United States
Environmental Protection
Agency
Office of Air and Radiation
(ANR-443)
Washington, DC 20460
EPA-450/3-84-012C
July 1987
Air
Evaluation of Air
Pollution Regulatory
Strategies for
Gasoline Marketing
Industry
Response to Public
Comments
-------
TECHNICAL REPORT DATA
(Please read Insiructiom on the reverse before completing)
. REPORT NO.
EPA-450/3-84-Q12C
3. RECIPIENT'S ACCESSION NO,
4. TITLE AND SUBTITLE
iii uc «rxu aua 111 uc
Evaluation of Air Pollution Regulatory Strategies
for the Gasoline Marketing Industry - Response to
Public Comments
S. REPORT DATE
July 1987
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME ANp.AOOaESS . ,-,_,,
Director, Office of Air Quality Planning and Standards
Director, Office of Mobile Sources
U.S. Environmental Protection Agency
Washington, D.C 20460
10. PROGRAM ELEMENT NO,
11. CONTBACT/GHANT NO.
68-02-3060
12. SPONSORING AGENCY NAME AND ADDRESS
Assistant Administrator for Air and Radiation
U.S. Environmental Protection Agency
Washington, D.C. 20460
13, TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The gasoline marketing industry (bulk terminals, bulk plants, service station
storage tanks, and service station vehicle refueling operations) emit to the atmo-
sphere several organic compounds of concern. These include: volatile organic
compounds (VOC), which contribute to ozone formation; benzene, which has been listed
as a hazardous air pollutant based on human evidence of carcinogencity, and gasoline
vapors, for which there is animal evidence of carcinogencity. This document provides
a summary of EPA responses to public comments on environmental and economic
analysis published by EPA in 1984 (EPA-450/3-84-012a and b). Changes made to EPA's
1984 analysis in response to public comments, additional analyses performed, and a
summary of the results are contained in a separate two-volume draft Regulatory
Impact Analysis document (EPA-450/3-87-001 a and b).
17.
KEY WORDS ANO DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Croup
Gasoline Refueling
Air Pollution Onboard
Pollution Control Stage II
Stationary Sources
Mobile Sources
Volatile Organic Compounds Emissions
Benzene
Air Pollution Control
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS fTMs Rtport/
Unclassified
21. NO, OF PAGES
327
Unlimited
20, SECURITY CLASS (Tllil page!
Unclassified
22. PRICE
EPA Form 2220-1 (R«». 4-77) PREVIOUS EDITION is OBSOLETE
-------
EPA 450/3-84-012c
Evaluation of Air Pollution
Regulatory Strategies for
Gasoline Marketing Industry
Response to Public Comments
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
AND
OFFICE OF MOBILE SOURCES
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Washington DC 20460
July 1987
-------
This report has been reviewed by the Office of Air Quality Planning and Standards and the Office of Mobile
Sources, EPA, and approved for publication. Mention of trade names or commercial products is not intended
to constitute endorsement or recommendation for use. Copies of this report are available through the Library
Services Office (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, N.C. 27711, or
from the National Technical Information Services, 5285 Port Royal Road, Springfield, Virginia 22161,
-------
TABLE OF CONTENTS
Section Page
1.0 INTRODUCTION 1-1
1.1 REFERENCES 1-20
2.0 DISCUSSION AND COMMENTS ON ONBOARD CONTROLS 2-1
2.1 ONBOARD CONTROL TECHNOLOGY 2-1
2.1.1 General Description of an Onboard System. . . 2-1
2.1.2 Recent Developments in Onboard Control. . . . 2-4
2.1.3 Description of the Onboard System Evaluated
by EPA 2-11
2.1.4 Summary and Analysis of Comments 2-18
2.1.b Safety Concerns 2-28
2.1.6 Canister Purge 2-34
2.1.7 Excess Evaporative Emissions 2-37
2.1.8 Test Procedure 2-38
2.1.9 Miscellaneous 2-39
2.2 EFFECTIVENESS OF ONBOARD CONTROLS 2-41
2.2.1 Onboard Controls for HDGV's 2-42
2.2.2 Emptying Losses 2-43
2.2.3 Gasoline Spillage 2-43
2.2.4 Onboard in Current Stage II Areas 2-45
2.2.5 Control System Tampering 2-46
2.2.6 Canister System Deterioration 2-48
2.2.7 Purge Effects on Efficiency 2-48
2.2.8 Overall Control System Efficiency 2-49
2.3 EMISSION FACTOR FOR REFUELING LOSSES 2-56
2.4 LEAD TIME 2-58
2.5 PHASE-IN OF CONTROLS 2-63
2.6 COST OF ONBOARD CONTROLS -. . 2-59
2.6.1 Onboard Control System Costs 2-72
2.6.2 Comparison of Cost Estimates 2-103
2.6.3 Manufacturer Overhead and Profit 2-1U5
2.6.4. Excess Evaporative Emissions 2-107
2.6.5 Heavy-Duty Gasoline Fueled Vehicles 2-109
-------
TABLE OF CONTENTS
(continued)
Section Page
2.7 FUEL CONSUMPTION BENEFITS OF ONBOAKU 2-114
2.7.1 Recovery Credits for Liyht-Uuty Vehicles. . . 2-114
2.7.2 Recovery Credits for Heavy-Duty Vehicles. . . 2-136
2.8 ENFORCEMENT REQUIREMENTS 2-143
2.9 REFERENCES 2-150
3.0 STAGE II CONTROLS 3-1
3.1 STAGE II TECHNOLOGY 3-1
3.2 STAGE II DESIGN AND SAFETY CONSIDERATIONS 3-2
3.3 STAGE II CONTROL EFFICIENCY 3-4
3.4 COSTS OF STAGE II 3-6
3.5 COST EFFECTIVENESS OF STAGE II 3-17
3.6 STAGE II ECONOMIC IMPACTS 3-19
3.7 STAGE II MAINTENANCE . 3-28
3.8 ENFORCEMENT OF STAGE II REQUIREMENTS 3-30
3.9 SCOPE/COVERAGE OF STAGE II REQUIREMENTS 3-35
3.10 CONSUMER REACTION TO STAGE II 3-37
3.11 REFERENCES 3-41
4.0 TRADEOFFS BETWEEN STAGE II AND ONBOARD 4-1
4.1 GENERAL ISSUES 4-1
4.2 FACILITY EXEMPTIONS FROM STAGE II 4-3
4.3 IN-USE CONTROL EFFICIENCIES . 4-6
4.4 REFERENCES 4-8
5.0 EPA's 1984 CONTROL STRATEGY EVALUATION . . b-1
5.1 GENERAL METHODOLOGY b-1
5.2 EMISSION ESTIMATES b-12
5.3 ENERGY IMPACT ANALYSIS b-13
5.4 GASOLINE CONSUMPTION PROJECTIONS b-14
5.5 SIZES'AND DISTRIBUTION OF FACILITIES 5-16
5.6 REFERENCES 5-19
6.0 REASONABLENESS OF CONTROL COSTS VERSUS HEALTH RISK
REDUCTION 6-1
6.1 NEED FOR STANDARDS 6-1
-------
TABLE OF CONTENTS
(concluded)
Section Page
6.2 COST/BENEFITS OF CONTROLS . .' ...6-8
6.3 ALTERNATIVE CONTROLS 6-9 .
6.4 REACTION TO GASOLINE PRICE INCREASE (STAGE II). . . . 6-11
6.5 REACTION TO INCREASE IN PRICE OF A NEW
VEHICLE (Onboard) 6-11
6.6 REFERENCES 6-14
7.0 STAGE I CONTROLS 7-1
c.O EFFECTS ON STATE IMPLEMENTATION PLANS (SIP's) 8-1
8.1 OZONE NAAQS ATTAINMENT DEADLINE 8-1
8.2 EMISSION CREDITS 8-4
8.3 EFFECT OF STATE ADOPTION OF CONTROLS 8-7
8.4 EPA'S ROLE IN SELECTING CONTROLS 8-9
8.5 MISCELLANEOUS 8-10
9.0 EXPOSURE/RISK ANALYSIS 9-1
9.1 UNIT RISK FACTORS 9-1
9.2 RISK ASSESSMENT METHODOLOGY 9-15.
9.2.1 General 9-15
9.2.2 Exposure Measurements 9-18
9.2.3 Incidence 9-18
9.2.4 Lifetime Risk 9-23
9.3 EXPOSURES DURING SELF-SERVICE REFUELING . 9-24
9.3".r Gasoline Pumping Rate 9-24
9.3.2 Number of Tank Fillings 9-25
9.3.3 Exposure Concentrations 9-25
9.4 REFERENCES 9-29
10.0 OTHER METHODOLOGIES AND CONSIDERATIONS 10-1
10.1 LEGAL AND POLICY CONSIDERATIONS 10-1
10.2 SUGGESTED REGULATORY APPROACHES 10-12
10.3 REFERENCES 10-21
-------
LIST OF TABLES
Table Title Page
1-1 List of Commenters on Evaluation of Regulatory
Strategies for the Gasoline Marketing Industry . . . 1-2
2-1 Calculation of Fuel Tank Capacities for Various
Years .- 2-23
2-2 In-Use EF Test Program M&D Types, Rates of
Occurrence, and Diurnal/Hot Soak Emissions for
Fuel-Injected Vehicles 2-55
2-3 Phase-In of Onboard Refueling Controls 2-66
2-4 Comparison of Phase-In Control Effectiveness .... 2-68
2-5 Onboard Control System Costs 2-73
2-6 Calculation of Carbon Bed Volume and Cost 2-78
2-7 Calculation of Canister Shell Costs 2-81
2-8 Summary of Canister Shell Costs 2-83
2-9 Vapor Line Cost Estimates 2-88
2-10 Calculation of Fillpipe Extension Cost 2-95
2-11 Comparison of Cost Estimates 2-104
2-12 Onboard Costs - HDGV's 2-111
2-13 Computation of Onboard Component Weights 2-114
2-14 Liquid Composition of Several Gasoline Samples
(Volume Percent) 2-126
2-15 Vapor Composition from Several Gasoline Samples
(Weight Percent) 2-128
2-16 Hydrocarbon Properties 2-129
2-17 LDV Mileage Accumulation 2-131
2-18 LOT Mileage Accumulation 2-132
2-19 Calculation of NPV of Recovery Credits - LDV, 1989. . 2-1-34-
2-20 Calculation of NPV of Recovery Credits - LDT, 1989. . 2-135
2-21 Fuel Economy Projections for HDGV's (mpg) 2-137
-------
LIST OF TABLES
(concluded)
Table Title Pa ye
2-22 HDUV Control System Weiyht Estimates (Ibs) 2-139
2-23 Sample Calculation of Weight Penalties 2-14U
2-2$ Weight Penalty Estimates (gallons/lifetime) (HDGV's). 2-141
2-25 Gross Refueling Recovery Credits for HUGV's (gallons) 2-144.
2-26 Net Fuel Consumption Credit for HDGV's (gallons). . . 2-145
2-27 Sample Calculation of Discounted Recovery Credits ($)
(All HDGV's, 1989) 2-146
2-28 Net Discounted Recovery Credits ($) (All HDGV's). . . 2-147
3-1 Weighted Average Stage II Costs 3-8
3-2 Stage II Revised Capital Cost Estimates for a "Typical"
35,DUO Gallon/Month Service Station 3-9
5-1 Stage II Average Total Cost of Control by
Model Plant 5-9
vn
-------
LIST OF FIGURES
Figure Title
2-1 Onboard System Mechanical Seal 2-2
2-2 In-Tube Trap Liquid Seal 2-7
2-3 Submerged Fill Liquid Seal 2-8
2-4 J-Tube Liquid Seal 2-8
2-b Integrated Evaporative/Refueling System 2-13
2-6 Vent Valve Activation Switch 2-14
2-7 Vapor Vent Valve 2-14
2-8 Possible Fill Limiter Design 2-16
2-9 Nozzle-Actuated Refueling Emissions Vapor Vent Valve.. 2-32
2-10 Onboard Leadtime 2-62
2-11 Float Valve 2-y3
-------
l.U INTRODUCTION
On August 8, 1984, EPA published in the Federal Register (49 FR
317U6) notice of the availability of a document evaluating regulatory
strategies being considered for controlling air pollutant emissions
from the gasoline marketing industry (I-A-5b). Public comments were
solicited and over 180 individual comment letters were received. The
purpose of this document is to respond to all of the major comments
made by the public. Due to the volume of comments received, EPA could
not respond to each comment letter or comment on an individual basis.
Thus, to facilitate responding, the comments have been summarized and
combined together under the following subject categories:
Chapter 2 - Onboard Controls
Chapter 3 - Stage II Controls
Chapter 4 - Tradeoffs Between Stage II and Onboard
Chapter 5 - EPA's 1984 Control Strategy Evaluation
Chapter 6 - Reasonableness of Control Costs Versus Health Risk Reduction
Chapter 7 - Stage I Controls
Chapter 8 - Effects on State Implementation Plans
Chapter 9 - Exposure/Risk Analysis
Chapter 10 - Other Methodologies and Considerations
Some of these categories reflect topic areas upon which the Agency
specifically requested comment in the 1984 Federal Register Notice;
these specific categories are denoted by an asterisk on the section
title and a footnote saying "1984 Federal Register topic". The list of
public commenters on the 1984 Federal Register notice and analysis
document and the EPA docket item number assigned to each comment
submittal are shown in Table 1-1.
Changes made to EPA's 1984 analysis in response to public comments,
additional analyses performed, and a summary of the results are contained
in a separate two-volume draft Regulatory Impact Analysis document.1
1-1
-------
Table 1-1. LIST OF COMMENTERS ON EVAL'IATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARK.TING INDUSTRY
Item Number in Date of
Docket A-84-07 Correspondence
I-H-ia
8-21-84
I-H-2
I-H-3
8-16-84
8-16-84
I-H-4
8-24-84
I-H-5
9-13-84
I-H-6
5-8-83
I-H-7
I-H-8
I-H-9
3-13-84
2-28-84
9-24-84
1130
Commenter and Affiliation
Mr. Roger F. Dreyer
Executive Vice President
Ohio Petroleum Marketers Assoc.
Inc.
50 West. Broad Street, Suite
Columbus, Ohio 43215
Ms. Jutta Schildknecht
BMW of North America, Inc-.
Montvale, New Jersey 07645
Mr. Raymond J. Grubbe
Senior Engineer
AT&T Teletype Corporation
5555 Touay Avenue
Skokie, Illinois 60077
Mr. Warren Cohen, President
American Car Wash Corp.
7333 Little River Turnpike
Annandale, VA 22003
Mr. Walter R. Quanstrom
General Manager
Standard Oil Company
200 East Randolph Drive
Chicago, Illinois 60601
(Extension requested)
Mr. Earl Harris Matheny
Box 195
R.R. #1
Stanford, KY 40484
Viola Amos
614 Riverside Drive
Holly Hill , FL 32017
Mr. Wilfred Szerenyi
15700 Olive Branch Drive
La Mirada, CA 90638
Mr. Milton Feldstein
Bay Area Air Quality Management
District
939 Ellis Street ' '
San Francisco, CA 94109
(Extension requested)
aAppendix A to this letter contains letters from 32 Ohio petroleum
marketers, designated in this document as items I-H-1A1 through I-H-1A32,
1-2
-------
Table 1-1. LIST
STRATEGIES
OF COMMENTERS N EVALUATION OF REGULATORY
FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence
I-H-10
9-25-84
I-H-11
9-26-84
I-H-12
10-1-84
I-H-13
10-1-84
I-H-14
10-1-84
I-H-15
10-2-84
I-H-16
10-1-84
I-H-17
10-2-84
I-H-18
9-28-84
Commenter and Affiliation
Mr. James D. Boyd
Air Resources Board
1102 Q Street
P.O. Box 2815
Sacramento, CA 9b812
(Extension Requested)
Mr. Jim Stokes, President
Greater Washington/Maryland
Service Station Association
9200 Edmonston Road, Suite 304
Greenbelt, MO 20770
J.C. Emmart
Emmart Oil Co.
P.O. Box 2247
Winchester, VA 22601
Robert J. Cutler, President
Energy Retailers, Inc.
P.O. Box 151
Ringham, MA 02043
Mr. W. G. Lyden, Jr.
Lyden Oil Company
P.O. Box 1854
Youngstown, OH 44501
Mr. William R. Deutsch
Illinois Petroleum Marketers
Association
P.O. Box 1508
Springfield, IL 62705
Mr. Cliff Brice
Cliff Brice Stations, Inc.
300 Moffat Avenue
Pueblo, CO 81003
Mr. C.D. Bolton
Bolton Oil Company
P.O. Box 397
Artesia, NM 88210
Mr. Wayne Binsted
Binsted's Exxon Service Center
4812 MacArthur Blvd., N.W.
Washington, D.C. 20007
1-3
-------
Table 1-1. LIST OF CO IENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR Tht GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence
I-H-19
10-1-84
I-H-20
10-5-84
I-H-21
10-8-84
I-H-22
10-4-84
I-H-23
10-8-84
I-H-24
10-8-84
I-H-25
10-4-84
I-H-26
10-2-84
Commenter and Affiliation
Mr. C.A. Fink
BAR-F Enterprises, Inc.
P.O. Box 129
Farmington, NM 87499
Mr. R.G. Roop
Petroleum Marketers, Inc.
1603 Santa Rosa Road
Richmond, VA 23288
Mr. George P. Ferreri, Director
Air Management Administration
Dept. of Health & Mental Hygiene
201 West Preston Street
Baltimore, MD 21201
Mr. Kenichi Chiku
Executive Vice President
Toyota Technical Center, U.S.A.,
Inc.
Ann Arbor Branch
Ann Arbor, IL 48105
Mr. Michael J. Dougherty
Manager, Environmental Control
Union Oil Co. of California
Box 7600
Los Angeles, CA 90054
Mr. Peter W. McCallum
Corporate Environmental
Specialist
The Standard Oil Company
Midland Building
Cleveland, OH 44115
Mr. Mike Hawkins,
Hawk Oil Company
P.O. Box
1050 So.
President
Medford,
1388
Riverside
OR 97501
Mr. Dave Fellers, CAE
Executive Vice President
Texas Oil Marketers Association
701 W. 15th Street
Austin, TX 78701
1-4
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIEr FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence Commenter and Affiliation
I-H-27 10-4-84 Mr. Doug Burton, Sales Manager
McGill Incorporated
P.O. Box 9667
Tulsa, OK 74157-96667
I-H-28 10-4-84 Mr. Braxton Hablof
2507 Central Avenue
Alexandria, VA 22302
I-H-29 10-4-84 Mr. Stephen J. Powers, Jr.
President
Maine Oil Dealers Assoc.
P.O. Box 536
Yarmouth, ME 04096
I-H-30 10-5-84 Mr. Herman L. Brummett
Apollo Oil Company
1200 W. Pioneer Parkway
Peoria, IL 61615
I-H-31 10-4-84 Mr. W. H. Hartley, President
. The Hartley Company
319 Wheeling Avenue
Cambridge, Ohio 43725
I-H-32 10-3-84 Mr. F.W. Englefield
Chairman of the Board
Englefield Oil Company
447 James Parkway
Newark, OH 43055
I-H-33 10-3-84 Mr. Don M. Ward
Executive Vice President
North Carolina Oil Jobbers
Assoc.
- - P.O. Box 30519
Raleigh, NC 27622
I-H-34 10-8-84 Mr. Gerald E. Wagner, President
Convenient Remote Services
P.O. Box 35580
Louisville, KY 40232
I-H-35 10-8-84 Mr. A. Tab Williams, Jr., President
A.T. Williams Oil Company
P.O. Box 7287
Winston Salem, NC 27109
I-H-36 10-6-84 Mr. Myron T. Holman, President
Gas Pumpers of America Corp.
Number One Valley Street
Corner Braen Avenue & valley St.
Hawthorne, NJ 07506
1-5
-------
Table M. LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence Commenter and Aff11iation
I-H-37 10-8-84 Mr. Thomas N. Allen, President
East Coast Oil Corporation
1420 East Commerce Road
Richmond, VA 23224
I-H-38 10-5-84 Mr. Gus Fleischli
Fleischli Oil Company
P.O. Box 487
Cheyenne, WY 82003
I-H-39 1U-8-84 Mr. Tendle L. Jones, Ptr.
Jones Oil Company
402 East First Street
Dehli, LA 71232
I-H-40 10-5-84 Mr. Francis C. Haviland
Executive Di rector
Fuels Merchants Association
of New Jersey
Gasoline Jobbers Division
P.O. Box 359
Springfield, NJ 07081
I-H-41 10-5-84 Mr. Michael Kirschner
Kirschner Bros. Oil Company
Marketers of Petroleum Products
569 W. Lancaster Avenue
Haverford, PA 19041
I-H-42 10-9-84 Mr. R.E. Germer, Vice President
Flying, Inc.
P.O. Box 678
Brigham City, UT 84302
I-H-43 10-9-84 Mr. David C. Waddell, Director
Supply and Transportation
Farmers Union Central Exchange,
Inc. (CENEX)
P.O. Box 43089
St. Paul, MN 55164
I-H-44 10-8-84 Mr. Don W. Myers, Sr.
Chairman of The Board
Swifty Oil Company, Inc.
P.O. Box 1002
Seymour, IN 47274
1-6
-------
''able 1-1, LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in
Docket A-84-07
I-H-45
I-H-47
I-H-48
I-H-49
I-H-50
I-H-51
I-H-52
I-H-53
Date of
Correspondence
10-5-84
10-8-84
Undated
10-4-84
10-7-84
10-5-84
10-10-84
10-8-84
10-8-84
Commenter and Affi1i ati on
Joseph R. Saxon, President
and Chief Executive Officer
Crystal - U.S.A. Oil, Inc.
P.O. Box 9128
Birmingham, AL 35213
Mr. Robert T. Welsh, Jr.
Welsh Oil, Inc.
P.O. Box 10725
Merrillvil le, IN 46411
Mr. David Maybelef
Mango Distributing Company
P.O. Box 69
Barnhart, MO 63012
Mr. Arthur Goldstein
Petrol Plus of Naugatuck, Inc.
P.O. Box 492
Derby, CT 06418
Mr. Richard M.L. Oxterman
Director of Development
Lockie Lee Services, Inc.
310 Chester Street
Painesville, OH 44077
Mr. J. Terry Ross
Musket Corporation
P.O. Box 26210
Oklahoma City, OK 73126
Mr. Thomas D. Hughes, President
Pride Petroleum Company, Inc.
P.O. Box 955
St. Charles, IL 60174
Mr. Reggie Dupree, President
Cajun Energy, Inc.
P.O. Box 878
Opelousas, LA 70570
Mr. Wataru Hayashibara, Manager
Certification Business Division
MAZDA (North America), Inc.
24402 Sincola Court
Farmington Hills, MI 48U18
1-7
-------
Table 1-1. LIST OF COMMENTED ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence Commenter and Affiliation
I-H-54 10-8-84 Mr. Roy A. Turner
Executive Vice President
Colorado - Wyoming - New Mexico
Petroleum Marketers Assoc.
4465 Kipling, Suite 104
Wheat Ridge, CO 80033
I-H-55 10-8-84 Mr. Maury S. Guttman
Executive Vice President
Guttman Oil Company
Speers Road
Belle Vernon, PA 15012
I-H-56 10-9-84 Mr. Shack Wimbish, Jr.
Colonial Oil Industries, Inc.
North Lathrop Avenue
Savannah, GA 31402
I-H-57 10-9-84 Mr. William T. Burkhart
Regional Air Pollution Control
- - Agency
451 W. Third Street
P.O. Box 972
Dayton, OH 45422
I-H-58 10-8-84 Mr. J.A. Stuart
Executive Officer
South Coast Management District
915U Flair Drive
El Monte, CA 91731
I-H-59 10-5-84 Mr. Robert L. French
Executive Vice President
California Target Enterprises,
Inc.
12739 Lakewood Blvd.
Downey, CA 90242
I-H-60 10-5-84 Mr. William Reilly
City of Philadelphia
Department of Public Health
500 S. Broad Street
Philadelphia, PA 19146
I-H-61 10-9-84 Mr. W. Carey Johnson
EZ-Go Foods, Inc.
P.O. Box 286
Lawton, OK
1-8
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
DocketA-84-07 Correspondence Commenter and Affiliation
I-H-62 10-12-84 . Mr, Jimmy Harrell
Inland Southwest Georgia Oil
Company, Inc.
'1711 E. Showell
P.O. Box 1510
Bainbridge, GA 31717
I-H-63 10-10-84 Mr. James A. Haslam, II
President
Pilot Oil Corporation
P.O. Box 10146
Knoxville, TN 37939-0146
I-H-64 10-9-84 Mr. Bill Hall
Independent Oil Men's Assoc.
of New England
25 Sea Breeze Lane
New Castle, NH 03854
I-H-65 10-9-84 Mr. James M. Lents, Director
Air Pollution Control Division
Colorado Department of Health
4210 E. llth Avenue
Denver, CO 80220
I-H-66 10-9-84 -Mr. R.S. Ivey, Treasurer
Apollo Oil Company
1200 W. Pioneer Parkway
peoria, IL 61615
I-H-67 10-16-84 Mr. W.B. Beaver
Executive Vice President
Southern Pumps & Tank Company
P.O. Box 31516
Charlotte, NC 28231
I-H-68 ( 10-12-84 Mr. Randy Castleberry
The Pantry, Inc.
P.O. Box 1410
Sanford, NC 27330
I-H-69 10-10-84 Mr. Bill Corning, President .
Winn Brothers, Inc.
P.O. Box 441
Weatherford, OK
I-H-70 10-7-84 Mr. Gil Arnold, President
Road Runner - Arnold Dist. Co.
South Robison Road
'P.O. Box 973
Texarkana, TX 75501
1-9
-------
Table 1-1. LIST OF CQMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence Commenter and Affiliation
I-H-71 10-15-84 Mr. L. Carl Adams
Executive Vice President
Florida Petroleum Marketers
Association
209 Office Plaza
Tallahassee, FL 32301
I-H-72 10-9-84 Mr. Barry W. Muller
State of Rhode Island and
Providence Plantation
Department of Environmental
Management
75 Davis St. - 204 Cannon Bldg,
Providence, RI 02908
I-H-73 9-21-84 Ms. Barbara Morin
State of Rhode Island and
Providence Plantation
Department of Environmental
Management
Providence, RI 02908
(Extension requested)
I-H-74 10-4-84 Mr. W.W. Hazlett
Hazlett Engineering Co.
1089 Indian Village Road
Pebble Beach, CA 93953
I-H-75 10-17-84 Mr. Samuel Chico, Jr.
Chico Dairy Company
331 Beechurst Avenue
Moryantown, WV 26505
I-H-76 10-17-84 Mr. Mike Sparkman, Vice
' . President
Modern Oil Co., Inc.
P.O. Box 218
Shawnee, OK 74801
I-H-77 10-18-84 Mr. Gerald J. Helfenkein
President - Treasurer
Marane Oil
501 Park Avenue
Worchester, MA 01610
I-H-78 10-17-84 Mr. R.G. Elmore
Executive Vice President
J&L Oil Company, Inc.
Box 214A
Mundelein, IL 60060
1-10
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence
I-H-79
10-22-84
I-H-80
I-H-81
10-22-84
10-23-84
I-H-82
10-23-84
I-H-83
10-19-84
I-H-84
10-24-84
I-H-85
10-22-84
I-H-86
10-29-84
Commenter and Affiliation
Mr. Erv Lackey, Chairman
Lackey, Distributing, Inc.
Chevron Petroleum Jobbers
5275 East 48th Avenue
Denver, CO 80216
Mr. Tom Slamans, Distributor
P.O. Box 603
Oktnulgee, OK 74447
Mr. Kenneth A. Baker
Vice President
J.D. Streett & Company, Inc.
144 Wei don Parkway
Maryland Heights, 110 63043
Mr. Milton Feldstei n
Air Pollution Control Officer
Bay Area Air Quality Management
District
939 Ellis Street
San Francisco, CA 94109
Mr. Phillip L. Youngblood
Director, Air Programs
Conoco, Inc.
P.O. Box 2197
Houston, TX 77252
Mr. Dave Fellers, CAE
Executive Vice President
Texas Oil Marketers Assoc.
701 W. 15th Street
Austin, TX 78701
Mr. Charles V. Stuckey, CAE
Executive Vice President
Oklahoma Oil Marketers Assoc.
5115 N. Western
Oklahoma City, OK 73118
Mr. J.J. van der Veken, Manager
Gas Stations Operations &
Maintenance
Northville Gasoline Corp.
P.O. Box 937
Melville, NY 11747
1-11
-------
Table 1-1.- LIST OF COMHENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in
Docket A-84-07
I-H-87
I-H-88
l-H-89
I-H-90
I-H-91
I-H-92
I-H-94
Date of
Correspondence
Undated
11-1-84
11-2-84
10-30-84
11-1-84
11-5-84
11-5-84
11-6-84
Commenter and A ff i 1i at i on
Mr. John S. Hough, President
Hough Fuel
340 Fourth Street
Trenton, NJ U8638
Mr. Carl C. Greer, President
Martin Oil Marketing Corp., Ltd.
P.O. Box 298
Blue Island, IL 60406
Mr. Thomas F. Wentworth
General Manager
Solar Oil Company, Inc.
P.O. Box 127
Hope, NJ U7844
Mr. Barnard R. McEntire
County of San Diego
Air Pollution Control District
9150 Chesapeake Drive
San Diego, CA 92123-1095
Mr. A.G. Smith
She!1 Oil Company
P.O. Box 4320
Houston, TX 7721U
Mr. Bill Stewart
Executive Director
Texas Air Control Board
6330 Highway, 29U East
Austin, TX 78723
Mr. Frank T. Ryan, Vice
President
Rubber Manufacturers Assoc.
1400 K. Street, N.W.
Washington, D.C. 20005
Mr. John W. Graves
Director of Environmental Safety
and Health Affairs
Pennzoil Company
P.O. Box 2967
Houston, TX 77252-2967
1-12
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in
Docket A-84-U7
I-H-95
I-H-96
I-H-97
I-H-98
I-H-99
I-H-100
I-H-101
I-H-102
Date of
Correspondence
11-5-84
Undated
Undated
11-7-84
11-5-84
11-7-84
11-5-84
11-8-84
Commenter and Affiliation
Mr. George P. Ferreri, Director
Air Management Administration
Maryland Dept. of Health & Mental
Hygiene
201 West Preston Street
Baltimore, MD 212U1
Mr. Robert L. Claibourne
Claibourne Oil Company
P.O. Box 787
Miami, OK 74354
Mr. Ray Reed
Bowling Oil Co., Inc.
P.O. Box 1282
Seminole, OK 74868
Mr. W.T. Danker
Manager, Environmental Programs
Chevron U.S.A., Inc.
P.O. Box 7643
San Francisco, CA 94120-7643
Mr. Paul D. Collier
Executive Vice President
Amoco Oil Company
2UO East Randolph Drive
Chicago, IL 60601
Mr. William Shapiro
Manager, Regulatory Affairs
Volvo - North American Operations
Rockleigh, NJ 07647
Chief Engineer
Fuel Economy
Mr. R.R. Love
Emissions and
Certification
Chrysler Corporation
P.O. Box 1118
Detroit, MI 48288
Ms. Barbara Faulkner
Vice President, Policy & Legal
Affairs
Petroleum Marketers Association
1707 H Street, N.W.
Suite 1100
Washington, D.C. 20006
1-13
-------
Table 1-1. LIST OF COMMENTED ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in
Docket A-84-07
I-ri-103
I-H-1U4
I-H-1U5
I-H-1U6
I-H-107
I-H-108
I-H-109
I-H-110
I-H-111
Date of
Correspondence
11-7-84
11-8-84
Undated
Undated
11-7-84
11-8-84
11-8-84
11-5-84
10-31-84
Commenter and Affiliation
Mr. S. William Becker
Executive Director
STAPPA/ALAPCO
444 North Capitol St., N.W.
Washington, D.C. 20001
Mr. Brian Gill, Manayer
Certification Department
American Honda Motor Co., Inc.
P.O. Box 50
Gardena, CA 9U247
Mr. Bedford M. Mitchell
Vice President
Carey Johnson Oil Co., Inc.
P.O. Box 286
Lawton, OK 73502
Mr. R.J. O'Halloran
O'Halloran Oil Corp.
Perry, OK
Mr. Glenn E. Moore
Vice President - Engineering
Dover Corp. OPW Division
P.O. Box 405003
Cincinnati, OH 45240-5003
Mr. Jeffrey L. Leiter
Collier, Shannon, Rill & Scott
1055 Thomas Jefferson Street, N.W.
Washington, D.C. 20007
{representing SIGMA)
Mr. J.C. Hi!drew, Manatje.*
Environmental Affairs
Mobil Oil Corporation
P.O. Box 1031
Princeton, NJ 08540
Mr. William G. Maxwel1
Red Rock/Pate Oil Distributing Co.
P.O. Box 82336
Oklahoma City, OK 73148
Mr. Detlev E. Hasselmann
Hasstech, Inc.
8821 Production Avenue
San Diego, CA 92121
1-14
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATION OF REG'LATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence . Commenter and Affiliation
I-H-112 11-7-84 Mr. S. Perkins, President
Perkins Petroleum, Inc.
P.O. Box 1078
Guymon, OK 73942
I-H-113 11-6-84 Mr. U.V. Henderson
Texaco, Inc.
P.O. Box 509
Beacon, NY 12b08
I-H-114 11-8-84 Mr. Donald R. Buist, Director
Automotive Emissions and Fuel
Economy Office
Ford Motor Company
Tha American Road
Dearborn, MI 48121
I-H-115 11-9-84 Mr. David D. Doniger
Natural Resources Defense
Council, Inc.
1350 New York Avenue, N.W.
Suite 300
Washington, D.C. 20005
I-H-116 11-8-84 Mr. W. Groth, Manager
Emissions Regulations &
Certifications
Volkswagen of America, Inc.
888 W. Big Beaver
P.O. Box 3951
Troy, MI 48007-3951
I-H-117 11-8-84 Mr. T.M. Fisher, Director
Automotive Emission Control
General Motors Ccrporation
General Motors Technical Center
Warren, MI 48090 "
I-H-118 11-13-84 Mr. James 0. Boyd
Executive Officer
California Air Resources Board
1102 Q Street
Sacramento, CA 95812
I-H-119 11-8-84 Mr. Jeffrey L. Leiter
Collier, Shannon, Rill, & Scott
1055 Thomas Jefferson Street, N.W,
Washington, D.C. 20007
(representing National Assoc. of
Convenience Stores)
1-15
-------
Table 1-1. LIST OF COMMENTERS ON EVALUATIO" OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in
Docket A-84-U7
I-H-120
I-H-121
I-H-122
I-H-123
I-H-124
I-H-125
I-H-126
I-H-127
Date of
Correspondence
11-8-84
Undated
11-14-84
11-5-84
11-5-84
11-8-84
11-19-84
11-8-84
Commenter and Affiliation
Mr. William F. O'Keefe
Vice President
American Petroleum Institute
122U L Street, NW
Washington, O.C. 20005
Mr. R.A. Davis, President
Denver Oil Company
P.O. Box 94597
Oklahoma City, OK 73143
Mr. James K. Hambright, Director
Bureau of Air Quality Control
Commonwealth of Pennsylvania
P.O. Box 2063
Harrisburg, PA 17120
Mr. R. Manning
Manning Oil Company
P.O. Box 576
Pauls Valley, OK 73075
Mr. Harry H. Hovey, Jr.
Director, Division of Air
New York State Department of
Environmental Conservation
50 Wolf Road
Albany, NY 12233-U001
Mr. Raymond J. Campion
Coordinator
Public Affairs Dept.,
Environmental Conservation
Exxon Company, U.S.A.
P.O. Box 2180
Houston, TX 77001
Ms. Jan W. Mares, Asst. Secretary
Policy, Safety and Environment
Department of Energy
Washington, D.C. 20585
Mr. Fred W. Bowditch, Vice
President
Technical Affairs
Motor Vehicle Manufacturers
Assoc.
300 New Center Building
Detroit, MI 48202
1-16
-------
Table 1-1. LIST OF CQMMERTERS ON VALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-84-07 Correspondence Commen t er and Affiliation
I-H-128 11-16-84 Mr. W.C. Jones, Manager
Governmental Regulations
American Motors Corporation
14250 Plymouth Road
Detroit, MI 48232
I-H-129 10-25-84 Mr. Robert D. Bradt
Hirt Combustion Engineers
931 South Maple Avenue
Montebello, CA 90640-5488
I-H-130 11-6-84 Mr. Vic Rasheed, Executive
Director
Service Station Dealers of
America
4UO North Capitol Street, N.W.
Suite 175
Washington, D.C. 20001
I-H-131 11-7-84 Mr. Ron M. Clark
President & General Manager
Emco Wheaton, Inc.
P.O. Box 688
Conneaut, OH 44030-U688
I-H-132 11-7-84 Mr. James W. Healy
Cambridge Engineering, Inc.
74 Faulkner Street
North Billerica, MA 01862
I-H-133 12-18-84 Mr. Fred W. Bowditch, Vice
President, Technical Affairs
Motor Vehicle Manufacturers
- - Assoc.
300 New Center 81 dg.
Detroit, MI 482U2
I-H-134 11-6-84 Ms. Cynthia Winklevoss
Slippery Rock University
Slippery Rock, PA 16057
I-H-135 1-10-85 Mr. T.M. Fisher, Director
Automotive Emission Control
General Motors Corp.
Warren, MI 48090
I-H-136 2-22-85 Mr. T.M. Fisher
General Motors Corporation
Warren, MI 48090
1-17
-------
Table 1-1. LIST OF COMMENT-IRS ON EVALUATION OF REGULATORY
STRATEGIES FOR THE GASOLINE MARKETING INDUSTRY
(continued)
Item Number in Date of
Docket A-B4-U7 Correspondence . Commenter and Affiliation
I-H-137 4-25-86 Mr. D.E. Hoag
Plymouth Oil Company
P.O. Box 27H7
Detroit, MI 48227
I-H-138 4-24-86 Mr. J. Edward Surette, Jr.
Executive Director
Bay State Gasoline Retailers
Association
574 Boston Road (Rt. 3A)
Billerica, MA 01821
i-0-51 3-25-85 Mr. H.G. Grayson
Mobil Oil Corporation
150 East 42nd Street
New York, NY 10017
I-D-52 5-16-85 Mr. B.E. Doll, Manager
Air Programs
Mobil Oil Corporation .
P.O. Box 1031
Princeton, NJ 08540
1-0-53 6-24-85 Mr. John M. Daniel, Jr.
Asst. Executive Director
Air Pollution Control Board,
Commonwealth of Virginia
P.O. Box 10089
Richmond, VA 23240
I-D-54 7-23-85 Mr. Charles J. DiBona, President
American Petroleum Institute
Washington, D.C. 20005
(signed by four other trade
representatives)
I-D-55 7-25-85 Mr. J.C. Hildrew, Manager
Environmental Affairs
Mobil Oil Corporation
Princeton, NJ 08540
I-0-56 8-6-85 Mr. C.L. Terlizzi, Manager
National Accounts
The B.F. Goodrich Company
500 South Main Street
Akron, OH 44318
1-18
-------
Table 1-1. LIST 01 COMMENTERS ON EVALUATION OF REGULATORY
STRATEGIES FO* THE GASOLINE MARKETING INDUSTRY
(concluded)
Item Number in Date of
Docket A-84-07 Correspondence
I-D-57, I-D-68
7-23-85
I-D-58
8-23-85
I-D-59
10-23-85
I-D-63
10-16-85
I-D-64
12-17-85
I-D-65
4-19-85
I-D-67
7-19-85
I-D-70
9-30-85
Commenter and Affiliation
Mr. Mark Cooper
Energy Di rector
Consumer Federation of America
1424 16th Street, N.W.
Washington, D.C. 20036
Mr. Robert A. Rogers, Director
Automotive Emission Control
General Motors Corporation
Warren, MI 48090
Mr. Robert A. Rogers, Director
Automotive Emission Control
General Motors Corporation
Warren, MI 48090
Mr. Howard H. Kehrl
Vice Chai rman
General Motors Corporation
3044 West Grand Boulevard
Detroit, MI 48202
Mr. J.J. Wise, Manager
Paulsboro Research Lab
Mobil R&D Corporation
Paulsboro, NJ 08066
Mr. Robert E. Hughey
Commissioner
Dept. of Environmental Protection
State of New Jersey
Trenton, NJ 08625
Mr. Phillip R. Chisolm
Executive Vice President
Petroleum Marketers Association
of America
1120 Vermont Avenue
Washington, D.C. 20005
Mr. James E. Benton
Executive Di rector
New Jersey Petroleum Council
170 West State Street
Trenton, NJ 08608
1-19
-------
All references used in this document are contained in EPA's Gasoline
Marketing Docket (No. A-84-07). Each document in this docket is assigned
a docket item number. These same docket item numbers are used as
reference numbers in this document. The docket is available for public
inspection and coping at EPA's Central Docket Section, West Tower Lobby,
Gallery 1, Waterside Mall, 401 M Street S.W., Washington, D.C. 20460
(phone number 202-382-7549). A reasonable fee may be charged for
copying.
1.1 REFERENCES
I-A-55 Evaluation of Air Pollution Regulatory Strategies for Gasoline
Marketing Industry, U.S. Environmental Protection Agency, Office
of Air and Radiation, Office of Air Quality Planning and Stan-
dards, and Office of Mobile Sources, EPA-450/3-84-012a (Executive
Summary - EPA-450/3-84-012b), July 1984, [NTIS # PB 84 231075
and PB 84 231083, respectively].
