alternative 1 - keeping the Phase 1 standards indefinitely (i.e., not adopting Phase 2 or Phase 3
standards)
Alternative 2 - adopting the snowmobile manufacturers' recommended phase 2 standards in 2010
(which provide a 50% reduction in HC but keep the CO standard at the phase 1 level), with no
Phase 3 standards
Alternative 3 - adopting Phase 2 standards in 2010 based on a large percentage of four-stroke
engines; (70% HC/30% CO) reduction
Alternative 4 - adopting more stringent Phase 2 in 2010 which would require optimized advanced
technology on every snowmobile; (85% HC/50% CO) reduction.
All of these alternatives were modeled assuming 100 percent compliance with the Phase
1 standards in 2006, whereas the final program includes a phase in with 50 percent compliance in
2006 and 100 percent compliance in 2007.
In addition to these alternative standards scenarios, we looked at what would happen if
four-stroke engine technology cost 25 percent more than we originally projected in order to
assess the sensitivity to four-stroke technology costs. This sensitivity analysis was done on
Alternative 4. This scenario will be referred to as Alternative 5 for the remainder of this
snowmobile section.
11-20
-------
Chapter 11: Regulatory Alternatives
11.3.3.1 Per unit Costs
The per unit costs for the various alternatives are shown in Tables 11.3.3-1 through
11.3.3-5. Also included in these tables are the technology mixes we used for each of the
alternatives. The per unit costs for alternative 1 (Phase 1 standards only) shown in Table 11.3.3-
1 are identical to the per unit costs for Phase 1 of the final program. The near term composite
incremental costs of all of the other alternatives can be compared to the near term incremental
cost of $89 for Phase 3 of the final program, as shown in Table 5.2.3-22 in Chapter 5.
11-21
-------
Table 11.3.3-1
Estimated Average Costs For Snowmobiles (Alternative 1 - Phase 1 only)
< 500 cc
(30%)
> 500 cc
(70%)
engine
modifications
modified
carburetor
direct
injection*
electronic fuel
injection
4-stroke engine
permeation
control
compliance
total
engine
modifications
modified
carburetor
direct
injection*
electronic fuel
injection
4-stroke engine
permeation
control
compliance
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$18
$18
$328
$175
$455
$7
$12
--
$25
$24
$295
$119
$770
$7
$12
--
--
--
Lifetime
Fuel
Savings
$0
$0
($512)
$0
($512)
($10)
--
--
$0
$0
($1,139)
$0
($1,139)
($10)
$0
--
--
--
Baseline
0%
0%
7%
12%
7%
0%
0%
--
0%
0%
7%
12%
7%
0%
0%
--
--
--
Phase 1
60%
60%
10%
15%
10%
100%
100%
--
60%
60%
10%
15%
10%
100%
100%
--
--
--
Incrementa
ICost
$11
$11
$10
$5
$14
$7
$12
$69
$15
$14
$9
$4
$23
$7
$12
$84
$80
$47
Incremental
Fuel Savings
$0
$0
($15)
$0
($15)
($10)
$0
($40)
$0
$0
($34)
$0
($34)
($10)
$0
($78)
($67)
($67)
-------
Table 11.3.3-2
Estimated Average Costs For Snowmobiles (Alternative 2 - Phase 2 HC standards with
Phase 1 CO standards)
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$175
$455
$2
--
$41
$295
$119
$770
$2
--
--
--
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
-
--
$0
($1,139)
$0
($1,139)
-
--
--
--
Phase 1
0%
10%
15%
10%
0%
--
0%
10%
15%
10%
0%
--
--
--
Phase 2
30%
35%
20%
15%
100%
--
30%
35%
20%
15%
100%
--
--
--
Incrementa
ICost
$12
$82
$9
$23
$2
$128
$12
$74
$6
$39
$2
$132
$131
$77
Incremental
Fuel Savings
$0
($128)
$0
($26)
$0
($154)
$0
($285)
$0
($57)
$0
($342)
($286)
($286)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
-------
Table 11.3.3-3
Estimated Average Costs For Snowmobiles (Alternative 3 - Four-stroke based Phase 2
Standards)
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$175
$455
$2
-
$41
$295
$119
$770
$2
--
--
--
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
--
-
$0
($1,139)
$0
($1,139)
--
--
--
--
Phase 1
0%
10%
15%
10%
0%
-
0%
10%
15%
10%
0%
--
--
--
Phase 2
25%
10%
65%
60%
100%
-
25%
10%
65%
60%
100%
--
--
--
Incrementa
ICost
$10
$0
$87
$228
$2
$327
$10
$0
$60
$385
$2
$457
$418
$260
Incremental
Fuel Savings
$0
$0
$0
($256)
$0
($256)
$0
$0
$0
($570)
$0
($570)
($476)
($476)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
-------
Table 11.3.3-4
Estimated Average Costs For Snowmobiles
(Alternative 4 - Phase 2 Standards based on broad application of advanced technology)
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$175
$455
$2
-
$41
$295
$119
$770
$2
--
--
--
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
--
-
$0
($1,139)
$0
($1,139)
--
--
--
--
Phase 1
0%
10%
15%
10%
0%
-
0%
10%
15%
10%
0%
--
--
--
Phase 2
0%
10%
90%
90%
100%
-
0%
10%
90%
90%
100%
--
--
--
Incrementa
ICost
$12
$0
$131
$364
$2
$497
$
$0
$90
$616
$2
$718
$652
$410
Incremental
Fuel Savings
$0
$0
$0
($410)
$0
($410)
$0
$0
$0
($911)
$0
($911)
($760)
($760)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
-------
Table 11.3.3-5
Estimated Average Costs For Snowmobiles (Alternative 4 with 25% higher 4-stroke costs)
< 500 cc
(30%)
> 500 cc
(70%)
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
pulse
air/recalibration
direct
injection*
electronic fuel
injection
4-stroke engine
certification
total
Near Term Composite
Incremental Cost
Long Term Composite
Incremental Cost
Cost
$41
$328
$218
$569
$2
-
$41
$295
$149
$963
$2
-
-
-
Lifetime
Fuel
Savings
$0
($512)
$0
($512)
-
-
$0
($1,139)
$0
($1,139)
-
-
-
-
Phase 1
0%
10%
15%
10%
0%
-
0%
10%
15%
10%
0%
-
-
-
Phase 2
0%
10%
90%
90%
100%
-
0%
10%
90%
90%
100%
-
-
-
Incrementa
ICost
$12
$0
$164
$455
$2
$621
$
$0
$112
$770
$2
$894
$812
$512
Incremental
Fuel Savings
$0
$0
$0
($410)
$0
($410)
$0
$0
$0
($911)
$0
($911)
($760)
($760)
* Direct injection costs are an average of the air-assisted and pump assisted system costs.