1*. Draft Regulatory Impact Analysis, Proposed Refueling Emission
Regulations for Gasoline-Fueled Vehicles -- Volume I - Analysis
of Gasoline Marketing Regulatory Strategies (EPA-450/3-87-001a),
Volume II - Additional Analysis of Onboard Controls (EPA-450/3-87-
OOlb), U.S. Environmental Protection Agency, Office of Air and
Radiation, Office of Air Quality Planning and Standards, and Office
of Mobile Sources, July 1987.
*Docket number not available. These documents will be published
at the same time as this report and will be assigned docket
numbers and placed in the docket at that time.
1-20
-------
2.0 DISCUSSION ANU COMMENTS ON ONBOARD CONlKOLS
2.1 ONBOARD CONTROL TECHNOLOGY*
The technological aspects of onboard control of refueliny emissions
were discussed in an EPA technical report that was included as Appendix C
of the July 1984 EPA analysis document (I-A-bb). Since that report was
issued, EPA has received numerous public comments on all aspects of
onboard controls and has re-evaluated many of its concepts and estimates.
The results of the Ayency's re-evaluation is presented in this chapter.
The first part of this chapter describes a yeneral onboard control
system. Then several desiyn considerations and recent onboard develop-
ment work are discussed. This is followed by the description of the
specific onboard system that EPA is consideriny in this analysis.
Costs and other aspects of this system are thorouyhly discussed.
Finally, Ayency responses to siynificant comments on onboard issues are
presented.
2.1.1 General Description of an Onboard System
In many ways a refueliny vapor control system is similar to the
evaporative emission control systems now in use on most automobiles.
Vapors that are displaced from the fuel tank during a refueliny event
are adsorbed onto a bed of activated carbon, where they are stored.
Uuriny vehicle operation, manifold vacuum is used to pull ambient air
over the carbon bed, desorbiny the stored hydrocarbons from the canister.
The hydrocarbon-rich purye yas is then routed to the enyine and the
hydrocarbons are burned in the enyine duriny combustion.
In some respects, however, the onboard refueliny vapor control
system differs from an evaporative control system. The biyyest physical
difference between the systems is caused by the need to prevent vapors
from escapiny via the fillneck duriny a refueliny event. This need
forces the introduction of some type of sealiny mechanism into the fuel
tank fillneck. Althouyh there are a variety of fillneck seals capable
of performiny the necessary task, the one discussed in the July 1984
analysis is a simple mechanical seal (such as that shown in Fiyure 2-1)
'1984 Federal Reyister topic.
2-1
-------
a.
sealed
sea?
w
Sea?
2-2
-------
located near the top of the fillneck. When the gasoline dispensing
nozzle is inserted into the fillneck, an interface is formed between the
seal and the fillneck, preventing displaced vapors from escaping into the
atmosphere. Instead of leaving the fuel tank via the fillneck, the dis-
placed vapors are routed to the onboard canister where adsorption occurs.
Another difference between the evaporative and refueling vapor
control systems is the frequency and magnitude of canister loadings.
Evaporative emissions are produced each time a vehicle is operated, as
well as in response to diurnal temperature cycles. This could mean
several evaporative loads per day, even if the vehicle were operated
only a few miles at a time. Refueling emissions, on the other hand,
are generally produced less frequently and can be much larger in magni-
tude on a per-event basis. A significant amount of mileage is generally
accumulated between refueling events. Because of the potential for a
large quantity of refueling emissions at any one time, more hydrocarbon
storage capacity would be needed for refueling emissions than is needed
for evaporative emissions.
This difference in the timing and magnitude of emission loads
results in the need for purge schedules different from those used
in current evaporative systems. The need to accommodate high vapor
flow rates and large hydrocarbon loads has several implications for
the requirements of the onboard system. First, a larger hydrocarbon
storage canister is needed to collect and store the refueling emissions.
Depending on cost and vehicle design considerations, a manufacturer
could choose to use an integrated, partially integrated, or separate
approach to the control of evaporative and refueling emissions. The
"separate" approach would involve two (or more) canister systems, each
of which would be loaded.with evaporative or refueling emissions.
A "partially integrated" system would route refueling and some evaporative
emissions to one canister, but only evaporative emissions to a separate
canister. An "integrated" system would use a single canister (or canisters)
into which both evaporative and refueling emissions would be loaded.
By using an integrated or even a partially integrated approach, a
manufacturer cou-ld limit the total amount of activated carbon needed . .
for both refueling and evaporative emission control, and reduce the
overall vapor handling system requirement. However, in either case,
additional storage capacity would be necessary.
2-3
-------
In addition to larger canisters, larger fuel tank vent lines would
also be needed. Refueling vapors typically leave the fuel tank at
rates of eight gallons per minute or more. In order to accommodate
these high vapor flow rates without an excessive rise in fuel tank
backpressure, a vapor line with an inside diameter of about b/8 inch
would probably have to be used. Since current evaporative vent lines
are smaller than this, a larger vent line would have to be added.
The need to provide a large, unrestricted vent line between the
fuel tank and canister during refueling could also lead to a potential
safety hazard that is not a serious problem for evaporative emission
control systems. Due to the large vent line needed between the
fuel tank(s) and canister(s), there is a potential for a significant
amount of fuel spillage in the event of an accident involving a vehicle
rollover. Current evaporative emission control systems utilize a
limiting orifice (40 to 6l) thousandths of an inch in diameter) in the
vent line to limit the flow of fuel from the tank in rollover situa-
tions. To prevent major fuel spills following vehicle rollover,, a
mechanism would have to be included as part of an onboard system which
would provide vent line closure in rollover situations. This mechanism
could provide vent line closure at all times other than refuelings, or
could be designed to close in response to rollover. Regardless of the
specific component design, some type of rollover protection would have
to be included in onboard control systems. This subject is discussed
further in Section 2.1.5.
Although there are other differences between evaporative and
refueling/evaporative emission control systems, the major distinctions
have been highlighted above. Details of the functioning of onboard
systems will be provided below. The next section of this chapter
examines recent developments in onboard control technology. Specific-
ally, these developments involve the method of sealing the fillneck
during refueling.
2.1.2 Recent Developments in Onboard Control
In the onboard control system described in Appendix C of the July
19B4 EPA analysis, the fillneck is sealed during refueling by means of
a mechanical elastomeric device (see Figure 2-1). This type of seal
was shown to control refueling emissions with a theoretical efficiency
2-4
-------
of greater than 98 percent in a demonstration program done by the
American Petroleum Institute (API) in iy?8 (I-F-17). Although the
demonstration program did prove the theoretical efficiency of this type
of seal to be adequate, the comments on the information in Appendix C
suggested a number of potential problems with mechanically sealed
systems.
A number of commenters voiced their doubts about the durability of
mechanical seals. Although the API demonstration vehicles were driven
up to 65.UUO miles, none was driven the light-duty vehicle (LDV) average
lifetime of 1UU,OUU miles, and a number of commenters questioned the
ability of the seal to retain its integrity over the life of the vehicle
in a wide range of environmental conditions.
Many commenters also stated that a pressure relief device would
be needed if the mechanical type fill pipe seal were used. This would
be needed to prevent damage to the fuel tank and other system compo-
nents if the automatic nozzle shutoff mechanism failed or if the
vapor line between the fuel tank and an onboard refueling canister
became blocked. In these situations, the fuel tank could come under
excessive pressure as more fuel was dispensed into the tank. Also, if
the fuel tank were slightly overpressured as described above, a pres-
sure relief device would also help to reduce fuel spit-back which
could occur when the nozzle was removed from the fillpipe. Although
an adequate pressure relief device could be designed without an unreason-
able amount of effort, the commenters argued that such a device would
add an additional level of cost and complexity to the system.
A number of commenters felt that adding a mechanical seal to the
fillneck would lead to tampering or abuse by the consumer, since the
nozzle would have to be inserted into the fillneck with a certain
degree of care. Although this would not require an unacceptable level
of additional effort, they felt it would involve a deviation from the
traditional refueling experience which could ultimately lead to fillpipe
tampering. Any such tampering would likely lead to a decrease in the
effectiveness of'the control strategy.
These purported problems (even if true) would not prohibit the use
of a mechanical seal, but they do highlight the need for improvements
in these devices. A possible alternative to the mechanical seal is a
2-5
-------
"liquid seal" device. These liquid seals employ modified fillneck
desiyns which route the incominy yasoline in such a way that a column
of yasoline prevents the fuel tanK vapors from escapiny to the atmosphere
duriny refueling. Liquid seals have been considered as part of onboard
systems duriny most of the history of the study of the control of
refueliny emissions. At least as early as iy?9, EPA did work to test
the control efficiency of an onboard system equipped with a liquid
fill neck seal (I-tf-18). In the 197y test proyram, a "submeryed fill"
liquid seal system was tested and showed a control efficiency of yy
percent. More recently, EPA performed a series of tests at the Motor
Vehicle Emissions Laboratory (MVEL) which examined the control efficien-
cies of a number of liquid seals. The liquid seal work done by EPA at
the MVEL is fully described in a technical report available in the public
docket (I-A-luy). Unly the hiyhliyhts of the study will be repeated here.
Three liquid seal systems were evaluated in the EPA study: the
"in-tube trap," the submeryed fill, and the "J-tube." The in-tube
trap (shown in Figure 2-2) is similar in concept to the standard sink
drainpipe. The submeryed fill system (Fiyure 21-3) employs a fillneck
that extends into the fuel tank and introduces yasoline near the bottom
of the tank. When the yasoline level rises in the tank duriny a fueling
event, the fillneck opening is submeryed in the yasoline and vapors are
trapped in the tank above the liquid level. The J-tube (Fiyure 2-4)
is a simplified form of the in-tube trap. The fuel being dispensed
into the fillneck is forced to pass through the "U"-shaped portion of
the fillneck. The liquid trap is formed as the yasoline passes from
the lowest point in the "U" to the openiny of the fillneck in the tank.
The submeryed fill and J-tube systems were identified as practical
alternatives to the mechanical seal systems. The J-tube was shown
to control refueliny emissions with an efficiency of at least y? per-
cent. The J-tube evaluation was conducted on a bench prototype, which
was later installed in a vehicle. It is clear that higher efficiencies
are achievable with more fully developed systems (I-H-lbb). The advantages
that these sealing approaches have over the mechanical type seals are
described below.
-------
Sl&E VISW
T'
4
11
Figure 2-2. In-Tube Trap Liquid Seal
2-7
-------
Pressure Rellef-
Figure 2-3. Submerged Ft!" Liquid Seal
Figure 2-4. J-Tube Liquid Seal
2-8
-------
Some commenters cited the purported lack of durability as a
possible deficiency of the mechanical seal. Clearly, a liquid seal
system would eliminate any questions concerning durability. A new liquid
seal is formed with each refueling and there are no mechanical parts to
wear with time. The same line of reasoning can be used to respond to
the questions about the susceptibility of the mechanical seal to environ-
mental extremes. Extremely hot or cold weather would have no effect
on the performance of the liquid seals. The possibility of fillneck
tampering would be eliminated by using a liquid seal, since the outward
appearance of a submerged fill or J-tube fillneck would be no different
from the appearance of the current fillneck. Therefore, the addition
of the liquid seal would not increase the incentive to tamper.
The other potential drawback to the mechanical seal is its
need for a pressure relief device. The liquid seal systems are an
improvement over the mechanical seal in this area as well. Both the
J-tube and the submerged fill systems avoid the problem of fuel tank
over-pressure during a refueling event. Any tank over-pressure
that occurred while fuel was being dispensed into the tank would cause
fuel to rise in the fillneck and automatic nozzle shutoff to occur.
Any pressure buildup is automatically released through the fillneck.
Similarly, failure of the nozzle automatic shutoff mechanism would
have no additional safety implications with liquid seal systems, since
failure would result in a fuel spit-back and subsequent manual shutoff
as now occurs.
Both of these systems avoid over-pressure during refueling, but
only the J-tube system can function safely without a pressure relief
mechanism of any kind. The submerged fill system is, however, poten-
tially susceptible to a potential safety problem of another kind and
may require a pressure rel.ief device. If the lines between the fuel
tank and the carbon canister were blocked between fueling events,
pressure could build in the tank during normal operation of the vehicle.
If the vehicle's gas cap was removed while the tank was still relatively
full, gasoline could be forced out of the fillnect (which would contain.
a standing column of gasoline) and possible onto the nozzle operator.
Although this situation might arise only on rare occasions, a pressure
relief mechanism may be required for some submerged fill systems to
2-9
-------
prevent the possibility of fuel spit-back. It should be noted that
mechanisms are available that could perform the functions required of
the pressure relief device. The adaptation of such mechanisms to this
application should be straightforward and inexpensive. This problem
can be easily avoided in the J-tube system by allowing the fillneck to
drain after the completion of.a refueling event. This can be accomplished
by drilling a small hole at the base of the liquid seal to allow a slow
drain of fuel into the tank following the refueling.
The technical report on the testing of the fillneck seals discusses
both the positive and negative aspects of each approach, and the discus-
sion will not be repeated here. The report concludes that both the
submerged fill and J-tube systems would be usable, and both would
eliminate problems related to durability and tampering. The J-tube
system does have one advantage over the submerged fill system in that
it does not require a pressure relief device. This tends to make the
J-tube approach more attractive than the submerged fill system for
both cost and safety reasons. For this reason, the primary onboard
control system considered by EPA in this analysis is equipped with a J-
tube fillneck seal. However, it should be noted that both the mechanical
seal and submerged fill approaches are also technologically feasible
and may be preferable in some applications.
Even though the liquid seal approaches may have some initial
advantages over the mechanical seal designs which have been demonstrated
thus far, this does not imply that mechanical seals cannot or should
not be used. Design and incorporation of a pressure relief device into
the system is a relatively straightforward engineering task. While
tampering could be viewed as a potential problem,.EPA would expect that
mechanical systems used by manufacturers would incorporate tamper-
resistant designs. Also, it is worth noting that much of the incentive
for fillneck tampering in the past was related to the consumers' desire
to save money through misfueliny. With the recent decline in the leaded
to unleaded fuel differential, this incentive has diminished. Finally,
with regard to the durability of mechanical seals, it should be noted
that the vehicles evaluated by API showed no deterioration in the
effectiveness of the mechanical seal during the test period. The seals
retained their effectiveness over a wide range of environmental conditions
2-10
-------
and with the use of alcohol blends. If needed, the option for usiny
improved .seal materials clearly exists. The rotary grease seal used in
the API demonstration project was off-the-shelf hardware which could be
improved specifically for this application if necessary. Thus, EPA
believes that either liquid seals or mechanical seals could be used as
part of onboard systems.
The development of a liquid seal alternative addresses many of the
key comments raised with regard to onboard control technology, and to
some degree changes the onboard system from that originally discussed
in the July 1984 analysis. In addition, the comments have raised other
issues which led to further revisions of the analysis of onboard systems.
Given these changes, the basic systems discussed in this document have
some substantial differences from the system discussed in the 1984
analysis. As an introduction to the analysis of comments, the onboard
systems as envisioned by EPA are described in detail. The description
of the systems should make the analysis easier to follow and will
provide an indirect response to many of the comments.
2.1.3 Description of the Onboard Systems Evaluated by EPA
The onboard systems are described here in terms of both function
and form. The reader is led through a typical refueling event and the
on-board hardware is described as each component performs its function.
The description of the refueling event is presented so that the reader
can get a feel for both the technical workings of the system and its
interaction with the operator of the gasoline nozzle.
The EPA envisions two types of onboard systems, one for fuel-
injected vehicles and one for carbureted vehicles. The difference in
the systems is brought about by the need for control of evaporative
hot soak emissions from the carburetor bowl(s) of carbureted vehicles.
Because of the current trend toward fuel injection and the associated
projections for an 88 percent fuel-injected light-duty fleet by 199U
(I-A-7U), the onboard system for fuel-injected vehicles is the primary
system described. The differences in the carbureted-type system are
also briefly noted. Because the system for fuel-injected vehicles . .
employs the use of a single canister for both refueling and evaporative
emission control, it is referred to as a "fully integrated" system.
The system for carbureted vehicles uses separate canisters to control
2-11
-------
emissions from the fuel tank and carburetor bpwl(s), and is therefore
referred to as a "partially integrated" system. Figure 2-5 shows a
typical fully integrated, onboard refueling and evaporative control
system.
The description of the system operation begins as the vehicle
operator parks the vehicle at the service station and the gas cap is
removed. The opening of the gas cap cover and removal of the gas cap
relays information to the electronic control unit (ECU), signaling the
start of the refueling event. In the system envisioned by EPA, the
switch is triggered by the removal of the gas cap as shown in Figure
2-6, but there are a number of other possible switch positions. Many
of these are described in a report written by Mueller Associates for
EPA in 1985. This report, entitled, "Costs of Onboard Vapor Recovery
Hardware," has been included as part of the public docket (I-A-77).
Therefore, these other possible locations will not be described here.
The removal of the gas cap throws a switch to indicate that a
refueling event is about to take place by sending a signal to a solenoid
valve located in the vapor line between the fuel tank and the carbon
canister. Figure 2-7 is a diagram of one such valve. A complete
description of the valve can be found in the Mueller Report referenced
above. The valve moves to the open position, allowing vapor to pass out
of the fuel tank at the high rates necessary during a,refueling event.
As was discussed earlier (Section 2.1.1), a relatively large
(approximately 1/2 to 5/8-inch inner diameter) vapor vent line would be
needed between the fuel tank and canister to accommodate tne high vapor
flow rates (>8 gpm) associated with refueling. An onboard control
system must include some provision to allow the necessary flow during
refueling and to prevent excessive fuel spillage should the vehicle
roll over. The solenoid valve discussed in the previous paragraphs is
one possible method of satisfying both of these requirements; the valve
is open during refueling events and closed during other modes of operation.
Another approach that could be used to solve this problem is discussed
in a recently published API report (I-H-158). In their work for API,
Mobil Research and Development used a valve that was mechanically
opened by the act of nozzle insertion to permit vapor flow during
refuelings. Although this and other approaches are feasible, the
2-12
-------
ro
i
CO
Figure 2-5. INTEGRATED EVAPORATIVE/REFUELING SYSTEM
TANK MSumiD VALVES
REAR MOUNTED CANISTER
J-TUBE SEAL
MECHANICAL OR SOLENOID
ACTUATED VENT/ROLLOVER
VALVE, LIMITING ORIFICE
& LIQUID/VAPOR SEPARATOR
5/8" DIA.
3' LONG
V *- J-TUBE SEAL
^CONTROIIfl) LEAK
3 LITER
CANISTER
VALVE
3/8" DIA.
5' LONG
TO PURGE
INDUCTION
POINT
15 GALLON fUEL TANK
-------
Figure 2-6. Vent Valve Activation Switch
SWITCH
Figure 2-7. Vapor Vent Valve
0.050
yo
MUELLER ASSOCIATES, INC.
«*oi . IDOIWOOD *T*MT
ALTIMOMB, MAMVLANO fttlt?
JANUAMV 81, 198S
-------
example presented here presumes the use of the solenoid type
system.
At about the same time that the solenoid valve opens the vapor
line, the operator inserts the fuel nozzle into the fillneck. Since
this system uses a J-tube liquid seal, the nozzle operator would
notice no difference between the fillneck of the controlled vehicle and
the fillneck on an uncontrolled .vehicle of the same make. As yasoline
flows down the fillneck and into the tank, a liquid trap is formed in
the fillneck almost immediately. As gasoline fills the tank, vapors
are displaced from the tank. Since the liquid trap has formed in the
fillneck, vapors cannot leave through the fillneck and are forced to
pass out of the tank via the refueling vent line which connects the
fuel tank to the carbon canister. After vapors pass out of the fuel
tank they pass through a liquid/vapor separator. The liquid/vapor
separator removes gasoline droplets from the vapor stream and returns
the liquid fuel to the tank. The separation of this liquid gasoline
from the vapor flow helps to reduce the hydrocarbon load reaching the
canister and prevents liquid gasoline droplets from poisoning the
canister.
The refueling vapors then pass through the vapor line (b/8-inch
diameter) and enter the carbon canister where the hydrocarbons in the
vapor stream are adsorbed onto the activated carbon. This canister
may be in the front or rear of the vehicle.
Inside the fuel tank, a float valve or some similar device is
connected to the vapor inlet orifice (see Figure 2-8). As the gasoline
level rises to the top of the tank, the float valve seats itself in a
a housing a^. the vapor orifice. As the float blocks the vapor orifice,
the pressure rises in the tank and a column of gasoline rises in the
fillneck. When the column of gasoline reaches the tip of the gasoline
nozzle, automatic shutoff is triggered and the refueling event is
completed. The float housing and float are designed to provide a soft
but effective close, so that the pressure in the tank does not rise
too high too quickly. This feature is included to help eliminate .
gasoline spillage at the end of a fueling event.
At the completion of the refueling, when automatic shutoff has
been triggered, the operator removes the nozzle from the fillneck and
2-15
-------
5/£
"
ro
i
o.oso" BYPASS
Figure 2-8. Possible Fill Limlter Design
MO
MUELLER ASSOCIATES. INC.
I40t «. eooewooD STREET
ALTIMOHF. MAHVLAND a
JANUARY a 1. 1088
-------
replaces the yas cap. As the yas cap is replaced, the solenoid that
opened the vent closure is reversed, and the vent line is closed.
Using a mechanical vent closure valve such as that developed by
Mobil, the vapor line would close when the nozzle was removed.
The ability of this system to open and close the refueliny vapor
vent line provides the fuel tank rollover protection required by FMVSS
3U1. The refueling vapor vent line would remain sealed as long as the
solenoid activating the vent closure valve was not activated. Since
actuation of the solenoid would require removal of the fuel cap, this
approach should provide reasonable certainty of vent closure even in
vehicle accidents. Using the mechanical vent closure valve, the system
would remain sealed during vehicle operation unless a serious colli-
sion destroyed the fillpipe area of the vehicle. In this case, fuel
could spill diractly from the fillpipe, as could occur on today's vehicles.
Also, when the refueling vent line is closed, the system could be
designed to provide a limiting orifice through which evaporative emis-
sions are metered during normal operation of the vehicle. Alterna-
tively, the current limiting orifice .system could be retained, with the
evaporative emission vapor line for this purpose being "teed" into the
refueling vapor line at a point beyond the vent closure valve. While
it is not absolutely clear that a limiting orifice is necessary, it
has been argued that the use of this orifice is desirable for several
reasons such as lower diurnal and running losses and improved vehicle
driveability.
After the vehicle has been started and warmed up, the onboard system
draws on manifold vacuum to pull air through the onboard canister and
purge" it of the hydrocarbon load. The integrated system handles
evaporative emissions in essentially the same way today's evaporative
control systems do. The one change for the integrated system is that
the evaporative (diurnal, hot soak, and running loss) emissions would be
loaded into a refueling/evaporative canister along with refueling
emissions.
As discussed earlier, there are three basic system designs that
could be used to control refueling and evaporative emissions. The
first is a fully integrated system (described above), in which all
evaporative emissions would be loaded into a canister (or canisters)
2-17
-------
that is also used for refueling vapor control. The second can be
called a partially integrated system. A partially integrated system
would route all emissions from the fuel tank (both evaporative and
refueling) to one canister. Any carburetor bowl (hot soak) emissions
would be loaded into a separate canister. This type of system would be
used only on carbureted vehicles. Based on the projection that 88
percent of the new car fleet will use fuel injection by 1990, EPA
assumed that 12 percent of the fleet will employ this kind of system
(I-A-70). The third system is a fully separate system. This type of
system would use separate canisters for refueling, evaporative fuel
tank, and carburetor emissions.* Because of the necessary complexity of
the purge system for such a separate system, EPA does not believe that
many of these systems would be used (one possible exception is heavy-duty
gasoline vehicles). Therefore, separate systems were not included in
this analysis.
The discussion above provides a brief synopsis of the onboard
control system envisioned by EPA. The system originally considered
by EPA is described in Appendix C of the July 1984 analysis. With a
clear image of both .the original and modified systems in mind, the
comments on technological aspects of onboard control can be analyzed.
2.1.4 Summary and Analysis of Comments
The onboard control system described in Appendix C of the 1984
analysis was based on the mechanical seal approach. Several of the
comments received described potential problems with the system described
in Appendix C. Some of the comments also contained suggestions for the
improvement of the system. As discussed above, after analyzing the
comments, EPA has developed revised versions of the basic onboard control
systems, which could avoid many of the purported problems raised in the
comments. In Chapter 2.0 of this report, the comments on the onboard
control system are summarized and the EPA responses to the comments are
given. First, the comments on specific hardware items are examined
(Section 2.1.4), followed by discussion of more general topics related
*A separate system for fuel-injected vehicles would use separate
canisters for refueling and evaporative fuel tank loads.
2-18
-------
to onboard controls. For the most part, only the comments opposing
various aspects of onboard controls are discussed.
a. Fill neck Seals
Comment: As discussed above, a number of commenters expressed
the belief that the mechanical fillneck seal would not be durable
enouyh to last the full life of an LDV or LUT (I-H-2, I-H-11, I-H-yy,
I-H-10U, I-H-101, I-H-116, I-H-120, I-H-127). This conclusion was based
on the fact that no vehicle had been operated with onboard controls
past 65,UUO miles. It was also based on concerns that the seals had
not been exposed to the variety of environmental extremes that might be
encountered in use. One commenter raised the question of whether an
onboard system with a mechanical seal would comply with existing regula-
tions concerning fillneck access. Several others felt that onboard
control systems would be subject to tampering (see Section 2.2.5).
Response: In the onboard control system described above, a liquid
seal is used to prevent vapors from leaving the fuel tank via the
fillneck. As was discussed previously, liquid seal designs such as
this avoid many of the potential problems associated with mechanical
seals. All problems of durability are eliminated because a new. seal is
formed with each refueling. The comments on the ability of the seal to
function in environmental extremes are also addressed by the use of the
liquid seal, since extremely cold or hot weather would have no effect
on the functioning of the liquid traps.
The liquid type seals also avoid concerns about tampering, because
the modifications to provide the liquid seal would be out of the nozzle
operator's sight.- From the nozzle operator's perspective, the controlled '
refueling event would appear identical to the current uncontrolled event.
If a mechanical seal were used, the possibilities of both tampering
and durability would have to be considered in seal design. Based on the
work done by AkCO, where mechanical seals were tested under adverse weather
conditions, it appears that a nozzle seal could be designed to retain its
integrity for the full life of the vehicle (I-F-17). The EPA also encoura-
ges the development of tamper-resistant designs that would be sturdy '
enough to discourage most efforts at tampering. However, as mentioned pre-
viously tampering is not expected to be a significant problem in future
years as the economic incentive to tamper decreases. Finally, the question
2-19
-------
of fillneck accessibility does not seem to be a major issue. The
modifications needed to provide a seal (mechanical or liquid) should
not impact current California fillpipe access standards. Onboard
equipped vehicles using mechanical seals or liquid seals could still
use vapor recovery nozzles.
b. Carbon Canister
Comment: A number of commenters questioned the adequacy of the
hydrocarbon storage capacities of the canisters for which costs were
given in the July 1984 EPA analysis. Some claimed that the working
capacity of the carbon was overestimated, and some felt that the refueling
emission load used to size the canister (4.54 g/gal) was underestimated.
Later comments stated that the average fuel tank sizes used by EPA for
sizing the canisters were too small (13 gal, LUV/18 gal, LUT). Some
commenters noted that the large onboard canister might create an unaccep-
table level of backpressure in the fuel tank during refueling (I-H-9U,
I-H-118, I-H-127). Also, a number of commenters expressed concern about
the durability of carbon canisters, some claiming that canister mainte-
nance would have to be allowed. 'These comments focused on: (1) carbon
aging, (2) carbon deterioration, and (3) the effects of alcohol blends
on activated carbon.
Response: The comments on the adequacy of canister sizing prompted
EPA to use a detailed methodology to calculate required working capaci-
ties and the associated canister sizes. This methodology consisted of:
(1) calculating an appropriate refueling emission rate for the test proce-
dure now being considered by EPA, (2) applying this emission rate to
projected fuel. t_ank sizes, and (3) calculating the amount of carbon
needed to control the emissions from these average fuel tank sizes.
The emission rate used to calculate average refueling emissions
and to size carbon beds was developed from a series of uncontrolled
refueling emission tests done at the MVEL in Ann Arbor in the winter
of 1984-85. Over one hundred uncontrolled tests were performed on a set
of six LDV's and two LDT's. The tests covered most applicable ranges of
residual tank and dispensed fuel temperatures as well as a significant
range of fuel volatilities as measured by Reid vapor pressure (RVP).
A multiple linear regression was done on the results of the uncon-
trolled tests, to develop an equation to predict the emission rate for
2-2U
-------
chosen conditions of temperature and KVP. The test program and the
analysis of the results are thoroughly described in an EPA technical
report, and will not be described here (I-A-69).
The emission rate regression equation uses three inputs to calcu-
late an emission rate: (1) TQ, the temperature of the dispensed fuel,
(2) AT, tank residual fuel temperature less Ty, and (3) RVP, fuel
volatility measured as RVP. The values for temperature chosen for
these variables are those given in the proposed refueling test proce-
dure (I-A-71). The values were chosen by using available data to
estimate 9Uth percentile conditions for TQ and AT for the ozone prone
regions and ozone prone months. In other words, the AT and 1$ conditions
were chosen to be worse (in terms of emissions) than yu percent of all
refuel ings during the ozone prone months of May through September in
those regions. The fuel volatility was chosen to be representative of
summertime in-use fuel in those regions. The draft recommended
practice for the measurement of refueling emissions details the deriva-
tion of each of these values (I-A-71), so the analysis will not be
repeated here. The parameter values used here are: TD = 88°F, AT =
5°F, and RVP = ll.b.psi. Using these values, the equation predicts a
refueling emission rate of 7.0 grams of hydrocarbon emitted per gallon
of fuel dispensed. This emission rate was used to size canisters.
If the RVP equals 9.U psi, the emission load is 5.8 grams per gallon.
This emission rate was then applied to average fuel tank sizes
for LOV's and LDT's. The fuel tank sizes were chosen by assuming an
average single tank driving range of 3UU miles, and applying the_average
vehicle fuel economy (from the MOBILES Fuel Consumption Model) for the
chosen year to the 300 miles (I-A-99, I-B-37).* This gives an average
tank size for each vehicle class and model year of interest. For the
refueling test procedure it has been assumed that a worst-case fueling
would require no more than a 90 percent fill and, therefore, carbon
*Fuel tank size projection data supplied by two manufacturers after the
close of the comment period on the gas marketing study indicate that
fuel tank sizes may not decrease significantly in the future as fuel
economy improves. This in turn argues that vehicle downsizing will not
be a significant source of fuel economy improvement in the future. If
this is the case, then average fuel tank sizes will be 10-20 percent
larger than those used here.
2-21
-------
canisters are sized to only 9U percent of the 3UU-mile driving range
capacity. The calculation of the fuel tank capacities is shown in
Table 2-1. For 1989, the average LDV fuel tank capacity was estimated
to be 12.2 gallons (9U percent = 11.U gallons). The average LDT fuel
tank capacity in 1989 is estimated to be 16.3 gallons (9U percent =
14.7 gallons).
The calculation of the necessary hydrocarbon storage capacity
can be completed by simply multiplying the predicted emission rate by
90 percent of the projected average fuel tank capacities. The pro-
jected gasoline working capacities (in grams) that went into the
calculation of onboard system costs are shown below:
LDT
LDV LDT (dual tank)
1989- 77 1U3 2U5
1994 71 1UO 198
200U 65 92 183
Some commenters felt that the working capacity of activated car-
bons might have been overestimated in sizing the canisters in Appendix C
of the Juiy 1984 analysis. In order to avoid this problem, EPA has
derived an appropriate gasoline working capacity from carbon manufactur-
ers' specifications for their carbons. The carbon EPA has chosen to
evaluate is Westvaco, extruded carbon. This carbon was chosen because
^
it has a relatively high butane working capacity (lU.b g/lUU ml) and.
can be produced to cause a low pressure drop through the canister (the
issue of system backpressure will be discussed in a following paragraph).
This is not meant to imply that other carbons (wood, coal, or coconut-
based) are not acceptable. They simply would have different working
capacities and different pressure drop characteristics.
The butane working capacity was corrected for the difference between
butane and gasoline vapors and for carbon aging. "Carbon aging" refers
to the process by which activated carbons lose working capacity with
successive load/purge cycles until a stabilized level is reached. The
fraction of the initial working capacity at which stabilization is
achieved is not known'with certainty. The EPA was quoted estimates
2-22
-------
Table 2-1. CALCULATION OF FUEL TANK CAPACITIES FOR
VARIOUS YEARS
1989
Driving Range
(mi ) 300
Fuel Economy
(mi /gal) 24.61
Fuel T.v . ;3city
(gao 12.2
LDV
1994 2000
300 300
26.64 29.13
11.3 10.3
LOT*
1989 1994
300 300
18.43 18.99
16.3 15.8
2000
300
20.56
14.6
*The fuel tank capacities for LDT's with dual fuel tanks were estimated by
doubling the LOT values.
2-23
-------
ranging from 60-75 percent by carbon manufacturers. To be conservative,
EPA used 60 percent in this analysis. At this point, an additional
safety factor of 10 percentage points was also applied. Therefore, the
virgin working capacity was reduced by 50 percent to reflect aging and
a moderate factor of safety. The aged butane working capacity is then
(10.5 grams HC/100 ml) x 0.5, or 5.25 g/100 ml.
Gasoline working capacities are generally expressed as some frac-
tion of butane working capacities. Depending on the source, the
fractions quoted range from 60-85 percent. In order to provide a
liberal estimate of the amount of carbon needed, the lower bound of
this range, 60 percent, was used. Thus, the aged gasoline working
capacity of the Westvaco extruded carbon has been estimated to be 3.15
grams HC/100 ml carbon (0.60 x 5.25 g/100 ml). The canister sizes were
found by dividing required hydrocarbon storage capacity from above by
the aged gasoline working capacity of the Westvaco carbon. Carbon
canister sizes, in milliliters, are shown next:
Year
1989
1994
2000
It should be noted here that in an attempt to estimate costs conservatively
the necessary refueling capacity was added to the existing evaporative
capacity. Therefore, the carbon bed sizes shown above would be larger
by the current evaporative control capacity.
The comments on the effects of alcohol on activated carbons are
related to the concept of carbon aging. Some commenters felt that the
increased use of alcohols in fuels could increase the size of the canister
heel and thus reduce the total canister vapor capacity in-use. There
is little evidence available to support this assertion. The EPA has
conducted two separate test programs to evaluate whether alcohols reduce
carbon working capacity (I-A-107, I-A-108). In each case the test programs
have led to the conclusion that alcohol has little or no effect on the
canister working capacity, and no long-term loss .in working capacity
can be traced to the presence of alcohols.- Based on the information
2-24
LDV
2,440
2,240
2,060
LOT
3,260
3,150
2,900
LOT
(dual tank)
6,520
6,300
5,800
-------
presently available, it is reasonable to conclude that alconols do
not cause a significant decrease in workiny capacity with time.
The commenters who felt that canister maintenance would be
required expressed concern over: (1) deterioration of carbon particles,
(2) reduction in working capacity through aging, and (3) canister
poisoning. The concerns about the deterioration of carbon particles
with time were expressed by a single commenter. This commenter's
concerns were based on examination of evaporative control systems from
in-use vehicles which showed that a number of them were inoperative.
In some cases, further inspection of the non-functioning systems showed
that pulverized carbon particles were blocking vapor lines or canister
orifices on some of the vehicles. The commenter claimed that similar
problems would arise in onboard systems and that canister maintenance/
replacement would be necessary to resolve this problem.
The key to avoiding problems with th.is type of carbon deteriora-
tion is proper canister design, internal packaging, and canister filling
technique. If the canister internal design is proper to prevent shaking
of the carbon bed during vehicle operation and to prevent any pulverized
carbon from reaching key orifices, then problems such as described by
the commenter should not occur. This involves proper filling technique
to prevent settling of the carbon granules in-use and the use of
internal foam or fiber separators to further pack and restrict the
carbon bed. The canisters examined in the study were from vehicles
produced in the first half of the 197U's, during the early stages of
evaporative emission controls. The canister filling process and internal
canister des.ign has improved since that time and deterioration of carbon
particles should not be a significant problem and has not been found to
be a problem on other vehicles tested in EPA's emission factors program.
Carbon aging was defined above as the decrease in total hydrocarbon
storage capacity with time, normally during the first one hundred or
so load/purge cycles of its use. The working capacity of a carbon bed
will decrease initially, as one commenter pointed out, but it will
stabilize and remain relatively constant at that level in the future.
As discussed in the paragraphs above, working capacity was reduced by
4U percent (based on carbon manufacturers' suggestions) to allow for
carbon aging. If carbon aging is accounted for in carbon canister
2-25
-------
designs, there.should be no need for canister maintenance or replace-
ment as long as the carbon is not poisoned.
Carbon poisoniny occurs when liquid fuel is allowed to flow
through vapor lines and into the carbon bed. If the larye hydrocarbon
molecules in the liquid fuel (which are only present in small quanti-
ties in vapor) reach the canister and are adsorbed onto the carbon
bed, they are unusually difficult to desorb. This problem can be
avoided by including a liquid/vapor separator in the vapor line between
the fuel tank and the carbon canister. If such a device is included,
canister poisoniny should not be a significant problem.
There were only a few comments that claimed that canister mainte-
nance would be required. There does not seem to be general concern over
the ability of carbon canisters to last the full life of a vehicle. In
a recent EPA final rule (bU FK 1U6U6, March Ib, iy«b), EPA increased
the LUT maintenance interval for evaporative control canisters to
1UU,UUU miles. The comments to the NPKM for this rulemakiny contained
only passiny reference to the increase in the maintenance interval
(I-U-318). The general lack of concern shown about the difficulty of
meetiny a maintenance interval of 1UU,UUU miles indicates a general
ayreement on the part of the auto industry that canister deterioration
for wel1-maintained vehicles is not an issue, and canisters can be
designed to last for the full life of an LUV, LUT, or HUGV.
Some commenters felt that the larye onboard canisters would induce
an unmanayeable fuel tank backpressure when refueling vapors are forced
through at 8 to 10 gallons per minute. Backpressure in the fuel tank
during refue.liny.is influenced by a number of factors. These include:
(1) activated carbon type, (2) mesh size of activated carbon, (3) vapor
line diameter, (4) vapor line configuration, (b) limiting orifice size,
(6) canister shape and the. associated vapor path, (7) canister configu-
ration - open or closed bottom, (8) canister location, and (y) fuel
fill rate. A low backpressure is essential to the proper functioning
of liquid seal systems, and the control of backpressure is discussed
thoroughly in the1 EPA technical report referred to previously (I-H-lbb).-
The work done in testing the liquid seal systems shows that through
proper material selection and system design it should be possible to
keep backpressure at a manageable level.
-------
c. Packaging
Comments: A number of auto makers expressed concern that adding
onboard refueling vapor controls would introduce packaging problems
and force them to make modifications to their vehicles. Some claimed
they would have to relocate a number of components to accommodate on-
board controls. Others claimed that the increased space requirements
of onboard controls would force them to sacrifice cargo space or fuel
tank capacity (I-H-22, I-H-53, I-H-100, I-H-1U4).
Response: Controlling refueling emissions from LDV's and LDT's
would require adding a carbon canister between 2-1/2 and b liters (for
11.5 psi RVP certification fuel) in volume,* as well as various other
related components. Mueller Associates examined packaging implications
in their report on the costs of onboard vapor recovery hardware (I-A-
77). Although the contractor's study was a limited survey of potential
impacts on current vehicles, several conclusions can be drawn. First,
for large and mid-size LUV's and most LDT's, packaging of onboard
control systems is not a significant issue. For compact and smaller
LDV's, some vehicle modifications might have to be made to accommodate
the onboard components. These modifications might be as insignificant
as relocating a few noncritical items or as involved as modifying fuel
tanks or otner vehicle sheet metal. It should be noted that these
smaller vehicles would also have the smallest refueling canisters, so
system packaging may not be as challenging as portrayed by the commenters.
Although tooling changes for current vehicle models that are not
scheduled for major design changes may be relatively costly, it poses
no threat to-the-feasibility of onboard controls. As time passes and
new vehicle designs are planned, onboard controls would be included as
one of many design criteria, reducing the impact of packaging concerns.
As a short-term option, manufacturers may choose to marginally reduce
cargo space or fuel tank volume on compact cars during the first few
years of the regulation, but in the long run, onboard systems could be
designed to use space efficiently and restore .capacity. In conclusion,
EPA recognizes that packaging of onboard systems might involve some.
*Uual tank trucks might require larger canisters. See Cost Section
2.6.1 for derivation of required canister sizes.
2-27
-------
engineering effort and expense and accounts for this in tne cost analysis,
but the problems it causes do not affect the technological feasibility
of onboard controls.
2.1.5 Safety Concerns*
Comment: The commenters on the safety of onboard control systems
addressed three main issues: (1) gasoline spillage and the need for a
pressure relief mechanism, (2) fuel system integrity during accidents
or vehicle rollover, and (3) potential hazards related to the carbon
canister. The comments related to these topics are summarized below.
Many of the commenters reiterated the safety problems mentioned
in the 1984 analysis regarding fuel spit-back and noted that it might be
more difficult to solve these problems than EPA suggested. Commenters
noted that pressure buildup could lead to fuel spillage and/or to the
poisoning of the carbon canister if the liquid/vapor separator failed.
The commenters pointed out that any spilled fuel could pose a fire
hazard and would reduce the effectiveness of the control system. The
commenters also noted that if gasoline were forced out through the
fillneck seal, the fuel spit-back may fall on tne nozzle operator.
Many commenters also noted that even the proper functioning of a pres-
sure relief device could cause some of these safety hazards.
A second major topic of comment was the integrity of the fuel
system during accidents or vehicle rollover. Commenters noted that
onboard systems would have to be designed to accommodate the high vapor
flow rates associated with refueling, and this would involve larger
vapor lines and associated orifices. Some changes would therefore have
to be made to ensure that Federal fuel tank integrity standards (FMVSS
3U1) could be met. Some commenters claimed that this would require
more than a simple enlargement of existing rollover valves (I-H-2,
I-H-53, I-H-90, I-H-100, I-H-101, I-H-114, I-H-118, I-H-127). Also,
one commenter noted that the connections needed to link the additional
onboard vapor lines and valve(s) from the fuel tank to the carbon
canister would create more chance for a loss of fuel system integrity
during a vehicle accident.
*1984 Federal Register topic.
2-28
-------
A third area of comment concerned the carbon canister used to
capture refueling emissions. These comments involved the possibility
of the canister creating a potential .fire hazard as a result of either
canister rupture during a vehicle accident or canister removal in-use
due to tampering.
Response: The following analysis summarizes the major engineering
and design issues raised in the comments. A more detailed discussion of
the safety implications of onboard controls, including applications of on-
board to HDlaV's, can be found in the EPA technical report available in the
public docket(I-A-112). In response to comments on onboard technology,
EPA has made a number of changes to the onboard control system concept
described in the July 1984 analysis. Concerns about the safety of the
original system were a major reason for many of these changes. Three
features of the new system led to substantial improvements in its
safety. These are the liquid fillneck seal, the refueling vent line
valve(s), and the "soft" closing fill limiter.
The EPA has duly noted that there are a number of potential, safety
problems associated with pressure buildup during refueling of vehicles
equipped with mechanical fillneck seals. While EPA still believes that
an adequate pressure relief mechanism can be developed for these systems
through direct engineering effort, it must be noted that the systems
currently being analyzed by EPA do not use a mechanical fillneck seal.
Rather, these systems use a liquid seal in the fillneck. As was discussed
earlier, the liquid seal systems avoid all difficulties related to pres-
sure buildup during refueling. Any abnormal increase in pressure will
cause gasoline to back up in the fillneck and trigger automatic nozzle
shutoff, as occurs on current vehicles. With a liquid seal system, >.!*>
probability of fuel spit-back is no greater than with present vehicles.
While there are good reasons to be concerned about the effects of
nozzle failure, there is no evidence to indicate that failures are or
will be common or widespread. Discussions with nozzle manufacturers
indicate that most failures of the automatic shutoff mechanism occur
at the very low dispensing rates (<2 gallons per minute) which sometimes
accompany persistent topping-off attempts at the end of a refill. Any
reasonably well-designed pressure relief valve should easily be able to
handle the excess fuel dispensed in this situation. Thus, EPA does not
2-29
-------
believe that fuel spit-back will present a significant design or in-use
problem for those manufacturers choosing to use mechanical seals.
A number of commenters suggested that the incidence of spillage
at the time of automatic nozzle shutoff would increase if onboard
controls .were required. Their contention is that the fill limiting
device needed for onboard systems would shut off abruptly when the
gas tank became full, causing pressure in the tank to rapidly increase.
The pressure would continue to rise in the tank until automatic nozzle
shutoff occurred. Although the time between fill-limiter closure and
nozzle shutoff would only be an instant, these commenters claimed that
fuel tank pressure could build to well above normal levels, which could
lead to fuel spillage even after automatic nozzle shutoff occurred.
This problem can be eliminated by using a fill limiter that closes
gradually. There are a number of mechanisms that could be used to
provide this "softer" close. One method is to use a less than perfect
seal in the housing into which a float seats. Then, when the tank is
full and the float "closes" the vent line, some vapor will still flow
out of the tank through the fill limiter. Fuel tank pressure will not
become excessive, gasoline will rise more slowly in the fillneck, and
automatic shutoff will be smoothly triggered. It should take only a
minimal amount of engineering effort to design a fill limiter with
this type of "soft" close.
With regard to fuel spillage, it is worth noting that the draft
onboard test procedure may actually lead to a reduction in the amount
of fuel spilled in-use, and thus improve the. overall safety of
refueling events. In the refueling test, vehicles would have to be
designed to accommodate a refueling dispensing rate of up to 10 gallons
per minute (near the high end of current in-use values) without any
fuel spit-back due to premature or final nozzle shutoffs. While shutoffs
are permitted in the test, any fuel spilled as a result of the spit-
back is considered as part of the test results (refueling emissions).
Since one tablespoon of gasoline evaporates to about 10 grams of vapor,
almost any spillage will result in a failure of the test. Thus, EPA .
believes that manufacturers will design fillpipes and fuel systems
that allow no spit-back of fuel at refueling rates up to 1U gallons
2-30
-------
per minute. This will lead to a reduction in the amount of fuel
spilled in-use and improve the safety of vehicle refueling.
Adjacent to the fill limiter in the onboard system discussed above,
is the refueling vent line closure valve. In the open position, this
device provides a large orifice (which can accommodate the high vapor
flow rates associated with refueling) and could also be designed.to
provide a limiting orifice when closed (see Figure 2-7). The valve is
open whenever the fuel cap is not in place on the fillneck and is closed
at all other times. When closed, the valve would serve to limit the
loss of gasoline from the fuel tank during a vehicle rollover. This is
the problem addressed by FMVSS 3U1, which requires a vehicle to restrict
fuel leakage to less than one ounce per minute when rolled on either
side or upside down following a front, rear, or side collision.
The refueling vapor vent closure valve must have one additional
feature. It must be designed so that failure can occur only in the
closed position. If the valve failed in the open position and rollover
occurred, gasoline might escape from the vehicle's fuel tank and cause
a fire hazard. Also, absent some other means, the vehicle owner would
not know the valve had failed. If the solenoid valve failed in the
closed position, it would be very difficult to fuel the vehicle. This
would cause an inconvenience for the vehicle operator, but the fuel system
integrity would be intact and there would be a strong incentive to have
the vent closure valve immediately repaired.
The mechanical approach to the refueling vent valve demonstrated
by API also addressed the need for vent line closure in the event of
vehicle rollover (I-H-158). This kind of valve (Figure 2-9) is closed
at all times except during refueling, and should therefore prevent the
escape of fuel in a rollover incident. Only when the nozzle is inserted
into the fillneck does the valve move to the open position, allowing
vapor to leave the tank. A mechanical valve located at the vehicle/
fillneck interface does present one safety concern that doesn't apply
to the solenoid type valve. Because the valve is located near the exterior
shell of the vehicle it could be susceptible to damage if struck in a
collision. Although this is- an issue that would have to be addressed
by a manufacturer, it seems that the vent valve would be no more sus-
ceptible than the fillpipe or gas cap. This valve would have to be
2-31
-------
Figure 2-9
EMISSIONS VAPOR VENT VALVE
OJ
ro
Vapor to
Canister
Vfepof from
Vapor/Lkfuld Separator
From "Vehicle Onboard Refueling Control", API Publication No. 4424, March 1986.
-------
designed with adequate structural integrity to fulfill its function in
this situation.
Similarly, the vapor line connections to and from this and any
other valves in the refueling control system would have to be designed
to withstand the stresses that may occur in an accident. For example,
when a force is applied during an accident, objects of different mass
(vapor lines, valves) are accelerated at different rates. These
connections and other vehicle component interface areas would have to
be designed to withstand the effects ot differential acceleration.
However, EPA does not believe that this design consideration presents
new or significant problems to manufacturers. Current vehicle fuel and
evaporative emission control systems meet this requirement, and EPA
believes that onboard systems can, with a reasonable amount of engineer-
ing effort, also meet the requirement.
Finally, several points can be made regarding potential safety
hazards related to the carbon canister. The EPA understands concerns
that a carbon canister ruptured during an accident may present a fire
hazard, but this potential problem is no greater with refueling canisters
than with evaporative emission canisters. There is no evidence that
current evaporative emission canisters present a fire hazard, and EPA
believes that refueling canisters can be used with the same degree of
safety. Any lingering canister safety concerns can be addressed through
placement of the canister in a more protected area such as the rear of
the engine compartment or in some other under body area. While refueling
canisters would be larger and more difficult to package in some cases,
on a unit volume basis they would contain no more activated carbon or
vapors than evaporative canisters.
Although not expected to occur, tampering that resulted in removal
of the refueling canister conceivably could lead to a fire/explosion
hazard under the vehicle hood immediately after the end of the refueling
event. While the vapor mixture reaching the underhood area in this .
situation is above the gasoline upper flammability limit (6 percent),
the mixture would briefly become flammable as the gasoline vapor dissipated.
If a spark or other ignition source were present at that time, the
mixture could burn briefly. While this problem is likely to be rare,
there are two means to address it. First, placement of the canister in
2-33
-------
a location distant from iynition sources would eliminate the problem.
Possible locations include the rear of the enyine compartment or some
underbody area. Second, a small label on the canister to discourage
tampering may also be helpful. Nevertheless, canisters should not be
placed in a location where tampering could create a safety problem.
Overall, since canister tampering is uncommon and steps to eliminate
potential problems do exist, EPA believes that this potential problem
can be easily addressed.
2.1.6 Canister Purge
Comment: A number of commenters claimed that purging refueling
vapors from the system canister would have a detrimental impact on
exhaust emission levels, and that the results of the 1978 API demonstration
program are not applicable to the vehicles of today because of more
stringent emission standards and changes in vehicle technology. Uther
commenters claimed that the added purge requirements would lead to
driveability problems (I-U-58, I-D-59, I-H-99, I-H-lUU, I-H-lUl, I-H-116,
I-H-127). One of these commenters felt that driveability problems
might induce onboard control system tampering (I-H-99). One commenter
claimed that EPA had failed to account for the interaction of purged
hydrocarbons and "live" vapor, produced during vehicle operation
(I-H-101). The commenter felt that vehicles' electronic controls
might have to be modified to interpret and adjust to the incoming vapor
mixture. Another of the commenters expressed a concern that vehicles
equipped with onboard controls would not meet emission standards and .
probably would not perform well.outside of the conditions encountered
in the Federal -Test Procedure (I-H-127).
Response: The EPA continues to believe that purge requirements
for vehicles equipped with onboard controls can be met without an
increase in exhaust or evaporative emissions. While many commenters
expressed concerns in this area, no data or detailed technical analyses
were provided to support these positions.
The American Petroleum Institute has sponsored two programs designed
to evaluate the feasibility of onboard control systems. The first one,
performed in 1978, is the one to which the commenters referred (I-F-17).
In this program, three vehicles, certified to 1978 California emission
standards, were equipped with prototype onboard control systems.
2-34
-------
Although all the vehicles in this demonstration program were carbureted,
and hence are not representative of the projected fleet makeup for the
late 198U's and beyond, the program does have some useful implications.
The vehicles, when equipped with onboard control systems, were able to
pass the 1978 California exhaust and evaporative emission standards
despite the fact that more hydrocarbons were being purged from the
refueling canister than from the stock evaporative canister. The API
did not demonstrate that purge was adequate to restore hydrocarbon
storage capacity between refuelings, but it did demonstrate that these
vehicles (equipped with older fuel system technologies) could handle
increased emission loads and still pass emission standards.
The second API program was performed in 1985 (I-H-158). In this
demonstration program, three 1985 model year vehicles were equipped
with onboard control systems and tested under the recommended test
sequence proposed in the EPA recommended practice published in July
(I-A-71). Two of these vehicles were equipped with feedback controlled
fuel injection systems and the third used a feedback controlled carbure-
tion system. Each of the vehicles was able to pass the current exhaust
and evaporative emission standards as well as the refueling requirements
described in the recommended practice. The EPA has, however, revised
the recommended test procedure since the publication of the July 1985
draft to more thoroughly test the adequacy of the vehicle purge systems
(I-A-76). The API vehicles were designed only to pass the requirements
in the July draft, however, and may not meet the more stringent purge
requirements in the April proposal. The results of the 1985 demonstration
do show that vehicles equipped with feedback fuel control can handle
hydrocarbon loads (aver the FTP) greater than those associated with
evaporative canisters without major system modifications, and that
required emission levels can be maintained.
The main technological issues surrounding the feasibility of
onboard controls for feedback-equipped vehicles are the range of
control and response time of the feedback systems. The 1985 API demon-
stration program showed that current feedback systems can adjust to
some increases in purged hydrocarbons while maintaining acceptable
exhaust emission levels. This demonstration supports EPA's belief
that vehicles equipped with feedback fuel control systems could be
2-35
-------
designed to purye refueling canisters efficiently through modified purge
systems and/or strategies.
A different situation arises for those vehicles that are not feed-
back controlled. As mentioned earlier, EPA projects that 88 percent
of the fleet will be fuel-injected by lyyu. Some of the carbureted
vehicles will also use feedback controls, so only a small percentage of
the fleet will be in this category. Discussions with a few auto
manufacturers suggest that most of these vehicles are compact or sub-
compact cars positioned as the lowest-priced cars in the manufacturers'
product line. The EPA believes that with some engineering effort,
these small cars can be equipped with onboard controls and still meet
exhaust emission standards without the addition of feedback fuel
control. The reasons behind this belief are outlined in the following
paragraphs.
First and foremost, the two 1978 API demonstration vehicles that
were not equipped with feedback fuel control systems were able to pass
the then current exhaust emission standards when retrofitted with on-
board controls. The emission standards have been tightened since 1978,
but vehicle emission control packages have improved to keep pace with
the more stringent standards. If the emission control system of the
mid-to-late seventies could meet exhaust standards with a retrofitted
onboard system, it is likely that the more advanced emission control
system of the eighties designed with the onboard purge requirement as
an integral part of the control strategy would be able to meet current
standards with some design modifications and system improvements.
Second, the size and weight of the typical non-electronically
controlled vehicle make it more likely to meet the exhaust emission
standards without substantial modification. As noted above, these vehicles
are typically very small cars with small fuel tanks and a lower total
refueling vapor load. Since the cars have better fuel economy than do
heavier cars, they have a relatively greater driving range in which to
purge a smaller hydrocarbon load. In addition, due to their lighter
inertia weight these smaller cars usually pass emission tests at a
lower emission level than do larger cars. That is, there is a wider
gap between their emission levels and the standards. Therefore, these
2-36
-------
cars should be able to emit at a greater rate than they would without
onboard controls, and still remain below applicable exhaust standards.
Because of the factors described, above, EPA still believes that
vehicles without electronic engine controls can be designed to include
onboard systems and meet exhaust emission standards. In some cases,
it will be an engineering challenge to design a non-electronic purge
control system that will provide adequate canister purge and keep
exhaust emissions and driveability within acceptable ranges. Although
difficult, it appears feasible and, as mentioned above, the task will
have to be undertaken only for a small fraction of vehicle models.
2.1.7 Excess Evaporative Emissions
Comment: Une group of commenters concurred with the EPA position
that it would be possible to control excess evaporative emissions using
the excess capacity of the refueling vapor control system. Although no
commenters denied the feasibility of controlling excess evaporatives
through expanded adsorptive capacity, one commenter noted that purye
rates would also have to be increased to purge the canisters of the
extra vapors. This commenter went on to state that not all excess
evaporative emissions can be controlled by increasing the purge and
storage capacities of an onboard system. The same commenter notes
that onboard systems with a separate carburetor canister would continue
to emit excess "hot soak" emissions, even though the refueling canister
had excess capacity. This commenter concludes that there are several
possible techniques available to control excess evaporative emissions,
and that the issue should be considered separately from onboard.
Response: At the time of the original publication of Appendix C,
the problem of excess evaporative emissions was just being identified.
The initial series of EPA evaporative emission tests seemed to show
that vehicles that had been certified as meeting appropriate evaporative
emission standards were exceeding this emission level in-use. Since
the publication of the July 1984 EPA analysis, the evaporative emis-
sions problem has been much more thoroughly investigated. All of this
work is discussed in great detail in a recently published EPA study-on
fuel volatility and hydrocarbon emissions (I-A-66). In that study a
number of sources of excess evaporative emissions were, identified.
2-37
-------
A large fraction of the excess emissions appears to be attributable to
inadequate hydrocarbon storaye (due at least in part to the increasing
volatility of in-use fuels) and purge capacities. This fraction of
the excess (hereafter referred to as "the controllable excess") can be
controlled through the use of increased hydrocarbon storage capacity
and purge system improvements. One method of requiring greater hydro-
carbon storage capacities and purge system improvements would be to
increase the volatility of the certification fuel and make other test
procedure changes, thereby forcing automobile manufacturers to design
control systems to capture and purge higher emission loads. The problem
could also be addressed by a certification fuel change as part of an
onboard regulation.
Although the excess evaporative emissions problem could be addressed
as part of an onboard emission control rulemaking, EPA is currently
considering several other strategies to obtain this control. Because
the excess evaporative emissions problem could be addressed outside of
an onboard regulation, EPA has focused the onboard analysis strictly on
refueling emission control. That is, the costs, benefits, and cost
effectiveness figures used in the evaluation of onboard controls as a
hydrocarbon control strategy are incremental to those associated with
the control of excess evaporative emissions. Related comments are
discussed in Sections 2.6.4 and b.l.
2.1.8 Test Procedure
Comment; A number of commenters stated that the lack of a proposed
test procedure made it difficult to comment on the technological
feasibility of-onboard controls. Commenters claimed that they could
not proceed with onboard control system development without some idea
of the demands of the test procedure. They also stated that they could
not accurately project what technological barriers might be encountered
in control system development without a complete description of the
refueling test conditions and some idea of what emission levels would
be allowed. A few commenters gave suggestions on possible test proce-
dure conditions and methods.
Response: The EPA agrees that there could be some difficulties
for a manufacturer in developing an effective onboard control system
without a knowledge of at least some of the likely specifics of the
2-38
-------
refueling test procedure. The EPA understood the desire for a recommended
test procedure at the time of publication of the July 1984 analysis,
but was still in the process of developing the details of the procedure.
In response to the manufacturers' concerns, EPA held a workshop on
October 17, 1984, to present some preliminary ideas for a refueling
test procedure. The recommended draft refueling test procedure was
completely described at the workshop (I-B-28). The EPA described the
methodology behind the development of the draft test procedure, the
selection of critical test procedure parameters, and the selection of
appropriate values for those parameters. Since the public comment
period closed on November 8, 1984, at least some commenters were able
to include some information on a possible test procedure requirement in
their comments on the July 1984 EPA analysis.
At the October 1984 workshop, EPA asked the automobile manufac-
turers to prepare comments on the recommended test procedure drafted in
1984. Several groups did comment on the procedure and the Motor Vehicle
Manufacturers Association also presented their comments orally to EPA
in February of 1985. The EPA then published a recommended test proce-
dure in July of 198b. Comments were accepted in response to the
July publication. Following EPA's evaluation of those comments and a
reanalysis of the test procedure requirements, the draft recommended
procedure was changed substantially. The changes in the procedure
were presented to the industry at the most recent test procedure work-
shop on April 10, 1986. Although the development of a refueling test
procedure has been an evolutionary process, and the procedure has not
been finalized, the intent and the basic structural components of the
procedure have been clear for some time. The EPA has been working to
improve the procedure to provide an adequate test of system capabilities
with minimum resource requirements. Although these improvements
may result in changes to the details of the procedure, they should not
affect the basic emphasis of the procedure, the necessary control techno-
logy, or the ability of auto manufacturers to proceed with control
system development.
2.1.9 Miscellaneous
The EPA received a number of isolated comments on minor issues
which are significant enough to merit response. This section summarizes
these comments and provides a brief response to each.
2-39
-------
a. Hydrocarbons Emitted at Fuel Cap Removal and Running Losses
Comment: Various comments were received which described the
phenomenon of "pop" and "hiss" emissions. Pop emissions are vapors
emitted when the fuel cap is removed. Hiss emissions are those emis-
sions that are released from the fuel tank by the gas cap pressure
relief during vehicle operation. Some commenters simply stated that
EPA should further examine the phenomenon. Some claimed that onboard
could control these emissions and felt that the magnitude of these
emissions should be determined and this amount should be credited to
onboard control strategies. Others felt that onboard could not
control these emissions and the efficiency of onboard controls should
be reduced to reflect this loss. A few commenters mentioned "hiss"
emissions as a source of emissions that had been excluded from the
original analysis without justification (I-H-33, I-H-35, I-H-4U,
I-H-44, I-H-46, I-H-69, I-H-76, I-H-85, I-H-87, I-H-89, I-H-y6, I-H-y7,
I-H-102, I-H-1U5, I-H-1U6, I-H-1U8, I-H-1U9, I-H-11U, I-H-119, I-H-120,
I-H-121, I-H-123).
Response: The EPA recently ran a series of tests to evaluate the
severity of the popping emissions problem. The test program and data
analysis are discussed in a technical memo which has been included in
the public docket (I-A-97). The evaluation showed that "pop" emissions
of a significant magnitude (i.e., 2 or more grams) do not occur frequently
enough to justify giving the problem further consideration at this
time. The analysis showed that only at high tank temperatures and with
high volatility fuels do appreciable pop emissions occur. It also showed
that these events occur for less than 5 percent of all refueling events.
Therefore, the total hydrocarbon inventory associated with "pop" emis-
sions is very small and can be excluded from the onboard analysis.
Currently, there is no test information available with which to
assess the magnitude of hiss emissions, although some work is planned
for the future. Because these emissions occur during operation of the
vehicle, it is difficult to isolate and measure them. The EPA has
conducted an engineering evaluation to determine if fuel caps are
likely to vent to the atmosphere during operation under extreme temp-
erature conditions (I-B-19). Although the limiting orifice between a
fuel tank and its evaporative canister is often no larger than forty
2-40
-------
thousandths of an inch, the study shows that tank venting is unexpected.
However, due to the wide variety of vehicle designs and higher in-use
fuel volatilities, some testing work is planned in this area.
b. Nozzle Modifications
Comment. Several commenters felt that gasoline dispensing nozzles
would have to be modified to include a pressure relief device to prevent
fuel tank over-pressure during refueling of onboard controlled vehicles
(I-H-74, I-H-9U, I-H-111, I-H-114, I-H-118, I-H-127).
Response: The EPA has discussed the existence of a potential
problem with fuel tank over-pressure during the refueling of vehicles
equipped with a mechanical fillnecK seal (Section 2.1.b). These
problems do not, however, arise during the fueling of vehicles equipped
with liquid fillneck seals. Therefore, no pressure relief device is
needed for liquid seal equipped vehicles.
If mechanical seal systems were to come into widespread use, it
would be possible to achieve pressure relief through nozzle modifica-
tion. Though possible, this does not appear to be the optimum approach.
If a manufacturer chose to use a mechanical fillneck seal, the manufac-
turer may have to include a vehicle-based pressure relief device, or
else a voluntary uniform nozzle modification would be required. This
does not seem likely, and a regulation requiring such modification would
not be desirable. Thus, a vehicle-based pressure relief device is
preferred. However, if deemed necessary, EPA is open to voluntary
uniform nozzle standards to address dispensing rate limits or standardize
nozzle geometries. Additional information concerning this issue can be
found in an EPA technical report available in the public docket (I-A-111).
2.2 EFFECTIVENESS UF ONBUAKD CUNTKULS
A number of commenters, representing the automobile industry and '
other Stage II proponents, felt that EPA had overestimated the effi-
ciency of onboard controls. These opinions were counterbalanced,
however, by a number of commenters, largely representative of the
petroleum industry, who accepted the EPA efficiency estimates presented
in the July 1984 analysis. The latter analysis projected a new system
efficiency of 98 percent, which might be reduced by tampering and
possibly by deterioration to an in-use efficiency of 92 percent.
Those commenters who disagreed with the EPA efficiency estimates based
2-41
-------
their objections on: (1) what they perceived to be portions of total
refueling emissions that would not be controlled by onboard controls,
and (2) factors affecting onboard control systems that they claimed
were not considered or were inadequately considered in the EPA analysis.
They predicted that the combination of all of these elements would
yreatly reduce the efficiency of onboard controls. Specific categories
of comments on the control effectiveness of onboard are summarized and
addressed separately below. These include: (1) onboard control for
HDGV's, (2) breathing losses, (3) gasoline spillage, (4) onboard in cur-
rent Stage II areas, (5) control system tampering, (6) canister system
durability, (7) purye effects on system efficiency, and (8) overall
control system efficiency.
2.2.1 Onboard Controls for HDGV's
Comment: Several auto industry commenters stated that since
onboard controls would not be applied to HUGV's and motorcycles, the
in-use efficiency of the onboard strategy would be reduced (I-H-2,
I-H-1U7, I-H-114). Another commenter stated that use of Stage II
controls would affect all kinds of vehicles and thus avoid any
"problems" resulting from the exemption of certain types of vehicles
(i.e., trucks and vans) from control (I-H-1UU).
Response: Although the July 1984 EPA analysis concentrated primarily
on LDV's and LDT's, EPA did not intend to imply that HDGV's would be
excluded from consideration if a decision were made to proceed with
onboard controls. In fact, Appendix C addressed onboard controls for
HDGV's, but implementation of an onboard requirement for these vehicles
was not included in the July 1984 analysis, since the focus of the
entire analysis was on the public gas marketing sector which covers
primarily LDV's and LDT's.
The EPA concurs with the commenters who claimed that not controlling
HDGV emissions would reduce the overall effectiveness of onboard by about
5 percent, since that percentage of the highway gasoline consumption
would remain uncontrolled. In response to these concerns, EPA has
included HDGV controls in the re-analysis; the basis for the costs and
other factors used in the reanalysis is discussed in Section 2.6.
It is clear that onboard controls could be applied to most if
not all HDGV's expected in the 199U's and beyond. Lighter GVW HDGV's
are similar in configuration and usage to LDT's. These lighter HDGV's
2-42
-------
comprise over 75 percent of the HDGV fleet and would be easily adapt-
able to onboard control utilizing LOT components. There is also no
reason that onboard controls could not be applied to the heavier HDGV's
as well, although the cost would be higher than those for the lighter
models since the necessary control system components would be larger
and possibly more complex. Also, since the emission control systems
on these vehicles will not be as sophisticated as for the lighter HUGV's,
purge control and exhaust interaction effects would be more difficult
to address. Thus, as discussed previously in the Strategies Document,
it may be wise to evaluate the cost effectiveness of onboard controls
for various GVW HDGV classes as part of the decisionmaking process.
2.2.2 Emptying Losses
Comment: Several commenters stated that onboard controls would
not control service station storage tank breathing (emptying) losses,
which would reduce the overall onboard efficiency (I-H-74, I-H-82,
I-H-111, I-H-114, I-H-118).
Response: Fuel tank breathing losses (more correctly referred to
as emptying losses) occur when liquid gasoline in the service station
underground tank evaporates. This evaporation occurs because ambient
air enters the underground tank during each refueling to replace the
dispensed fuel, and this air becomes saturated with vapors. The rising
pressure in the tank causes an escape of vapors through tank vents.
While data and information in this area are very scant, EPA has assumed
that Stage II controls would reduce this evaporation since the refueling
vapors returned underground would at least partially saturate the
ambient air entering the tank.
However, this should be viewed as an increase in reductions for
Stage II, and not a decrease in onboard efficiency. Onboard effi-
ciency in controlling refueling losses is not decreased by emptying
losses. In fact, emptying losses, to the degree that they do occur,
would not be affected by onboard controls at all.
2.2.3 Gasoline Spillage
Comment: A number of commenters stated that the efficiency of
onboard controls would be reduced by fuel spills. These concerns were
primarily related to the effects of nozzle failures and tank over-
pressures (I-H-111, I-H-114). One commenter expressed concerns about
2-43
-------
fuel spills resulting from tank overfills also floodiny the carbo
canister (I-H-111).
Response: The above comments are predicated on the assumption
that a mechanical seal will be used to seal the pump nozzle-to-filIpipe
interface. However, EPA believes that most manufacturers would opt for
the liquid seal approach to avoid these potential problems. In the event
of a nozzle failure, with a liquid seal system the motorist or service
station attendant could see the liquid backing up in the fillpipe and
release the trigger on the pump nozzle to stop the flow. As Ford
correctly pointed out, this is the current practice with present conven-
tional fillpipes and results in "very little spillage" (I-H-114). EPA
sees no reason why spillage should be any greater with liquid seal
onboard control systems than with current conventional fillpipes.
Consequently, spillage was not a part of the efficiency calculation.
It is true that with mechanical elastomer seals, the tank could
be pressurized in the event of a nozzle shutoff failure, and that manu-
facturers may want to incorporate a pressure relief valve in the system
to forestall such eventuality. Moreover, the use of a mechanical seal
does not inherently create the potential for spillage. Some other
event such as a persistent topping off, vent blockage, or nozzle failure
must occur. In these cases, the pressure relief valve would be activated,
the operator would notice the spill, and fuel flow would be stopped.
In fact, it could be argued that a mechanical seal could reduce overall
spillage, because spillage due to premature shut-offs would be contained
within the fillpipe.
It should also be noted that EPA believes that the potential for
fuel spills due to nozzle failures is greatly overstated by the commen-
ters. Discussions with several nozzle manufacturers indicate t'h?t: nozzle
"failures" occur at very low dispensing rates (<2 gpm) and not at the
12 gpm dispensing rate indicated by one commenter. With normal refueling
technique, spills should not be an issue. Failures caused by persistent
attempts at topping off should be minimized by the use of a pressure
relief device on mechanical seal systems.
Concerns about fuel spills are also addressed by the onboard test
procedure being considered by EPA. In that procedure, a vehicle must
be refueled from 10 percent to automatic shutoff at a gasoline
2-44
-------
dispensiny rate of up to ID gallons per minute. Under tlr.. procedure,
any spills which occur would be measured as part of the refueliny
emissions. Since even a tablespoon of spilled yasoline evaporates
to about ID yrams of vapor, onboard systems will have to be desiyned
to ensure no spillaye. For this reason, it miyht be aryued that
onboard would reduce yasoline spillaye. At this point, a potential
increase in onboard efficiency due to decreased spillaye (U.lb y/yallon)
has not been considered in the onboard efficiency estimate (I-A-1UU).
Finally, EPA also expects that most onboard confiyurations would
incorporate a float valve in the vent line to the canister to prevent
overfilliny the tank. This float valve, plus the liquid/vapor sepa-
rator, would also act as a check valve to prevent floodiny the canister
under the circumstances mentioned by Hasstech.
2.2.4 Onboard in Current Staye II Areas
Comment: Three commenters stated that the EPA analysis credited
onboard controls with emission reductions that should be attributed
to Staye II controls that are already in place (i.e., in California
and Washinyton, D.C.) (I-H-1U4, I-H-114, I-H-127).
Response: The commenters were correct in noting that the £PA
onboard analysis issued in July 1984 included control of yasoline
consumption in California and Washinyton, U.C. Since those areas
already have Staye II in most areas (some portions of California do not
have Staye II), it could be considered as unnecessary control or double
countiny to include onboard for these areas. In the reanalysis, onboard
control for California vehicles was not considered in one scenario;
thus, a 49-state onboard approach is also evaluated. Not equipping
California vehicles with onboard controls is considered a feasible
option since California accounts for more than lu percent of new
vehicle sales each year, and California already has different emission
control requirements for vehicles reyistered there.
Un the other hand, Washinyton, D.C. which is a much smaller
area and represents a tiny fraction of the total hiyhway yasoline
consumption (less than U.2 percent as opposed to almost 11 percent for
California) -- has no separate emission control proyram and would not
be exempted from an onboard requirement (I-F-134). This would not
result in any significant duplication of effort, however, since Staye
II equipment in the District of Columbia is largely first yeneration
2-4b
-------
and is not particularly efficient even if well-maintained. Any credit
given to onboard for emission reductions from Staye II would be very
small and would not noticeably affect the overall effectiveness calculation.
Onboard efficiencies would therefore not include any significant areas that
were subject to Stage II controls. Thus, there would also not be any
appreciable double-counting of costs as a result of implementation of
onboard controls, as stated by Honda.
Similarly, the impact of expected Stage II reductions in the St.
Louis, Missouri, area -- where Stage II is just beginning to be installed
is not now (and will not be) great enough to significantly affect the
overall (i.e., national) effectiveness of onboard.
2.2.b Control System Tampering
Comment: Commenters who disagreed with EPA's estimates of new
system efficiency and in-use efficiency stated that EPA had under-
estimated or neglected to consider the effects of a number of sources
of reduced efficiency.
A number of commenters felt that EPA had not sufficiently con-
sidered the effects of deterioration and tampering on fillpipe seals
(I-H-2, I-H-11, I-H-22, I-H-b7, I-H-99, I-H-1UU, I-H-101, I-H-1U4,
I-H-114, I-H-116, I-H-127). A number of commenters felt that the use
of alcohol blend fuels would increase seal deterioration (I-H-1U1,
I-H-114, I-H-116, I-H-127). One commenter felt that fillpipe seals
could encourage tampering by making refueling operations more difficult
(I-H-99). One commenter felt tampered fillpipe seals would be harder.
to repair (I-H-2).
On -the other hand, several commenters felt fillpipe seals would
make tampering more difficult (I-H-41, I-H-46, I-H-71, I-H-76).
Other commenters felt that fillpipe seals might not be compatible
with existing Stage II nozzles, e.g., in California, and that this
issue should be addressed (I-H-114, I-H-118, I-H-127).
Response: Most of these concerns noted in the comments apply to
elastomer seals located in the upper part of the fillpipe. As stated
previously, EPA believes that the use of this type of seal will not be
common, with most manufacturers opting for the simpler and inherently
more reliable'liquid seal approach. The liquid seal approach would
obviate almost all of the aforementioned objections. Nevertheless,
2-46
-------
some manufacturers may choose the elastomer seal approach, so a response
to these comments will be made.
Design of an efficient seal would not be difficult. The API
system demonstrated in 1978 used an off-the-shelf seal with only minor
adaptation to the fillpipe (I-F-17). These seals have shown excellent
initial efficiency, on the order of 98 to 99 percent, and little if any
loss of sealing efficiency to 65,1)00 miles in a range of climatic
conditions.
There is no evidence that alcohol blend fuels would affect the
durability of an elastomer seal. In fact, one vehicle in the afore-
mentioned 1978 API program ran 12,OOU miles on an alcohol blend with
no effect on the elastomer seal. If contradictory evidence is found,
EPA believes that any such problems could be eliminated through proper
choice of seal material.
The EPA agrees with the commenters who stated that fillpipe
seals would make tampering more difficult. The elastomer seal would
be more difficult to remove than a leaded fuel restrictor flap, for
example. In some designs, the presence of the seal in the fillpipe
would essentially eliminate the need for a leaded fuel restrictor as
the fillpipe seal guide would have to be sized to be compatible with
the unleaded fuel nozzle. With the liquid seal approach, the seal
would be located in the tank rather than in the fillpipe, making it
impossible to tamper with or remove.
The EPA agrees that repairing a tampered elastomer seal would be
difficult, but no more so than repairing a tampered leaded fuel
restrictor would be, since replacement of the fillpipe would be
required in almost any imaginable instance.
Finally, seal compatibility with Stage II pump nozzles would be
no problem in California, if onboard were required in that State. With
the liquid trap approach, the top part of the fillneck would be virtually
identical to current uncontrolled fillnecks, so there would be no change
from current new vehicles. Even if an elastomer seal were to be used,
proper design and placement would assure compatibility with Stage II
refueling equipment. There is no reason why a properly designed onboard
system would be incompatible with Stage II pump nozzles.
2-47
-------
2.2.6 Canister System Deterioration
Comment: Several commenters stated that the adsorption capacity
of the activated carbon in an onboard system would deteriorate with
use, necessitating frequent replacement (I-H-58, I-H-90, I-H-111). One
commenter stated that canisters would either fail or their efficiency
would be reduced after repeated adsorption/desorption cycles due to
buildup of heavier hydrocarbons that are not easily desorbed (I-H-b8).
Another commenter stated that the efficiency of the canister would
decrease 5 percent with each adsorption/desorption cycle, thus necessi-
tating replacement at lU.UUO-mile intervals (I-H-111).
Response: The EPA recognizes that charcoal canisters undergo an
initial break-in period wherein a portion of the virgin working
capacity can be lost. However, following the initial buildup of this
"heel," data indicate that the working capacity of the canister
stabilizes and deteriorates very little, if at all, thereafter. As
explained'elsewhere in this document, EPA expects that manufacturers
will design their canisters with sufficient excess capacity to compen-
sate for heel buildup, and this requirement has been considered in the
onboard cost estimates. Claims of total canister failure after
1U,UOU-20,OOU miles or after 20 Adsorption/desorption cycles are clearly
refuted by both manufacturers' certification testing programs for evapora-
tive emissions and the results of EPA in-use evaporative emissions
testing programs. After development of the initial heel, canister deter-
V
ioration is negligible unless poisoning or malmaintenance/abuse occur.
2.2,7 Purge Effects on Efficiency
Comment: "A few commenters stated that a requirement for onboard
controls could add to the present problems involved in canister purging,
thereby reducing system efficiency (I-H-53, I-H-95, I-H-101). One
manufacturer maintained that, since desorption is a nonlinear process,
it is difficult to achieve control over the purging of present evapora-
tive emissions, particularly with the higher RVP fuels prevalent in
recent years. An onboard requirement would further compound the prob-
lems involved (I-H-101). Two commenters felt that more precise and
sophisticated control of purging would have to be developed, likely
including modifications to the control logic system (I-H-b3, I-H-1U1).
2-48
-------
Response: The EPA agrees that the purge requirements for onboard
systems will be different from, and somewhat greater than for, current
evaporative emission control systems, due to the increased volume of
vapors involved. However, given the current state-of-the-art in feedback
control systems, the Agency does not believe the problem is as difficult
as has been stated by Chrysler (I-H-101). Since the problem is essen-
tially one of purge air volume and mixture control, EPA does not believe
that new vapor sensor technology needs to be developed. Most present
sensors are capable of achieving the necessary control, and although
some modifications to purge control logic systems may be required, the
problems are essentially calibration problems rather than new design
problems. The fact that only two manufacturers commented on the issue
indicates that the industry as a whole does not consider the problem
involved to be a particularly troublesome one. The EPA believes that
sound engineering can eliminate any driveability or excess exhaust
emission problems and that there should be no increased incidence of
tampering.
For further discussion on the effects of purge on exhaust emissions
and driveability, the reader is referred to the onboard technological
feasibility analysis contained in this document (Section 2.1.6).
2.2.8 Overall Control System Efficiency
Comment: Several commenters felt that EPA's estimated onboard
efficiency levels were too niyn in view of recently disclosed problems
with in-use evaporative emissions (I-H-101, I-H-114, I-H-127). One
commenter stated that EPA's projected in-use efficiency of 98 percent
without tampering was inconsistent with the current evaporative control
system performance as modeled in MOBILES. The commenter suggested an
efficiency of 88 percent would be more appropriate because the EPA-
predicted in-use efficiency of 92 percent did not account for produc-
tion variation, defects not remedied through warranty or recall,
atypical operation which may produce system leaks, and canister poison-
ing by alcohol blend fuels (I-H-114).
Another commenter said that system effectiveness depends on fill-
pipe-to-nozzle interface and control technology design and that effi-
ciency could not be calculated until system design was finalized.
Nevertheless, this commenter predicted an efficiency of only about
2-49
-------
77 percent (I-H-116). Alony the same lines, another commenter main-
tained that EPA's projected control efficiency was based on tne results
of a few API tests of a vehicle with an incomplete system. This com-
menter suyyested an efficiency of yu percent or less, based on iy7b GM
comments, which in turn were based on 1973 SHED tests (I-H-127).
Kesponse: Some of the commenters did not make a clear distinction
between new system efficiency (the 98 percent fiyure projected by EPA),
which is the averaye efficiency of a new control system absent tamperiny
and deterioration, and in-use efficiency (the 92 percent EPA fiyure used
in the July 1984 analysis (I-A-bb), which does include potential deteri-
oration and tamperiny. Each of these factors is discussed below.
a. Theoretical Efficiency
The July 1984 EPA analysis projected a new system efficiency of
98 percent based on 1978 API and EPA test data on a system that used an
elastomer seal. Subsequent 1984-1985 EPA tests usiny a liquid trap
indicate new system efficiencies from 9b to 99 percent, with the majority
of the data falliny into the 9b to 98 percent ranye (i-A-93). Based on
this more recent and more substantial body of test results as well as
on the earlier data, it is reasonable to assume an averaye new system
efficiency of at least 97 percent. New control system efficiencies
niyher than this are clearly feasible (I-H-lb8). In their vehicle
demonstration proyrams, API has demonstrated prototype system
efficiencies of 99 percent.
For purposes of comparison, it is interestiny to note that these
results are consistent with new system efficiencies for evaporative
control systems. Krom the 1984 model year certification data, which
indicate certification levels for evaporative emissions ranyiny from
U.3 to 1.8 yrams per test, it appears that the mean certification
level would be about 1.4 to l.b yrams per test (hot soak plus diurnal).
Deterioration factors are included in these levels where applicable.
Given the MUBILE3 level for uncontrolled vehicles of about 4U yrams
per test, these certification levels would equate to new system
efficiencies of 96 to 97 percent, or possibly yreater, since the
certification data include deterioration.
The EPA recognizes that there will be a certain amount of produc-
tion variability, which could affect new system efficiency, but it will
2-bU
-------
be random in nature, with some vehicles exhibiting efficiencies that
are hiyher tnan the norm and others showiny lower efficiencies. As
lony as the vehicles in question are capable of meetiny the established
standard, EPA is not concerned about such variability. The effects of
such variability are accounted for in the Ayency's Selective tnforcement
Audit (SEA) and in-use audit proyrams and would also be considered in
the level of any onboard refueliny emission standard. If systematic
bias due to production defects should manifest itself, there are other
remedies available to the Ayency to ensure compliance.