-------
Chapter 11: Regulatory Alternatives
11.3.3.2 Aggregate Cost Estimates
Based on the above per unit costs, we have estimated the aggregate costs for the
alternatives. The aggregate costs for the alternatives are presented in Table 11.3.3-6, along with
the aggregate cost estimates for the final snowmobile program, which are estimated in Chapter 5.
The fuel savings result in varying degrees of switching from current two-stroke technology to
direct injection two-stroke and four-stroke technology.
Table 11.3.3-6
Summary of Annual Snowmobile Aggregate Costs and Fuel Savings (millions of dollars)
Final program
Alternative 1
Alternative 2
Alternative 3
Alternative 4
Alternative 5
Fuel savings (Final
program)
Fuel Savings (Alt 1)
Fuel Savings (Alt 2)
Fuel Savings (Alt 3)
Fuel Savings (Alt 4)
Fuel Savings (Alt 5)
2006
$6.58
$13.17
$13.17
$13.17
$13.17
$13.17
$0.78
$0.78
$0.78
$0.78
$0.78
$0.78
2010
$37.55
$12.07
$38.99
$98.99
$148.68
$182.23
$11.81
$4.31
$8.81
$11.81
$16.31
$16.31
2015
$41.91
$11.08
$28.65
$70.03
$104.08
$127.25
$58.23
$9.13
$38.59
$58.23
$87.68
$87.68
2020
$41.56
$11.73
$30.32
$74.13
$110.17
$134.69
$103.00
$12.33
$66.73
$103.00
$157.40
$157.40
2025
$41.56
$11.73
$30.32
$74.13
$110.17
$134.69
$123.66
$13.51
$79.60
$123.66
$189.75
$189.75
11.3.3.3 Emissions Reductions
In Chapter 6, we estimated the emissions reductions for the final program. We have
estimated the emissions reductions for the alternatives using the same methodology. The results
for HC are shown in Table 11.3.3-7 and in Figure 11.3.3-1, while the results for CO are shown in
Table 11.3.3-8 and in Figure 11.3.3-2.
As can be seen in Tables 11.3.3-7 and 11.3.3-8, there are cases where the emissions
reductions for a given pollutant are different for different alternatives even though the numerical
limits for that pollutant are the same for those alternatives. For example, the final program and
11-27
-------
Draft Regulatory Support Document
Alternative 2 would both require 50 percent reductions in HC, but the HC reductions shown in
Table 11.3.3-7 are different for these two options. The reason for this difference in HC
reductions is that under these two options the CO limits are different. Under the final program
the CO limit would require a 50 percent reduction in CO, while in Alternative 2 the CO
reductions would only be 30 percent. This difference in CO limits results in the need for a
different technology mix being needed under the two alternatives. The more aggressive
application of technology needed under the final program to meet the CO limit has the effect of
producing somewhat higher HC reductions. Similarly, the different HC limits for Alternatives 1
through 3 result in different technology mixes for the these alternatives. These different
technology mixes result in different CO reductions for each alternative even though the CO
limits are the same for all three alternatives. This can be seen in Tale 11.3.3-8.