In-use efficiency is calculated by applyiny appropriate factors to
the new system efficiency to account for deterioration, tamperiny, and
malmaintenance and defects. The July 1984 EPA analysis did not calculate
a separate deterioration factor for onboard emissions, under the presump-
tion that any deterioration would be insiynificant and would likely be
overwhelmed by the tamperiny rate, which was a composite of both canister
and fillpipe tamperiny.
As stated earlier, EPA now believes many manufacturers will utilize
the liquid trap approach, rather than an elastomer seal, which would
virtually eliminate fillpipe tampering as a source of control perfor-
mance deyradation. The likelihood of diminished fillpipe tamperiny is
also reinforced by the recent nationwide trend toward decreases in the
unleaded to leaded fuel price differential caused by EPA's lead phase-down
requirements (I-F-146). A lower differential reduces the incentive to
tamper. Therefore, only the charcoal canister and hose tamperiny rates
plus malmaintenance and defects in these areas will be used in calcu-
latiny the revised in-use efficiencies. Since the canister and hose
tamperiny rates are considerably lower than the composite rates used
previously, it is now deemed appropriate to consider usiny separate
deterioration rates.
b. Deterioration
Since onboard systems have only been demonstrated in vehicle pro-
totypes, there is still a scarcity of data reyardiny the deterioration
of onboard control systems. However, yiven the similarity of onboard
control systems to current evaporative systems, it would be reasonable
to use evaporative system deterioration factors (UF's) as a basis for
modeliny the deterioration in onboard systems. It was decided to limit
2-bl
-------
the data for modeling onboard UF's to evaporative DF data from fuel-
injected vehicles. Fuel-injected vehicles have no float bowl hardware,
as do carbureted systems, and consequently generate no significant hot
soak emissions. Therefore, the DF's for such vehicles would represent
only canister and hose deterioration and would not include any induction
system component deterioration.
The 1984 model year certification data indicate that the majority
of evaporative control systems from fuel-injected vehicles exhibit no
deterioration and that the DF's are very small for those families that
do show some loss of control (I-A-73). Of a total of 59 fuel-injected
LDV evaporative emission families and four fuel-injected LUT families,
only six had evaporative emission DF's. The average DF at midlife
(50,UUU miles for LDV's and 6U5UUU miles for LDT's) is 0.08 grams per
test for LDV's and U.052 grams per test for LDT's. The LDT DF was
extrapolated from the 50,000-mile data, since the useful life period
for LDT's changed for the 1985 model year. For all 59 families con-
sidered the average DF was less than u.Ul g/test. This clearly
demonstrates that canister and hose system deterioration is negligible
for wel 1-maintained vehicles. Over 9U percent of the families in
this sample had no evaporative emissions deterioration, which indicates
that zero deterioration is feasible and a reasonable assumption.
The potential effects of alcohol blends on canister durability
are discussed in detail in the technological feasibility discussion
presented earlier (see Section 2.1.4). The basic conclusion presented
there is that no evidence exists that alcohols have any significant
impact on canister efficiency or capacity. Thus no effects of alcohol
fuels on deterioration are considered here.
In addition to the possibility of system deterioration for well-
maintained vehicles, the effects of control system tampering and
malmaintenance and defects (M&D) must also be considered. These are
addressed separately below.
c. Tampering
Turning first to tampering, since many manufacturers are expected
to use one of the liquid seal approaches to seal the fillpipe, the only
other possible area for tampering involves the canister system and
related feed and purge hoses. Given the similarity of these portions
2-52
-------
of onboard refueling and evaporative control systems, canister and nose
system tampering for onboard systems was modeled using similar informa-
tion from tamperiny surveys for evaporative control systems. The
canister and hose system tampering rates used here are based on EPA
in-use tampering surveys (I-A-4U).
When analyzed using simple linear regression techniques, the
canister and hose tampering rates for LUV's and LUT's can be modeled as
presented below:
LUV's: Tamp = -1.U1 + U.U736(M)
LUT's/HUGV's: Tamp = 1.2U + U.U34y(M)
Tamp = Tampering incidence expressed in percent at
a particular mileage (M)
M = Mileage/lU.UUO
Absent any other information, the LL)T tampering rate can be used
to model HUGV tampering, since HUGV tampering data are not available
in the surveys. This is due to the fact that HUGV evaporative emissions
were uncontrolled prior to iy«b.
Assuming average lifetime periods of 1UU.UUU miles for LUV's,
12U.UUU miles for LUT's, and IIU.UUU miles for HUCiV's, the average (mid-
point) tampering rate for vehicles in each class would be 2.67, 3.2y,
and 3.12 percent, respectively. Un a fleet-weighted basis this
averages to about 2.B percent.
For purposes of emissions modeling, it was assumed that tamperiny
would completely disable the control system and the resultant control.
efficiency would be zero. This would happen if the canister were
removed 'or if the feed or purge lines were cut completely. This is
clearly conservative, since less severe forms of tampering would not
reduce system efficiency as completely. The effect of this tampering
on fleetwide emissions control was considered inherently in the modified
MUBILE3 fuel consumption model runs performed to support the onboard
analysis (I-B-37). The model considers, in addition to tampering, the
effects of fleet turnover, scrappage rates, changing vehicle fuel
economies, non-linear mileage accumulation rates, and all the other
factors that affect the percentage of consumption. As can be seen in
Table 2-3 in the portion of this document addressing phase-in of
controls (Section 2.b), tampering reduces the in-use efficiency of
2-b3
-------
onboard controls by about 3.8 percent over the long term. This is not
the same as doubling the 2.8 percent mentioned above, primarily because
more fuel is consumed early in the fleet life when tampering rates are
low. Finally, it should be noted that these tampering rate effects may
be substantially greater than actually occur, because placing the
canister in the rear of the vehicle or in the underbody will reduce
accessibility and therefore tampering incidences.
d. Malmaintenance/Defects
The final consideration with regard to onboard efficiency is
related to the effects of malmaintenance and defects (M&D) on refueling
emission control efficiency. The EPA has evaluated the rates of occur-
rence and evaporative emission effects of M&D as part of the in-use
emission factors test program, and this is presently the best information
available that can be used to estimate the impact on refueling emissions
control. This analysis will be limited to M&D rates and effects on the
fuel-injected vehicles tested, since the new motor vehicle fleet is
expected to be almost 90 percent fuel-injected into the 1990's and
beyond.
As shown in Table 2-2, the emission factor test program conducted
by EPA has identified seven M&D types, their rates, and effects on
emissions. Of these seven, two would have no effect on refueling emis-
sions controls and a third would be very unlikely to occur in onboard
equipped vehicles because a liquid/vapor separator will be used as
part of the system. This leaves only four categories for further eval-
uation: purge system problems, purge hose disabled, canister filter
dirty, and canister broken. As is shown below, these would have little
effect on the fleetwide efficiency of onboard controls.
One hundred sixty-three fuel-injected vehicles were tested in
the emission factors program. Of these 163, 12 had the M&D effects
mentioned above. The remainder were problem-free for purposes of this
analysis. Thus for commercial fuel, the effect is a 10.9 percent
increase in emissions over the problem-free vehicles:
3 (28.98) + 43.07 + 7 (20.15) + 1(15.2) + 151(9.63)
153 = 10.68 g/test
2-54
-------
Table 2-2. IN-USE EF TEST PROGRAM M&D TYPES, RATES OF OCCURRENCE, AND
DIURNAL/HOT SOAK EMISSIONS FOR FUEL-INJECTED VEHICLES
Defect
Gas Cap Leak
Air Cleaner
Gasket Broken/
Missing
Canister Filter
Di rty
Canister Saturated
Canister Broken
Purge System
Plugged/Damaged
Purge Hose Disabled
Problem-Free
Emissions
No. of
Vehicles
4
1
7
2
1
3
1
Rates*
%
2.5
0.6
4.3
1.2
0.6
1.8
0.6
Avg. Evap. Emissions (g/test)
Commercial (11.7 RVP)
DJ_ HS Total
No effect on refueling emissions
No effect on refueling emissions
14.49 5.66 20.15
Future problems fixed by liq/vap. sep.
2.14 13.06 15.2
18.50 10.48 28.98
4.92 30.15 43.07
7.56 2.07 9.63
*0ne hundred sixty-three fuel-injected vehicles tested.
2-55
-------
10.68 g/test with H&D vehicles
9.63 g/test problem-free = 1-109: a 10.9% increase in
emissions
The actual fleetwide increase could be less than 10.9 percent since
this approach assumes all canisters have filters which require replace-
ment. This is clearly not the case, since many manufacturers use closed
bottom canisters.
Since the 97 percent onboard efficiency is considered a problem-
free value, the increase in emissions shown above can be used to
estimate the effect on refueling.
Refueling emissions
after control: 5.9-g/gal (1-0.97) = 0.177 g/gal
With a 1U.9 percent increase, these emissions increase to 0.196 g/
gallon. A 0.196 g/galIon emission rate for controlled vehicles is
equivalent to an onboard efficiency of 96.7 percent instead of 97 per-
cent; that is,
0.196
Efficiency = 1 -"579 = 0.967
This demonstrates that while M£D may affect emissions on some
vehicles, the effect is negligible on a fleetwide basis.
e. Summary
In summary, it can be stated that onboard theoretical control
efficiency (whether the 97 percent used here or the higher value demon-
strated by others) may be reduced somewhat by in-use tampering effects.
However, given the data currently available on canister deterioration
and malmaintenance and defects, these effects are negligible overall.
2.3 EMISSION FACTOR FOR REFUELING LOSSES*
Comment; Several commenters felt that the refueling emission
factors used in the July 1984 EPA analysis were too low and suggested
alternative values.
One commenter stated that the AP-42 emission rate of 4.1 grams
per gallon (g/gal) was too low, considering that RVP values for gaso-
line have been rising steadily for a number of years and that this
*1984 Federal Register topic.
2-56
-------
would tend to increase the emission factor (I-H-21). Another commenter
(BMW) felt that the emission factor of 5.U to b.5 y/yal developed by
API was also too low, stating that the correct value was 7 y/gal for
straight gasoline and even higher for alcohol blends. If a manufacturer
designed a- canister based on the 5 gram value, the canister might not
be able to control the higher in-use emissions (I-H-2). Two auto
industry commenters stated that the value of 4.54 g/gal used by the
California Air Resources Board (GARB) was more representative than the
AP-42 value (I-H-114, I-H-127).
Response: The EPA concurs with the comments that the AP-42 and
API emission factors are both low, due largely to the fact that RVP
values have increased considerably in recent years, as the Maryland
Department of Health correctly pointed out. The EPA also agrees that
the CARS value is too low as a nationwide value, for the same reason.
Further investigations in this area uncovered a number of other liter-
ature sources on refueling emission rates. Some sources provided models
(equations) for calculating refueling emission rates based on fuel RVP,
dispensed fuel (TnJ, and fuel tank temperatures (Tt). However, in most
cases the models were developed using low RVP fuel and were not valid
in the range of interest for dispensed and fuel tank temperature.
Since there were no usable published data or models available, EPA
developed a formula for calculating refueling emission rates. This
formula (discussed above in Section 2.1.4) is based on a multiple
regression analysis of recent baseline test data for uncontrolled
vehicles (I-A-69).
In order to calculate emission factors, further information on
regional and national average values for the three main parameters
that determine refueling emissions, i.e., dispensed temperature (TQ),
tank temperature minus dispensed temperature (T^ - Tp, = T), and RVP
was necessary. Regional average values for these parameters were
determined from available sources and were weighted according to
regional fuel consumption in the U.S. in order to determine represent-
ative national average values. Using the emission factor formula and
these data, EPA has calculated an average nationwide uncontrolled
2-57
-------
emission factor of 5.9 yrams per yallon.* Tnis emission factor will be
used for all subsequent air quality and health exposure risk calculations.
In response to comments, BMW appears to confuse the emission factor
described above with the emission rate that would be produced by the
conditions specified in the draft refueliny test procedure. The company
is correct that the parameters found in the EPA test procedure,
representiny near worst-case conditions, would result in refueliny
emissions of approximately 7 yrams per yallon. This value is represen-
tative of the canister desiyn capacity necessary to enable manufacturers
to meet the certification standard for refueliny emissions. It would
not be appropriate to specify averaye conditions for the refueliny test
procedure, since to do so would mean that the resultiny systems would
be inadequate to control refueliny emissions in siynificant portions of
the country duriny peak emission periods. The test procedure require-
ments were chosen so that additional system capacity would be provided
(over that which would be required to control emissions in the ranye of
b.y y/yal). This additional capacity would be needed to ensure that
onboard systems are capable of control liny the majority of refueliny
emissions under some.of the most adverse conditions likely to be found
in the United States.
2.4 LEAD TIME
Comment: Several auto manufacturers and the California Air
Resources Board (CAKB) felt that 2 years lead time would be insufficient
to implement onboard controls, and that from 3 to 6 years would actually
be required. Some therefore felt that implementation by the 198b model
year was not feasible.
Une manufacturer stated that 3 to 4 years lead time would be
required due to system packayiny problems and the "ripple effect,"
i.e., a redesiyn of other components in the system beiny necessitated
whenever one component is chanyed (I-H-1UU). Two manufacturers stated
that 4 years lead time would be necessary (I-H-2). One of these
said the additional time would be required for development, repackaying
the system, and for retooliny all of the fuel tanks in the product -line
*Further discussion and information on how the refueliny emission fac-
tor was calculated can be found in I-A-69.
2-b»
-------
(I-H-114). The other commenter stated that additional time would be
required for development and testiny of necessary components that do
not currently exist (I-H-1U1). Another manufacturer stated that they
would need 6 years lead time due to the necessity for large scale
design chanyes and because of the difficulty of changing over all of
their models simultaneously (I-H-22). CAKb also felt that 4 years of
lead time would be required, based on tneir experience with a fill pipe
access rulemaking for Stage II (I-H-118). Une manufacturer felt that
lyay would be the earliest possible model year for implementation
(I-H-114). Another commenter stated that onboard controls could be
implemented by the 1987 model year, since the technology involved is
basically an extension of current evaporative emissions control systems
(I-H-llb). Petroleum industry interests who commented on the issue
felt that 2 years was adequate lead time, since onboard technology
already exists in current evaporative control systems (I-H-liy,
I-H-12U).
Response: With the exception of the five auto manufacturers noted
above, most manufacturers did not take exception to the 24 months lead
time put forth in the July 1984 analysis. The EPA recognizes that the
lead time requirements among the manufacturers may vary somewhat due to
differences in areas such as product line design, testing capabilities,
and the number of product offerings. However, as is discussed below, no
strong arguments were presented against the 24-month lead time estimate.
As was discussed in the comments, onboard systems are in many
ways similar to current evaporative emissions technology. With the
exception of the fillpipe seal and fuel tank valve(s) (rollover, vent
closure, fill limiter), no new technology is required and the design
and implementation issues are similar to those faced in meeting the
previous evaporative emission standards. In terms of currently avail-
able hardware, the components required are relatively uncomplicated and
many of them need only to be sized up from current equipment. With
regard to the two new components needed, three fillpipe seal approaches
have already been demonstrated by EPA, API, and various auto manufac-
turers, and a number of different fuel tank valve designs have already
been proposed by the auto industry and their suppliers (I-U-3iy, I-H-lb8)
The EPA therefore does not believe that an extensive amount of system
2-by
-------
design and development work would be required to bee n implementation
of onboard controls.
Further, while EPA recognizes that some vehicle redesign may be
necessary to accommodate refueling canisters on smaller vehicles, it
does not appear that an extensive vehicle redesign program will be
required overall. It should be noted that smaller vehicles generally
have smaller fuel tanks and would thus require smaller refueling
control canisters. Thus, the vehicle redesign burden may not be as
great as portrayed in the comments.
Given these arguments, EPA believes a lead time period of 24
months is reasonable. All previous evaporative emission standards have
been implemented with 24 months of lead time (both light- and heavy-
duty) and, given the complexity and magnitude of the task, 24 months
for implementing an onboard requirement appears reasonable.
The question of retooling was given consideration in the July 1984
EPA analysis. A period of up to 12 months was assumed, depending on
whether current tooling must be modified or whether new tooling must be
procured. The EPA did not consider retooling to be a critical path
item, however, since it could be carried on in parallel to final in-
vehicle testing of components and to certification. The 1984 analysis
referred primarily to tooling for control system components such as
fillpipe seals and fill-limiter valves, but there is no reason it could
not also be applied to necessary fuel tank modifications or to minor
vehicle modifications necessary for packaging the system. It should be
noted that many routine tooling changes are accomplished in periods of
far less than 12 months. For example, tooling changes accompanying
model year changeover are accomplished in only a few months.
The longer lead time estimates provided by CARB and several other
auto manufacturers appear to allow some accommodation for normal model
change, retirement, and new model introduction. This is clearly the
case in the Toyota and CARB comments (I-H-22, I-H-118), as evidenced by
other materials provided by these commenters (I-D-263, I-F-78). The
EPA expects similar considerations were included by the other commenters
as well, since the wide range in lead time estimates received in the
comments cannot be explained .by differences in product design, testing
capacity, etc. The CARB estimate was based on their experience in
implementing the fillpipe access requirement for Stage II. As was
2-60
-------
stated previously, EPA does not expect that widespread vehicle redesiyn
will be necessary as was required to comply with the California fill-
pipe access requirements. Therefore, EPA believes the lead time esti-
mates of 4 to 6 years are unrealistically long.
One lead time item in the July 1984 analysis may merit further
consideration in the future as more information becomes available. The
EPA assumed that 1U-12 months would be required for certification. This
estimate may need to be re-examined in the future since manufacturers
would be facing recertification of their entire product lines and may
also need to conduct some vehicle safety crash testing to assure
compliance with FMVSS 301. While it is not certain, some extension of
lead time may be warranted in order to accommodate the concentrated
recertification effort on the part of the industry. Alternatively, it
may be desirable to phase in controls by vehicle class or by some other
similar approach in order to distribute the certification burden over
more than one model year.
It should be noted that the date of final rulemaking could also
significantly affect the actual amount of lead time provided. Depend-
ing on the timing of the promulgation of the final rule, manufacturers
may receive up to 11 months additional lead time over the 24 months
that EPA suggested in the July 1984 analysis. Since the model year
usually begins in September, publication of the FRM after that month
would allow the manufacturers the remaining fraction of the model year
of publication, plus two full model 'years thereafter. Manufacturers
might, therefore, have considerably more than the 24-month minimum,
depending on the timing of the rulemaking process.
As was alluded to above, the actual model year of implementation
for an onboard requirement depends on when a final rule would be
published. Given the time necessary for developing a final rule plus
the minimum lead time, the earliest possible model year for an onboard
requirement is probably 1990 (see Figure 2-10). The NRDC comment that
onboard controls should be implemented by the 1987 model year appears to
be based on NROC'
-------
Safety-up to 14 MO,
Tooling!
NDM 1B-12 HA.
No Jit, 6-8 m. is
CUPf, 3M HO,
in n illinium
HI n n i in i
ro
Certification 18-12 MO,
DevelopMent;
Fro tote
Lab
IHelilcle
-6 HO,
HO,
-6 no,
i i i
FRH
4
8
12
Months
16
24
i::::::::::::::: Ignntgi
fl(liililitl|.
Figure 2-10. Onboard Leadtine
-------
on the implementation of onboard controls under Section 202(a)(6) of
the Act. Moreover, Section 202(a)(6) itself requires that any onboard
regulation provide the industry adequate lead time.
In summary, EPA does not believe that the comments provided
adequate basis to change the estimated lead time of 24 months. The EPA
estimates that the 1990 model year would be the earliest practicable
model year for implementation. As stated above, the industry might
also be afforded some additional lead time beyond the 24 months
estimated minimum, depending on when the final rule is promulgated.
Thus, the exact amount of lead time available would depend on the timing
of the final rulemaking. Nevertheless, EPA remains open on the leadtime
issue and is willing to consider a phase-in or an increase in the amount
of leadtime provided based upon information provided in response to the
proposal.
2.5 PHASE-IN UF CONTROLS
Comment: Automotive industry and other commenters who favored
Stage II pointed to the slower implementation of onboard controls
(compared to Stage II) and stated that additional problems could cause
the phase-in rate for onboard to be even slower than that stated in
the July 1984 EPA analysis. Several commenters aryued that onboard
controls were less effective than Stage II because the full benefits of
onboard would not occur until 10 to 14 years after implementation
(I-H-22, I-H-93, I-H-99, I-H-100, I-H-101, I-H-104, I-H-114, I-H-115,
I-H-117, I-H-118, I-H-124, I-H-127, I-H-128). One commenter felt that
phase-in would be even slower due to tampering and the increasing
age of the fleet (I-H-57). Another felt that implementation of
onboard controls would result in an unwarranted delay in the control
of refueling emissions .(I-H-74).
Conversely, petroleum marketers and others who favored onboard
controls generally felt that EPA had been too conservative in esti-
mating the rate of onboard phase-in, stating that Stage II programs
have their own implementation problems and might not be phased in any
more quickly than onboard controls. Two commenters stated that the
accelerated lead phasedown rule would virtually eliminate fillpipe
tampering, which in turn would speed the phase-in of onboard controls
due to higher in-use efficiency. They further stated that EPA had
2-63
-------
overestimated the fuel economy of future vehicles. Since the amount of
control varies inversely with fuel economy, lower, more realistic fuel
economy figures would result in more yasoline consumption subject to
control (I-H-10b, I-H-119, I-H-12U). Some commenters also felt that
problems in develop ny State and local standards and shortayes of
equipment and contractor expertise would delay the implementation of
nationwide Staye II controls, possibly to as lony as 7 years (I-H-y4,
I-H-99, I-H-1U2, I-H-1U8, I-H-1U9, I-H-119, I-H-12U, I-H-124).
Response: The comments on the phase-in of refueliny controls
can be separated into two areas: (1) Staye II phase-in rate versus
onboard phase-in rate and (2) assumptions for determininy the phase-
in rate for each control technoloyy.
In yeneral, EPA concurs with the commenters who claimed that
Staye II could be implemented more quickly than onboard. It is esti-
mated that, under optimum conditions, Staye II could be implemented
fully in about 3 years. In response to comments on the lacK of avail-
ability of this equipment, the Ayency contacted Staye II equipment
manufacturers, who indicated that meetiny the expected demand could be
a done (I-E-1U, I-E-11, I-E-12). However, yiven the need for enabliny
leyislation in some cases, plus the potential for some shortayes of
trained installers and needed equipment, it is very possible that Staye II
implementation could take considerably lonyer than 3 years. Thus, a
period of 7 years was considered as a lonyer time estimate (3 years for
non-independents, 7 years for independents) for nationwide strateyies
and b-1/2 years for nonattainment area strateyies for both independents
and non-independents. An analysis conducted by Radian Corporation
under contract to API supports the implementation ranye presented above
(I-H-lsl).
Since the phase-in of onboard controls is a function of fleet
turnover, the period to achieve full implementation is lonyer than
Staye II. In an effort to better evaluate the phase-in rate of on-
board, EPA has developed a new, more sophisticated hiyhway fuel con-
sumption model that incorporates updated fuel economy, sales, and
scrappaye rate data. This new model, based on EPA's MUBILE3 emissions
2-64
-------
model, provides a more accurate estimate of the rate of increase of
controlled consumption than was provided by the less sophisticated
model used in the July 1984 EPA analysis (I-A-99, I-B-37). As shown in
Table 2-3, the new model indicates that control of bu percent of total
highway gasoline consumption would be achieved within 5 years of im-
plementation and that more than 75 percent control would be achieved
within 10 years. These results are similar to those originally projected
in the July 1984 analysis.
The new analysis assumes no fillpipe tampering for two reasons.
First, many manufacturers are likely to utilize the tamper-proof
liquid seals. Second, for those onboard systems that do employ mechanical
seals, the economic incentive for fillneck tampering has decreased
substantially as a result of the decrease in the unleaded/leaded price
differential. This leaves disablement of the canister system as the
only possible form of tampering. Canister system tampering includes
tampering with the canister as well as all valves and vapor hoses
related to the load and purge functions of the system. Canister
system tampering rates of 2.67 percent and 3.29 percent, respectively,
were assumed for LDV's and LDT's, based on the 1982 through 198b NEIC
Tampering Surveys (I-A-40). These replace the composite tampering
rates used for the calculations in the 1984 analysis.
Details on the fuel economy, scrappage rates, dieselization rates,
and other values used in the model are shown in documents contained
in the public docket (I-A-99, 1-8-37). While the accuracy of any
projection which is carried many years into the future, such as these
are, is problematic, the impact of any errors on the onboard phase-in
rate is not significant. For example, EPA considered a scenario where
fuel economies were 1U percent worse than projected in the model, yet
this had no effect on the .phase-in rate (I-B-21). The primary reason
that the model is not sensitive to small errors in the projected values
is that both the controlled and total consumption are affected about
equally. Thus, the controlled fraction is affected only a small amount.
Given this information on the revised analytical approach for
determining onboard and Stage II phase-in rates, the phase-in
rates and effectiveness for Stage II and onboard controls can now be
compared. For onboard, the phase-in rate is the same for a nationwide
2-65
-------
Table 2-3. PHASE-IN UF ONBOA-D REFUELING CONTROLS
Year
1989
19 9L)
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2U01
2UU2
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Annual
Consumption,
109 gal
81.009
79.901
78.797
77.686
76.575
7b.498
74.384
73.425
72.641
71.986
71.442
71.023
70.996
70.998
71.048
71.089
71.151
71.205
71.247
71.301
71.346
71.395 - ...
71.439
71.481
71.521
71.561
71.598
71.663
71.668
71.701
71.735
71.768
Tampered
Consumption,
109 gal
0.000
0.062
0.180
0.344
0.537
0.742
0.952
1.163
1.368
1.563
1.747
1.914
2.063
2.197
2.311
2.405
2.481
2.534
2.576
2.609
2.633
2.650
2.660
2.662
2.665
2.666
2.667
2.669
2.669
2.669
2.669
2.669
Control led
Consumption,
109 gal
0.000
9.749
18.276
.25.741
32.202
37.781
42.502
46.617
50.220
53.389
56.113
58.465
60.715
62.647
64.226
65.508
66.490
67.187
67.736
68.136
68.429
68.642
68.737
68.802
68.844
68.884
68.922
68.959
68.995
69.030
69.066
69.099
Potential
Fraction
Control led
0
0.122789
0.234222
0.335775
0.427542
0.510252
0.584185
0.650732
0.710177
0.763371
0.809888
0.850133
0.884247
0.913322
0.936508
0.955324
0.969361
0.979159
0.986877
O.y92202
.0.996019
0.998557
0.999692
0.999762
0.99832
0.999846
0.999874
0.999512
0.999665
0.999972
0.999986
1
Actual
Fraction
Control led
0
0.122013
0.231938
0.331347
0.420529
0.500424
0.571386
0.634893
0.691345
0.741658
0.78b434
0.823184
0.855189
0.882377
0.903980
0.921493
0.934491
0.943571
0.950721
0.955611
0.959115
0.961440
0.962457
0.962522
0.962570
0.962591
0.962625
0.962268
. 0.962434
0.962748
0.962780
0.962810
2-66
-------
or nonattainment area program. For Stage II, the implementation schedule
is different for nationwide and nonattainment area programs. Also, it
should be noted that the Stage II analysis allows exemption for non-
independent stations with a throughput less than 10,000 gallons per
month and independent stations with a throughput of less than 50,000
gallons per month.
Table 2-4 provides a comparison of control effectiveness for on-
board and Stage II in both nationwide and nonattainment areas. For
onboard, the control effectiveness rates shown in Table 2-3 were used.
For Stage II, the implementation periods discussed above were used,
together with the implementation rate from the draft Regulatory Impacts
Analysis (RIA) Vol. I. A control efficiency range of 62-86 percent was
used for Stage II.
Table 2-4 shows that for a nationwide program, Stage II has no
long-term advantage over onboard. For the more optimistic phase-in
schedule and higher in-use Stage II efficiency, Stage II has a higher
control effectiveness than onboard until the eighth year. At that
point, the lower efficiency of .Stage II and the exemptions assumed yield
less reductions than onboard. For the longer Stage II phase-in and
lower in-use efficiency, onboard is always better.
For a nonattainment area program, 3-year and 5-1/2-year phase-in
schedules were used for Stage II. For the higher Stage II efficiency,
Stage II clearly has a short-term advantage. Onboard achieves greater
reductions in the eighth-year. However, for the lower Stage II effi-
ciency, onboard will achieve greater reductions from the outset. For a
longer implementation period in nonattainment areas, Stage II has no
short-term benefit over onboard.
Thus, while Stage II controls could be implemented more quickly
than the motor vehicle fleet could be equipped with onboard controls,
it is not clear that this provides any meaningful long-term advantage.
Onboard control effectiveness reaches the most optimistic Stage II
values about 4 years after Stage II and from that point, the reductions
from onboard surpass Stage II. Onboard eventually captures about 93
percent of all refueling vapors, while Stage II captures only 48-66
percent, including some emptying losses. Thus, in the long-term, the
onboard strategy has the advantage.
2-67
-------
Table 2-4. COMPARISON OF PHASE-IN CONTROL EFFECTIVENESS
(% OF AVAILABLE REDUCTION)
Years
1
2
3
4
5
6
7
8
y
10
Onboard3
12%
22
32
41
49
55
62
67
72
76
Nationwide Stage IIb
3 yrs/86%
15%
45
62
64
64
64
64
64
64
64
7 yrs/62%
7%
20
34
41
43
44
46
46
46
46
61 NA Area Stage IIb
3 yrs/86%
28%
55
63
65
66
66
66
66
66
66
b-1/2 yrs/62%
5%
13
22
3U
39
46
47
48
48
48
aControlled consumption from Table 2-3 x 97% theoretical efficiency.
bAssumes a 10/50 exemption level and a linear phase-in rate.
2-68
-------
2.6 COST UF UNBOARD CONTRULS*
Comment: Many of the comments addressed the costs of onboard
control systems. The comments on cost can be broadly grouped into two
categories. The first category consists of those comments claiming that
the cost range estimated by EPA ($lb-$2b) adequately represents the cost
of control. Most of the commenters in this group have a direct interest
in the gasoline marketing industry (e.g., service station owners, oil
companies). The second category consists of those comments that claimed
that the EPA estimate was low and should be modified upward. Most of
these commenters are automobile manufacturers or are representing areas
of the country where Stage II controls are already in place.
The comments that generally agreed with the EPA cost estimate
varied widely in terms of complexity and detail. The large majority
of these commenters simply claimed that the onboard system cost would
fall within or below the range proposed by EPA and gave little
supporting evidence. The cost estimates that were given ranged from $9
(I-H-1) to $17 (I-H-120). The comments submitted by the American
Petroleum Institute (API) (I-H-120) were the most detailed of those in
this group, and the API work is referenced in many of the other comments.
The API comments rtferred to the cost of refueling emission control
systems assembled and tested on vehicles during their 1978 demonstra-
tion program. A comparison of the cost estimates made by EPA, API,
and L.H. Lindgren is also presented as part of the API comment package.
This cost comparison generally supports the EPA estimates.
The comments that disagreed with the EPA cost estimate also varied
widely in complexity and detail. The Ford Motor Company comment (I-H-
114) addressed the issue of the cost of onboard control in more detail
than did any other comment. In their comments, Ford proposed a
refueling control system which they believed would function adequately
under all in-use conditions and attempted to give a detailed estimate
of costs involved in equipping their cars and trucks with such a control
system. Ford's weighted average cost to the consumer for onboard con-
trol is $b'3 per vehicle.
*1984 Federal Register topic.
2-69
-------
The comments submitted by General Motors (I-H-117) also included
a relatively well supported estimate of the costs of control. GM
provides a component-by-cornponent cost comparison as well as an overall
pen-vehicle cost to the consumer. GM estimates the consumer cost of
control to be at least $30. This is the lowest estimate submitted by
an auto manufacturer.
Chrysler Corporation included a very detailed description of a
prototype onboard control system in their comments (I-H-1U1). The
function of each component in the system and of the system in general
is clearly outlined in the Chrysler document. Chrysler also gave an
estimated per-vehicle cost of control - $8b. Chrysler did not attempt
to provide a component-by-component breakdown of this cost, however,
and it is thus very difficult to assess the appropriateness of this
estimate.
Seven automobile manufacturers, other than the three referred to
above, submitted comments which included an estimate of the per-
vehicle consumer cost of onboard control of refueling emissions.
None of these commenters, however, supplied a detailed derivation of
its cost estimate. Therefore, the commenters and their respective cost
estimates are simply listed below:
Commenter Cost ($/Vehicle)
American Honda Motor Company (I-H-1U4) <90
American Motors Corporation (I-H-128) _<50
BMW of North America (I-H-2) 65
Mazda (North America) (I-H-53) 60-75
Toyota Technical Center, USA (I-H-22) 70-1UO
Volkswagen of America (I-H-116) 72
VoTvo-North'American Car Operations
(l-h-100) 60
Within the general area of control system costs, certain issues
are noteworthy because they were raised by more than a single com-
menter. One comment frequently made was that EPA should have included
the cost of a device or devices to provide fill-limiting capabilities
and rollover protection. The comments noted that the cost of the system
described in the July 1984 analysis did not include any provisions for
these requirements.
2-70
-------
Several commenters claimed that adding onboard control systems to
vehicles could negatively impact the performance of exhaust emission
control systems. They went on to assert that there would be costs
associated with recalibration or other changes needed to insure that
onboard equipped vehicles would comply with exhaust emission standards.
The cost estimate in the 1984 EPA analysis did not include any cost
directly associated with exhaust emission compliance.
Several commenters also discussed the refueling test procedure.
Commenters pointed out that at the time of the publication of the July
1984 analysis, EPA had not proposed any test procedures or emission
standards to be used in determination of compliance with an onboard
regulation. These commenters claimed that without this information on
test procedures and standards, it is difficult for them to accurately
assess the complexity of the system that would be needed to control
refueling emissions. Because of the uncertainty in system designs,
they claim it is very difficult to estimate system costs. One auto
manufacturer made the related comment that onboard system costs would
be heavily influenced by the values chosen for the critical test
procedure parameters such as fuel temperature and volatility (I-H-127).
Costs associated with ensuring the safety of onboard systems
using a mechanical fillneck seal was another area of frequent comment.
Commenters noted that overpressurization of a fuel tank during a
refueling event (caused by a fueling nozzle malfunction) could lead to
fuel being forced from the fillneck during or after the refueling
event. This could result in fuel squirting onto the nozzle operator
or onto the ground. The commenters claimed that a pressure relief
de"ice would be needed to avoid these safety concerns. They went on
to point out that no such component was included in EPA's original
onboard cost estimate.
In addition to the areas of comment listed above, some important
cost issues were raised by only one or a few commenters. These included
costs for: (1) onboard system maintenance, (2) taxes and insurance,
and (3) enforcement of an onboard regulation. . .
One final area that was frequently raised in the comments was the
magnitude of the factor used by EPA to mark system costs at the vendor
level up to the Retail Price Equivalent (RPE), or cost to the consumer.
2-71
-------
Because this issue was raised quite frequently, the comments on the
markup factor will be addressed separately from the other cost-related
comments.
Response: The EPA has examined and evaluated all of the comments
related to the cost of onboard control of refueling emissions and has
developed a revised cost estimate based on this evaluation. The most
thorough and well documented comments on cost were submitted by com-
menters claiming that the original EPA cost estimate was too low. As
was discussed in the technological feasibility section of this document,
these comments, in combination with the results of EPA testing of
refueling vapor control systems at the Motor Vehicle Emissions Labora-
tory, have convinced EPA that the refueling control system described in
the July 1984 EPA analysis could be improved in many ways.
The technological feasibility section of this study (Section 2.1.3)
contains a description of an onboard control system that was designed
as an improvement to the system described in the July 1984 analysis.
The improvements in the system responsd to comments received on the
cost and feasibility of onboard control systems. The remainder of the
response to comments on onboard system cost (excluding comments on
markup factors which are discussed separately) is a detailed breakdown
of the cost of the system described in the technological feasibility
section of this chapter. First, the system costs are given for three
vehicle categories (LUV's, single and dual tank LDT's) along with a
breakdown of component cost estimates. Then, for each cost component,
the source of the cost estimate is given, along with a description of
any calculations used in arriving at the estimate. Finally, if available,
alternate cost estimates are quoted and evaluated.
2.6.1. Onboard Control System Costs
Table 2-5 summarizes the calculation of the onboard control cost
estimates for light-duty vehicles (LDV's), light-duty trucks (LDT's),
and light-duty trucks with dual fuel tanks. For each vehicle cate-
gory, the values of cost components for both integrated and separate
systems are presented. Two long-term costs of onboard control are
also presented at the bottom.of Table 2-5. The 1994 cost estimate
reflects three differences from the 1989 cost estimate: (1) the absence
2-72
-------
Table 2-5. ONBOARD CONTROL SYSTEM COSTS
LDV
Activated Carbon
Canister Body
Vapor Vent Valve
Vapor Line
Fill Limiter
Liquid/Vapor
Separator
Fillneck Extension
ECO/Purye
Modifications
Tank Modifications
Packaging
Assembly
Certification
Facilities
Modifications
Vendor Cost
Markup by 1.26
Systems Engineering
Engineering
LOT - Dual
Total R.P.E. Cost
198y
1994
2000
Integrated
2.20
2.24
4.60
1.44
0.82
0.73
1.21
0.07
0.50
0.50
___
0.61
0.3U
15.22
19.18
0.45
TO!
Separate
2.20
3.21
4.60
2.40
0.82
0.73
1.21
0.22
0.50
0.50
0.75
0.61
0.30
18.05
22.74 .
0.45
7J7T9"
20.00
17
16
.70
.30
LDT
Integrated
2.94
2.84
4.60
1.44
0.82
0.73
1.21
0.07
0.50
0.50
0.77
0.30
16.72
21.07
0.69
21.76
22
19
18
Separate
2.94
3.84
4.60
2.40
0.82
0.73
1.21
0.22
0.50
0.50
0.75
0.77
0.30
19.58
24.67
0.69
2!>-36
.20
.60
.10
DOAL
Integrated
5.88
6.67
9.20
5.12
1.64
1.46
2.42
0.14
1.00
0.50
0.75
0.77
0.30
35.85
45.17
0.69
0.25
46.11
46
42
40
Separate
5.88
7.85
9.20
5.76
1.64
1.46
2.42
0.44
1.00
0.50
1.50
0.77
0.30'
38771
48.79
0.69
0.25
49.73
.60
.70
.50
2-73
-------
of modification costs, which would drop out as vehicle designs incorpor-
ate onboard controls and previous costs are amortized, (2) the absence
of certification and systems engineering costs after these costs have
been amortized, and (3) improvements in fuel economy. The cost for the
year 2UUU represents the cost of control after a number of the first
vehicles with onboard control have left the fleet and reflects further
improvements in fuel economy and the absence of facilities modification
costs. The onboard cost items in Table 2-5 are discussed below.
The hardware costs shown in Table 2-b can be characterized as EPA's
best estimate at this time. To be conservative, the higher end of the
cost ranges were used when a range was given for a component.
a. Activated Carbon
One of the vital functions of an onboard control system is the
collection and storage of hydrocarbon vapors displaced during a refuel-
ing event. The first item in Table 2-5, activated carbon, is used as
the hydrocarbon storage medium for current evaporative emission control
systems.
The activated carbon used in the onboard control system tests at
MVEL was Westvaco extruded activated carbon. This type of carbon was
chosen because it has a relatively low cost, a high working capacity,
and provides a low backpressure during adsorption.* The cost estimates
shown in Table 2-5 are based on the characteristics of this carbon and
include capacity for refueling vapors only. The specifications for
this carbon are shown below:
Carbon base Wood
. Butane working capacity lUb g/liter C
Apparent density 3UU g/liter = U.661 Ib/liter
High volume cost to vehicle $1.4U/lb
manufacturer
The cost and working capacity figures were quoted by Westvaco
representatives. The apparent density figure is a compromise between
the density quoted by Westvaco (32U-340 g/liter) and the apparent density
as measured at MVEL (26U g/liter, loosely packed). It should be noted
*This is not meant as an endorsement of this product or its manufacturer.
2-74 .
-------
that each of these figures will vary with the carbon base, mesh size,
and manufacturer chosen.
Ford Motor Company asserted that the cost of activated carbon was
$2.01) per pound. Ford may be using a more dense product (more working
capacity per pound of product) or a more expensive product. Other car-
bons have been quoted at only $1.00 per pound. Since EPA has been
given a quote of $1.40/1b for a carbon that appears to function ade-
quately, this cost has been used in this cost estimate (I-B-27).
The first step in finding appropriate carbon bed sizes for various
fuel tank sizes was to develop a refueling emission rate. This rate,
when multiplied by an appropriate fuel tank size, gives the total mass
of hydrocarbon that is expected to be displaced from a typical fuel
tank during the refueling capacity portion of EPA's proposed test pro-
cedure. The emission rate used by EPA was found using a regression
equation generated from a series of over 100 uncontrolled refueling
tests done on a number of vehicles. The uncontrolled test program and
the development of the regression equation are fully discussed in an
EPA technical report entitled "Refueling Emissions from Uncontrolled
Vehicles" (I-A-6y). When the refueling test procedure conditions
(dispensed temperature = 88°F,AT = 5°F, 11.5 psi RVP fuel) are evaluated
in the regression equation, the emission rate it,provides is 7.0 g/gal
of fuel dispensed. In other words, the equation predicts that for each
gallon of 11.5 psi fuel dispensed at the conditions specified in the
proposed refueling test procedure, approximately 7.0 grams of hydro-
carbon vapor would be displaced. (At 9.0 psi RVP the emission rate.is
5.8 grams per gallon.)
The emission rate described above is used in combination with the
gasoline working capacity of the chosen carbon to find necessary
carbon bed size. The working capacity quoted above is a virgin butane
working capacity and thus reflects neither carbon aging nor the differ-
ence between adsorption characteristics of butane gas and gasoline
vapor. The working capacity for aged carbon is approximately 60 percent
of the virgin working capacity. This figure was supplied by a manufac-
turer of activated carbon. In this analysis, the virgin working
capacity was reduced by 50 percent to reflect aging and a moderate (lu
percent) factor of safety. Gasoline vapor working capacities are generally
2-75
-------
estimated as some fraction of butane working capacities. The fractions
quoted, depending on the source range from 60-85 percent (I-A-74, I-B-27).
Sixty percent was used in this analysis in the interest of being conser-
vative. Using the two fraction discussed above, the gasoline working
capacity for aged carbon was calculated as shown below:
105 g/liter x 0.50 x 0.60 = 31.5 liter
The amount of carbon needed to control the vapor associated with
each gallon of fuel dispensed was then estimated by dividing the emission
rate (7.0 g/gal) by the aged, gasoline working capacity:
7.0 g/gal = 0.222 liter C
31.5 g HC/liter C gal
The cost of this amount of carbon was found using the apparent
density and price per pound for this carbon:
(0.22 liter C/gal) x (0.66 Ib C/liter C) x $1.40/lb = $0.20/gal.
Per-vehicle carbon costs were estimated using average LDV and LUT
fuel tank sizes. Fuel tank sizes were calculated by assuming a single
tank driving range of 300 miles and dividing by fuel economy estimates
for the years 1989, 1994, and 2000. These fuel economy estimates were
taken from EPA's MOBILES fuel consumption model (I-A-99, I-B-37).
Table 2-6 shows the calculation of activated carbon cost for LDV's and
LDT's for 1989, 1994, and 2000. Because the test procedure requires
only a 90 percent fill, nominal fuel tank capacities were multiplied
by 0.9.
The comments submitted by Ford indicate a higher carbon cost than
that developed-by EPA. Ford Motor Company uses a lower working capacity
(6.7 g HC/100 mg carbon vs. 10.5 g HC/100 g carbon), a lower refueling
emission factor (5.2 g HC/gal vs. 7.0 g HC/gal) and a higher carbon
cost ($2.00/lb vs. $1.40/15.) than does EPA. The difference in the
working capacities could be due to the use of a different carbon, or to
a different approach to calculating working capacity, or a combination
of both. Ford also uses a larger fuel tank capacity (20 gal) than does
EPA (LDV-12.2, LDT-16.3).
The methodology used in this analysis to estimate the size of the
canister needed for onboard control systems is somewhat irregular in that
it was assumed that the activated carbon needed for refueling emission
2-76
-------
control would be added to the existiny evaporative control capacity. Tne
canister siziny was done very early in the analysis process when liPA had
little information witn which to estimate necessary carbon bed volumes,
and the additive methodoloyy was used to ensure that an ample carbon
volume was used.
Since the time when the carbon bed volumes were estimated, tPA has
collected more information on carbon canisters. This information, which
was presented to industry representatives at an April ID, 1986 workshop,
has some implications on the canister siziny question, and clearly shows
that a tradeoff exists between the air flow rate used to purye the
canister and the effective workiny capacity of a given canister. The
tPA used the information on the purye response characteristics of
several activated carbon canisters to estimate the ranye of canister
size needed for control of refueliny and evaporative emissions. The
canister sizes developed previously by EPA (Table 2-b) fell within this
ranye.
b. Canister body
Haviny estimated carbon bed sizes, it is now possible to develop
costs for carbon canister shells. The cost estimates made in this
analysis are based on work done for API and MVMA by Leroy H. Lindyren in
1983 (I-U-269). Lindyren beyan this work for API in 1983 and revised it
under contract to MVMA in 1984 (the work for MVMA was included in their
comment packaye (I-H-127)). In the 1983 study, Lindyren estimates the
manufacturiny cost of an 8bU ml evaporative canister. In this study
it is assumed that Lindgren has accurately estimated the cost of an
8bu ml canister shell and that the cost of larger canisters can be
found by scaling the costs by the ratio of the canister surface areas.
In order to simplify this analysis, it was assumed that all
canisters are cylindrical in shape with equal dimensions of height
and diameter. Usiny this assumption, it was possible to calculate a
surface area for any canister volume, usiny only the formulas for the
volume and surface area of a cylinder with closed ends:
2 d3 (d=h)
V = TTT h =^-4 (volume)
y ? ^ 2
As = 2 TT r + rrdh = n
-------
Table 2-6. CALCULATION OF CARBON BED VOLUME AND COST
Calculation
Parameter
Driving Range (mi)
Fuel Economy (mi /gal )
Fuel Tank Capacity (gal)
90% Fill (gal)
Carbon Bed Volume (ml)
Carbon Cost ($)
1989
300
24.61
12.2
11.0
2,440
2.20
LDV
1994
300
26.64
11.3
10.1
2,240
2.02
2000
300
29.13
10.3
9.3
2,060
1.86
1989
300
18.43
16.3
14.7
3,260
2.94
LOT*
1994
300
18.99
lb.8
14.2
3,150
2.84
2000
300
20.56
14.6
13.1
2,900
2.62
*The capacity and cost of carbon needed to control refueling emissions for LDT's
with dual fuel tanks was estimated by doubling the LOT values.
2-73
-------
The volume formula was used to solve for the diameter of the
canister (equal to the height according to the assumption discuss, d
above). The area formula is shown to demonstrate that the ratio of
two surface areas is equal to the ratio of the square of their diameters,
or equivalently their radii:
l.Sud2, d2 r2
l.STTd'2 = d1"2 = "P"2
In the Lindyren cost analysis, the vendor selling price of the
canister shell can be broken down into three parts; component costs,
assembly costs, and overhead and profit markups. In scaling the
cost of the 850 ml canister, the component cost is the portion of the
vendor cost that was scaled by the ratio of the surface areas. The
EPA does not expect the cost associated with assembling a large canister
to be any different from that associated with assembling a smaller
canister and has accordingly applied an assembly charge of $U.24 to the
component cost of each canister. Finally, because the markup factor
used here is multiplicative, the change in canister costs associated
with the difference in canister size will be magnified by the vendor
markup factor of 1.4.
A final complication in the estimation of the canister cost is
that a vehicle with an integrated evaporative/refueling system will
replace the smaller evaporative emission control camster(s) with a
single larger canister to control both evaporative and refueling emis-
sions. The removal of the evaporative canister implies a lower overall
cost, since the cost for an integrated evaporative/refueling canister
would be less than the cost of two smaller separate canisters whose
combined volume equaled that of the integrated canister.
In"order to calculate the value of this credit, evaporative emis-
sion control canister sizes must be evaluated for each vehicle class
of interest. In this analysis, it has been assumed that light-duty
vehicles and light-duty trucks currently use an evaporative emission
control canister of 850 milliliters, and light-duty trucks with dual
fuel tanks use a single canister of 1.5UO milliliters volume. Although
these canister sizes may be smaller than the average evaporative
canisters in use today, the use of a small canister leads to a more
conservative estimate of the onboard system cost. This is because
the smaller evaporative canister size leads to a lower value for
2-79
-------
the credit. This is offset somewhat by the fact that the cost of the
integrated canister shell would be marginally less with a smaller
evaporative capacity (see Table 2-7). Therefore, the assumption of
small evaporative canister sizes is reasonable.
The calculation of canister shell costs is detailed in Table 2-7.
In general, there were 6 steps used in the calculation of canister
shell costs. First, the calculated refueling capacity was added to the
assumed evaporative canister capacity to give a total onboard canister
capacity. Then, the square of the radius of a cylinder of the appropriate
volume was calculated. The ratio of this squared radius to that of the
85U ml. canister costed by Lindgren (26.35 cm2) was then used to scale
the component cost of the 850 ml canister to the appropriate amount. To
this calculated component cost, $0.24 was added for assembly to give a
total manufacturing cost. Next, the manufacturing cost was multiplied
by 1.4 to account for overhead and profit at the vendor level. Finally
the total vendor selli.ng price was multiplied by 1.U3 to express the
cost in terms of 1984 dollars. Although the calculation of the canister
shell cost for separate systems .is somewhat different, the detailed
calculations in Table 2-7 should adequately demonstrate the differ-
ences. Table 2-8 shows the canister costs for each system evaluated in
this analysis.
Ford Motor Company and General Motors were the only two commenters
that gave estimates of total canister costs (including carbon). Ford
estimated the total canister costs to be about $7.50 per vehicle (LDV
and single tank LUT) assuming a refueling emission factor of 5.2 grams
per gallon. General Motors estimated the canister costs for a typical
passenger car to be $5.81. The EPA's estimated average canister costs
are $4.56 and $5.91) for light-duty vehicles and single tank light-
duty trucks, respectively. Ford's canister costs are higher than EPA's
for several reasons. Much of the difference can be attributed to the
fact that Ford assumed a fuel tank size of 20 gallons. The EPA cal-
culated average fuel tank sizes of 12.2 and 16.3 gallons for LDV's and
single tank LDT's, respectively, based on an assumption of a 300-mile
driving range. The Ford estimate is almost 65 percent larger than the
LDV estimate, and over 20 percent larger than the LOT estimate. Although
EPA has received some information suggesting that some manufacturers
2-80
-------
Table 2-7. CALCULATION OF CANISTER SHELL COSTS
Integrated
Vehicle
Class
LDV
LOT (Sing!
LOT (Dual)
Systems
Calculated
Refueling
Capacity (ml )
2,440
e) 3,260
6,520
Assumed
Evap
Volume (ml )
850
850
1,500
Onboard
Canister
Required (ml)
3,290
4,110
3,260 + 4,760
Typical calculation of canister shell cost:
o Onboard Canister Cost
- Solve for square of radius from necessary onboard volume.
3,290 cc = r2irh = 2trr3
r = 8.06 cm
r2 = 64.96
Scale component cost by ratio of square of radii - 850 ml
canister r2 = 26.35, component cost = $1.06
Component cost = $1.06 (64.96/26.35)
= $1.06 (2.47)
= $2.61
- Add in assembly cost $0.24
$2.61 + $0.24 = 2.85
- Markup by factor of 1.4
$2.85 (1.4) = $3.99
o Subtract cost of 850 ml evaporative canister ($1.82)
$3.99 - $1.82 = $2.17
o Convert to 1984 dollars - rr-u'tiply by 1.03*
$2.17 (1.0291) = $2.24
""Average new car CPI for 1984 (208.5) divided by average for 1983 (202.6)
2-31
-------
Table 2-7. CALCULATION OF CANISTER SHELL COSTS
(concluded)
Separate Systems
Vehicle
Class
LDV
LOT (Single)
Calculated
Refueling
Capacity (ml )
2,440
3,260
Assumed
Evap
Volume (ml )
850
850
Onboard
Canister
Required (ml)
2,585 + 425
3,685 + 425
LOT (Dual) 6,520 2,500
Typical canister shell cost calculation:
o Onboard canister costs
4,010 + 4,010 + 1,000
2,865 ml = TTr2h = 2 TT r3
r = 7.70
r2 = 57.24
425 ml = 2 * r3
r = 8.3
r2 = 16.60
Scale component costs by ratio of square of radii - 850 ml
canister r2 = 26.35, component cost = $1.06.
Component
Cost = $1.06 (59.24/26.35)
= $1.06 (2.25)
= $2.38
= $1.06 (16.60/26.35)
= $1.06 (0.63)
= $0.67
- Add in assembly cost of $0.24
$2.38 + $0.24 = $2.62 $0.67 + $0.24 = $0.91
- Mark up by factor of 1.4
$2.62(1.4) = $3.67 $0.91(1.4) = $1.27
- Total canister costs
$3.67 + $1.27 = $4.94.
o Subtract cost of 850 ml evaporative canister ($1.82)
$4.94 - $1.82 = $3.12
o Convert to 1984 dollars - multiply by 1.0291
$3.12(1.0291) = $3.21.
2-82
-------
Table 2-8. SUMMARY OF CANISTER SHELL COSTS
($/vehicle)
Year
1989 1994 2000
Integrated Systems
LUV 2.24 2.08 1.94
LOT (single) 2.84 2.76 2.58
LOT (dual) 6.67 6.43 6.07
Separate Systems
LDV 3.21 3.05 2.90
LOT (single) 3.84 3.76 3.57
LOT (dual) 7.85 7.69 7.33
2-83
-------
may not reduce fuel tank sizes as vehicle fuel economy improves in the
future as EPA expected, 20 gallons clearly does not represent a fleet-
wide average tank size. The remaining difference between the EPA and
Ford canister cost estimates is due to differences in carbon prices
($1.40 vs. $2.00 per pound) and working capacities (6.7 g HC/1UU g
vs. 10.5 g HC/1UU g).
There is insufficient detail in the General Motors comment to
attempt to explain the difference between the GM and EPA estimates.
However, it is expected that the carbon canister costs differ for
reasons such as those discussed above for Ford.
As stated earlier, EPA has observed a tradeoff between working
capacity and purge rate that could not be quantified because of the
many design variables involved. Any upward adjustment in the EPA
estimates of carbon volume as a result of this tradeoff could bring the
EPA cost somewhat nearer to the General Motors and Ford estimates.
c. Refueling Vent Line Valve
The next component examined is the refueling vent line mechanical
or solenoid valve. The function of this valve is to open prior to
the start of a refueling and provide a 5/8-inch diameter orifice for
vapor flow during refueling, and to close when the event is completed
and provide rollover protection. The function of two possible valves
(and related components) which could serve this purpose are further
detailed in the technological feasibility section of this document.
The EPA expects the cost of this item to fall somewhere in the range of
$3.00-$4.60. The EPA envisions this valve to be similar in complexity
to an air management valve such as is currently used on some vehicles
to feed additional air to the engine during starts. The EPA has been
quoted a vendor selling price of approximately $3.50 for this valve at
high production volumes. The EPA also had a contractor estimate the
cost of a solenoid valve built to serve this purpose. In the report
"Costs of Onboard Vapor Recovery Hardware" prepared by Mueller Associates
for EPA in 1985, Mueller estimated the cost of an electronically acti-
vated solenoid valve designed to open and close in response to the . .
start and completion of a refueling event, respectively (I-A-77). The '
cost of this solenoid valve, an actuator located at the fillcap, and
the necessary wiring and connectors was estimated at $4.60.
2-34
-------
Ford Motor Company and General Motors were the only two commenters
to estimate the cost of a component (or components) needed to perform
all the functions required of such a valve. Ford estimates the cost
of a mechanical valve designed to allow for venting of refueling vapors
and closure during other modes of operation at about $5.UU. The system
described by Ford does not presume the removal of the U.05-inch
diameter orifice of the current evaporative emission control system.
Rather, Ford expects they would add the refueling control valve to the
fuel tank and retain the evaporative emission orifice (which also acts
to provide rollover protection in the current configurations).
General Motors estimated the cost of an electronic system similar
to that described as part of EPA's cost estimate as $6.UO. This
includes $4.0(J for the solenoid valve and $2.CD for some type of
refueling sensor for which they had no particular design in mind when
their comments were written.
d. Vapor Line
Refueling vapors are transported from the fuel tank to the onboard
canister and from the canister to the vehicle's fuel metering system
via vapor lines. In order to calculate estimated vapor line costs,
three sets of information were needed: (!) vapor line material costs,
(2) evaporative emission vapor configurations of current vehicles, and
(3) evaporative emission vapor line configurations for onboard equipped
vehicles. The vapor line used in this cost analysis is epichlorohydrin
(ECO) tubing. The costs quoted in Table 2-5 are taken from the Mueller
report (I-A-77). The Mueller cost estimate for ECO tubing is the cost
to the automobile manufacturer for high volume purchases as quoted by a
Detroit area supplier. Mueller also got a cost estimate for acryloni-
trile butadiene rubber tubing, but the lower ECO cost was used in this
analysis because it would also be effective. Cost estimates are given
(on a unit length basis) for tubing sizes of 1/4-inch, 3/8-inch,
and 5/8-inch inner diameter. The tubing costs, including any markup
at the vendor level, are shown below:
1/4" $U.27/ft
3/8'1 $U.32/ft
5/8" $0.bU/ft
2-35
-------
Depending on the layout of a particular vehicle, and considerations
of available space, vehicle safety, and cost, a manufacturer may choose
to locate a refueling canister in the front of the vehicle (typically
in the engine compartment) or in the rear (near the fuel tank). Because
of the differences between the size of the vapor lines needed to route
refueling vapors to the onboard canister and the size of those needed
to transport purge vapors from the canister, it appears that it would
be marginally less costly to locate the canister near the fuel tanks.
Because it is not possible to identify the preferred canister location
for all vehicles, the vapor line cost was specified as a range bounded
by EPA's best estimate of the vapor line cost for front and rear canister
locations.
In this analysis, it was assumed that the amount and size of vapor
lines used in the evaporative emission control system of a typical LDV
would be identical to those of the typical LOT. Hence, the LDV and LDT
(single tank) cost estimates are identical.
The costs do vary, however, with system configuration (partially
integrated vs. integrated systems). The difference between the partially
integrated and integrated system costs is due to the incremental nature
of the cost estimates in this report. If an integrated system were
used, some smaller evaporative system vapor lines would have to be
replaced with larger refueling vapor lines. If a partially integrated
system were used, not only would lines have to be replaced but, in most
cases, an entirely new purge line would have to be added. In this
analysis it was assumed that all vapor lines are 3/8-inch I.D. except
for the refueling vent line, which must be b/8-inch I.D. Vehicles
equipped with fuel injection systems have negligible "hot soak" type
emissions emanating from the fueling system, whereas carbureted vehicles
would have evaporation from the carburetor bowl(s). Therefore, carbureted
vehicles are equipped with a vapor line not needed for fuel-injected
systems. The estimated vapor line lengths for current evaporative
systems of carbureted and fuel-injected vehicles are shown below.
2-86
-------
Description Fuel-Injected Carbureted
Fuel Tank Vent 8 ft. 8 ft.
Carburetor Vent NA 3 ft.
Purge Lines 3 ft. 3 ft.
Since virtually all evaporative emission control canisters are
located in the engine compartment, it was assumed for costing purposes
tnat they are always located there.
The final set of data needed to calculate vapor line costs is the
configuration of vapor lines needed for onboard (refueling and evapora-
tive) control systems. As discussed above, the end points of the range
of the vapor line cost estimate will be defined by the vapor line costs
associated with front and rear canister locations. The assumed onboard
vapor line configurations are shown below:
Canister Located in Engine Compartment
Integrated Part. Integrated
Description (Fuel-Injected) (Carbureted)
Refueling/Evap Vent Line 8 ft. 8 ft.
Carburetor Vent Line NA 3 ft.
Purge Lines 3 ft. 3 ft. + 3 ft.
Canister Located Near Fuel Tank
Integrated Part. Integrated
Description (Fuel-Injected) (Carbureted)
Refueling/Evap Vent Line 3 ft. 3 ft.
Carburetor Vent Line NA 3 ft.
Purge Lines 8 ft. 3 ft. + 8 ft.
Using the vapor line sizes and cost estimates given above, total
incremental vapor line cost estimates can be found for integrated and
partially integrated systems for both front and rear onboard canister
locations. The table below shows the vapor line costs for each system
evaluated in this analysis. Table 2-9 details the calculation of all
vapor line cost estimates and provides information on the assumption
used to calculate the vapor line costs for LOT's equipped with dual
fuel tanks.
2-87
-------
Table 2-9. VAPOR LINE COST ESTIMATES
LDV/LDT - Integrated Systems
o Canister located in engine compartment
- Onboard system
Refueling vent 8 ft (? $0.50/ft = $4.00
Purge line 3 ft @ $0.32/ft = $0.96
Total W79~6"
- Credit for current evap system
Fuel tank vent 8 ft G> $0.32/ft = $2.56
Purge line 3 ft @ $0.32/ft = $0.96
Total I775T
Incremental vapor line cost $1.44
o Canister located near fuel tank
- Onboard system
Refueling vent 3 ft @ $0.50/ft = $1.50
Purge line 8 ft G> $0.32/ft = $2.56
Total $4.06
- Credit for current evap system
Total $3.52
Incremental vapor line cost $0.54
LDV/LDT - Partially Integrated
o Canister located in engine compartment
- Onboard System
Refueling vent 8 ft @ $0.50/ft = $4.00
Carburetor vent 3 ft @ $0.32/ft = $0.96
Purge line 3 ft 9 $0.32/ft = $0.96
Total IBT^Z
- Credit for current evap system
Fuel tank vent 8 ft G) $0.32/ft = $2.56
Carburetor vent 3 ft @ $0.32/ft = $0.96
Purge lines 3 ft @ $0.32/ft = $0.96
Total $4.48
2-88
-------
Table 2-9. VAPOR LINE COST ESTIMATES
(continued)
Incremental vapor line cost $1.44
o Canister located near fuel tank
- Onboard system
Refueling vent 3 ft @ $0.50/ft = Si.50
Carburetor vent 3 ft @ $0.32/ft = $0.96
Purge lines 8 ft @ $0.32/ft = $2.56
3 ft @ $0.32/ft = $0.96
Total ~
- Credit for current evap system
Total $4.48
Incremental vapor line cost $1.50
Tank LPT - Integrated System
o Canisters located in engine compartment
- Onboard system
Refueling vent 8 ft ? $0.50/ft = S4.00
8 ft (3 $0.50/ft = $4.00
Purge lines 3 ft @ $0.32/ft = $0.96
3 ft @ $0.32/ft = $0.96
Total $9.92
- Credit for current evap system
Fuel tank vents 8 ft @ $0.32/ft = $2.56
4 ft @ $0.32/ft = $1.28
Purge line 3 ft @ $0.32/ft = $0.96
Total $4.80
Incremental vapor line cost $5.12
o Canisters located near fuel tank
- Onboard system
Refueling vents 3 ft @ $0.50/ft = $1.50
3 ft (3 $0.50/ft = $1.50
Purge lines 8 ft G> $0.32/ft = $2.56
8 ft (3 $0.32/ft = $2.56
Total $8.12
- Credit for current evap system $4.80
Incremental vapor line cost $3.32
2-89
-------
Table 2-9. VAPOR LINE COST ESTIMATES
(concluded)
Dual Tank LPT - Partially Integrated System
o Canisters located in engine compartment
- Onboard system
Refueling vent 8 ft 0 $0.50/ft = $4.00 '
8 ft @ $0.50/ft = $4.00
Carburetor vent 2 ft I? $0.32/ft = $0.64
Purge lines 3 ft 0 $0.32/ft = $0.96
3 ft 0 $0.32/ft = $0.96
3 ft @ $0.32/ft = $0.96
Total 511.52
- Credit for current evap system
Fuel tank vent 8 ft I? $0.32/ft = $2.56
4 ft 0 $0.32/ft = $1.28
Carburetor vent 3 ft P $0.32/ft = $0.96
Purge line 3 ft IP $0.32/ft = $0.96
Total $5.76
Incremental vapor line cost $5.76
o Canisters located near fuel tank
- Onboard system
Refueling vent 3 ft IP $0.50/ft = $1.50
3 ft P $0.50/ft = $1.50
Carburetor vent 2 ft 0 $0.32/ft = $0.64
Purge lines 8 ft 0 $0.32/ft = $2.56
8 ft 0 $0.32/ft = $2.56
3 ft 0 $0.32/ft = $0.96
Total $ 9.72
- Credit for current evap system
Total $5.76
Incremental vapor.line cost $3.96
' 2-90
-------
Ranges for Incremental Vapor Line Costs
LDV/LDT (single) LPT (dual)
Integrated $0.54 - 1.44 $3.32 - 5.12
Part. Integrated $1.44 - 1.50 $3.y6 - 5.76
Ford Motor Company estimated the incremental vapor line costs as
about $3.86 for LUV's and $4.10 for LDT's. General Motors estimated
the incremental vapor line cost for passenyer cars at $2.38. Both of
these estimates fall above the cost ranges developed by EPA. The vari-
ations could be due to differences among vapor line materials, unit
costs, or assumptions on system configurations. The vapor line material
estimates were quoted by a Detroit area supplier who claimed that
epichlorohydrin tubing is impermeable to gasoline vapors, so the unit
costs should be reasonable. The other probable source of differences
is in the assumed vapor line configurations. Because Ford and GM
provided no details on vapor line configuration, it is impossible to
directly compare methodologies. The EPA has assumed that all vapor
lines (excluding the refueling vent line) are 3/8" inner diameter. In
some cases these lines may be smaller than 3/8" inner diameter, which
could lead to a smaller credit and higher incremental cost. On the
other hand, in those situations where EPA assumed that vent and/or
purge lines were added to the system, these were also assumed to be
3/8" lines that might be oversized as well, leading to a higher incre-
mental cost. Overall, the assumption of 3/8" vapor lines should not
critically impact the analysis.
e. Fill Li miter
The next component of cost in Table 2-b is the "fill limiter." The
fill limiter is a device that closes the refueling vent line when the
fuel tank becomes full and indirectly forces automatic nozzle shutoff.
The simplest device that could be used to accomplish this task is some
kind of flotation device that would be buoyed up by the fuel rising in
the tank and would seat itself in the opening to the refueling vent
line as the fuel tank became full. The requirements of the fill limiter
are more thoroughly described in the technological feasibility portion
of this document (Section 2.1.3).
The cost of the fill limiter was taken from a 1984 report written
by the American Petroleum Institute (I-F-98). Although the device for
which a cost estimate is given in the API document is not identical to
2-91
-------
the fill limiter envisioned by EPA, the two appear to be of similar
complexity. The cost quoted in the API document is $U.8U. Since the
report was written in January of 1984, the figure has been marked up by
3 percent to reflect inflation since that time. The inflation adjusted
vendor cost of the fill limiter as shown in Table 2-b is $u.82. This
is basically the same cost used in the API estimate. Neither General
Motors nor Ford estimated the cost of a fill limiter directly. Ford
did include the fill-limiting capability in their "Kol lover/Vent/Fil 1
Valve." The cost of this valve ($b.UU) falls within the range estimated
by EPA for the refuel iny vent valve plus the fill lirniter ($3.82 -
A commercially available valve, which is similar in concept to the
fill limiter and seems to meet the basic functions needed, is shown in
Figure 2-11. Conversations with the manufacturer indicate that this
valve has a vendor price of less than $1.UU.
f . Liquid/Vapor Separator
The cost of a liquid/vapor separator is also included in the on-
board cost estimate. The function of . the liquid/vapor separator is to
remove entrained liquid droplets from the vapor stream flowing to the
canister during a refueling event and return the fuel to the tank. The
cost of a liquid/vapor separator was originally developed by Leroy
Lindgren in his 1983 report to API (I-U-2fay). The original Lindgren
cost ($U.71) has been marked up by 3 percent to account for inflation
since the publication of the report. The liquid/vapor separator cost
estimate is $U.73 in 1984 dollars,
g. Fill neck Extension - Liquid Seal
The next hardware item for which a cost is given in Table 2-b is
the J-tube fillneck extension. The standard fillneck in use today
extends only a short way (if at all) into the fuel tank. In order to
provide the liquid trap that is integral to refueling vapor control,
it would be necessary to add a curved fillneck extension to this stan-
dard fillneck. It was assumed that all fillnecks are 2 inches in
diameter and it was estimated that a 7-inch extension would be
necessary. The cost for the extension was developed using cost esti-
mations found in a report written by Leroy Lindgren for EPA in 1978. 38
In this report, Lindgren estimates the cost of a tailpipe section 3b.7
2-92
-------
STANDARD
VERSIONS
ORIFICE
FLOAT
SPRING
HIGH FLOW t*J
FILTERED
The tank mounted spring balanced float valve is a low cost unit designed for venting
fuel tank vapor to the carbon canister. The device employs a float which remains open
under normal conditions. Should the tank level reach a critical height, the float will
close the canister vent line. In the event of extreme vehicle attitude or roll-over, the
float will close the canister vent line.
A filtered tank mo.un.ted spring balanced float valve is available that performs the
same functions as the above sketches except the tank side of the part is filtered to
prevent contaminates from entering the part which might effect float closing of the
canister vent line.
For high flow applications that require a large volume of vapor venting, such as fuel in-
jection applications, a high flow valve has been developed that has more than twice
the present flow capacity without loosing other critical performance parameters.
Figure Ml. FLOAT VALVE
jj^£ Borg-Warner Automotive, Inc.
K^ 707 Southside Dr., Decatur, Illinois 62525
UPSS Phone 217/428-463t
437
SKETCH
NUMBER
-------
inches in length and 1.4 inches in diameter. The cost of the fillneck
extension was found by scaling the cost of this tailpipe section by the
ratio of the tubing weights. The pipe material used in the cost
estimate was also changed from the stainless steel used in the tailpipe
to standard steel. The manufacturing cost of the extension was then
marked up by 4U percent to account for overhead and profit at the
vendor level and by 35 percent to account for inflation since the time
of the original Lindgren report. Table 2-1U details the calculation of
the cost of the fillneck extension. It should be noted that for safety
reasons a manufacturer may choose to make the J-tube from plastic or
rubber. However, costs would be comparable, if not less.
The API recently completed a test program in which several liquid
seal systems were tested for refueling vapor control efficiency (I-H-158).
One idea evaluated was to leave the stock fillneck in its standard
configuration (with the stock vent tube plugged) and rely on the incoming
fuel to prevent vapor from escaping from the fillneck. With an adequately
sized vent line to route vapors to a canister, the systems examined
achieved greater than 98 percent efficiency in all cases. Although
this concept may not be applicable in some cases, it does represent an
area where cost savings could be made. The API also experimented with
mechanical seals and found very high control efficiencies. Although
onboard systems with a mechanical seal (and the necessary pressure
relief valve) appear more costly than the liquid seal approaches, EPA
still considers them a feasible alternative. The costs of a mechan-
ical seal would be comparable to a liquid seal, but adding a pressure
relief device if needed would increase costs.
h. ECU Modifications
Table 2-b also contains a cost item for "ECU modifications." Due
to the difference between evaporative and onboard system purge strate-
gies, a cost has been included to cover the modification of vehicle
purge control systems. It was assumed that 88 percent of all vehicles
will be equipped with electronic controls by 1989. For those vehicles
equipped with electronic controls, ECU modification would include
research and development to determine the optimal purge cycle for the
given vehicle and the necessary reprogramming of the unit. Mueller
Associates has estimated the cost of this modification in their 1985
2-94
-------
Table 2-10. CALCULATION OF FILLPIPE EXTENSION COST
o Calculate the weight per uni-t of surface area for the tailpipe
section costed in the Lindgren Report.
L = 35.7" D = 1.4" wt = 5.0 Ib
As=LxC = Lx1TxD = 35.7 (ir)(1.4)
As = 35.7 (*)(!.4)
= 157 in?
Wt/unit As = 5.0 lb/157 in2 = 0.0318 Ib/unit A
o Calculate surface area of fillneck extension needed
L = 7 in. D = 2 in.
As = L * * D = 7(2)( TT)
= 43.4 in2
o Calculate weight of material needed
(43.4 in2)(0.0318 Ib/unit A) = 1.4 Ibs
o Calculate cost of fill pipe extension using Lindgren "complex"
formula for.manufacturing cost (I-A-101)
manufacturing cost = (0.367 x weight) + 0.13
= (0.367 x 1.4) + 0.13
= $0.64
o Calculate vendor selling price
vendor selling price = manufacturing cost x 1.4
= $0.64(1.4)
= $0.896
o Amount for inflation since 1978 (35.6%)
1984 dollars = 1978 dollars x 1.356
vendor selling price (1984 dollars) = $0.896(1.356)
$1.21
*Reference I-A-101.
2-95
-------
report to EPA (I-A-77). Amortized at ID percent per annum over 5 years,
this cost would be approximately $U.07 per vehicle.
For systems without electronic controls, a vacuum operated purge
control system would have to be used. The same sort of research and
development effort needed for the ECU-equipped systems would be needed
for those without an ECU. In addition, the non-ECU vehicles could
need a modified purge valve. Mueller has estimated the cost of this
modification to be $0.15 per vehicle. The total purge control cost for
non-ECU vehicles, amortized at 1U percent per year over 5 years, is $U.U7,
General Motors estimated the cost of ECU modifications as $0.2b/vehicle.
i. Tank Modifications
During the first few years of a possible onboard requirement, on-
board control hardware would have to be added to vehicles that were not
designed specifically to accommodate this equipment. Fuel tanks would
have to be modified to accept the necessary fill limiter and to provide
larger orifices where necessary. In some cases, an entirely new orifice
might have to be added. Although it is difficult to accurately quantify
the cost of these modifications, EPA recognizes that these modifications
may be necessary. Therefore, a cost of $0.50 per fuel tank has been
added to cover the cost of these modifications. After the first few
years of regulation, vehicle designs would inherently include provisions
for onboard equipment. The $0.50/vehicle for tank modification included
in the system cost covers amortization of the total investment required
to modify fuel tanks over a 5-year period.
j. Packaging Cost
In order to accommodate an onboard refueling vapor control system,
a number of components would hev to be added to each new vehicle
produced, including either a larger canister or an additional one, and
some sort of fillneck modification. The addition or onboard equipment
may also require modification of existing body parts of packaging
hardware. For example, the underbody or underhood area on some small
vehicles might have to be modified to make room for the larger onboard
canister. For small vehicles, some cost would be incurred in solving .
such packaging problems. For large and mid-size LDV's and LDT's,
virtually no costs may be incurred (I-A-77). The EPA recognizes that
such a cost would be incurred in some cases and has tried to' include a
2-96
-------
reasonable value to cover this cost. As in the case of fuel tank
modifications, these modification costs would no lonyer be incurred
after the first few years of reyulation. Therefore, the $U.bU packayiny
allowance represents an investment cost amortized over a b-year
period. Tne $U.bU per vehicle included represents over $2b million
industry-wide.
k. Assembly/Installation
The next cost item listed in Table 2-5 is an assembly cost. A
component of cost for assembly has been included only for systems
requiriny more than one canister (separate systems or systems for
vehicles with dual fuel tanks). This is because, on the maryin, it
should cost no more to install an inteyrated refueliny and evaporative
emission control system than it costs to install an evaporative system
on a vehicle. L'nce again, the fiyure used for assembly costs is EPA's
best estimate cost, lackiny a significantly better estimate. Ford,
(ieneral Motors, and API have all included an assembly or installation
cost in their incremental cost estimates ($2.UU, $0.7U, and $2.42,
respectively) without discussion or justification. The EPA dues not
believe that inteyrated onboard control systems would require any more
assembly time than current evaporative systems. Virtually every new
onboard component replaces one now present on the vehicles in some form,
with the possible exception of the fill lirniter and rollover valve.
1. Certification
The next two components of cost in Table 2-b are for certification
and systems enyineeriny. The certification cost estimate is based on a
iy?b EPA memo authored by Daniel P. Hardin, Jr. (I-b-38). The memo
suggests that certification costs can be estimated as the sum of the
costs of: (1) vehicle procurement and modification, (2) testiny, and
(3) mileage and maintenance. The derivations of the costs of these
items are discussed below.
The average cost to produce and modify a production vehicle was
estimated at $10,UUU in 197b, or about $16,UUU in 1984 dollars. This
cost would be incurred for each durability or emissions data vehicle
tested. Because durability vehicles require more maintenance than do
emissions data vehicles, the mileage accumulation and maintenance costs
for durability vehicles are higher than those for emissions data vehicles,
2-97
-------
Specifically, the mileage accumulation costs for durability and emissions
data vehicles were estimated to be $3.10 and $2.64, respectively, in
1984 dollars. Durability vehicles accumulate 50,000 miles at a cost of
$155,000 and emissions data vehicles accumulate about 4,QUO miles at a
cost of about $1U,500.
The final part of the certification cost is a testing cost. The
Hardin memo estimates the cost of an in-house test as $300 and the
cost of a contracted test as $600. It was assumed that 75 percent of
all automobile emissions tests would be performed in-house, so the
weighted average test cost is $375. In terms of 1984 dollars, the
cost is $610. The test in 1975, however, did not include the refueling
portion of the test, and the testing cost was increased to reflect this
change. The changes to the Federal Test Procedure to measure refueling
emissions control, as now considered by EPA, would substantially
increase the per-vehicle test cost.
The test procedure changes now being considered would require two
or three additional tanks of fuel ($40) and an estimated 5 extra techni-
cian hours per test ($150/test). The total cost per test would be:
$610 + $40 + $150 = $800.*
A durability vehicle will be tested 13 times on the average,
including voids, during its mileage accumulation at a cost of $10,400.
The emissions data vehicles will be tested prior to and after the 4,000-
mile accumulation at a cost of $1,600 per vehicle.
The following table shows the calculation of total certification
costs for a durability vehicle and an emissions data vehicle. Similar
costs were assumed for LDV's and LDT's, even though formal certification
protocols differ.
Durability Emissions Data
Vehicle Procurement and
Modification $ 16,000 $ 16,000
Mileage and Maintenance 155,000 10,400
Testing Costs 10,400 1,600
Totals ' $181,400 $ 28,000
*Necessary changes in facilities to accommodate additional testing are
estimated later in this section.
2-98
-------
The promulgation of an onboard rule would result in all vehicle
models being recertified for the first year of the regulation. Since
about ID percent of all vehicle models undergo exhaust and evaporative
certification in any given year due to vehicle changes regardless of
new emission standards, the cost of certifying yu percent of the fleet
for exhaust and evaporative emissions would be directly due to an on-
board regulation. The per-vehicle certification cost was found by
first estimating the industry-wide certification cost and then applying
it to the projected 1989-93 sales volume.
The industry-wide certification cost was estimated by counting
the number of durability vehicles and emissions data vehicles tested
as part of the certification process for MY1984 LDV's and LUT's, and
multiplying by the estimated certification costs above. The totals*
were then multiplied by 0.9U, amortized at 1U percent over 5 years, and
spread among the projected LDV and LOT new car fleet for 1989. The
table below shows the calculation of the per-vehicle certification
cost for both LDV's and LDT's.
LUV LPT
PUR ED PUR ED_
Number of test vehicles 109 307 45 133
Cost per vehicle $ 181K $ 28K $ 181K $ 28K
Total cost $19,729K $8,596K $ 8.145K $3,724K
Sum of durability and emissions
data cost $28,325K $11,869K
90 percent of total $2b,492K $10,682K
Amortize over 5 years at 10%
(amortization factor = 0.2638) $ 6,72bK $ 2.818K
Sales projection for 1989 ll.OOOK 3.640K
Cost/vehicle $ 0.61 $ 0.77
m. Facility Modifications
A number of comments received in response to the Strategies
Document suggested that costs related to the modification and/or con-
struction of test facilities would be incurred should an onboard regula-
*Technically, only the exhaust and the evaporative portion of the cer-
tification costs should be multiplied by 0.90 since all new vehicle
models will have to undergo-refueling certification. However, since
the exhaust and the evaporative tests constitute the majority of the
certification costs, including refueling certification costs in the
total leads to a cost difference of less than 1 cent per vehicle.
2-99
-------
tion be adopted. The EPA recognizes that some expenditures will nave
to be made in order to improve and/or expand testing facilities to
accommodate the added demands of the refueling test procedure. At the
present time, EPA has not conducted a detailed analysis of what facili-
ties modifications might have to be made. However, the Motor Vehicle
Manufacturers Association (MVMA) has estimated the cost of facilities
modifications on a per-manufacturer basis in a recent submission
responding to the proposed refueling test procedure (1-0-46). MVMA has
estimated the cost to each manufacturer to be $734,OUU. This figure
includes two SHED's, two dynamometers, two fuel conditioniny units, two
FIU's, two strip chart recorders, safety equipment, air conditioning/
ventilation, and construction and installation costs. The EPA has
taken this estimate at face value and used it in this analysis.
Assuming that 35 manufacturers (total number of manufacturers
certifying LDV's in 1984) will undertake such modifications, an invest-
ment of about $26 million would have to be made prior to 1989. Although
probably fewer than 35 manufacturers would perform modifications as
extensive as those described above, some of the larger manufacturers
might incur a cost higher than $734,OUU. The total of $26 million is
expected to represent a generous estimate of industry-wide investment.
Amortizing this amount over 1U years at 1U percent and spreading the
cost among an average of 14.6 million light-duty gas vehicles and
trucks gives a cost of about 3U cents per vehicle.
n. Systems Engineering and Development
The final component of cost listed in Table 2-5 is for systems
engineering/development (SE/D). Here the term "systems engineering/
development" refers to any developmental effort involved in combining
the components of the onboard control system to form a unit that
interacts appropriately with the rest of the automobile. For example,
the installation of onboard controls could affect a vehicle's exhaust
emissions. The mileage accumulation, testing, and engineering costs
incurred in altering the vehicle to adequately control the exhaust
emissions would be termed "systems engineering/development" costs.
The SE/D cost estimate w.as made in much the same way the certifi-
cation cost estimate was made. Four components went into this estimate:
(1) vehicle procurement and modification, (2) mileage accumulation and
2-100
-------
maintenance costs, (3) testing costs, and (4) salary costs. These SE/D
costs would be incurred for each refueling family in a manufacturer's
product line. Since the parameters defining refueling families have not
yet been fully defined, costs were developed on an evaporative family
basis under the assumption that the number of evaporative families
approximates the probable number of refueling families. An industry-
wide cost was then developed and applied over the projected LDV and LOT
sales.
The vehicle procurement and modification cost for SE/L) is the same
as that for certification, $16,DUO. The mileage accumulation and mainte-
nance cost used for SE/D vehicles is the same as that used for durability
vehicles in the certification cost calculation, $3.U9/mile. The mileage
that each vehicle will accumulate was estimated to be 8,000 miles, twice
the mileage accumulated by an emissions data vehicle. The total mileage
cost is $24,720 per family. Testing costs were also calculated for SE/U
as they were for certification. The cost per test is $800; 25 tests
were assumed for each evaporative family for a total of $20,000.
The final component in the SE/U cost is salary costs. It was
estimated that salary expenditures incurred in designing and testing
the onboard system for a single family would net exceed about 60 per-
cent of the full time salaries of one engineer and one technician.