Table 11.3.3-7
Summary of Snowmobile HC Reductions (thousands of tons)
Final Program
Alternative 1
Alternative 2
Alternative 3
Alternatives 4 and 5
2006
4.0
7.9
7.9
7.9
7.9
2010
42.9
44.9
47.3
52.1
55.8
2015
123.3
98.4
114.2
146.8
172.4
2020
196.1
135.1
165.2
227.6
276.4
2025
230.4
148.5
185.6
262.4
322.4
Table 11.3.3-8
Summary of Snowmobile CO Reductions (thousands of tons)
Final Program
Alternative 1
Alternative 2
Alternative 3
Alternatives 4 and 5
2006
9.9
19.9
19.9
19.9
19.9
2010
105.3
112.7
116.2
120.1
123.1
2015
285.0
246.6
270.1
296.6
317.4
2020
442.2
338.7
383.6
436.8
476.8
2025
513.4
372.3
427.7
493.1
544.0
11-28
-------
Chapter 11: Regulatory Alternatives
Figure 11.3.3-1 Snowmobile HC Emissions Inventory
Nationwide HC Emissions (tons)
T3
CO
400
300
200
100
2005
2010
• Baseline
• Final program
2015 2020
Year
Alt. 1 —*—
* Alt. 2
2025
Alt. 3
2030
Alt. 4 and 5
Figure 11.3.3-2 Snowmobile CO Emissions Inventory
Nationwide CO Emissions (tons)
2005
•^— Baseline
0 Final Program
2010
2015 2020
Year
2025
Alt. 1
Alt. 2
Alt. 3
2030
Alt. 4 and 5
11-29
-------
Draft Regulatory Support Document
11.3.3.4 Cost Per Ton
Chapter 7 provides the cost per ton estimates for the final program. Using the same
methodology, we have estimated the cost per ton of HC and CO reduced for the alternatives, as
shown in Table 11.3.3-9. The results for alternative 1 (Phase 1 standards only) are shown first.
All other scenarios, including the final program, are base on the incremental change from the
Phase 1 standards to whatever Phase 2 standards are considered in the particular scenario.
Table 11.3.3-9
Estimated Snowmobile Average Cost per Ton of HC and CO Reduced
(7 percent discount rate)
Alternative 1
Final Program0
Alternative 2a
Alternative 3a
Alternative 4a
Alternative 5ab
Lifetime Reductions
per Vehicle
(NPV tons)
HC
0.40
n/a
0.10
0.28
0.49
0.49
CO
1.02
0.25
n/a
n/a
0.50
0.50
Discounted per Vehicle Cost Per
Ton without Fuel Savings
($/ton)
HC
$90
n/a
$1,370
$1,480
$670
$840
CO
$40
$360
n/a
n/a
650
$810
Discounted Per Vehicle Cost Per
Ton with Fuel Savings ($/ton)
HC
$20
n/a
($1,610)
($210)
($110)
($50)
CO
$10
($410)
n/a
n/a
($110)
($50)
a. Shown based on incremental change from Phase 1 standards.
b. Alternative 4 with 25% higher 4-stroke cost.
c. Shown based on incremental change from Phase 2 standards
11.3.3.5 Economic Impacts Discussion
The economic costs of the regulatory alternatives for snowmobiles are presented. Net
social costs (or gains) of the alternatives in the year 2030 are shown on Table 11.3.3-10, while
the net present value of these costs through 2030 are reflected on Tables 11.3.3-1 la and 11.3.3-
1 Ib. The methodologies used to estimate the economic costs of these alternatives are discussed
extensively in Chapter 9. Each of the alternatives, is modeled based on a 30 percent reduction in
HC and CO, respectively during Phase 1 of the regulation.
11-30
-------
Table 11.3.3-10
Economic Costs of Alternative Snowmobile Standards—
Values in 203013 ( millions of 2001$)
Scenario
Alternative 1
Alternative 2
Final Program
Alternative 3
Alternative 4
Alternative 54
Engineering
Costs
$11.7
$30.3
$43.1
$74.1
$111.2
$134.7
Economic Costs
(Surplus Losses)
$11.6
$29.8
$41.9
$70.5
$102.1
$122.7
Fuel Efficiency
Cost Savings
$18.2
$88.0
$135.0
$134.5
$204.3
$204.3
Economic Gains or
Costs2
$6.6
$58.2
$93.1
$64.0
$102.2
$81.6
1. Assumes the final program Phase 1 standards as the first phase in each alternative
2. Economic costs or net economic costs shown in parenthesis.
3. Dollar values are rounded to the nearest 10 million.
4. Same standards as Alternative 4, but assumes a 25% increase in the cost of a 4-stroke engine.
Table 11.3.3-lla
Economic Costs of Alternative Snowmobile Standards—
Net Present Value 2002 through 20301
(millions of 2001$, using 3 percent discount rate)
Scenario
Alternative 1
Alternative 2
Final Program
Alternative 3
Alternative 4
Alternative 53
Engineering
Costs
$183.7
$426.9
$569.6
$987.6
$1,450.1
$1,763.8
Economic Costs
(Surplus Losses)
$182.1
$418.9
$553.1
$885.0
$1,335.0
$1,591.8
Fuel Efficiency
Cost Savings
$174.7
$697.7
$999.6
$1,046.3
$1,569.3
$1,569.3
Economic
Gains or
Costs2
($7.4)
$278.8
$446.5
$161.3
$234.3
($22.5)
1. Assumes the final program Phase 1 standards as the first phase in each alternative
2. Economic costs or net economic costs shown in parenthesis.
3. Same standards as Alternative 4, but assumes a 25% increase in the cost of a 4-stroke engine.
-------
Draft Regulatory Support Document
Table 11.3.3-1 Ib
Economic Costs of Alternative Snowmobile Standards—
Net Present Value 2002 through 20301
(millions of 2001$, using 7 percent discount rate)
Scenario
Alternative 1
Alternative 2
Final Program
Alternative 3
Alternative 4
Alternative 53
Engineering Costs
$106.6
$235.7
$305.7
$531.5
$775.7
$941.1
Economic Costs
(Surplus Losses)
$105.7
$231.1
$296.9
$470.0
$713.1
$847.6
Fuel Efficiency
Cost Savings
$86.8
$327.2
$459.7
$487.4
$727.8
$727.8
Economic Gains
or Costs2
($18.9)
$96.1
$162.8
$17.4
$14.7
($119.8)
1. Assumes the final program Phase 1 standards as the first phase in each alternative
2. Economic costs or net economic costs shown in parenthesis.
3. Same standards as Alternative 4, but assumes a 25% increase in the cost of a 4-stroke engine.
11-32
-------
Chapter 11: Regulatory Alternatives
11.3.3.6 Discussion
Alternative 1 (Phase 1 standards only) would require relatively minimal additional use of
advanced technologies beyond what we project as a baseline. These advanced technologies
(direct injection two-stroke, and four-stroke technologies) have been shown to be both feasible
and capable of emissions reductions well below those required of the Phase 1 standards. Thus,
we do not believe that this alternative would meet the basic criteria of the Clean Air Act which
requires us to set standards based on the greatest degree of emissions reductions achievable.
Alternative 2 (Phase 2 HC standards with Phase 1 CO standards) would require roughly
half of new snowmobiles to have advanced technology beginning with the 2010 model year, with
the emphasis on direct injection two-stroke technology. The remaining snowmobiles would have
a combination of engine modifications, recalibration and electronic fuel injection. We believe
that a higher level of advanced technology than 50 percent penetration is certainly feasible
beyond 2010 and therefore do not believe that in the absence of more stringent Phase 3 standards
this alternative would meet the basic criteria of the Clean Air Act which requires us to set
standards based on the greatest degree of emissions reductions achievable.
Alternative 3 (more stringent Phase 2 HC standards than final program in conjunction
with Phase 1 CO standards) would require more advanced technology. We modeled 60 percent
of the snowmobiles produced would be powered by four-stroke engines in 2010 and an
additional ten percent would utilize direct injection two-stroke technology. The remainder would
require some other technologies such as recalibrations and electronic fuel injection. We believe
that these alternative standards strike a reasonable balance for allowing four stroke engines to be
a primary Phase 2 technology, and have adopted these standards as an alternative to our primary
Phase 2 standards on an engine family by engine family basis. Further discussion of our reasons
for offering these standards as a Phase 2 option can be found in the preamble to the final rule.
Alternative 4 would require advanced technologies on all snowmobiles, beginning in
2010. We modeled 90 percent requiring four-stroke engines and the remaining ten percent
requiring direct injection two-stroke technology. As discussed in detail in the preamble, given
the number of snowmobile models and engine model offerings for each snowmobile model, and
the fact that snowmobiles have not previously been regulated or used these advanced
technologies in large numbers, we do not believe that it is feasible to apply and optimize
advanced technology to every snowmobile by the 2010 model year. Thus we are not confident
that this option is would be feasible in the time frame provided. We will, however, monitor the
development and application of advanced technology and will in the future consider the adoption
of snowmobile standards that would require advanced technology on every snowmobile.
Alternative 5 is simply a sensitivity analysis to look at how the cost of four-stroke engines
might impact the consideration of Phase 2 standards which are based largely on four-stroke
technology. This alternative has the same standards as Alternative 4, but with 25 percent higher
11-33
-------
Draft Regulatory Support Document
costs for four-stroke engines.
11.4 Recreational Vehicle Permeation Emission Standards
While developing the fuel tank and hose permeation standards, we analyzed alternative
approaches both more and less stringent than the final standards. These alternative approaches
are discussed below.
11.4.1 Fuel Tanks
The final permeation standard for fuel tanks is 1.5 g/m2/day when tested at 23°C on a test
fuel with 90 percent gasoline and 10 percent ethanol. This standard represents approximately an
85 percent reduction from baseline HDPE fuel tanks. We considered an alternative standard
equivalent to about a 60 percent reduction from baseline. This could be met by fuel tanks
molded out of nylon. We also considered requiring metal fuel tanks which would essentially
eliminate permeation emissions from fuel tanks.
11.4.1.1 60 Percent Reduction (Nylon Fuel Tanks)
One manufacturer commented that we should relax the fuel tank standard to a 55-60
percent reduction so that other technologies could be used. Specifically, they point to injection-
molded nylon. Therefore, for this analysis, we consider the costs and emissions reductions
associated with molding the fuel tank out of nylon.
As discussed in Chapter 5, nylon costs about $2.00 per pound while HDPE costs about
$0.50 per pound. Depending on the shape of the fuel tank and the wall thickness, recreational
vehicle fuel tanks weigh about 1-1.3 pounds per gallon. Including a 29% markup for overhead
and profit, the increased cost for using nylon fuel tanks would be about $21 for snowmobiles (11
gallons), $10 for ATVs (4 gallons), and $8 for off-highway motorcycles (3 gallons). This is
actually 5-10 times higher than our projected costs for using sulfonation to meet the final
standard which represents about an 85 percent reduction.
Based on the data presented in Chapter 4, the use of nylon could achieve more than a 95
percent reduction in permeation compared to HDPE when gasoline is used. However, if a 10
percent ethanol blend is considered, then the reduction is only 40-60 percent depending on the
nylon composition. On a 15 percent methanol blend, the permeation rate through nylon can
actually be several times higher than through HDPE.