Including benefits, these annual salaries were chosen as $50,000 and
$36,000, respectively. The calculation of the SE/U cost for both LUV's
and LUT's is shown below:
Vehicle Procurement and Modification $16,000
Mileage Accumulation - 8,000 miles 24,72o
Testing 20,000
Salary Expense . 51,000
Safety Compliance Testing 34,000
Total SE/D cost per family $146,000
The total cost was multiplied by the total number of evaporative
families for both LUV's and LDT's. The projected number of refueling
families was used here because it best represents the number of separate
engineering and development actions that will be needed. Some refueling
families may involve two or more exhaust families and thus could incur
additional exhaust emissions development costs. While these would likely
2-101
-------
be small on an overall basis, an estimate was not included in this
analysis because the degree of exhaust emissions development overlap
and extrapolation between families could not be determined.
These costs were then amortized over a 5-year period and annual
costs spread over the projected LDV and LOT sales. The calculation of
cost on a per-vehicle basis is shown below. Information on safety
testing costs is listed below but is described in a separate memo
(I-B-20).
j_DV LPT
Total cost per family $146,UUU $146,DUD
Number of families
(approx.) 140 65
Industry-wide cost
(thousands) $ 20,440 $ 9,490
Amortize over 5 years at 10%
(0.2638) (thousands) $ 5,392 $ 2,503
Sales projection, 1989
(million) 11.0 3.64
Cost per vehicle $ 0.49 $ 0.69
As can be seen in Table 2-5, an additional $0.25 has been added
to the cost of control for LDT's with dual fuel tanks and is listed as
an "engineering" cost. This amount has been added in recognition of
the possibility that certain problems may be encountered in equipping
these vehicles with onboard controls that are unique to those vehicles
with very large fuel capacities. Extra R&D testing costs may be
incurred in the development of systems capable of capturing and purging
hydrocarbon emissions of the magnitude produced by this type of vehicle.
o. Costs Not Included in the EPA Estimate
Some items that commenters suggested be included in the onboard
system cost were not included. Specifically, these are costs for main-
tenance, enforcement, and for taxes and insurance. As was discussed
in the technological feasibility portion of this study, there is general
agreement that deterioration of activated carbon canisters is not an
issue and canister maintenance will not be needed. Activated carbon
does lose some fraction of its working capacity through successive
load/purge cycles, but this phenomenon is well known and accounted for
in canister sizing. For these reasons no maintenance cost was included
in the onboard cost estimate.
2-102
-------
Although an item labeled "enforcement" does not appear in
Table 2-b, enforcement costs have been indirectly included in the
onboard cost estimate. The most powerful tool used to enforce motor
vehicle emissions regulations is the new vehicle certification program.
A substantial cost has been included in the onboard cost estimate to
account for the need to recertify the LDV and LOT fleets in response to
an onboard regulation. The other tools used to enforce motor vehicle
emission regulations, such as selective enforcement audits and recall
testing, have a negligible cost when compared to the costs involved with
certification. Because of the very low marginal cost of these programs,
a separate enforcement cost item has not been included in the onboard
system estimate.
One commenter (I-H-1U1) claimed that increases in taxes and insur-
ance (resulting from vehicle price increases caused by the addition of
onboard controls) should be included as part of the societal cost of
onboard control systems. In response to the question of taxes, EPA
contends that taxes can be considered to be transfer payments and
should therefore be excluded from the system cost. Increases in insur-
ance expenses could be considered a societal cost, but because the
fractional increase in vehicle cost caused by adding onboard controls
is so small it was difficult to envision an incremental cost item for
insurance costs. Therefore, insurance costs were not included in the
incremental cost estimate.
2.6.2 Comparison of Cost Estimates
Each of the components of cost listed in Table 2-5 has been dis-
cussed on the previous pages. Although the subject of the factor used
to account for overhead and profit at the vehicle manufacturer and
dealer levels has not yet beer addressed (the markup factor is discussed
in the next section), onboard system cost estimates can be compared at
the vendor selling price level (total system costs less manufacturer
and dealer overhead and profit). The only cost estimates that contain
enough component-by-component detail to suggest a comparison with
the EPA estimates are those compiled by Ford, General Motors, and API.
(Confidential estimates submitted by several manufacturers could not be-
analyzed here.) Table 2-11 shows a comparison of the incremental onboard
system cost for LDV's made by EPA, GM, Ford, and API. The LOT estimates
are shown in total for EPA, Ford, and API.
2-103
-------
Table 2-11. COMPARISON OF COST ESTIMATES
Activated Carbon
Canister Body
Refueling Vent Line Valved
Vapor Line
Fill Limiter
Liquid/Vapor Separator
Fillneck Extension
ECU Modifications
Tank Modifications
Packaging
Certification
Assembly
Facilities Modifications
Engineeri ng
Fillneck Seal
Pressure Rel ief
Purge Control Valve
Canister Brackets
Other
Subtotal
Credits
TOTAL
LOT Totals
EPAa
2.20
2.36
4.60
1.56
0.82
0.73
1.21
0.09
0.50
0.50
0.61
0.09
0.30
0.45
16.00
$16.00
$27.00
Ford
7.50b
5.00
3.85
__f
--
--
--
--
--
2.00
--
--
2.40
2.50
1.00
24725"
(0.80)
$23.45
$35.00
GM
5.81
6.00e
2.38
1.30
0.25
0.70
1.00
0.73
0.25
TO2~
$18.42
API
6.19C
2.66
0.809
--
--
--
--
2.42
--
--
2.40
--
--
--
--
14.47
$14.47
$16.52
aWeighted average of integrated and separate system costs.
bTotal of activated carbon ($5.50) and canister ($5.00) less credit for removing
evap canister (-$3.00).
cTotal of activated carbon ($5.77) and canister ($3.62) less credit for removing
evap canister (-$3.20').
^Includes any necessary wiring, fuel cap sensors, etc.
Includes $2.00 for a "refueling sensor."
fFill limiting capability is built into Ford's "Rollover/Vent/Fill Valve."
9Cost of a float valve/fill limiter - same as EPA cost of 0.82 (not increased
for inflation).
2-104
-------
Most of the differences between estimates for individual compo-
nents have been discussed in the sections dealing with those components,
but there are some areas that merit comment here. First, all of the
cost estimates in Table 2-11 show a higher canister cost than does EPA.
These differences are primarily due to assumptions on average fuel tank
size, and/or to carbon cost estimates. Second, all of these commenters
estimated costs for systems incorporating the mechanical-type fillneck
seal. The modifications needed to provide a liquid fillneck seal are
marginally less costly than those needed for a mechanical fillneck seal
and a pressure relief device. Third, all of these comments included
assembly costs which EPA believes may be overstated. Fourth, the API
fuel tank valve is not equipped to handle vent line closure during
driving modes, and does not provide rollover protection. Finally, a
major discrepancy between cost estimates arises from the differences
in the manner in which manufacturer overhead and profit are included.
This issue is addressed in the following section.
2.6.3 Manufacturer Overhead and Profit
Comment: A number of commenters claimed that the markup factor
of 1.27 used by EPA in the July 1984 analysis was too low, and they
saw no reason fo- EPA to abandon the markup factor of 1.8 developed
by Leroy Lindgren in his Retail Price Equivalent (KPE) cost analyses.
General Motors used a markup factor of about 1.6 and claimed that EPA
should use a factor in the range of 1.6 to 2.U. Ford did not use a
multiplicative markup factor, but claims that a factor of about 1.8 is
justified (I-H-114, I-H-117, I-H-127, I-H-132).
Response: In its 1984 analysis, EPA used the basic methodology
used by Leroy -indgren to estimate the "retail price equivalent" or
cost to the consumer of an onboard system. According to Lindgren, the
retail price equivalent can be obtained from the "manufacturer" level
cost* by accounting for: (1) vehicle manufacturer profit, (2) vehicle
manufacturer overhead, and (3) dealer overhead and profit. In the
*The "manufacturer" level refers to the price an automobile manufac-
turer would have to pay to internally produce or to purchase a
part from a vendor or supplier.
2-105
-------
Strategies Document, EPA used a factor of 1.27 to account for all of
these factors. The justification for the use of a markup factor of
this magnitude is provided in a previous EPA analysis (I-A-2y).
In choosing a markup factor, EPA attempted to account for over-
head costs actually incurred and profit margins actually achieved in
the automobile industry. In response to the comments claiminy that
EPA had underestimated overhead and profit by assuminy a markup factor
of 1.27, EPA had a contractor further evaluate an appropriate markup
factor (I-A-1U6). In this report Faucett takes a weiyhted averaye of
the publicly reported overhead and profit markups of the three laryest
American automobile manufacturers over the past lu years. The averaye
markup factors as a percentage of cost of sales for the light-duty
market as estimated by Jack Faucett Associates are shown below:
Manufacturer Overhead lb.4
Manufacturer Profit 3.8
Manufacturer Net ly.2
The RPE must include dealer overhead and profit as shown below:-
Dealer Interest Expense 1.7
Uealer Profit . 2.U
Dealer Sales Commission 2.U
Dealer Net 5.7
The retail price equivalent of the onboard control system was calcu-
lated from the vendor selling price in the followiny manner:
KPE = [Vendor Selliny Price (excluding enyineeriny)
x (1.192)(1.U57)] + engineering costs.
As mentioned above, many of the automobile manufacturers claimed
that a much higher markup factor was appropriate. It is clear from the
Faucett report that overhead as a percentage of cost of sales has been
declining in recent years, and the averaye manufacturer profit maryin
experienced over the past 1U years is only about 3.8 percent. Manu-
facturers and dealers would prefer to see hiyher profit margins, and
they have experienced periods when overhead represents a larger percentage
of the cost of sales than estimated in the Faucett report. The EPA,
however, is attempting to use an industry-wide, lony-term average
2-1U6
-------
value, and the 1.26 markup factor is representative. For further dis-
cussion of markup factors, the reader is referred to the Faucett report.
2.6.4 Excess Evaporative Emissions
Comment: Several commenters (e.g., I-H-114, I-H-118, I-H-127)
noted that, in the July 1984 EPA analysis, onboard controls were evaluated
both with and without substantial emission reduction credits associated
with the control of excess evaporative emissions. These commenters
went on to argue that the excess evaporative emissions issue is separate
from the refueling emissions question, and should not be considered in
the evaluation of onboard controls. The commenters stated that onboard
controls should be evaluated only in terms of the costs and benefits
associated with the control of refueling emissions.
Response: During the past few years, EPA has tested several in-
use vehicles for their compliance with evaporative emission standards.
The results of this testing have suggested that many vehicles are
exceeding these standards by substantial margins. The reasons for the
in-use failures can basically be traced to the upward trend in the
volatility of commercially available gasoline coupled with vehicle
purge control system underdesign (I-A-66). Preliminary engineering
analysis indicates that properly designed onboard control systems
coupled with test procedure and certification fuel changes would help
control these'excess evaporative emissions with no significant marginal
impact on onboard costs. Therefore, the 1984 analysis included a
scenario in which onboard control systems were credited with the control
of excess evaporative emissions.
In response -to the comments concerning excess evaporative emissions,
EPA attempted to evaluate the costs and benefits of onboard control
incremental to those associated with the control of excess evaporative
emissions. Because properly designed onboard systems could control
excess evaporative emissions, the total onboard system costs given
previously in this section are not incremental to the costs of vehicle
excess evaporative emission controls. In order to get the incremental
hardware cost, the cost of the vehicle excess evaporative emission '
controls should be deducted from the onboard hardware cost.
2-107
-------
In a recent EPA study on gasoline volatility and evaporative emis-
sions, several strategies for the control of excess evaporative emis-
sions were evaluated (I-A-66). One of these strategies, referred to as
the vehicle control approach, is to increase the volatility of the
certification fuel from 9.U to 11.5 psi and make other test procedure
changes to require improved purge capability. This would require
vehicle manufacturers to improve evaporative emission control systems
to capture and purge a larger evaporative load. This would have
implications for onboard costs, since onboard systems and the refueling
test procedure would accomplish, at least to some extent, the same ends.
For example, vehicle excess evaporative emission controls implemented
concurrently with onboard would reduce costs in areas such as canister
size, purge control, and certification. Therefore, the cost of the
onboard system incremental to the costs associated with the control of
excess evaporative emissions can be estimated by subtracting the costs
associated with the vehicle excess evaporative control approach (taken
from the volatility study mentioned above) from the onboard hardware
estimates provided earlier. The calculations of the incremental, onboard
costs are shown below. HDGV costs are discussed more fully in the
following section.
Onboard Hardware Costs Incremental to
Excess Evaporative Emission Control
LDV LOT* HUGV*
1989-1993
1994-1999 '
2000+
The incremental onboard hardware costs (bottom line values above)
are the costs used to calculate the total incremental onboard cost and
cost effectiveness values discussed elsewhere in this document.
$20.00
2.90
17.10
17.70
2.00
15.70
16.30
2.00
14.30
$27.20
3.80
23.40
24.20
2.60
21.60
22.60
2.60
"ZOU
$34.80
4.20
30.60
29.80
3.50
26.30
29.40
3.50
25.90
*Since only a single cost is given for the vehicle evaporative emis-
sion control system improvements for LUT's and HDGVs, the weighted
average LOT and HDGV onboard hardware costs are shown.
2-108
-------
Subtracting the excess evaporative emission control cost from the
total onboard system cost separates the two areas with regard to costs.
To be consistent, the benefits of the control strategies must also be
separated. In the cost effectiveness calculations presented elsewhere
in this study, only the benefits associated with capturing refueling
vapors are included in the incremental cost effectiveness of onboard
controls. The benefits in terms of excess evaporative emission reduc-
tions associated with vehicle controls are not included.
Another benefit of evaporative emission control is a recovery cre-
dit for the consumption of previously escaping evaporative hydrocarbons.
In the volatility study, the cost of the vehicle changes needed to
control excess evaporative emissions were reduced to reflect a cost
savings associated with the use of the controlled evaporative hydrocar-
bons (I-A-66). This evaporative emission recovery credit has also been
omitted from the incremental analysis of onboard controls.
2.6.5 Heavy-Duty Gasoline Fueled Vehicles
a. Introduction
As discussed in other portions of this document, a number of
commenters stated that onboard controls should be applied to HDGV's,
and not applying such controls reduces the overall efficiency of the
onboard approach.
In Appendix C of the July 1984 EPA analysis, the technological
feasibility of onboard controls for heavy-duty gasoline vehicles (HDGV's)
was briefly discussed in qualitative terms. In that document, it was
postulated that HUGV's could be equipped with onboard controls. Although
the feasibility of onboard controls for HDGV's has not been physically
demonstrated since the publication of the 1984 analysis, there is no
information that suggests the contrary.
Une possible area of concern surrounding the feasibility of onboard
controls for HDGV's is the ability of the HDGV to purge the refueling/
evaporative canister(s) of a large amount of hydrocarbons without unaccep-
table impacts on exhaust emissions and dri veabi lity.' For many of the
smaller HDGV's, the task will be nearly identical to that of equipping
LDT's with onboard controls. The fuel system and emission control techr-
nologies used on the light HDGV's in the early 1990s are expected to be
quite similar to those used on larger LDT's. Fuel tank capacities for
these vehicles are also expected to be quite similar.
2-109
-------
There are more areas of uncertainty, however, for the heavier trucks.
First, the laryer trucks generally have laryer fuel tanks and hence a
laryer refueliny load to handle. Also, and of greater concern, the fuel
and emission control systems of these heavier trucks are generally less
advanced than those of the lighter HUbV's. For example, many of the
truck models may not use feedback fuel control systems and some vehicles
may not even be catalyst equipped. Although these heavier HUGV's present
a more difficult engineering challenge, there is no reason to conclude
that they could not be equipped with onboard controls. Indeed, other
control technologies such as the fuel bladder could be used to reduce
the refueling and evaporative emissions vapor load, but costs may be an
issue (I-B-41).
Since it appears feasible to equip HDUV's with onboard controls
and there would be air quality and health benefits associated with
such control, EPA included HUGV's in the onboard control option evaluated
in this document. The EPA is currently preparing a report that looks
at the feasiDility, costs, and cost effectiveness of onboard controls
for HDGV's in more detail, but the report was incomplete at the time
this analysis was performed. In order to include HUlaV's in the onboard
analysis, a preliminary estimate of the cost of onboard systems for
heavy-duty gasoline vehicles was made. The cost estimate is briefly
discussed in the following paragraphs.*
b. Onboard System Costs - HUGV's
Table 2-12 shows the detailed estimates for the costs of onboard
control systems for HDGV's. Most of the estimates were derived using
the same methodologies used in the light-duty analysis.. In fact,
many of the costs are identical to those used there. Because of the
similarities in the costing methodologies, the heavy-duty cost analysis
is not discussed in detail in this section. In order to provide a
basis for analyzing the validity of the cost estimations, some of the
key assumptions used in the analysis are described below.
*A preliminary estimate of the cost effectiveness of onboard controls
for HUGV's can be found in the Public Docket (I-B-42).
2-11U
-------
Table 2-12. ONBOARD COSTS - HDGV's
Fuel Tank Configuration
Fleet Composition (%)
Activated Carbon
Canister Body
Refuel iny Line Valve
Vapor Line
Fill Limiter
Liquid/Vapor Separator
Fil Ineck Extension
Mechanical Seal
Pressure Relief
Tank Modifications
Packaging
Assembly
Certification
Facilities Modifications
Vendor Cost
Markup by 1.27a
Systems Engineering Develop-
ment
Purge System Improvements
Total RPE Cost
Total Costs, 1989
1994 - 1999b
2000 and Later MY Costs0
Class I
Single
80
3.60
3.16
4.60
1.54
0.82
0.73
1.21
--
--
0.50
0.50
0.00
2.00
0.30
18.96
24.08
0.69
0.25
25.02
$30.
$25.
$24.
Ib
Dual
20
7.20
7.52
9.20
5.24
1.64
1.46
2.42
__
--
1.00
0.50
0.75
2.00
0.30
39.23
49.82
0.69
0.50
51.01
22
29
91
Class VI
Single
85
5.40
7.17
4.60
5.36
0.82
0.73
1.21
1.12
2.50
0.50
--
2.00
0.30
31.71
40.27
1.50
0.25
42.02
Dual
15
14.40
15.74
9.20
10.08
1.64
1.46
2.42
2.24
5.00
1.00
0.00
2.00
0.30
65.48
83.16
1.50
0.50
85.16
$48.49 .
$43.39
$43.04
aSee docket item (I-A-74).
bFollowing amortization of the "5-year" fixed costs.
cFollowing amortization of the "10-year" fixed costs.
dCosts shown here differ slightly from those in docket item (I-B-37) due to
correction of computational errors.
2-111
-------
(1) Fleet Composition
No data were available at the time of this analysis with'which to
accurately cateyorize the HUGV fleet in terms of chassis designs,
vehicle lengths, and other Key vehicle parameters necessary to charac-
terize HUGV's and their fuel systems by weight class. It was known,
however, that about 9U percent of all HUGV's fall in Classes lib
(8,501-10,000 Ibs), VI (19,501-26,UUU Ibs), and VII (26,001-33,000 Ibs),.
with about three-fourths of these falliny in Class lib. In order to
simplify the analysis, it was assumed that 75 percent of all HUGV's can
be classified as Class lib trucks and all remaininy HUGV's fall in
Class VI. A number of local vehicles were evaluated to yet a preliminary
idea of the Class lib and VI HUGVs' vehicle and fuel/vapor system
characteristics needed for this analysis.
(2) Fuel Tank Capacities
In order to calculate carbon bed volumes and their costs it was
necessary to estimate fuel tank capacities for the various vehicle
confiyurations analyzed. The fuel tank capacities selected for Class
lib trucks were 20 and 40 yallons for single and dual tank trucks,
respectively. The sinyle tank capacity of 20 yallons was chosen based
on recent certification information for Class lib trucks. The dual
tank usaye rate of 2U percent used in the LL)T analysis was applied to
the lib trucks as well.
The sinyle fuel tank capacity for Class VI vehicles (30 yallons)
was based on the inspection of a number of Class VI HUGV's used in
local service. The dual tank fuel capacity was estimated by assuminy
a 500-mile driving range and applying the projected 1994 fuel economy
for Class VI HUGV's (6.17 mpy) (I-A-99). This gives an estimated fuel
tank capacity of about 81 yallons, which was rounded to 8U gallons (two
40-gallon tanks). The dual tank usaye rate for the Class VI trucks
was approximated by the fraction of the Class VI HOGV's that are engaged
in short and long range travel (i.e., not strictly in local service).
Information from the 1977 Census of Transportation suggests that less
than 15 percent of the Class VI HUGV's are used for non-local
business (I-F-142). Therefore, 15 percent was chosen as the dual tank
usage rate.
2-112
-------
(3) Separate vs. Integrated Systems
Because of differences in the assumptions concerning the number of
canisters needed to control refueling and evaporative emissions (out-
lined in the light-duty analysis), the cost of an integrated system
differs from that of a separate system. As in the light-duty analysis,
the proportion of HDGV's using integrated systems was estimated by the
proportion of fuel-injected vehicles in the fleet. Due to the expected
similarities between the fuel and emission control technologies of
LDT's and Class lib HDGV's, the fuel-injected fraction for LDT's
(88 percent) was applied to the lib trucks. All Class VI HDGV's were
assumed to use separate systems, because in the near term few are
expected to use fuel injection.
(4) Emission Rate and Canister Sizing
Refueling emission rates are somewhat sensitive to fuel tank and
fillneck configurations. Since the fuel tank configurations of the
heavier HDGV's differ significantly from those of LDV's and LDT's, it
would be desirable to have a refueling emission rate for HDGV's alone.
No data base with which to estimate this emission rate was available
at the time this analysis was performed, however. The only data avail-
able were the data used to estimate an emission rate for LDV's and LDT's.
Therefore, the emission rate used to size canisters for LDV's and LDT's
(7.U g/gal) was used in this analysis.
(5) Miscellaneous
The cost estimations for activated carbon, canister body, and
vapor lines were made using the techniques used in the light-duty
cost analysis and estimates of typical Class lib and VI truck configura-
tions. The costs estimated for refueling vent line solenoid valvs?,
fill limiters, liquid/vapor separators, fillneck extension, fuel tank
modification, packaging, as.sembly, and facilities modifications (and
engineering/development costs for the lib trucks) are identical to
those used in the light-duty analysis. Due to the limited fillneck
height available on most of the heavier HDGV's,* it was assumed that
*The gas cap on the fuel tanks of most of the heavier HDGV's is typically
integral with the tank (i.e., little or no fill height).
2-113
-------
manufacturers of the heavier trucks would choose to use mechanical fill-
neck seals. The cost included for each fillneck seal was that used in
the July 1984 analysis, $1.12. In conjunction with the use of a
mechanical seal, a pressure relief device of some sort would have to be
used. In their comments on the 1984 EPA analysis (I-H-114), Ford Motor
Company estimated the cost of a pressure relief device to be $2.50.
To be conservative, this cost was used in the HUGV analysis. The
remaining costs are EPA best estimates based on available informa-
tion.
2.7 FUEL CONSUMPTION BENEFITS OF ONBOARU
In the July 1984 EPA analysis, the costs of onboard control are
developed in Appendix C. In the analysis of onboard control costs
presented in that appendix, the fuel economy impacts of onboard control
are briefly considered. The discussion presented in Appendix C states
that the implementation of an onboard regulation would have no net
impact on fuel economy. Although -the discussion asserts that the
recovery of refueling vapors could marginally reduce fuel consumption,
it also recognizes that the added weight of the onboard control system
would reduce the vehicle's fuel economy. The assumption was made that
these two factors roughly offset one another.
2.7.1 Recovery Credits for Light-Duty Vehicles
Comment: A number of commenters stated that a net fuel consumption
benefit would be realized if onboard controls were required (I-H-11,
I-H-40, I-H-70, I-H-102, I-H-108, I-H-119).
Response: In response to these comments, EPA examined the fuel
economy impacts of onboard control in more detail. The general conclu-
sion of the analysis is that the addition of onboard cont.'ois to light-
duty vehicles (LDV's) and light-duty trucks (LDT's) would marginally
decrease their lifetime fuel consumption. Although the magnitude of
the benefit is not large as a percentage of fuel consumed over the
life of the vehicle, the benefit could reduce the consumer's cost of
control substantially. The remainder of this section details the calcu-
lation of the net fuel consumption improvement.
2-114
-------
The fuel consumption analysis can be conveniently broken into three
parts. The first part is the calculation of the fuel economy penalty
caused by the additional weight of the onboard system. This calculation
involves the average weights of the various classes of LUV's and LUT's,
the estimated weiyht of the onboard system, and the sensi.tivity of fuel
economy to weight for the various vehicle classes. The second part of
the analysis is the calculation of a gross fuel consumption improvement
derived from the recovery of refueling vapors. This calculation involves:
(1) calculation of the average total weight of hydrocarbons recovered
over the life of a vehicle, (2) calculation of the energy value of the
fuel recovered, and (3) conversion of the results of these calculations
to equivalent gallons of fuel saved. The final step in the analysis is
the calculation of the net fuel consumption benefit from onboard con-
trol (gross savings - weight penalty) and tr.e conversion of the results
to appropriate units. In this analysis, the results are presented in
terms of both quantity of fuel saved over the life of the vehicle and the
net present dollar value of the lifetime fuel savings.
2.7.1.1 Weight penalty. . Four pieces of information are needed in
order to calculate the fuel economy penalty caused by the addition of
onboard controls for each class of vehicle: (1) average vehicle weight,
(2) estimated average control system weight, (3) average fuel economy,
and (4) sensitivity of fuel economy to changes in vehicle weight. The
sources of these pieces of information are documented below. Then a
detailed description of the calculation of the penalty in fuel economy
associated with the added weight of an onboard system is given.
a. Average Vehicle Weight
The average vehicle weights for LDV's and L!)T's were taken from SAE
Technical Paper 85055U entitled "Light-Duty Automotive Fuel Economy ... '
Trends thru 1985" (I-A-1U4). This paper gives a fleetwide salesweighted
average vehicle inertia weight for both light-duty vehicles (3,1)82
IDS., Table 1) and light-duty trucks (3,832 Ibs., Table B-2) for 1985.
The weighted average vehicle weights were calculated using manufacturer-
specific sales projections for 198b.
b. Average Control System Weight
The average control system weights for LUV's and LUT's were developed
by estimating the incremental increase in vehicle weight associated with
2-115
-------
each component of the refueling control system. The components for which
weights were estimated are those that make up the onboard control system
discussed elsewhere in this study. The components and their estimated
incremental weights are shown below:
LDV LPT LPT (dual)
Carbon iTfTlb. 2.2 1b. 4.3 Ib.
Canister 0.8 0.9 1.8
Vapor Line* 0.7 0.7 2.2
Fillneck Extension 1.4 1.4 2.8
LDV LPT LPT (dual)
Liquid/Vapor Separator 074~~ 074~~ (579
Rollover/Vent Valve 0.4 0.4 0.8
Fill Limiter 0^4 £^4 0.9
LOT Weighted Average** 5.7 6.3 13.6
The carbon weights were found by applying the activated carbon
density used in the canister sizing analysis (300 g/liter) to the volume
of carbon needed for each vehicle class. The canister shell, vapor
line, and rollover valve weights were found by weighing representative
components and, when necessary, scaling the weights to reflect size/
capacity differences. The computation of these estimated weights is
detailed in Table 2-13. The fillneck extension weight (1.4 pounds)
was estimated as part of the calculation of the cost of that compo-
nent. The weight of the liquid/vapor separator was taken from the
Lindgren report on onboard costs (1-0-269). Due to a lack of better
information, the fill limiter weight was taken to equal the weight of
the liquid/vapor separator under the assumption that these components
would be of similar material, size, and complexity.
The system weights shown above are intended to cover the total
incremental weight increase associated with equipping LPV's and LDT's
with onboard refueling control systems. One large component of the
* Vapor line weight estimates are weighted 88 percent for integrated and
12 percent for separate systems.
**As outlined in the section on onboard system costs, approximately 80
percent of the vehicles in-the LPT fleet are equipped with a single
fuel tank and 20 percent are equipped with dual fuel tanks. Throughout
this section, the "weighted average" figures for LDT's. are based on
this 80-20 weighting.
2-116
-------
Table 2-13. COMPUTATION OF ONBOARD COMPONENT WEIGHTS
Carbon - 300 grams/liter = 0.66 Ib/liter
.LDV 2.44 liter x 0.66 = 1.6 Ib
LOT 3.26 liter x 0.66 = 2.2 Ib
Dual 6.52 liter x 0.66 = 4.3 Ib
Canister (Body, Grid, Filters)
GM 1,500 ml canister - nass measured @ EPA
202 g - 200 g = 0.445 Ib
scale by ratio of surface area - or equivalently square of areas.
1,500 ml canister w/h = d r'2 = 38.5
LDV - 2,440 + 850 = 3,290 r"2 = 65.0
r1'2 65.0
7^~ = ssT? = l'7 IJ x °-445 lb- = °-76 * °-8 lb-
LOT 3,260 + 850 = 4,110 r'"23 = 75.4
P'''2 75.4
= 2-° 2.0 x 0.445 lb. = 0.87 * 0.9 lb.
r ' 38.5
dual tank trucks - 1.8 Ib
Vapor Line
5/8" I.D. line - mass of 1 ft. section measured - 91 g/ft
3/8" I.D. 3/5 x 91 g/ft » 55 g/ft
LDV/LDT
Integrated system - replace 8 ft of 3/8" with 8 ft of 5/8"
8 ft (91 g/ft - 55 g/ft) = 288 g <* 0.6 lb
Separate System:
- replace 8 ft of 3/8" with 8 ft of 5/8"
- add 3 ft of 3/8"
288 g + (3 ft x 55 g/ft) = 453 g
= 1 lb
- weighted average - 0.7 lb
2-117
-------
Table 2-13. COMPUTATION OF ONBOARD COMPONENT WEIGHTS
(concluded)
Dual tank trucks
Integrated System:
- Add: 16 ft. of 5/8" - 1,456 g
6 ft. of 3/8" - 330 g
1,786 g
- Remove: 15 ft. of 3/8" - 825 g
Net Mass 961 g * 2.1 Ib
Separate System:
- Add: 18 ft. of 5/8" - 1,456 g
11 ft. of 3/8" - 605 g
?70"6T g
- Remove: 18 ft. of 3/8" - 990 g
Net Mass 1,071 g = 2.4 Ib
Weighted Average 2.2 Ib
2-118
-------
increase in vehicle weight is additional activated carbon used to
collect and store refueling emissions. It appears that the additional
hydrocarbon storage capacity could be used to collect the "excess
evaporative emissions" that have recently been measured on in-use
vehicles. The EPA is currently evaluating strategies to control these
excess emissions in a rulemaking separate from an onboard rulemaking.
In this analysis, EPA is attempting to separate the costs and benefits
of the two problems, i.e., attempting to evaluate the costs and bene-
fits of an onboard regulation incremental to those of excess evaporative
emission controls.
In order to evaluate the weight penalty associated with refueling
controls incremental to that of excess evaporative emission controls,
it was necessary to subtract the incremental weight of the excess
evaporative control system expansion from that of the onboard system.
The increases in vehicle weights associated with vehicle control of
excess evaporative emissions were estimated as part of the analysis
in the recent EPA study of fuel volatility (I-A-66). The weight
increase used in the volatility study for light-duty vehicles was
0.8 pound and that for light-duty trucks was 1.1 pounds. The calcu-
lation of the incremental onboard system weights are shown following:
LDV LPT LPT (dual tank)
Onboard system weight incremental
to current evap system (Ib) 5.7 6.4 13.7
Excess evap control system weight
incremental to current evap
system (Ib) 0.8 1.1 1.1
Onboard system weight incremental
to excess evap control system
weight (Ib) 4.9 5.3 12.6
As part of their comments on the final volatility study, General
Motors raised an issue with EPA regarding the weight impact associated
with adding components to a vehicle. GM claimed that adding "x" pounds
of equipment wouVd lead to a total increase in vehicle weight of 1.7(x)
pounds. The factor of 70 percent added to the component weight covers
2-119
-------
the weight of the additional vehicular structure that would have to be
added to carry the added components. Although vehicle structure might
not be improved in equipping vehicles with onboard controls during the
first few years of a regulation, GM claims that as vehicle models
changed and new models were designed to include onboard controls, the
vehicular structure would be designed to carry the added weight of the
onboard system. This assertion suggests that the EPA estimate of the
weight impact of adding an onboard system is lower than it ought to be.
Due to the late date at which GM presented this information to EPA,
it has been more thoroughly evaluated in a separate analysis (I-B-22).
There is some conservatism built into the weight estimate. For
purposes of the analysis of the weight penalty, it has been assumed
that all canisters would be located in the engine compartment of the
vehicle. However, some manufacturers could choose to locate the onboard
canister(s) near the fuel tank, which could lead to a somewhat lower
total vapor line weight, since less 5/8" hose would be used. This
current assumption leads to the need for a longer refueling vent line
(5/8" inner diameter) and a marginally higher vapor line weight.
In this section.of the analysis, incremental weight penalties
for a system to control refueling emissions have been developed. This
information will be used to estimate the potential fuel consumption
benefits of onboard controls. The potential fuel consumption benefits
of vehicle excess evaporative controls were addressed in the recent EPA
study of gasoline volatility (I-A-66), and will not be further evaluated
here.
c. Fuel Economy Estimates
Fuel economy estimates are those used in EPA's MOBILES fuel con-
sumption model (I-A-99, I-B-37). The table below shows estimated
average in-use fuel economy for LDV's and LDT's for model years 1989,
1994, and 200U. These three model years have been chosen because the
costs of onboard control have been estimated for these 3 years and this
makes it possible to compare the recovery credits and costs of onboard
control in the same time frame.
2-120
-------
Year LDV LPT*
1989 24.61 18.43
1994 26.64 18.99
2UUU 29.13 20. b6
d. Fuel Economy/Weight Sensitivity
The final component used in calculating the fuel economy penalty
was the sensitivity of fuel economy to changes in vehicle weight for
LDV's and LDT's. These sensitivities have been calculated by Energy
and Environmental Analysis, Inc., for the U.S. Department of Energy
(I-T-135). The sensitivities are presented in terms of the quotient of
percent change in fuel economy per percent change in vehicle weight.
The sensitivities are presented for both LDV's (-0.329) and LDT's
(-0.402). The sensitivities do vary a great deal from manufacturer to
manufacturer, but for a fleetwide average the sensitivities presented
above are the best information available to EPA. The sensitivities
should provide a reasonable estimate of the decrease in fuel economy
caused by increasing vehicle weights.
These sensitivities were used in the analysis presented in the EPA
study on fuel volatility referred to earlier (I-A-66). In that report,
the sensitivities were used to calculate the fuel economy penalty that
would be incurred as a result of the vehicle weight increase associated
with the vehicle based control of excess evaporative emissions.
e.. Calculation of Weight Penalties
The fuel economy penalty associated with the weight of the onboard
hardware can now be calculated using the information presented above.
The steps in the calculation of the weight penalties are described
below and summarized in the table which follows the discussion.
The percent change in vehicle weight was calculated by dividing
the average incremental weight of the control system by the average
vehicle weight for each vehicle class of interest. The percent change
in vehicle weight was then'multiplied by the sensitivity factor to give
a percent change in fuel economy. Next, the projected average lifetime
fuel consumption was calculated for the various vehicle classes. This
*The fuel economies of single and dual tank trucks have not been pre-
sented separately even though dual tanks are used predominantly on
the larger LDT's. Using the control system weights and applying them
to the weighted average LDT vehicle weight will produce representative
LOT fuel economy penalties.
2-121
-------
was done by dividing the estimated average lifetime vehicle miles
travelled by the projected average fuel economy for each vehicle class.
The percent change in fuel economy was then applied to the lifetime
vehicle fuel consumption to produce the fuel consumption penalty asso-
ciated with adding onboard controls.*
LUV LPT Dual
1) Average control system weight (Ibs) 4.9 5.3 12.6
2) Average vehicle weight (Ibs) 3,082 3,832 3,832
3) Percent change in weight (%) 0.159 0.138 0.329
4) Sensitivity factor
(dimensionless) -0.329 -U.402 -0.402
5) Percent change in fuel economy (mpg)
(%) -0.052 -0.056 -0.132
6) 1989 projected fuel economy (mpg) 24.61 18.43 18.43
7) Projected average lifetime
mileage (miles) 100,000 120,000 120,000
8) Lifetime fuel consumption (gal) 4,063 6,511 6,511
9) Lifetime weight penalty (gal) 2.1 3.6 8.6
Using similar calculations, the weight penalties for each model
year and vehicle class of interest are shown next:
Weight Penalties
(gal Ions/lifetime)
1989 , 1994 2000
LDV
LOT
LDT (Dual )
2.1
3.6
8.6
2.0
3.5
8.4
1.8
3.3
7.8
2.7.1.2 Gross fuel consumption credit. This section details the
calculation of the gross fuel consumption benefit related to onboard
control. The gross fuel consumption benefit related to the recovery of
refueling vapors was found using: (1) the uncontrolled refueling emis-
sion factor, (2) the theoretical control efficiency of onboard systems,
(3) the energy content of typical gasoline expressed on a volume basis
(Btu/gal), and (4) the energy content of the refueling vapors expressed
on a weight basis (Btu/lb). Each of these factors is discussed below.
*Although the sensitivity factor is presented in terms of fuel economy,
it can be applied equally well to fuel consumption because it is
expressed in terms of percentage change. Since fuel consumption and
fuel economy are inversely related, the direction of change is oppo-
site, i.e., an increase in fuel economy leads to a decrease in fuel
consumption.
2-122
-------
a. Uncontrolled Emission Kate
The uncontrolled emission rate was developed from a series of
refueliny SHED tests conducted at the EPA Motor Vehicle Emissions Labora-
tory in Ann Arbor. The emission rate was developed by applyiny the
national averaye conditions of temperature and fuel volatility to a
multiple regression equation developed from the results of the uncon-
trolled refueliny emission test proyram. The details of this test proyram
and the development of the reyression equation are presented in an EPA
technical report (I-A-69). The national averaye conditions were chosen
so that the emission rate calculated would yive an estimate of the total
potential refueliny emissions that could be controlled by onboard systems,
yiven perfect control efficiency. The emission rate that results when
the national average conditions (Tu = fc>8.9°F, T = 4.4°F, KVP = 12.6) are
inserted in the reyression equation is 5.9 yfatns of hydrocarbons emitted
per gallon of fuel dispensed.
b. Control Efficiency
The theoretical control efficiency of onboard control systems has
been well demonstrated over the past several years. In 1978, the
American Petroleum Institute (API) evaluated the efficiency of an onboard
control system equipped with a mechanical seal built into the vehicle
fillneck and observed efficiencies of greater than 98 percent (I-F-17).
In a more extensive set of tests performed on these vehicles using a
nozzle-based seal, refueling emissions were also controlled to between
96 and 99.8 percent. The API supplemented the work done in 1978 with a
study performed in 1985 (I-H-lt>8). In the 198b onboard demonstration,
Mobil Kesearch and Development experimented with a number of "liquid
seal" concepts and observed control efficiencies of greater than 98
percent. Exxon Research and Engineering equipped two vehicles with
onboard control systems usiny liquid seals and measured refueliny
emissions usiny 9, 10.5, and IZ psi RVP fuels. Both Exxon vehicles
achieved refueling emission control efficiencies of yreater than 98
percent. For a further description of the API work see docket items
(I-F-17) ana (I-H-158). . .
The EPA has also done some developmental work with an onboard
equipped vehicle utilizing a J-tube type fillneck seal (I-A-93, 1-A-1U9).
In refueliny emission control tests performed at the conditions outlined
2-123
-------
in EPA's draft Refueling Test Procedure (I-A-71), refueling emissions
were controlled to between 95 and 99 percent with an average of
97.6 percent. In each of the test programs discussed above, the average
efficiency of refueling emission control was shown to be greater than
97 percent. Also, these were prototype systems which could be expected
to perform less efficiently than systems designed for the vehicle and
installed during manufacture. Therefore, EPA has chosen a theoretical
efficiency of 97 percent under the assumption that this is a readily
attainable value, and may be conservatively low.
The theoretical control efficiency of the system does not reflect
tampering effects. These factors are accounted for, nowever, in the
fuel consumption figures used to calculate the nationwide, long-term
cost of control and cost effectiveness presented in the regulatory
impact analysis document (Reference 2 of Chapter 1). The effect of
reduced control due to tampering is accounted for in that manner.
c. Energy Content of Gasoline and Refueling Vapors
In order to calculate the net fuel consumption benefit associated
with onboard controls, the weight penalty and lifetime recovery credits
have to be expressed in equivalent units (e.g., gallons of gasoline). This
is not as simple as it initially appears, however. The difficulty arises
because of the fact that hydrocarbon vapors collected during a refueling
event are composed of a different combination of constituents than is
the liquid fuel dispensed in that same event. About 75 percent of the
vapor is comprised of paraffins with carbon numbers of 4 and 5. The
liquid gasoline, on the other hand, has a much lower concentration of
these lighter ends and a much higher concentration of aromatic hydrocar-
bons witn carbon numbers from 7 to ID. The difference in composition
means that the vapor has a different density (in a condensed form) and
energy content than the same mass of liquid fuel.
Since the composition of refueling vapor is not identical to the
composition of the gasoline from which it comes, it was necessary to
estimate the difference in their energy contents. It would have been
ideal if the energy content and density of a "typical" fuel and its
vapor could be developed. This requires a detailed knowledge of the
properties of fuels and their vapors as consumed throughout the United
2-124
-------
States for all seasons of the year. No data base of this sort is cur-
rently available to EPA. However, EPA did have a study that charac-
terized liquid fuel and associated vapors for a number of fuels dispensed
in northeastern cities during the month of July lybO (I-A-68). The
study was done by Midwest Research Institute (MKI) of Kansas City under
contract to EPA's Office of Air Quality Planniny and Standards.
Although the information is not ideal, it is the best information cur-
rently available to EPA, and it was used in these calculations.
The MKI study reports the weight- and volume-percent of hydro-
carbons by category (paraffin, olefin, or aromatic) and carbon number
for 20 fuels obtained from retail outlets. Since iy8y and later model
year vehicles will all burn unleaded fuels, the results from the seven
leaded fuel samples were not included in this analysis. Uf the 13
unleaded samples remaining, 8 are regular and b are premium. Since
survey data show that about 7b percent of all unleaded fuel sold is
regular, the data have been weighted 7b percent regular and 25 percent
premium (I-F-144).
Table 2-14 shows the weighted average volume percent of each hydro-
carbon constituent in the liquid phase of the unleaded samples. Table
2-lb shows the average weight percent of the hydrocarbon constituents
in the vapor phase of the samples. Given the weightings in Tables 2-14
and 2-lb, the properties of the typical fuel and its vapors can be
estimated from the properties of the constituents. The density and
heat of combustion for each of the hydrocarbons in these tables are
presented in Table 2-16. The values in Table 2-16 were taken from the
Technical Uata Book - Petroleum Refining, Volume I as compiled by the
American Petroleum Institute (I-F-143). Table 2-16 shows heats of
combustion in terms of Btu/gal only. In order to use this information
with the weightings in Table 2-15, the heats of combustion had to be
converted to btu/lb. This conversion was done usiny the density fiyures
also shown in Table 2-16. The liquid unleaded fuel samples in the
"Northeast Corridor ..." study (I-A-58) had a calculated weiyhted
average heat of combustion of 113,80U Btu/gal and a density of 6.13
Ib/gal. The vapor associated with these samples had a calculated
weiyhted averaye heat of combustion of y7,yuu Btu/lb and a density, in
condensed form, of b.U7 Ib/yal.
2-125
-------
Table 2-14. LIQUID COMPOSITION OF SEVERAL
GASOLINE SAMPLES (VOLUME PERCENT)*
c#
Paraffins
3
4
5
6
7
8
9
10+
Olefins
4
b
6
7
Aromatics
6
7
8
9+
Regular
0.1
6.0
14.6
10.9
8.6
13.5
3.8
2.4
Regular
0.9
3.4
2.0
1.0
Regular
1.0
6.6
8.9
16.3
Premium
0.1
4.6
14.4
7.1
5.4
11.9
4.8
1.6
Premium
l.b
b.2
2.9
U.9
Premium
0.7
lb.4
9.4
14.1
Wtd 75/25
0.1
5.6
14.6
10.0
7.8
13.1
4.0
2.2
Wtd 75/25
1.1
3.8
2.2
1.0
Wtd 7b/2b
0.9
8.8
9.0
15.7
99.9%
Unleaded fuel only.
2-126
-------
Table 2-15. VAPOR COMPOSITION FROM SEVERAL GASOLINE SAMPLES
(WEIGHT PERCENT)*
c#
Paraffins
3
4
5
6
7
8
Olefins
4
5
6
Aromatics
6
7
8
Regular
0.4
38.6
40.7
8.6
1.8
1.1
Regular
3.1
3.7
0.6
Regular
0.5
0.8
0.1
Premium
....
31.4
42.5
6.2
1.0
1.1
Premium
7.6
5.1
2.7
Premium
0.2
2.1
0.1
Wtd 75/25
0.3
36.8
41.2
8.0
1.6
1.1
Wtd 75/25
4.2
4.0
1.1
Wtd 7b/2b
0.4
1.1
0.1
99.91
Unleaded fuel only.
2-127
-------
Table 2-16. HYDROCARBON PROPERTIES
c#
Paraffins
3
4
5 '
6
7
8
9
10
Olefins
4
5
6
7
Aromatlcs
6
7
8
9
10
Density,
Ib/gal
4.22
4.77
5.14
5.51
5.75
5.94
6.1
6.3
5.07
5.43
5.70
5.80
7.37
7.26
7.27
7.3
7.3
Heat of Combustion,
Btu/gal (Net)
19,767
93.10U
100,000
106,000
110,000
114,000
116, 5UO
121,000
98,000
104,000
107,000
112,500
127,270
126,640
128,000
129,000
129,000
2-128
-------
The final factor that must be considered is the combustion effi-
ciency of the purged vapors relative to the fuel from the vehicle fuel
system (i.e., the degree to which the engine uses the vapors from the
canister as if they were vaporized gasoline from the fuel system). As
discussed elsewhere, EPA expects that most vehicles will have feedback
control which will allow a very high combustion efficiency. The EPA's
current best estimate used here is a 1UU percent efficiency. However, .
for sensitivity a combustion efficiency of 80 percent was also evaluated,
With the information collected above, equivalent gallons of gaso-
line recovered by onboard systems were calculated as shown below:
Equivalent gallons of gasoline saved = (EF)(EV)(VMT)0.97 x CE
(MPG)(Eg)
EF = Refueling emission rate (5.9 g/gal) ^
Ey = Heat of combustion for gasoline vapors (42.5 8tu/g)
VMT = Lifetime vehicle miles travelled (100,000 mi LDV,
120,000 mi LOT)
MPG = Fuel economy for the vehicle class and model year of
interest (miles per gallon) [Section 2.7.1.c]
En = Heat of combustion for the liquid gasoline (113,800 Btu/
gal)
CE = Combustion efficiency of purged vapors (U.8 - 1.0)
Using the fuel economies projected for LDV's and LDT's for the
years 1989, 1994, and 2000, which were presented previously, and the
information presented above, the gross equivalent gallons of fuel
recovered for LDV's and LDT's for the model years of interest were
calculated from this equation as shown below:
1989 1994 2000
LDV
LDT*
100%
80%
100%
80%
8
7
13
11
.7
.0
.9
.1
8
6
13
10
.0
.4
.5
.8
7.3
5.9
12.7
10.2
*Same for single and dual tank trucks.
*97,900 Btu/gal T 5.07 1b/gal + 545 g/lb = 42.5 Btu/g.
2-129
-------
2.7.1.3 Net fuel consumption credit. Both the weight penalty and
the gross consumption credit are now expressed in terms of equivalent
gallons of gasoline saved (or consumed) over the life of the vehicle.
The net fuel consumption credit can now be found by simply subtracting
the weight penalty from the gross consumption credit. The net, undis-
counted, lifetime refueling vapor recovery credits, in gallons, for
LDV's and LDT's are shown below:
1989 1994 2000
LDV 6.6 6.0 5.5
LOT (single) 10.3 10.0 9.2
LOT (dual) 5.3 5.1 4.7
LOT Weighted Average 9.3 9.0 8.3
2.7.1.4 Dollarvalue of recovery credits. The dollar values of
the refueling recovery credits were found using the net equivalent
gallon recovery credits presented above. In order to calculate these
dollar values, both the vehicle mileage accumulation rates and the time
value of money must be considered. In discounting recovery credits, it
is necessary to estimate the timing .of vapor recovery. Since the
vapor will be recovered and used as the vehicle is fueled and operated,
the timing of the recovery credits can be approximated by the average
rate at which a vehicle is expected to accumulate mileage.
Mileage accumulation rates for LDV's and LDT's were taken from a
recently published EPA study in support of the 1988 and later model year
LDT/HDE NOX and particulate standards (I-A-105). Table 2-17 shows the
average mileage accumulation and proportion of LDV mileage accumulated
during each year in the vehicle's life. Table 2-18 shows the mileage
accumulation for LDT^ and LDT2, as well as the weighted average LOT
mileage accumulation. The mileage accumulation is sales weighted as
61.9 percent LDTj and 38.1 percent LDT£. The sales weighting is consis-
tent with the splits used in EPA's MOBILES fuel consumption model.
Also shown in Table 2-18 is the proportion of LOT lifetime mileage
accumulated in each year. The mileage accumulation figures shown in
Tables 2-17 and.2-18 represent the average amount of mileage expected
to be accumulated by a "model vehicle" during the first 20 years of
use. The mileage accumulation figures reflect both changes in usage
patterns over time and expected vehicle scrappage rates. Although some
2-130
-------
Table 2-17. LDV MILEAGE ACCUMULATION
Year
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
2U
Miles
14,580
13,200
10,870
9,830
8,440
7,630
7,070
5,980
4,810
4,110
3,490
2,780
1,860
1,400
1,080
72U
680
580
550
340
Proportion of LDV
Mileage Accumulation
0.1458
0.1320
0.1087
0.0983
0.0844
0.0763
0.07U7
0.0598
0.0481
0.0411
0.0349
0.0278
0.0186
O.U140
0.0108
0.0072
0.0068
0.0058
0.0055
0.0034
Source - "Regulatory Impact Analysis, Oxides of Nitrogen
Pollutant Specific Study and Summary and Analysis of
Comments," U.S. EPA, OAR, OMS, March 1985.
2-131
-------
Table 2-18. LOT MILEAGE ACCUMULATION
Year
1
2
3
4
5
6
7
8
9
1U
11
12
13
14
15
16
17
18
19
2U
LDT1
(miles)
17,394
lb,373
13,553
11,917
1U.447
9,127
7,944
6,884
5,937
4,986
4,239
3,574
2,983
2,459
1,996
1,587
1,226
91U
633
479
LDT2
(miles)
18,352
16,149
14,175
12,409
10,831
9,421
8,164
7,044
6,048
5,058
4,281
3,594
2,987
2,451
1,980
1,568
1,206
891
617
465
LOT
(miles)
17,759
15,669
13,790
12,104
10,593
9,239
8,028
6,945
5,979
5,013
4,255
3,582
2,985
2,456
1,990
1,580
1,218
903
627
474
Proportion of LOT
Mileage Accumulation
0.1419
0.1252
0.1102
0.0967
0.0846
0.0738
0.0641
0.0555
0.0478
0.0400
0.0340
0.0286
0.0238
0.0196
0.0159
0.0126
0.0097
0.0072
0.0050
0.0038
Source - "Regulatory Impact Analysis, Oxides of Nitrogen Pollutant
Specific Study and Summary and Analysis of Comments," U.S.
EPA, OAK, OMS, March 1985.
2-132
-------
vehicles may survive past the 2Uth year, the total mileage accumulation
is negligible.
Two more pieces of information are needed to calculate the dis-
counted dollar value of the fuel recovery credits: (1) the dollar
value to the vehicle operator of the fuel recovered and (2) the appro-
priate discount rate. The first piece of information needed is an
estimate of the price of a gallon of gasoline. Gasoline prices include .
the wholesale price for the fuel itself, profit/overhead, and taxes.
In terms of the costs to society, the fuel tax can be considered a
transfer payment, so EPA has decided not to include taxes on gasoline
in the recovery credit gasoline price. Therefore, the March 1985 gas-
oline price of $1.2U/gallon (I-F-133) was reduced by $0.22/gallon
(consumption-weighted Federal and average State taxes) (I-B-23) to result
in a recovery credit gasoline price of $U.98/galIon. The discount rate
used in this analysis is the standard EPA discount rate of 10 percent
per annum.
Given the total amount of fuel recovered, the timing of that
recovery, the recovery credit gasoline price and the discount rate, it
is a simple matter to calculate the discounted dollar value of the
recovery credits. The calculation for LDV's is described below and
summarized in Table 2-19. Table 2-2U shows the LOT calculations.
The first step was to calculate the total amount of fuel recovered
during each year of the vehicle life. This was done by multiplying
the fractional mileage for each year of life by the total equivalent
gallons of fuel recovered (calculated previously). Then the gallons
recovered during each year were multiplied by the recovery credit
gasoline price, $U.98/gallon. Finally, the dollar value of the fuel
recovered for each year was discounted at 1U percent per annum back to
the beginning of year one, assuming the credit was received at mid-
year.
The total discounted dollar values of the recovery credits are
shown below:
Refueling Vapor Recovery Credits - 1984 Dollars
LDV LUT (single tank) LPT (dual tank)
1989-1993 4.26 6.5U 3.3b
1994-1999 3.85 6.31 3.22
2UUU + 3.53 5.93 3.U9
2-133
-------
Table 2-19. CALCULATION OF NPV OF RECOVERY CREDITS -- LDV, 1989
A
Fraction
B
Dollar Value
of Fuel
of Recovered in
C
Fuel Recovered Given Year
Year
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
During Gi
Year*
0.1468
0.132:0
0.1087
O.U983
0.0844
0.0763
0.0707
0.0598
0.0481
0.0411
0.0349
0.0278
0.0186
0.0140
0.0108
0.0072
0.0068
0.0058
0.0*155
0.0034
Net Present Value
credi
t comes at
ven (A x 6.6 Gal/
Life x $0.98/yal)
0.943
0.854
0.703
U.636
0.546
0.494
0.457
0.387
0.311
0.266
0.226
0.180
0.120
0.091
0.070
0.047
0.044
0.038
0.036
0.022
at beginning of year 1,
midpoint in year
Discount
Factor
0.953
0.867
0.788
0.716
0.651
0.592
0.538
0.489
0.445
0.404
0.368
0.334
0.304
U.276
0.251
0.228
0.208
0.189
0.171
0.156
assuming
D
Discounted
Dollar Value
of Fuel
Recovered (B x C)
0.899
0.740
0.554
0.455
0.35b
0.292
0.271
0.189
0.138
0.107
0.083
0.060
0.036
0.025
0.018
0.011
.009
.007
.006
.003
$4.24
Taken from Table 2-17 - column labeled "Proportion of LDV Mileage
Accumulation."
2-134
-------
Table 2-20. CALCULATION OF NPV OF RECOVERY CREDITS
LOT, 1989
Year
1
2
3
4
b
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Fraction of
Fuel Recovered
During Given
Year9
0.1419
0.1252
0.1102
0.0967
0.0846
0.0738
0.0641
0.0555
0.0478
0.0400
U.034U
0.0286
0.0238
0.0196
0.0159
0.0126
0.0097
0.0072
O.OObO
0.0038
Dollar Value
of Fuel
Recovered in
Given Year
(A*0.10.3*0.98)
1.43
1.26
1.11
0.98
0.85
0.74
0.65
0.56
0.48
0.40
0.34
0.29
0.24
0.20
0.16
0.13
0.10
0.07
0.50
0.04
Discount
Factor
0.9b3
0.867
0.788
0.716
0.651
0.592
0.538
0.489
0.445
0.404
0.368
0.334
0.304
0.276
0.251
0.228
0.208
0.189
0.171
0.156
Discounted
Dollar Value
of Fuel
Recovered (B*C)
1.36
1.09
0.87
0.70
0.55
0.44
0.35
0.27
0.21
0.16
0.13
0.10.
0.07
0.06
0.04
0.03
0.02
0.01
0.01
0.01
Total NPV of Fuel Recovered
$6.50
aFrom Table 2-18 - column labeled "Proportion of LOT Mileage Accumulation.
2-135
-------
2.7.2 Recovery Credits for Heavy-Duty Vehicles
As is the case for light-duty trucks (LL)T's), EPA expects that on
the average, the addition of onboard controls to heavy-duty yasoline
vehicles (HUGV's) would result in a net improvement in fuel economy.
The net improvement in fuel economy is the difference between a yross
reduction in fuel consumption brought about by the combustion of fuel
vapors previously lost to the atmosphere, and an increase in fuel
consumption resulting from the added weight of the onboard system. The
methodology used to calculate the estimated change in fuel consumption
resulting from the addition of onboard controls to heavy-duty vehicles
is virtually identical to that used previously in this section for
light-duty vehicles. Therefore, the methodology used for the calculation
will only be outlined here. The reader should refer to the portion of
this section on LUV's and LUT's (Section 2.7.1) for a more detailed
description. As in the heavy-duty vehicle cost analysis, it has been
assumed that all HUGV's fall into one of two weight classes: Class lib
(8,501 to 10,000 Ib) or Class VI (19,501 to 26,000 Ib). This is a
reasonable approach since these HUGV GVW classes represent about 8b
percent of 1984 HuiiV sales. Therefore, recovery credits were calculated
only for vehicles representing these two classes.
2.7.Z.I Weight penalty. Four pieces of information are needea
to calculate the increase in fuel consumption, or weight penalty caused
by the added weight of the system. These are: 1) average vehicle
weight, 2) average fuel economy, 3) average control system weight, and
4) sensitivity of fuel economy to changes in vehicle weiyht.
a. Average Vehicle Weights
It was assumed that the average weiyht of the trucks in each of
these classes fall at or near the midpoint of the weight class range.
The average weights of the trucks in Class lib and Class VI were assumed
to be 9,2bO and 22,bill) pounds, respectively.
b. Fuel Economy
Vehicle fuel economy estimates were taken from the MOt5lLE3 fuel
consumption model (I-A-99). The fuel economy projections are a function
of the model year and weiyht class of the vehicle. Table 2-21 shows
the HDGV fuel economy estimates used in this analysis.
2-136
-------
Table 2-21. FUEL ECONOMY PROJECTIONS FOR HDGV'S (MPG)
Vehicle Weight Class
Model Year lib VI
1989 11.16 6.07
1994 11.53 6.17
2UOU 11.98 6.31
2-137
-------
c. Average Control System Weights
Control system weights vary with the amount of activated carbon
needed for refueling vapor storage. Necessary carbon bed size, in
turn, is a function of vehicle fuel tank volume. Fuel tank sizes were
estimated to be 20 gallons for Class lib trucks (40 gallons dual) and
30 gallons for Class VI trucks (8U gallons dual). As discussed in the
onboard system cost analysis, dual tank usage rates of 20 percent and
15 percent were used for the Class lib and Class VI truck fleets,
respectively. Control system weight estimates are based on essentially
the same set of data as used in the light-duty analysis, and a detailed
explanation of the derivation is not presented here. Control system
weight estimates are shown in Table 2-22.
d. Sensitivity of Fuel Economy to Vehicle Weight
There are no readily available figures which specifically relate
changes in average vehicle weight to changes in fuel economy for HDGV's.
Some work has been done in this area, however, and this work suggests
that the fuel economy of heavy-duty vehicles is less sensitive to
changes in vehicle weight than is'the fuel economy of LDT's. Therefore,
the LDT sensitivity value was used in the HDGV analysis, in an attempt
to be conservative in evaluating the effects of the weight penalty and
provide a conservative estimate of the recovery credit. Actually, on a
sales weighted basis, about 80 percent of HUGV's are in Class lib;
those vehicles are similar to LDT's in many ways (drivetrain, body
design, etc.). Thus, their fuel/weight sensitivity would also be
similar. The sensitivity factor used is -0.402 (percent change in fuel
economy .per percent change in vehicle weight).
e. Calculation of Weight Penalties
Table 2-23 presents samples of the calculation of the penalties in
fuel economy associated with the added weight of onboard control systems.
For a more complete explanation of the calculation, the reader should
refer to the light-duty section. The weight penalty estimates are sum-
marized in Table 2-24.
2.7.2.2 Gross fuel consumption credit. The gross fuel consumption-
benefit was found using the following information: 1) the uncontrolled
refueling emission rate, 2) the theoretical control efficiency of on-
board systems, 3) the energy content of typical gasoline expressed on a
2-133
-------
Table 2-22. HOGV CONTROL SYSTEM WEIGHT ESTIMATES* (Ibs)
Vehicle Class Single Tank Dual Tank
lib 5.0 13.1
VI 5.9 19.0
incremental to weight of evaporative emission control system components
for system certified with 11.5 psi RVP fuel.
2-139
-------
Table 2-23. SAMPLE CALCULATION OF WEIUHT PENALTIES
Class lib Class VI
Fuel tank configuration S* (80%) U** (20%) S* (86%) U** (15%)
Average incremental control
system weight (Ib) 5.0 13.1 b.9 19.0
Average vehicle weight (Ib) 9,250 9,250 22,500 22,500
Percent change
in weight (%) 5.41 x 10~4 1.42 x 10'3 2.62 x 10"4 8.44 x 10'4
Sensitivity Factor
(dimensionless) -0.402 -0.402 -0.402 -0.402
Percent change in fuel
economy (%) 2.17 x 10'4 5.69 x 10'4 1.05 x 10'4 3.39 x 10'4
1989 projected fuel
economy (mpg) 11.16 11.16 6.07 6.07
Projected average lifetime
mileage (miles) 110,000 110,000 110,000 110,000
Lifetime fuel consumption
(gallons) 9,857 9,857 18,122 18,122
Lifetime weight penalty
(gallons) 2.1 5.6 1.9 6.1
*Single fuel tank.
**Dual tanks.
2-140
-------
Table 2-24. WEIGHT PENALTY ESTIMATES (gallons/lifetime)
(HDGV's)
Class
Class
lib - Single tank
Dual tanks
VI - Single tank
Dual tanks
(80%)
(20%)
(85%)
(15%)
1989
2.1
5.6
1.9
6.2
Year
1994
2.1
5.4
1.9
6.0
2000
2.0
5.3-
1.9
6.0
2-141
-------
volume basis (Btu/gal), 4) the energy content of refueling vapors
expressed on a weight basis (Btu/lb), and 5) a factor to reflect the
engine's efficiency in burning purged hydrocarbons relative to its
ability to burn fuel delivered from the fuel tank.
Ideally, the uncontrolled emission rate used in this calculation
would be based on the results of a series of refuel ings performed on
heavy-duty vehicles. No data of this nature were available to EPA at
the time of this analysis. Therefore, the refueling emission rate used
in the calculation is that used in the LDV and LOT calculations, approx-
imately 5.9 grams/gallon. As previously discussed, the theoretical
control efficiency of onboard control systems has been shown to fall
between 97 and 1UO percent. Although these figures are based on tests
performed on light-duty vehicles, there is no reason to believe that
systems for HDGV's would be any less efficient. As was done in the
light-duty analysis, the conservative end of the efficiency range, 97
percent, was used. The energy content and density of a "typical"
unleaded fuel and its vapor were estimated in the light-duty analysis,
and the calculation will not be repeated here. (While some HDGV's may
still be certified for leaded fuel in the lyyu's, refueling vapor from
"leaded" fuel will be similar to "unleaded" fuel due to lead phase-down
requirements.) The energy content and density of the liquid were esti-
mated to be 113,800 Btu/gal and 6.13 Ib/gal, respectively. The vapor
associated with this fuel was estimated to have an energy content of
97,900 Btu/lb and a density (in condensed form) of b.07 Ib/gal.
The final factor used in the calculation of the gross refueling.
recovery credits is included to reflect the fact that the truck engine
may not burn purged vapors with the same efficiency that it uses liquid
fuel sent from the fuel tank. Depending on the level of sophistication
of a vehicle's fuel system, EPA estimates that the relative combustion.
efficiency of the purged vapors will fall between 0.6 and 1.0. The
gross recovery credits are calcu-lated here using both the low and high
ends of the efficiency ranges, since the credit is expected to fall
between these values. However, a value of 1.0 was carried through in
further calculations. The value of 1.0 was chosen rather than a lower
figure for two reasons. First, there is currently a trend toward the
use of feedback controlled fuel injection systems for HDGV's, which
2-142
-------
should result in improved fuel metering. Second, the new requirement
that the refueling/evaporative emission canister(s) be purged during
exhaust emission testing will force the manufacturers to design control
systems which deal efficiently with purged vapors.
The equation used in the calculation of the gross refueling recovery
credits is shown below:
Equivalent gallons of gasoline saved =
where:
(EF)(Ev)(E)(CE)(VMT)
MPG (£g)
EF = refueling emission rate (5.9 g/gal)
Ev = Heat of combustion for gasoline vapors (42.5 Btu/g)
e = Theoretical efficiency of onboard control systems (0.97)
CE = Relative combustion efficiency of purged vapors (0.6-l.U)
VMT = Lifetime vehicle miles traveled (110,000 mi)
MPG = fuel economy for the vehicle class and modelyear of
interest (miles per gallon)
£q = Heat of combustion for liquid gasoline (113,81)0 Btu/gal)
Table 2-26 shows the estimated gross refueling recovery credits for
heavy-duty vehicles.
2.7.2.3 Net fuel consumption credit. The final step in the
calculation of the net fuel consumption benefit associated with the
addition of onboard control systems is to subtract the weight penalty
effects from the gross fuel consumption credits. Table 2-26 shows the
net fuel consumption credits for HDGV's in gallons. Table 2-27 shows
an example calculation of the discounted dollar values of the gross
recovery credits. The dollar values of the recovery credits are
summarized in Table 2-28.
2.8 ENFORCEMENT REQUIREMENTS
All of the six commenters who submitted comments on this issue
felt that onboard refueling control would require more enforcement, and
consequently would incur greater enforcement costs, than the Regulatory
Strategies Document estimated would be necessary.
Comment: Several commenters felt that existing enforcement measures
would not be adequate to ensure compliance. Some of these commente'rs
stated that canister tampering would not be checked on most I/M inspec-
2-143
-------
Table 2-25. GRUSS REFUELING RECOVERY CREDITS FUR HUGV's
(gallons)
Relative
Combustion
Efficiency 1989 1994 20UQ
Class lib 100% 21.2 20.4 20.0
60% 12.2 12.2 12.0
Class VI 100% 38.8 38.1 37.9
60% 23.3 22.9 22.7
2-144
-------
Table 2-26. NET FUEL CONSUMPTION CREDIT FUR HDGV's (gallons)
Year
1989 1994 2UUO
Class lib 18.3 17.6 17.3
Class VI 36.2 35.6 35.4
Weighted Average . 22.8 22.1 21.8
2-145
-------
TaDle 2-27. SAMPLE CALCULATION OF UlbCOUNItU KdCOVtKY CKtUlTb ($)
(AH HUUV'S, 1989)
Year
1
2
3
4
5
6
7
8
y
1U
11
12
13
14
15
16
17
18
19
2U
A
Fraction of
Fuel Kecovered
Uuriny Year*
U.141b
0.1274
0.1146
0.0998
U.U9U4
U. 1)745
U.U641
0.0649
0.0469
U.U394
0.0324
0.0269
U.0219
U.U178
U.U152
U.U113
O.UU89
U.UU68
O.UU47
O.UU31
B
Dollar Value
of Fuel
Kecovered in
Year (A x 22.8)
Gal/Life x $U.98/Gal)
3.162
2.847
2.661
2.23U
2.U2U
1.666
1.432
1.227
1.U48
U.88U
U.724
U.601
U.489
0.398
0.340
0.252
0.199
0.152
0.105
0.069
_C
Uiscount
Factor
0.953
0.867
0.788
0.716
0.651
0.592
0.638
0.489
0.446
0.404
0.368
0.334
0.304
0.276
0.251
0.228
0.208
0.189
0.171
0.156
Discounted
Value of Fuel
Kecovered (bxC)
3.013
2.468
2.018
1.597
1.315
0.986
0.770
0.600
0.466
0.366
0.266
0.201
0.149
0.110
0.085
0.058
0.041
0.029
0.018
0.011
$ 14.56
*Source: I-A-1U5.
2-146
-------
Table 2-28. NET DISCOUNTED RECOVERY CREDITS ($)
(ALL HDGV's)
Class lib
Class VI
Weighted Average
i989
11.70
23.10
14.50
1994
11.20
22.70
14.10
2000
11.00
22.60
13.90
'.-.jir-ed assuming 75 percent Class lib and 25 percent Class VI.
2-147
-------
tions and others maintained that even if canisters and associated
systems were subject to such inspections, it would be very difficult to
verify compliance or detect tampering (I-H-2, I-H-57, I-H-58, I-H-1U1).
One commenter stated that the kind of "spot inspections" (presumably
referring to tampering surveys) described in the July 1984 EPA analysis
would not be as extensive or effective as the field inspections already
in place for Stage II. The commenter therefore questioned the disparity
between the projected efficiency of Stage II with no enforcement and
onboard with tampering (I-H-57). One commenter maintained that assembly
line testing would be required on every vehicle for safety reasons
(i.e., to check for backpressure and leaks) (I-H-118). Finally, two
commenters stated that the onboard enforcement costs presented in the
1984 analysis were too low (I-H-90, I-H-101).
Response: Canisters are not now inspected during the course of
I/M inspections because current tampering surveys do not indicate a
strong need for such inspections. The July 1984 analysis originally
included fillpipe tampering as a source of diminished onboard control
efficiency. However, EPA now expects that many manufacturers will
utilize liquid seals rather than mechanical seals in the fillneck, so
removal of the lead restrictor in the fillpipe would have no effect
on onboard efficiency. Also, as was mentioned previously, tampering
rates are expected to diminish as the unleaded fuel price differential
drops as has been the recent trend. That being the case, canister
tampering remains the only potentially significant source of diminished
onboard efficiency. The 1985 NEIC tampering survey indicated average
canister tampering rates of 2.67 percent for LDV's and 3.29 percent for
LDT's. The low canister tampering rates determined by the surveys
indicate to EPA that tampering would not present a significant problem
if onboard systems were used to control refueling emissions.
In the event that I/M inspections of onboard systems should prove
to be necessary at some future time, EPA also does not believe that
identification of tampering would prove to be particularly difficult.
Tampering is unlikely to be a difficult-to-identify act such as hose.
splitting, but would typically take the form of canister and/or hose
removal or cutting of hoses. This type of tampering would be readily
identifiable through even the most cursory of visual inspections.
2-148
-------
Given the low tampering rates described above, EPA does not find
the disparity between the projected onboard and Stage II efficiencies
under minimal enforcement scenarios unreasonable. Rather, it seems
indicative that onboard is not particularly dependent on rigorous
field enforcement measures as appears to be the case with Stage I and
Stage II systems. If onboard vapor recovery system in-use inspections
are implemented in the future, both the costs and benefits would have to
be included in this analysis. In addition, since either fully or
partially integrated refueling/evaporative systems are expected, the
costs would either have to be shared between refueling and evaporative
emissions or the benefits from evaporative control improvements included
in this analysis.
While some selective assembly line testing of onboard systems or
components may be conducted by vehicle assemblers or component manufac-
turers, this would not be conducted for every vehicle system or compo-
nent. This has never been the practice for any emission control or
safety system and there is no reason it would be required for onboard
controls. Manufacturing techniques and controls are sophisticated
enough to provide assemblers and component manufacturers statistical
confidence in their performance. In addition, onboard control systems
are relatively simple and any likely system failures can be designed
out with ease.
The costs estimated by EPA were Agency costs for certification and
Selective Enforcement Audits. This estimate is not low or misleading,
however, because no other enforcement costs are likely to be incurred.
As explained above, EPA does not anticipate that canisters and hoses
would routinely be checked on I/M inspections. Nor does the Agency
foresee any need for performance testing. Thus, no additional enforce-
ment costs are likely to be incurred.
2-149
-------
2.9 REFERENCES (Comment letters are not repeated here. See Chapter 1,
Table 1-1, for a complete list of comment letters.)
I-A-29 Summary and Analysis of Comments on the Notice of Proposed
Rulemaking for the Control of Light-Duty Diesel Particulate
Emissions from 1981 and Later Model Year Vehicles, Chapter V -
Economic Impact. U.S. EPA. October 1979.
I-A-40 Motor Vehicle Tampering Survey - 1982, 1983 and 1984. U.S.
EPA, Office of Enforcement and Legal Counsel/Air and Radiation.
EPA-330/1-83-001. April 1983, July 1984 and September 1985.
I-A-55 Evaluation of Air Pollution Regulatory Strategies for Gasoline
Marketing Industry, U.S. Environmental Protection Agency,
Office of Air and Radiation, Office of Air Quality Planning
and Standards, and Office of Mobile Sources, EPA-450/3-84-012a,
July 1984; Including Appendix A: "The Feasibility, Cost and
Cost Effectiveness of Onboard Vapor Control," Glenn W. Passavant,
U.S. EPA Report EPA-AASDSB-84-01, EPA, OAR, QMS, ECTD, SDSB,
March 1984.
I-A-58 Northeast Corridor Regional Modeling Project - Determination
of Organic Species Profiles for Gasoline Liquids and Vapors.
Prepared by MRI, Kansas City, MO, for U.S. EPA, Research
Triangle Park, NC. EPA-450/4-80-036a. December 1980.
I-A-66 Study of Volatility and Hydrocarbon Emissions from Motor
Vehicles. U.S. EPA, Office of Mobile Sources. EPA-AA-SDSB-
85-5. November 1985.
I-A-69 Refueling Emissions from Uncontrolled Vehicles. U.S. EPA,
Office of Mobile Sources. EPA-AA-SDSB-85-6. 1985.
I-A-70 Tech IV Credit Model: Estimates for Emission Factors and
Inspection and Maintenance Credits for 1981 and Later Vehicles
for MOBILES. U.S. EPA, Office of Mobile Sources. EPA-AA-IMG-
85-6. October 1985.
I-A-71 Draft Recommended Test Procedure for the Measurement of
Refueling Emissions. U.S. EPA, Office of Mobile Sources.
EPA-AA-SDSB-85-5. July 1985.
I-A-73 Control of Air Pollution from New Motor Vehicles and New
Motor Vehicle Engines. Federal Certification Test Results
for 1984 Model Year. U.S. EPA, Office of Mobile Sources.
1984.
I-A-74 Control Characteristics of Carbon Beds for Gasoline Vapor
Emissions. U.S. EPA. EPA-600/2-77-057. February 1977.
2-150
-------
I-A-76 Refueling Test Procedures. Workshop held at Motor Vehicle
Emissions Laboratory. U.S. EPA, Ann Arbor, MI. April 10,
1986.
I-A-77 "Costs of Onboard Vapor Recovery Hardware," Jack Faucett
Associates and Mueller Associates for U.S. EPA, 1985.
I-A-93 "Transmittal of Onboard Test Data," Series of U.S. EPA Memoranda,
William Pidgeon to Addressees, EPA, OAR, QMS, ECTD, TEB, August 27,
1985 to January 23, 1986.
I-A-97 "Cap Removal Emissions," U.S. EPA Memorandum, Gary Holladay to
The Record, EPA, OAR, OMS, ECTD, SDSB, November 22, 1985.
I-A-99 Wollott, M.A. U.S. EPA and D.F. Kahlbaum, Computer Sciences,
Corp. Mobile 3 Fuel Consumption Model. EPA-AA-TEB-EF-85-2.
Office of Mobile Sources. U.S. EPA, Ann Arbor, MI. February
1985.
I-A-100 "Investigation of Passenger Car Refueling Losses," APTD-1453,
Scott Research Laboratories for U.S. EPA, September 1972.
I-A-101 Cost Estimations for Emission Control-Related Components/Systems
and Cost Methodology Description. U.S. EPA, Ann Arbor, MI.
EPA-460/3-78-002, 1978. (Available in Docket A-80-18.)
I-A-104 "Light-Duty Automotive Fuel Economy ... Trends Thru 1985,"
Heavenrich, Murrell, Cheng, Loos, U.S. EPA, SAE Technical Paper
Series #850550, 1985.
I-A-105 "Regulatory Impact Analysis, Oxides of Nitrogen Pollutant Spe-
cific Study and Summary and Analysis of Comments," U.S. EPA, OAR,
OMS, March 1985.
I-A-106 "Update of EPA's Motor Vehicle Emission Control Equipment Retail
Price Equivalent Calculation Formula," Jack Faucett Associates
for U.S. EPA, September 4, 1985.
I-A-107 "In-Use Evaporative Canister Evaluation," U.S. EPA Report
, EPA-.460/3-85-003, EPA, OMS, ECTD, December 1985.
I-A-108 "Additional Mini-Canister Evaluation," U.S. EPA Report, EPA-
460/3-85-010, EPA, OMS, ECTD, December 1985.
I-A-109 "Evaluation of the Feasibility of Liquid Fillneck Seals," U.S.
EPA Report EPA-AA-SDSB-86-03, EPA, OAR, OMS, ECTD, SDSB, December
1986.
I-A-111 "Investigation of the Need for In-Use Dispensing Rate Limits and
Fuel Nozzle Geometry Standardization," U.S. EPA Report EPA-AA-
SDSB-87-07, EPA, OAR, OMS, ECTD, SDSB, May 1987.
I-A-112 "Safety Implications of Onboard Vapor Recovery Systems, U.S.
EPA Report EPA-AA-SDSB-87-05, EPA, OAR, OMS, ECTD, SDSB, June
1987 Final Report.
2-151
-------
I-B-18 Memorandum from Kelneman, Martin, U.S. EPA/SUSB, to Gray,
Charles, U.S. EPA/ECTU. March 6, iy?y. Status of In-House
Refueling Loss Measurements.
1-B-iy Memorandum from Heiser, Daniel , U.S. EPA/SUSB, to G
Charles, U.S. EPA/ECTU. January 1984. Significance of
'Popping and Hissing1 Emissions from an LUV Gasoline Tank.
I-B-2U Memorandum from Johnson, Robert, U.S. EPA/SUSB, to the Record.
September 2, 1986. Cost of Crash Testing to Assure Fuel
System Integrity for Unboard Systems.
I-B-21 Memorandum from Passavant, Glenn, U.S. EPA/SUSB, to Cristofaro,
Alexander, U.S. EPA, Regulatory Policy Ui vision. January 14,
1986. Fuel Consumption Sensitivity Analysis.
I-B-22 Memorandum From France, Chester, U.S. EPA/SUSB, to Wilson,
Richard, U.S. EPA/UMS. July 28, 1986. Review of General
Motors' Comments on Fuel Economy Impacts of Unboard Regulation.
I-B-23 Memorandum from Shedd, Steve and Robson, John, U.S. EPA/UAQPS,
to Weigold, Jarnes, U.S. EPA/UAQPS. Uctober 17, iy8b. Recovery
credit gasoline prices for the gasoline marketing analyses.
I-B-27 "Cost of Activated Carbon Used in Evaporative Canisters," U.S.
EPA Memorandum, Uaniel Heiser to The Record, EPA, UAR, UMS, ECTU,
SUSB, July 1U, iy84.
I-B-28 Workshop on Refueling Test Procedure held at the Motor Vehicle
Emissions Laboratory on Uctober 17, iy84.
I-B-37 Memorandum from Passavant, Glenn, U.S. tPA/SUSB, to Shedd,
Stephen, U.S. EPA/ESEU. January 28, iy87. Updated Computer
Outputs for Stage II and Unboard Analysis.
I-B-38 "Light-Duty Vehicle Certification Cost," U.S. EPA Memorandum,
Uaniel P. Hardin, Jr., to Edward J. Brune, EPA, UANR, UMSAPC,
CSU, March 13, iy7b.
I-B-41 "Investigation of Bladder Tank Costs," U.S. EPA Memorandum,
Robert- J. Johnson to The Record, EPA, UAR, UMS, ECTU, SUSB,
February 26, iy87.
I-B-42 "Incremental Costs and Cost Effectiveness of Unboard Controls for
HUGVs," U.S. EPA Memorandum, Robert J. Johnson to The Record,
EPA, OAR, OMS, E'CTU, SUSB, February 26, iy87.
I_U-46 Letter plus enclosures from Gobis, L.P., MVMA, Uetroit, MI,
to France, C.J., U.S. EPA/SUSB. March 7, 1985. Transmittal
of graphics and list of major MVMA concerns on. refuel ing test
proposal .
I-U-263 Letter from Ito, Kenji, Toyota Technical Center, Ann Arbor,
MI, to Gray, Charles, U.S. EPA/ECTU. June 23, 1986. Toyota's
additional information regarding onboard refueling control
system.
2-162
-------
i-U-269 Draft of "Manufacturiny Cost and Ketail Price Equivalent of On-
board Vapor Recovery System for Gasoline Filling Vapors," Leroy
H. Lindyren for tne American Petroleum Institute, 1983.
I-U-318 Comments on "Notice of Proposed Rulemakiny for Gaseous Emission
Regulations for 1987 and Later Model Year Liyht-Uuty Vehicles
and for 1988 and Later Model Year Liyht-Uuty Trucks and Heavy-
Uuty Enyines; Particulate Emissions for 1988 and Later Model Year
HUUE." Available in Public Docket A-8U-18.
I-u-319 Memorandum, David E. Martin, Director, Automotive Safety Enyineer-
iny, GM, to Barry Felrice, NHTSA, March 24, 1986.
I-F-17 On-Board Control of Vehicle Refueliny Emissions Demonstration
of Feasibility. American Petroleum Institute (API) Publication
No. 43U6. API. washinyton, D.C. October 1978.
l-F-78 A Report to the Leyislature on Gasoline Vapor Recovery Systems
for Vehicle Fueliny at Service Stations (Staye II Systems).
State of California Air Resources Board. March 1983.
I-F-98 Cost Comparison for Staye II and Un-Board Control of Refueliny
Emissions. American Petroleum Institute. Washinyton, u.C.
January 1984.
I-F-133 1984 National Petroleum News (NPN) Factbook Annual Issue.
I-F-134 "Hiyhway Statistics 1984," U.S. DUT Report, FHWA-HP-HS-84, DOT,
FHWA, 19 8b.
I-F-135 "Analysis.Memorandum: Desiyn Factor Update," Eneryy and Environ-
mental Analysis, Inc.,. for U.S. DOE, 1982.
I-F-142 "1977 Census of Transportation - Truck Inventory and Use Survey,"
U.S. Department of Commerce, 1980.
I-F-143 Technical Data Book - Petroleum Refining, Volume I. American
Petroleum Institute, 1982.
I-F-144 "Petroleum Marketiny Monthly," Eneryy Information Administration,
October 1984 throuyh September 1985.
I-F-146 "Gasoline Averaye Prices per Gallon, U.S. City Averaye and
Selected Areas," U.S. Department of Commerce, Bureau of Labor
Statistics, 1987.
I-H-lbl "Analysis of Selected Staye II Issues for Dallas, Texas, and
Philadelphia, Pennsylvania," Radian Corporation for API, July Ib,
19 8b.