About one third of the gasoline sold in the U.S. today is blended with ethanol or some
other oxygenate. In addition, the trend in the U.S. is towards using more renewable fuel and
ethanol may be the leading choice. Therefore, it is important that the permeation control strategy
used for recreational vehicles be effective on ethanol fuel blends. For this analysis, we consider a
11-34
-------
Chapter 11: Regulatory Alternatives
10 percent ethanol blend when calculating emissions reductions.
Table 11.4-1 presents the projected national emission reductions for this approach. These
figures can be compared to the anticipated reductions presented in Chapter 6 for the final
standards (Table 6.2.6-3). Table 11.4-2 presents the cost per ton of permeation emissions
reduced per fuel tank, using a 7 percent discount rate, with and without fuel savings. These
figures can be compared to the cost per ton presented in Chapter 7 (Table 7.1.5-1).
Table 11.4-1
Projected Fuel Tank Permeation Emissions from Recreational Vehicles
for the Alternative Approach of a 60 Percent Reduction [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
3,389
3,389
0
3,985
3,985
0
882
882
0
8,255
8,255
0
2005
4,181
4,181
0
6,751
6,751
0
1,303
1,303
0
12,234
12,234
0
2010
5,032
4,106
92
9,275
8,072
1,202
1,710
1,492
218
16,016
13,671
2,345
2020
6,456
2,737
3,719
11,109
5,455
5,654
2,061
1,239
821
19,626
9,431
10,194
2030
7,061
2,824
4,236
11,231
4,539
6,692
2,248
1,315
933
20,539
8,678
11,862
Table 11.4-2
Estimated Cost Per Ton of HC Reduced (7 percent discount rate)
for the Alternative Approach of a 60 Percent Reduction from Fuel Tanks
Snowmobiles
ATVs
OHMC
Total
Cost Per
Vehicle
$21
$10
$8
Lifetime Fuel
Savings Per
Vehicle
(NPV)
$o
3
$2
$1
Lifetime
Reductions Per
Vehicle
(NPV tons)
0.0084
0.0047
0.0027
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$2,541
$2,065
$2,819
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$2,178
$1,702
$2,456
Constructing fuel tanks out of nylon would be significantly more expensive than
constructing them out of HDPE and applying a barrier treatment such as sulfonation to control
11-35
-------
Draft Regulatory Support Document
permeation. Therefore, we believe that most manufacturers would choose the lower cost option
of applying a barrier treatment even if we were to set a standard based on a 60 percent reduction.
In addition, we believe that they would target the maximum effectiveness of the barrier
treatment. Designing for a 60 percent reduction would not have meaningful cost savings over
designing for a 95 percent reduction. As a result, while this option could result in less emission
control than the standard, we do not believe that it would lower costs for manufacturers.
11.4.1.2 Metal Fuel Tanks
One commenter pointed out that essentially a 100 percent reduction in fuel tank
permeation emissions could be achieved by replacing plastic fuel tanks with metal fuel tanks.
However, they stated that a performance standard approaching this amount of emission reduction
would be appropriate because it would allow industry flexibility on how to meet the standard.
For this scenario we consider the use of metal fuel tanks in recreational vehicles.
Today, most if not all recreational vehicles use plastic fuel tanks. According to
manufacturers plastic fuel tanks are desirable because they weigh less than metal fuel tanks, are
more durable, can be formed into more complex shapes, are non-corrosive, and cost less. In
recreational vehicle applications, weight is an issue because the vehicles must be light enough to
be manipulated by the rider. However, more importantly, durability is an issue because of the
rough use of these vehicles and because many of the fuel tanks are exposed. For example, if a
dirt bike were to fall over, a metal tank could be dented on a rock which would damage the
integrity of the fuel tank. A plastic tank, however, would likely be undamaged. In addition metal
fuel tanks have seams due to the manufacturing process which are weak point and could result in
leaking. Fuel tanks on recreational vehicles, are designed to maximize the fuel stored in a
limited space. Current plastic fuel tank designs are molded with contours that match the vehicle
chassis. Manufacturers have stated that these complex shapes cannot be stamped into metal parts
and that using metal tanks could cause them to need to redesign the fuel tank geometry and could
require modifications to the chassis in order to maintain the same fuel capacity.
For the purposes of this analysis we use a cost increase of 30 percent for metal tanks
versus plastic fuel tanks. This is based on pricing seen for marine applications which use metal
fuel tanks in some cases. Because metal fuel tanks are not used in recreational vehicle
applications, direct costs cannot be used. This cost does not include research and design costs
that would be required for developing metal tanks or costs of modifying production practices.
Dealer prices for plastic fuel tanks, of the size used in recreational vehicles, range from 3 to 9
dollars per gallon of capacity.1 Using an average cost of 6 dollars per gallon and a typical dealer
markup, we get a cost of about 2 dollars per gallon for plastic fuel tanks. This cost estimate for
plastic fuel tanks was confirmed in conversations with recreational vehicle manufacturers. Based
on this analysis and a markup of 29%, we estimate a cost increase of about $9 for snowmobiles,
$3 for ATVs, and $2 for non-competition off-highway motorcycles.
11-36
-------
Chapter 11: Regulatory Alternatives
Table 11.4-3 presents the projected national emission reductions for this approach. These
figures can be compared to the anticipated reductions presented in Chapter 6 for the final
standards (Table 6.2.6-3). Table 11.4-4 presents the cost per ton of permeation emissions
reduced per fuel tank, using a 7 percent discount rate, with and without fuel savings. These
figures can be compared to the cost per ton presented in Chapter 7 (Table 7.1.5-1).
Table 11.4-3
Projected Fuel Tank Permeation Emissions from Recreational Vehicles
for the Alternative Approach of a 100 Percent Reduction [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
3,389
3,389
0
3,985
3,985
0
882
882
0
8,255
8,255
0
2005
4,181
4,181
0
6,751
6,751
0
1,303
1,303
0
12,234
12,234
0
2010
5,032
3,489
1,542
9,275
7,271
2,004
1,710
1,347
363
16,016
12,107
3,909
2020
6,456
258
6,198
11,109
1,685
9,424
2,061
692
1,369
19,626
2,635
16,991
2030
7,061
0
7,061
11,231
78
11,153
2,248
692
1,556
20,539
770
19,769
Table 11.4-4
Estimated Cost Per Ton of HC Reduced (7 percent discount rate)
for the Alternative Approach of a 100 Percent Reduction
Snowmobiles
ATVs
OHMC
Total
Cost Per
Vehicle
$9
$o
3
$2
Lifetime Fuel
Savings Per
Vehicle
(NPV)
$5
$3
$2
Lifetime
Reductions
Per Vehicle
(NPV tons)
0.0140
0.0078
0.0046
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$668
$435
$509
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
$305
$72
$146
Although this approach appears to be cost effective, we did not chose to set standards that
would require manufacturers to use metal fuel tanks. We believe that there may be safety
concerns with metal fuel tanks on recreational vehicles because of the rough use and likelihood
of damage to the fuel tanks. Because some applications may be able to use metal fuel tanks, we
11-37
-------
Draft Regulatory Support Document
will accept a metal tank for design-based certification to our standard. In addition, we believe
that the final tank permeation standard can achieve nearly the same level of reduction as metal
tanks while providing manufacturers very important flexibility in their design and manufacturing.
11.4.2 Hoses
The hose standard is 15 g/m2/day when tested at 23°C on a test fuel with 90 percent
gasoline and 10 percent ethanol (E10). For hoses we considered basing the standard on testing
with an alcohol-free test fuel. We also considered a standard that would require the use of fuel
tubing, such as used in automotive applications, which is fairly rigid in comparison to fuel hoses
because tubing is generally constructed out of fluorothermoplastics while hoses are primarily
constructed out of rubber.
11.4.2.1 Alcohol-Free Test Fuel
Manufacturers commented that we should specify ASTM Fuel C (50% toluene, 50% iso-
octane) for the hose permeation testing, stating that this is the fuel used for measuring permeation
under the SAE J30 recommended practice for R9 hose. Under SAE J30, R9 hose must meet a
permeation rate of 15 g/m2/day when tested at 23°C. Manufacturers noted that fuels with
ethanol-gasolines blends would have a higher permeation rate than if they were tested on
gasoline. Therefore, R9 hose would not necessarily meet the hose permeation standards. As
noted in Chapter 4, barrier materials typically used in R9 hose today may have permeation rates 3
to 5 times higher on a 10 percent ethanol blend than on straight gasoline. In this section, we
analyze the alternative of basing our hose permeation standard on testing using an alcohol-free
test fuel.
For the purposes of our benefits analysis, as described in Chapter 6, we estimated that a
hose designed to meet 15 g/m2/day on E10 fuel would permeate at half of that rate when tested
on gasoline. This estimate considers the entire hose construction and not just the effect of
alcohol on the barrier materials. To model this alternative, we doubled the estimated permeation
rates for hoses meeting the permeation standards. Based on costs of hose available today, R9
hose would cost about $0.75/ft which represents a $0.50/ft increase from R7 hose used in most
applications today. For the same reasons as discussed in Chapter 5, we are conservatively adding
a cost of hose clamps ($0.20 each). As with the analysis in Chapter 5, we include a 29 percent
markup in costs for profit and overhead.
Table 11.4.1-5 presents the projected national emission reductions for this approach.
These figures can be compared to the anticipated reductions presented in Chapter 6 for the final
standards (Table 6.2.6-4). Table 11.4-6 presents the cost per ton of permeation emissions
reduced per fuel tank, using a 7 percent discount rate, with and without fuel savings. These
figures can be compared to the cost per ton presented in Chapter 7 (Table 7.1.5-1).
11-38
-------
Chapter 11: Regulatory Alternatives
Table 11.4-5
Projected Fuel Hose Permeation Emissions from Recreational Vehicles for
the Alternative Approach of Using an Alcohol-Free Test Fuel [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
4,471
4,471
0
4,243
4,243
0
1,878
1,878
0
10,592
10,592
0
2005
5,516
5,516
0
7,189
7,189
0
2,774
2,774
0
15,478
15,478
0
2010
6,638
4,659
1,979
9,876
7,800
2,076
3,642
2,890
751
20,156
15,349
4,806
2020
8,517
564
8,074
11,829
2,068
9,761
4,389
1,553
2,836
24,735
4,184
20,550
2030
9,315
254
9,061
11,959
407
11,552
4,787
1,565
3,222
26,061
2,225
23,835
Table 11.4-6
Estimated Cost Per Ton of HC Reduced (7 percent discount rate) for
the Alternative Approach of Using an Alcohol-Free Test Fuel [short tons]
Snowmobiles
ATVs
OHMC
Total
Cost Per
Vehicle
$4
$1
$2
Lifetime Fuel
Savings Per
Vehicle
(NPV)
$7
$3
$3
Lifetime
Reductions
Per Vehicle
(NPV tons)
0.0179
0.0081
0.0095
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$212
$144
$157
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
($151)
($219)
($206)
We also received comment that we should use the most permeable fuel blend on the
market for testing the permeation rates through hoses. As discussed above, we believe that the
use of ethanol-blended gasoline is too significant today to ignore and could increase in the future.
For this reason, we believe that it is appropriate to base the standards on testing using E10 fuel.
We do not believe it is necessary to relax the standards to allow R9 hose to be able to pass on
E10 fuel. Several materials are available today that could be used as a low permeation barrier in
rubber hoses that are resistant to permeation on alcohol fuel blends. In fact, SAE J30 specifies
Rl 1 and R12 hose which are low permeability hoses tested on 15 percent methanol blend.
Chapter 4 presents data on low permeation hoses developed for automotive applications that
easily meet the final hose permeation standards that we believe could be used on recreational
11-39
-------
Draft Regulatory Support Document
applications. Finally, the incremental cost is small ($0.10/ft) between hose that would meet 15
g/m2/day on straight gasoline versus gasoline with a 10 percent ethanol blend.
11.4.2.2 Automotive Plastic Fuel Tubing
In developing emission standards for nonroad vehicles, the Clean Air Act requires us to
first consider standards for comparable on-highway applications. In automotive applications,
manufacturers generally use very low permeation plastic fuel tubing to meet our evaporative
emission requirements. Recommended practice specified by SAE J2260 defines a Category 1
fuel line which must meet a permeation requirement of 25 g/m2/day at 60°C on a test fuel with 85
percent gasoline and 15 percent methanol (Ml5). This is roughly equivalent to meeting a limit of
2 g/m2/day at 23°C. In addition, based on the data in Chapter 4, permeation rates for most
materials used in hoses tend to be at least twice as high for M15 than E10 fuel. This plastic
tubing is generally made of fluoropolymers such as ETFE or PVDF.
Manufacturers commented that fuel hose standards based on automotive fuel lines such as
specified in SAE J22602 as Category 1 would be inappropriate for recreational vehicles.
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
recreational vehicle design. Manufacturers have commented that they would need flexible hose
to fit their many designs, resist vibration, and to simplify the hose connections and fittings.
Plastic fuel tubing would likely cost less than multilayer barrier fuel hoses, but we
estimate that it would cost about $0.50 per foot more than the rubber hoses currently used on
recreational vehicles. This additional cost includes a markup to form the tubing to the tight
bends that would be required for recreational applications. Although the fluoroplastics are more
expensive than the materials used in hoses on a per pound basis, plastic automotive tubing is
constructed with thin walls (approximately 1 mm on average). An additional cost associated
with automotive fuel tubing would be for more sophisticated connectors for the plastic tubing.
On recreational vehicles using rubber fuel hose, the hose is generally just pushed on to
connectors formed into the fuel tank and carburetor. In some cases, these are push on fittings
without the use of a clamp. In automotive applications, quick connects are generally used which
cost about $0.50 each.3 For ATVs and OHMCs, we include the costs of two quick connects for
each vehicle. Snowmobiles can require 4 to 8 quick connects depending on the fuel pump
configuration, number of carburetors, and if a fuel return line is included. We include the cost of
six quick connects in this analysis.
Table 11.4-7 presents the projected national emission reductions for this approach. These
figures can be compared to the anticipated reductions presented in Chapter 6 for the final
standards (Table 6.2.6-4). Table 11.4-8 presents the cost per ton of permeation emissions
reduced per fuel tank, using a 7 percent discount rate and a 29 percent markup for overhead and
profit, with and without fuel savings. These figures can be compared to the cost per ton
11-40
-------
Chapter 11: Regulatory Alternatives
presented in Chapter 7 (Table 7.1.5-1).
Table 11.4-7
Projected Fuel Hose Permeation Emissions from Recreational Vehicles for
the Alternative Approach of Basing the Standard on Automotive Fuel Tubing [short tons]
Vehicle
Snow-
mobiles
ATVs
OHMCs
Total
Scenario
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
baseline
control
reduction
2000
4,471
4,471
0
4,243
4,243
0
1,878
1,878
0
10,592
10,592
0
2005
5,516
5,516
0
7,189
7,189
0
2,774
2,774
0
15,478
15,478
0
2010
6,638
4,605
2,033
9,876
7,744
2,132
3,642
2,870
772
20,156
15,219
4,936
2020
8,517
348
8,169
11,829
1,804
10,026
4,389
1,476
2,913
24,735
3,627
21,107
2030
9,315
8
9,306
11,959
93
11,865
4,787
1,478
3,310
26,061
1,579
24,481
Table 11.4-8
Estimated Cost Per Ton of HC Reduced (7 percent discount rate) for
the Alternative Approach of Basing the Standard on Automotive Fuel Tubing
Snowmobiles
ATVs
OHMC
Total
Cost Per
Vehicle
$6
$2
$2
Lifetime Fuel
Savings Per
Vehicle
(NPV)
$7
$3
$4
Lifetime
Reductions
Per Vehicle
(NPV tons)
0.0184
0.0083
0.0097
Discounted Per
Vehicle Cost Per Ton
without Fuel Savings
($/ton)
$333
$233
$232
Discounted Per
Vehicle Cost Per Ton
with Fuel Savings
($/ton)
($30)
($130)
($131)
Although this approach appears to be cost effective, we did not choose to set standards
that would require manufacturers to automotive type fuel tubing. We are concerned that the
tubing is too rigid for the tight installation spaces and radii in recreational vehicle applications.
Hoses on these vehicles today often have tight bends and are subject to high amounts of shock
and vibration The above analysis does not include costs of adding additional length that may be
required for molding in spirals or other bends for vibration resistance. Because some
applications may be able to automotive fuel tubing, we will accept fuel lines conforming to
SAE J2260 Category 1 for design-based certification to our standard. In addition, we believe that
11-41
-------
Draft Regulatory Support Document
the final hose permeation standard can achieve nearly the same level of reduction as metal tanks
while providing manufacturers flexibility in their design.
11.5 Incremental Cost Per Ton Analysis
The above discussion analyzes several options for the different engine categories. For
completeness, we have also examined the cost per ton associated with the incremental steps in
standards changes. The table below provides a summary of the incremental cost per ton for the
differences in the alternatives analyzed above. Details of the alternative are provided above for
each program.
11-42
-------
Chapter 11: Regulatory Alternatives
Table 11.5-1: Incremental Cost Per Ton Estimates
Change in Standards
Off-highway Motorcycles
(change in g/km HC+NOX
standard)
Baseline -> 4.0 g/km b
Baseline -> 2.0 g/km
2.0 g/km -> 1.0 g/km
ATVs (change in g/km
HC+NOX standard)
Baseline -> 2.0 g/km
2.0 -> 1.5 g/km
1.5 -> 1.0 g/km
Snowmobiles
(HC/CO percent
reduction)
Baseline -> 30/30
30/30 -> 50/30
50/30 -> 50/50
50/30 -> 70/30
70/30 -> 85/50
Large SI
Baseline -> Phase 1
Phase 1 -> Phase 2
Average Cost
w/o fuel
savings
$210
$158
$70
w/o fuel
savings
$73
$11
$48
w/o fuel
savings
$80
$131
$89
$287
$234
w/o fuel
savings
$611
$55
w/fuel
saving
$122
$105
$70
w/fuel
saving
$50
$11
$48
w/fuel
saving
$13
($155)
($102)
$97
($50)
w/fuel
saving
($3,370)
$55
Lifetime
Reductions per
Vehicle
(NPVtons)"-
HC+NOx
0
0
0
.50
.38
.02
HC+NOx
0
0
0
HC
0.40
0.10
n/a
0.19
0.14
.20
.01
.01
CO
1.02
0.16
0.25
n/a
0.15
HC+NOx
3
0
.07
.80
Discounted per
Vehicle Cost Per
Ton without Fuel
Savings ($/ton) a
HC+NOx
$420
$410
$3,590
HC+NOx
$370
$1,010
$4,740
HC
$90
$1,370
n/a
$1,540
$820
CO
$40
n/a
$330
n/a
$780
HC+NOx
$240
$80
Discounted Per
Vehicle Cost Per
Ton with Fuel
Savings ($/ton) a
HC+NOx
$210
$280
$3,590
HC+NOx
$250
$1,010
$4,740
HC
$20
($1,610
)
n/a
$520
$180
CO
$10
n/a
($430)
n/a
($170)
HC+NOx
($1,150)
$80
a. Calculated using a discount rate of 7 percent.
b. The 4.0 g/km alternative requires manufacturers to certify competition off-highway motorcycles whereas the other
alternative does not.
11-43
-------
Draft Regulatory Support Document
Chapter 11 References
1.White, JJ and Carroll, JN. (1998) Emissions from snowmobile engines using bio-based fuels
and lubricants. Final Report. Prepared for State of Montana Department of Environmental
Quality. Southwest Research Institute report #73 83.
2.U.S. EPA Guidelines for the Health Risk Assessment of Chemical Mixtures. EPA Report No.
630/R-98/002. September 24, 1986. Federal Register 51(185) : 34014 - 34025.
http ://cfpub. epa. gov/ncea/
3.White, JJ and Carroll, JN. (1998) Emissions from snowmobile engines using bio-based fuels
and lubricants. Final Report. Prepared for State of Montana Department of Environmental
Quality. Southwest Research Institute report #73 83.
Buckingham, JP; White, JJ; and Carroll, JN. (1996) Development of snowmobile test cycle -
final report. Southwest Research Institute report #7574.
4.Benignus, V, W Boyes, and P. Bushnell, US Environmental Protection Agency. National
Health and Environmental Effects Research Laboratory. Memorandum to the Docket. Acute
Behavioral Effects of Exposure to Toluene and Carbon Monoxide from Snowmobile Exhaust.
September 2002.
1. www.marinepart.com/fuetmold, A copy of this has been placed in the Docket A-2000-01,
Document IV-A-87.
2. SAE Recommended Practice J2260, "Nonmetallic Fuel System Tubing with One or More
Layers," 1996, Docket A-2000-01, Document IV-A-18.
3. 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 A-2000-01, Document No. IV-B-05.
11-44
------- |