I-H-lb8 "Vehicle Onboard Refueliny Control," American Petroleum Institute
Publication No. 4424, March 1986.
2-lb3
-------
3.0 STAGE II CUNTRULS
3.1 STAGE II TECHNOLOGY
Comment: Several commenters stated that the recently certified
Stage II equipment and nozzles have proven to be durable and reliable,
and have worked well in the hands of both service station attendants
and customers. The commenters felt early technical problems with Stage
II nozzles used in California in the mid-1970's have been substantially
reduced through debugging and redesiyn (I-U-b6, I-D-59, I-H-2, I-H-58,
I-H-99, I-H-107, I-H-127, I-H-131). Other commenters noted that the
controls are thoroughly demonstrated and are being used in California
and Washington, D.C., with satisfactory results (I-H-22, I-H-b3, I-H-1UU,
I-H-101, I-H-115, I-H-128). One of these commenters pointed to implemen-
tation of Stage II controls in California and Washington, U.C., and
proposals to implement controls in seven other States as indicative of
the confidence that can be placed in Stage II for controlling regional
ozone formation (I-H-100).
A control equipment manufacturer added that, although present
production nozzles work quite satisfactorily, the company has developed
a new Stage II nozzle that is almost as light and small as a regular
nozzle. The manufacturer indicated this low maintenance nozzle'was due
to be placed in service in 198b (I-H-131).
Several commenters claimed that Stage II is not a satisfactory
method of vapor recovery and that the technology is relatively experi-
mental and unproven (I-H-1, I-H-1A4, I-H-1A28, I-H-13, I-H-52, I-H-67,
I-H-84). Three commenters indicated that Stage II vapor recovery has
not been endorsed by the U.S.-EPA since no CTG has ever been published
for Stage II (I-H-1, I-H-lb, I-H-84).
Response: The Agency, in reviewing the operating experience
reported with newer Stage II equipment, feels that Stage II is a tech-
nologically feasible approach to vapor emission control. In addition,
the trend is toward improved, more acceptable components, as referred
to by one of the commenters. To the extent that States adopt Stage II
in ozone nonattainment areas,, it is possible that further improvements in
the systems would occur as the increased market led to further competition
among manufacturers. Regardless of the improvements, Stage II equipment
3-1
-------
has the potential (real or perceived) for beiny less convenient to use
then conventional equipment. The following sections discuss various
aspects of Staye II controls.
Comment: Une comrnenter felt that closed Staye II systems, such as
the Hirt vacuum-assist system, would not suffer the deyree of spit-back
hazard that characterizes pressurized refueliny concepts such as the
Staye II balance system (I-H-74).
Response: All three types of Staye II systems, the vapor balance,
hybrid, and vacuum assist, were evaluated on an equal basis with no
attempt made to establish a ratiny of their performance. Specific
information was not available on the deyree of spit-back or spillaye
associated with each one of these Staye II systems. As a result, a
comparison of the spit-back and spillage hazard of the three systems
could not be made. It should be noted, however, that all three system
types have been certified and are operating in California. The CAKB
has indicated that, while spit-back was a problem with early Staye II
systems, the problem has been solved in current systems throuyh nozzle
modifications and throuyh the use of hiyh hany hoses and hiyh hose
retractors (I-E-b3).
3.2 STAGE II DESIGN AND SAFETY CONSIDERATIONS
Comment: Une commenter stated that one particular model of
Staye II nozzle in current use does not latch securely to approximately
bu percent of the domestic and imported automobiles, as well as motor-
cycles, yas cans, vans, trucks, and many off-the-road vehicles. The'
commenter also stated there is no uniformity amony automobile manufacturers
as to the fillpipe location and openiny, unleaded inserts are not always
correctly installed, and it is difficult to determine whether the tank is'
really full due to the variable shutoff sensitivity of nozzles. This com-
menter also indicated that duriny refueliny, some fuel can recycle back
throuyh the vapor recovery hose. The commenter stated that in order to
purye the aboveyround system of yasoline (as reportedly recommended by the
manufacturer), the fuel is often allowed to drain to the pavement,
which allows vapors to escape and exposes the employee (I-H-18). A
number of commenters based their opposition to Staye II on the problems
and expense actually experienced in California and the District of
Columbia (I-H-1A4, I-H-1A22, I-H-1A27, I-H-18, I-H-bl, I-H-84).
-------
Response: Stage II controls are currently hanaliny about y percent
of the nation's yasoline in California and the District of Columbia
with yenerally satisfactory results. Staye II systems in some areas of
California have been installed for over ID years and have been demonstrated.
tquipment manufacturers claim that preliminary technical problems with
Staye II have been reduced by design improvements, and that customer
acceptance problems also are beiny reduced.
The California Air Resources Board (CARB) has analyzed incidents
of spillaye and other types of yasoline liquid loss due to incompati-
bility between nozzles and vehicle fillpipes. Leyislation, passed in
California in 1976, requires vehicle manufacturers to desiyn uniform
vehicle fillpipes to allow a tiyht seal and to accept specified gaso-
line dispensiny rates without gasoline spillaye and spit-back, and
without shuttiny off prematurely. In response to reports of yasoline
recirculation in Staye II systems, California issued a new standard
requiring vapor recovery nozzles to have parallel shutoff mechanisms
(a shutoff activated by the liquid level in the fillnecK and a shutoff
activated by sensiny liquid in the vapor hose). Also, hiyh-retractor
twin and coaxial hose or high-hang coaxial hose configurations have
been found to allow a yasoline dispensiny rate as high as ID gallons
per minute. This increase in dispensiny rate may improve the reli-
ability of shutoff mechanisms, thus reduciny the potential for yasoline
recirculation. It is asserted that this hiyher rate should not result
in increased incidences of spillage and spit-back of yasoline during
refueling (I-F-78).
Comment: - Kour commenters stated the belief that Staye II systems
present a fire/safety problem (I-H-1, I-H-1A3, I-H-18, I-H-84). Unu
of the commenters said that the L).C. Fire Marshal had offered testimony,
at a roundtable discussion held by the U.C. City Council Transportation
Committee, that Staye II vapor recovery nozzles are unsafe (I-H-18).
Another commenter referred to a iy81 study by Radian Corporation for
the Texas Air Control Board, which claimed that the vacuum assist Staye
II system has potential safety hazards with explosive conditions in
piping and vehicle tanks, al-ong with leakage of vapors from recovery
lines and units (I-H-84). Une manufacturer of Stage II systems commented
that vacuum assist is an effective fire hazard reduction technique (I-H-12y).
3-3
-------
Response: In response to these comments, the California Air
Resources Board (CARB) was contacted to determine whether there had
been any incidents of fire or explosion related to Staye II systems in
California or whether CARB considered the systems to be unsafe in any
way (I-E-b4). A CARB representative related that the systems in use
in California are considered safe, and no incidents of fire or explosion
have occurred. Refueliny is actually considered safer with Staye II in
place than without it because fumes around the islands are reduced and
the storaye tank vapor spaces are kept saturated (above the explosive
ranye). The State Fire Marshal approves all systems as safe before
they are certified for use.
3.3 STAGE II CUNTRUL EFFICIENCY*
Comment: A number of commenters thouyht EPA's assumed Staye II
in-use efficiency was too low (I-H-2, I-H-b7, I-H-58, I-H-74, I-H-82,
I-H-93, I-H-101, I-H-114, I-H-127). Two of these commenters referred
to tests performed on Staye II systems in customer service in California
(I-H-2, I-H-127). Uther commenters state that the experience of the
South Coast Air Quality Manayement District (SCAQMU) of California is
that Staye II equipment attains more than yb percent efficiency the
majority of the time, and is taken out of service if there is a problem
(I-H-b'7, I-H-58). Stage II efficiency estimates are claimed to be too
low because they were based on the initial California proyram administration
and on data for first-generation equipment that is now obsolete and
illegal to install (I-H-82, I-H-114). The b6 percent efficiency cited
by EPA assumes that there is no field enforcement. A commenter thinks
this is unrealistic since an in-use efficiency of yreater than 8U
percent'can be" insured with annual inspections (I-H-118). Another cites
results from the SCAijMU that show a 98.6. percent efficiency at the nozzle./
fillneck interface for a Hirt system usiny a Rudolph nozzle (I-H-74).
Two commenters thought the Staye II in-use efficiency assumed
in the evaluation was too high (I-H-84, I-H-120). Une of them pointed
out that a July 1981 study by Radian Corporation for the Texas Air
Control Board estimated that the balance system would provide 8b" to 88
percent control with a standard nozzle, compared with 93 to 9b percent
with a no-seal, no-flow nozzle; and the vacuum system would achieve
*1984 Federal Register topic.
3-4
-------
bb to 9u percent with a standard nozzle and up to 97 ercent witn a
no-seal, no-flow nozzle. This commenter mentioned that CAKB had
estimated that Stage II equipment defects may reduce California's 13U
tons per day of emission reductions by 3 to 16 percent (I-H-84).
Another commenter referred to a report to the California Leyislature
that showed actual in-use efficiencies to be only bU to 92 percent.
The commenter felt that, although enforcement data from actual experience
in California was used, California may not provide a useful "average"
from which to predict nationwide in-use efficiencies for Stage II,
because it assumes a vigorous, costly, and difficult enforcement program
(I-H-12U). Four commenters thought Stage II would cause serious public
acceptance problems that would drastically offset in-use effectiveness.
In fact, the commenters did not consider it to be certain that an b6
percent efficiency level of compliance could be obtained for any feasible
enforcement effort (I-H-1U2, I-H-1U8, I-H-119, I-H-12U).
Response: All three types of Stage II systems were assigned a
theoretical (certification-level) control efficiency of 9b percent
(Table 3-1 of the strategies evaluation document Reference 1 in Chapter
1). In-use efficiencies in the July 1984 analysis were determined
based on various assumed frequencies of enforcement, i.e., facility
inspections to determine the need for and to enforce necessary repairs.
The balance type system was estimated to have an in-use efficiency
range of b4 percent (minimum enforcement) to b6 percent (annual enforce-
ment), the hybrid system a range of 62 to b8 percent, and the vacuum
assist system a range of bb to 86 percent. The weighted average for
all types of systems ranged from bo to b6 percent.
The £PA evaluated new data in an effort to update the in-use
efficiency estimates (see Appendix A of the Vol. I KM referenced in
Chapter 1). The Agency examined a recent report on inspection of all
Stage II service station installations in the Washington, U.C., area,
and revisions were subsequently made to the estimates for the frequency
and types of defects affecting Stage II systems. Usiny this information,
the Agency's estimate for the lower end of the Stage II efficiency
range was adjusted from b6 to 62 percent.
The EPA also evaluated the latest California Air Resources Board
data, which were presented in the 19b3 Report to the Legislature (I-F-78),
However, the data were insufficient to differentiate between first-
3-b
-------
based solely on third-generation systems could not be estimated.
Additional data were obtained from randomly selected service stations
in the Bay Area of California, which indicated an in-use efficiency
of yu to y2 percent; however, the data were considered inadequate to
update the in-use fiyure for the entire State of California. There-
fore, the upper end of the in-use efficiency ranye was not cnanyed
from the previously used value of 86 percent. Enforcement costs associ-
ated with this upper efficiency value were calculated and included in the
costs evaluated in the Volume I KIA for Staye II reyulatory strateyies.
Comment: One commenter claimed that spills and leaks reduce the
effectiveness of Staye II (I-H-by). Several other commenters indicated
that customer avoidance actions, such as toppiny off, would reduce the
control effected by Staye II (I-H-33, I-H-4U, I-H-44, I-H-68, I-H-71,
I-H-76, I-H-77, I-H-78, I-H-84, I-H-85, I-H-8y, I-H-y6, I-H-y7, I-H-1U2,
I-H-lUb, I-H-1U6, I-H-11U, I-H-121, I-H-123).
Response: It is true that spills and leaxs would reduce the
effectiveness of Staye II or any other vehicle refueliny control system.
Recent desiyn chanyes implemented in California are intended to reduce
the spillaye and spit-back problems found in first-yeneration systems.
Uesiyn chanyes include new nozzle desiyns, hi.jh hose retractors, and more
strinyent nozzle certification procedures. Customer avoidance actions
also can reduce Staye II efficiency. California officials contacted ad-
vise that, when undertaking Staye II, an active initial consumer education
proyrarn be initiated at an early staye in order to limit such problems.
3.4 CUSTS UF STAGE II*
Comment: Several commenters felt that the cost estimates for Staye II
controls used in the EPA study were too low., since they were based on
California data from iy78. Some of these commenters stated that the
costs in 1978 were based on first-yeneration systems, while current
2nd- and 3rd-generation costs are much hiyher (I-H-2U, l-H-34, I-H-37,
I-H-44,'I-H-62, I-H-66, I-H-68, I-H-89, I-H-91, I-H-1U8, I-H-lUy, I-H-llU,
I-H-126). One of the commenters believed that the Staye II costs developed
by EPA severely underestimate the costs actually experienced with real
systems, indicating that annualized cost estimates made by CARb are about
*iy84 Federal Reyister topic.
-------
57 percent yreater than EPA's estimates (I-H-109). Two other commenters
said that a recent American Petroleum Institute (API) cost study (l-F-98)
contained data on the actual construction and maintenance costs of
Stage II, showing that the costs derived from 1978 California data are
unreasonably low (I-H-20, I-H-34). One commenter noted that EPA's
Staye II cost analysis relied heavily on the Luken report (I-A-22).
The commenter felt the Mueller comparison of API, EPA, and Lundyren
cost estimates, and, in particular, API's analysis of CARB data showed
that Staye II is substantially more expensive than the EPA or Luken
estimates (I-H-12U).
Several commenters cited various Staye II costs from studies by
the Petroleum Marketers Association of America (PMAA), Sierra Research
Corporation, API, and individual personal research and experience.
Many of the commenters referred to the study performed by the PMAA that
estimated the cost of Staye II controls at $16,DUD for each service
station (I-H-19, I-H-33, I-H-54, I-H-56, I-H-64, I-H-71, I-H-77, I-H-79,
I-H-80, I-H-86, I-H-96, I-H-97, I-H-lUb, I-H-1U6, I-H-11U, I-H-121,
I-H-123). Three other commenters referred to the January 1984 API
study which estimates that for a Stage II balance system the cost
ranges from $7,74U for a three-nozzle to $23,780 for an 18-nozzle
outlet, and for a Stage II vacuum assist would range from $12,860 for a
three-nozzle outlet to $29,770 for a laryer outlet (I-H-24, I-H-84,
I-H-91). Other commenters indicated that the cost to the averaye
marketer would ranye between $1,200 and $2,000 per hose on an installation
of Staye II. The commenter's estimates reportedly were based on personal
research and on actual experience in California and Washington, O.C.
(I-H-1, I-H-15, I-H-1A17, I-H-1A24, I-H-1A30, I-H-102). Several other
commenters stated that the necessary capital cost for compliance with
Stage II would mean an investment of approximately $10,1)00 to $22,000
per service station (I-H-15, I-H-1A13, I-H-1A20, I-H-1A25, I-H-87, I-H-108,
I-H-130). One of these commenters supplied a $70.1 million total cost
figure for Staye II equipment within an 8-county area of Illinois
based on an averaye cost for Stage II equipment per station of
$lb,000 (I-H-15). One commenter, a manufacturer of control equipment,
stated that the costs for Stage II systems presented in Table 7-lb of
the EPA report were much different than his experience indicated,
saying that typical installation costs are $1,800 per station (I-H-129).
3-7
-------
This commenter and two others (I-H-108, I-H-119) pointed out that EPA's
Stage II cost estimates, on a per-nozzle basis, increase after nine
nozzles and that such a result is contrary to practical experience
where, once certain expenses are fixed, the marginal costs of a larger
system on a per-nozzle basis should be less. One commenter indicated
that replacement nozzles cost hundreds of dollars apiece and that some
suppliers will not handle an order of less than $500 (I-H-18).
Response: The 1984 analysis drew almost entirely upon published
data to estimate Stage II costs. Costs were put on a common basis
(1982 dollars) using cost indices. Under the new analysis, Stage II
costs have been updated by completing a new and detailed cost analysis.
Table 3-1 indicates the differences in the Stage II costs between the
previous analysis and the new analysis.
Table 3-1. WEIGHTED AVERAGE STAGE II COSTS3
(Retrofit of Existing Stations)
Model
Plantb
1
2
3
4
5
Previous
Capital
Cost, $
5,70U
6,100
6,600
9,800
14,800
Analysis
Annual
Cost, $
1,400
1,300
1,300
1,400
bOO
New Analysis0
Capital Annual
Cost, $ Cost, $
5,700 1,300
7,300 1,400
12,200 2,500
16,100 3,200
23,200 3,100
Weighted average - 80 percent balance systems, 15 percent hybrid
systems, 5 percent vacuum assist systems.
bThe model plant parameters ere described in the Draft Vol. I RIA.
GAnnual costs reflect annual enforcement.
Detailed information on certified systems currently used in
California, manufacturers' cost data, and engineering estimates were
used to develop the new Stage II cost estimates. Retail costs (with
and without quantity discounts) were obtained for all hardware (i.e.,
nozzles, hoses, swivels, piping, etc.) from the manufacturers. Engi-
neering estimates were used to calculate trenching and backfill costs.
These cost data were used to make estimates for the installation of a
3-8
-------
Stage II vapor recove.y system at an existing facility, and also the
additional cost of Stage II equipment associated with the construction
of a new facility. The California data were used because these systems
are currently operational and have been shown to achieve the assumed
control efficiencies, and because detailed third and fourth quarter
1984 cost information was available for each of the individual compo-
nents comprising each of the California-certified systems; i.e., the
individual balance system, the manifolded balance system, the hybrid
system, and two assist systems.
Table 3-2 presents the capital cost estimates from the new
analysis for Stage II systems installed at existing or new facilities
for a "typical" 35,000 gallon per month facility (Model Plant 3). The
selection of a "typical" station was based upon a throughput weighted
average of stations that would require controls under a nationwide
Stage II regulatory strategy. Comparing the costs for this "typical"
station to those provided by the commenters is difficult because the
commenters supplied insufficient descriptive information (e.g., number
of nozzles, system type, other costs included, etc.) for an accurate
comparison to be made. A detailed accounting and cost breakdown of
capital and annualized costs used in the new analysis for each type of
Stage II control system is contained in Appendix B of the Vol. I RIA.
The EPA agrees that the cost per nozzle should decrease as station
size increases because certain fixed costs are divided among more
nozzles. The 1984 analysis reflects this trend as long as the calcula-
tions are based consistently on the upper or lower end of the nozzle
range for each model plant. In the new analysis, specific nozzle
quantities were attributed to each model plant in order to identify
plumbing configurations. As .in the original analysis, the trend of
lower per-nozzle cost as the station size increases is repeated.
Table 3-2. STAGE II REVISED CAPITAL COST ESTIMATES FOR A
"TYPICAL" 35,000 GALLON/MONTH SERVICE STATION ($)*
Type of
System
Balance
Hybrid
Vacuum Assist
Weighted Avg.
Existing
Capital
11,900
12,600
15,400
12,200
Facility
Annual
2,470
2,690
3,470
2,550
New Facility
Capital Annual
6,540 ' 1,740
6,940 1,880
10,190 2,760
6,780 1,810
a
Annual costs reflect annual enforcement.
3-9
-------
Comment; Une commenter stated that any major effort to fix or
replace leakiny underground storaye tanks could significantly lower
Stage II plumbing costs if they were done at the same time. This
should be considered in the final analysis (I-H-127).
Kesponse: It is unlikely that a Stage II and an underground storaye
tank (UST) proyram would coincide. However, in the unlikely event that
they would coincide, EPA analyzed the affects on Stage II installation
costs (See Appendix K of the Draft Volume I, RIA). In this analysis, it
was assumed that Stage II installations and UST repairs would occur simul-
taneously over a 6 year period. Depending on the type of UST repair required,
some or all of the Stage II trenchiny costs could be saved. Using a con-
servative estimate that 36 percent of the existing tank systems leak, the
analysis indicated only a small savings (less than 6 percent) in the
Stage II costs.
Comment: The same commenter felt that the cost analysis should
include higher gasoline consumption, greater emissions controlled, and
different exemption levels. This commenter estimated that incorporatiny
these changes would decrease nationwide NPV costs by 21 percent and
would yield a Stage II cost effectiveness of $442/Mg (I-H-127). This
commenter and one other also felt that EPA cost estimates were over-
stated because the Agency had projected a constant number and size
distribution of facilities, whereas trends are toward fewer stations
and a higher percentage of large stations (I-H-114, I-H-127).
Response: In response to comments received on the initial analysis,
the Agency h.as incorporated several changes in its revised analysis of
Stage II costs-. These changes include a new projection of future
gasoline consumption, a revised projection of the number of service
station facilities, and an analysis of additional facility exemption
levels.
The EPA MUBILE3 computer model was used to project total domestic
gasoline consumption out to the year 20UU. This model utilizes the
best Agency estimates of changes in vehicle miles traveled, fleet miles
per gallon, fleet size, dieselization, and vehicle scrappage rates.
Due to the uncertainty associated with several variables, the original
assumption that consumption between the years 2UUU and 2U2U would
remain constant was retained. The model projected a 16 percent decline
3-10
-------
in total gasoline consumption between 1986 and 2UUU, compared to tne
oriyinal projection of a 23 percent decline during tnis period
(Figure 4-2 in the iyy4 evaluation document). The Draft Volume I KlA
contains details on the new gasoline consumption projections.
The revised projection of service station population was based on
several key assumptions: (1) the total throughput of all types of
stations would decrease in proportion to the decrease in gasoline
consumption; (2) the average throughput per public facility would
remain constant, i.e., the number of public stations would decline; and
(3) the number of private stations would remain constant, i.e., the
average throughput per facility would decline. In addition, this
projection also considered the trend toward larger, more cost effective
facilities. In order to reflect this trend, some throughput at public
stations was shifted from smaller to larger model plants. A more
detailed description of the model service station projections is given
in Appendix D of the Vol. I RIA. The results indicated an overall 23
percent decrease in the total station population, with a 28 percent
decrease in the smallest stations and only an 11 percent decrease in
the largest size stations.
Two additional service station exemption levels were examined as
part of nationwide and nonattainment strategies, and included a determi-
nation of total costs. Under these strategies, all service stations
having throughputs of: 1) <2,UOU gal/month, or 2)
-------
should be lb years, and not 8 years, as in the evaluation document,
since this system uses the same underground pipiny as the balance,
hybrid, and simple closed systems (I-H-74).
Response: In the 1984 analysis, no attempt was made to divide
the capital costs for aboveyround and underground equipment to reflect
different component lives. However, in the new analysis, the capital
cost data for al1 Staye II vapor recovery systems were re-evaluated
and broken down into aboveyround costs, which include the cost for
dispenser components, and underyround costs, which include the install-
ation cost of an underyround pipiny system. As a result, the capital
recovery cost factor used to calculate annualized costs was based on an
equipment life of 8 years for the dispenser and auxiliary equipment,
and 3b years for the underyround pipiny system. Discussions with
dispenser equipment vendors and the manufacturers of the two assist
systems indicated that the additional dispenser equipment needed for
Staye II (retractors, flow limiters, hanyer kits) and the processing
unit equipment will last a minimum of b years, and should last up to
ID years (I-E-19, I-E-2U, I-E-22, I-E-23, I-E-25, I-E-26, I-E-27).
Thus, 8 years was assumed for the capital recovery calculation. Several
vendors of fiberylass piping, which is often used for vapor recovery
pipiny in California, were contacted about the expected life of fiber-
glass pipe used in underyround systems (I-E-37, I-E-38). All vendors
contacted indicated that there was no reason that the pipe should ever
need replacement because of deterioration of materials. One vendor
indicated that it would last at least 3D years (I-E-38). For purposes
of the capital recovery calculation, the underyround piping lifetime
for fiberylass pipe was estimated to be 3b years.
Comment: Une commenter felt that revisions to assumptions in
the Staye II analysis should include: 1) prioritization of phase-in
(laryest stations first), 2) revision of equipment life to 21) years,
and 3) quarterly inspections instead of annual (I-H-114).
Response: The Ayency re-evaluated all of the assumptions made
in its original analysis in light of public comments and other new
information obtained since the evaluation document was issued.
Prioritization of phase-in (largest stations first) would presumably
lead to the fastest emission reduction by controlling the laryest
V
3-12
-------
emitters earlier in the Staye II program. However, for the purpose of
simplifying the analysis, it was assumed that all station sizes would
install equipment under the phase-in schedule used in the analysis.
In the iy«4 analysis, an equipment lifetime of lb years was assumed
for balance and hybrid systems and'8 years was assumed for a vacuum
assist system. As indicated in the previous response, equipment lifetime
was modified for all Stage II systems to 3b years for the underground
fiberylass piping and 8 years for aboveyround dispenser equipment.
The new analysis considers a ranye for the efficiency of
Staye II based on various enforcement levels. The upper end of the
range is based on the enforcement program used in California, which is
as active as any expected in other States. The California AKts esti-
mates that annual enforcement best represents the enforcement program
in effect for the entire State of California (I-E-47). In addition,
an evaluation of quarterly enforcement yielded greater emission reduc-
tions but at a much higher cost, resulting in a worse cost effective-
ness than for annual enforcement. Therefore, analysis of enforcement
levels beyond "annual" was not considered appropriate.
Comment: Une commenter pointed out that with the stimulus of a
larger market, the Staye II equipment manufacturers could develop even
more effective products (than the current fourth-generation Stage II
equipment) at less expense. The commenter said this cost savings was
not reflected in tPA's assessment of Stage II (I-H-118).
Response: The stimulus of a larger market would most likely
result in Stage II equipment vendors developing more effective and less
costly equipflie-nt-. However, the Agency has no way to reliably predict
and quantify or project any effects that an expanded market might have
on the costs and effectiveness of Staye II equipment. Therefore, for
the purpose of comparing strategies, it is considered appropriate to
assume current, known costs in the analysis.
Comment: Une manufacturer of service station control systems
stated that extensive operation proves that the Hirt VCS-2UU Stage I/
Stage II system pays for itself, saying that the saleable gasoline
generated and conserved by the recovered vapors can be seen through the
gasoline inventory records of each-station with the system. He claimed
3-13
-------
typical yasoline recoveries of U.3 percent, or 3UU yallons per month
for a typical (1UU,UUU yal/month) station. This commenter stated that
3UU yallons per month is equivalent to approximately $3UU per month,
which provides a 3-year payoff for the controls (I-H-12y).
Response: It is not the Agency's intention in this analysis to
recommend or differentiate amony specific brands of Staye II systems,
particularly on the basis of cost or cost recovery. The yasoline
recovery credit calculation for both vacuum assist systems (Hirt and
Hasstech) assumed a bt) percent recovery of both displacement losses and
breathiny losses. This resulted in recovery credit cost savinys of
$2UU/month for a Model Plant b station (lab.UUU yal/month).
Comment: Several commenters stated that the additional cost of
Stage II maintenance would add a yearly expense of $1,UUU to $2,UUU per
location (I-H-1A13, I-H-1A24, I-H-1A3U, I-H-lb, I-H-6b, I-H-87). Une
commenter indicated that additional maintenance costs would be $26U per
year per pump (I-H-1). Two commenters claimed that an increase of
about 1/2 cent per yallon of purchased yasoline would cover the cost of
maintenance (I-H-1A4, I-H-65). A commenter felt that operation and
maintenance costs would exceed EPA's estimates since the operation of
Staye II systems will be a continual problem, with costs that cannot be
accurately foreseen at present. Also, the commenter stated the cost to
replace nozzles on hybrid and vacuum systems would be $1UU per year per
nozzle (I-H-24).
Kesponse: Maintenance costs for Staye II systems have been re-
evaluated. The maintenance requirements and the associated annual
costs were obtained from equipment manufacturers and vendors based on
their experience in existing Stage II areas. The annual maintenance
costs were estimated to ranye from $476 for a two-nozzle balance
outlet to $3,U7U for a Ib-nozzle outlet. The hybrid system has the
hiyhest ranye of annual maintenance costs, from $497 for a two-nozzle
station to $3,23U for a Ib-nozzle station. These maintenance cost
estimates were based on several assumptions: 1) that nozzles on all
systems are repl'aced every 2 years; 2) that the vapor hoses on al 1
systems are replaced every 2 years; 3) that the boot/faceplate assembly
on a balance system is replaced three times per year; 4) that the
boot/faceplate assemblies on hybrid and assist systems are replaced
3-14
-------
twice a year; and b) that maintenance of the processing unit on an
assist system is required once a year. Further details reyardiny
maintenance costs for each model plant and the various Staye II systems
can be found in Appendix b of the Vol. I KIA. For a "typical" 6-
nozzle service station, the total maintenance costs have been estimated
as $l,2k:6/yr. This is equivalent to a maintenance cost of about U.3
-------
to be considered under maintenance costs and are not treated as capi-
tal costs in any future replacements. Appendix B of the Draft Volume I
K1A discusses the revised costs and assumptions relating to Staye II
system equipment.
Comment: Une commenter questioned EPA's assumption that tne price
of yasoline will remain constant over the analysis period (until the
year 2U2U). The commenter felt if the price of yasoline increases, as
many forecasters predict, the cost effectiveness of Staye II will
appear more attractive (I-H-118).
Response: Cost data for Staye II were obtained and developed on a
pen-facility basis for each model plant size of each source cateyory
in the yasoline marketiny network. These per-facility costs were
then combined with data on the number of.facilities requiriny control
within each source cateyory and with projections of yaso'line consump-
tion, in order to determine nationwide costs for the period 1987-2U2U.
Unooard costs were determined usiny per-vehicle costs and projec-
tions on the number of new vehicle reyistrations for the period 1987-
2U2U. The nationwide cost estimates for this time period were not
based on a projection of an increase in the price of yasoline. In the
analysis, all prices were treated uniformly, assuminy constant real
prices. It is true that if the price of gasoline were considered to
escalate at a rate yreater than inflation (i.e., the actual value was
increasiny), the recovery credits associated with Staye II and onboard
would increase, and thereby the cost effectiveness of both systems
would appear more attractive.
It'shou-ld-also be pointed out that if yasoline prices decrease, as
they did in iy8b-86, the recovery credits would decrease and the cost
effectiveness would appear less attractive. Since all cost fluctuations
cannot be predicted, especially over a time span as lony as that evalu-
ated in this analysis, tne simplifyiny assumption was to assume a
constant yasoline price.
Comment: Three commenters felt that recovery credits do not
reduce the net cost of Staye II because the prevalent gasoline distri-
bution practice involves forciny the vapors through Staye I equipment
back into truck transports and deliveriny them to a terminal for dis-
3-16
-------
posal. The commenters concluded the gasoline marketer receives no
economic benefit from the collected vapors (I-H-1U2, I-H-1U8, I-H-liy).
Response: AS discussed in Chapter 7, recovery credits are shown
in the analysis for Staye II at service stations because vapors piped
to the underyround storaye tanks from refueliny automobiles serve to
prevent evaporation and subsequent loss of stored product. Thus, the
value of the product not lost throuyh evaporation is applied toward
loweriny the net cost of installiny and operatiny a Staye II system
(Tables 7-12, 7-14, and 7-16 of the iy«4 EPA evaluation report). As
pointed out on paye 7-2U of the evaluation report, and indicated in
Table 7-1U, no yasoline recovery cost credit for the service station
owner is assumed to result from the Staye I system, since the displaced
vapors in this case are piped to the delivery truck tank and do not
influence the product supply controlled by the service station.
3.b CUST EFFECTIVENESS UF STAliE II
Comment; Several commenters felt that Staye 11 would have a
considerably hiyher cost per unit of emission reduction than onboard
(I-H-1, I-H-84, I-H-%, I-H-97, I-H-1U9, I-H-122, I-H-123). One
presented an analysis that included the company's own estimates of the
cost effectiveness of four reyulatory options. The comrnenter indicated
that the cost effectiveness of onboard controls, usiny the hiyher
"extra evaporative emissions" factors, was estimated to be $b3u/My;
Staye II in nonattainment areas (with "creep" to additional areas) was
estimated at $l,6UU/My; and Staye II nationwide, with and without
exemptions, was calculated to be $l,yi)U/My and $3,7l)U/My, respectively
(I-H-luy).
Another of the commenters felt that EPA should- revise its cost
effectiveness estimates, considering that nationwide Staye II would
require nationwide Staye I and, thus, Staye I control cost effective-
ness must also be considered (I-H-122).
Response: The EPA has re-evaluated the cost effectiveness of
the reyulatory strateyies usiny revised values for per-facility costs
and emission factors. For Staye II, a component-by-component cost
analysis was conducted, based on actual systems currently certified in
California (see Draft Volume I, RIA). In the revised analysis, EPA's
3-17
-------
calculations are based primarily on an assumed annual averaye RVP of
12.6 (rather than IU.0), which is more representative of the current
trend in gasoline volatility. However, EPA has also analyzed the
effects of a range of in-use RVP levels, as low as y RVP in the summer,
in light of the Agency's plans to propose volatility limits. The
assumed averaye in-use KVP of gasoline, 12.6 (projected future averaye),
caused an increase in some emission factors of as much as bu percent
(including the factor for excess evaporative emissions) (see Draft
Volume I, KIA). Increasing the emission factors increases both the
emission reductions achievable by the regulatory strategies and the
amount of gasoline recovery credits achievable by all control systems
(thereby reducing net annualized costs). The Draft Volume I RIA contains
a discussion of the revised nationwide cost analysis.
Technically, Stage II controls operate independently of Stage I
controls, and neither affects the emission reduction efficiency of the
other. Therefore, nationwide Stage II does not require nationwide
Staye I. If Staye I were considered in conjunction with Staye II con-
trols, the cost effectiveness of the combined strategy would be better
than for Stage II alone, since the cost effectiveness of Staye I is
better than that of Stage II.
Comment: One commenter pointed out that, while EPA had not drawn
any firm conclusions in its analysis, the text of the evaluation report
appeared to suggest that onboard vapor recovery is the preferred alter-
native. The commenter further stated that, on the other hand, the
table on paye l-2b indicates a better cost effectiveness for Stage II.
The commenter-fe.lt that this contradiction was not explained (I-H-21).
Response: In both the text and tables, the evaluation report sug-
gests that the relative attractiveness of onboard and Stage II controls
depends on a variety of assumptions; i.e., on what particular option is
being considered. Exemptions, discount rates, levels of enforcement,
and in-use efficiency estimates influence cost effectiveness. However,
cost effectiveness is only one of many factors involved in makiny the
decision. Other-factors can include costs, emission reduction, risk . .
reduction, energy impacts, and equitable concerns. Results and recommen-
dations depend on these and numerous other factors. The Agency does
not believe that the material presented in the evaluation report represented
3-la
-------
a contradiction. The new analysis (described in the Volume I KIA)
examines additional options, and many of the results in the original
analysis are superceded.
3.b STAGE II ECONOMIC IMPACTS
Comment: Several commenters pointed out the neyative economic
consequences to be expected if facility exemptions from Stage II
requirements were granted. One commenter cited a hearing held in
December 1982, by the District of Columbia Council, at which a number
of gasoline marketers presented testimony concerning customer reactions
to Staye II controls. The commenter presented four instances where
station owners noticed a significant drop in monthly throughput after
the installation of these controls (I-H-125). This commenter and three
others (I-H-y2, I-H-102, I-H-137) felt that similar marketplace distortions
(customer flight) can be expected if exemptions from Stage II are
allowed in other areas requiring Stage II. The first commenter said
that allowing exemptions would, over time, undoubtedly increase the
fraction of gasoline sold without refueling emission controls; consequ-
ently, no exemptions should be allowed because of their "anti-competitive"
effects.
Three commenters suggested that Stage II control requirements
would create market distortions because larger retailers would have to
carry the whole burden of protecting against air pollution, and might
adopt otherwise inefficient marketing approaches to avoid control costs
(I-D-b4, I-H-106, I-H-liy, I-H-120). One of the commenters said a
large retailer might attempt to split his operation into two smaller
exempt stations rather than maintain a single nonexempt station that
operates more efficiently (I-U-b4, I-H-120). The two other commenters
felt that, to the extent that exemptions were granted to smaller businesses,
Stage II would place an unfair burden on high volume, self-service
retailers, by giving them alone the job of controlling air emissions
(I-H-108, I-H-119).
Two commenters pointed out that smaller retailers just above
an exemption cutoff would be at a competitive disadvantage because .of
their greater pen-unit control costs than the larger facilities.
The commenters felt these facilities could face bankruptcies and unemployment
(I-D-b4, I-H-102, I-H-120).
3-19
-------
Une commenter stated that the majority of "major" oil company
service station operators not eliyible for exemptions are, in fact,
small independent businessmen who lease or rent the facility from the
oil company. The commenter felt, while the intent of the Clean Air Act
may have been to exempt "independents," the proposed Stage II exemptions
would hurt many small independent retail gasoline businesses, since
they would have to bear the increased operatiny costs of Staye II and
the probable loss of business (I-H-12b). Another commenter felt that a
considerable seyment of small ousiness yasoline marketers would be
lost if no exemptions from Staye II based on size were yranted (I-H-1U2).
Response: Under any control strateyy where some facilities are
controlled and others are not, there could be some flight of customers
to uncontrolled facilities, if Staye II equipment were considered
more difficult to use. To what extent this occurred would depend on
the location of alternative facilities, the importance of other services
offered by various facilities, and other factors.
More flight miyht result from the choice of exemption options based
on facility size than on nonattainment strategies. Where entire popula-
tion areas must install Staye II controls, the distance to non-Staye II
stations generally would be yreater, and the associated costs of making
the trip to these facilities often would be hiyher, than where a smaller,
exempted facility may be located on the next block.
Under size exemptions, some increase in sales for non-Stage II
retailers is to be expected; however, this increase would be limited.
When exempted stations gain sales, their throuyhputs increase. If
throughputs-increase sufficiently, the stations could lose their
exemptions and be required to install controls.
The Agency agrees that facilities receiving exemptions would not
face tne cost burden associated with Stage II control. Exemptions are
intended to protect smaller facilities whose ability to pay for con-
trols may be restricted, and such exemptions would often improve the
competitive position of smaller facilities with respect to larger,
controlled facilities. The Agency further agrees that, under an exemp-
tion scenario, those potentially experiencing the greatest economic
impact would be retailers just above the size cutoff. The net employ-
ment effects of a regulatory strategy including exemptions are unknown.
-------
As Appendix E in the Draft-Volume I, KI . indicates, the potential service
station closure rate would be considerably smaller under the options
including exemptions than under the no-exemption options. Also, there
are several factors that would influence a firm's viability other than its
business volume, such as location and local competition, sales of
other items, etc. Many smaller facilities with hiyher per-unit costs
are optimal, for the smaller markets they serve, and would likely be
able to pass through the control costs.
The exemption levels used in the analysis were selected to reflect
the requirements of the Clean Air Act and to allow a determination of
the effects of other possible exemption scenarios. Kor example, actual
levels of exemption from Staye II could be determined for a NESHAP
during further regulatory analyses examining a number of specific
levels of exemption. With a CTG, or if a State chose to be more stringent
than a NESHAP program, exemption levels could be prescribed by individual
States for implementation in their own jurisdictions. These specific
levels cannot be predicted with certainty. Whatever their eventual
levels may be, the purpose of the exemptions would not be to hurt, but to
provide relief to, facilities .facing.significant economic burden due to
control costs.
Comment: Numerous commenters objected to Stage II controls on the
grounds that the costs would be prohibitive and would place a great
financial strain on service station owners/operators. Several claimed
that these costs would probably force them to halt gasoline operations
or shut down their facility(ies). Many commenters pointed out the
particular burden that would be imposed on small independent dealers
(I-U-bb, I-H-1-, -I-H-1A1 to I-H-1A32, I-H-4, I-H-8, I-H-11 to I-H-17,
I-H-iy, I-H-2U, I-H-2b, I-H-26, I-H-2^ to I-H-32, I-H-34 to I-H-4U,
I-H-44 to I-H-b2, I-H-b4 to I-H-bfa, I-H-by, I-H-bl to l-H-64, I-H-bb to
I-H-71, I-H-75 to I-H-81,.I-H-84 to I-H-fcy, I-H-137, I-H-138).
Three of the commenters felt that under Stage II requirements
they would stop developing new stations and/or expanding and upgrading
their existing facilities (I-H-1A8, I-H-1A17, I-H-44). Four of them
stated that the 'cost of Stage II systems would have a serious finan-
cial impact in light of the expenses already being incurred to comply
with other government regulations (I-H-1A3, I-H-1A8, I-H-3U, I-H-31).
Another commenter from the District of Columbia said he is forced to
3-21
-------
man his self-service operations (at additional expense) in order to
prevent customer abuse of Staye II equipment (l-H-18).
Response: For some retailers, tne costs associated with Staye II
control will be prohibitive. The financial strain (a reflection of limited
profit potential) could lead to cessation of yasoline operations for some
retailers and station closure for others. However, not all stations in any
yiven size class will fail. While there is potentially a yreater buroen
imposed on small businesses, an examination of financial ratios for
various size classes of yasoline marketiny firms indicates that many
small firms are at least as financially sound as their larger counterparts
(Appendix E in Volume I, KIA discusses station closures in more detail).
Recently, the vehicle manufacturing and yasoline marketiny indus-
tries have been subject to economic fluctuations, and reduced profit-
ability has been evident within each. Moreover, each industry is
already incurriny reyulatory costs. Considering this variability and
the relatively modest economic impacts, there is no reason to delay
further regulation until markets stabilize, especially in view of the
extent of the ozone problem and the potential for hazardous exposure.
There may be some problems with consumer abuse of equipment, but this
would occur primarily during the transition period from familiar equip-
ment to Staye II devices.
Comment: Two commenters believed any action by EPA that resulted
in Stage II controls would result in the cost of recoveriny refueling
emissions beiny borne by the yasoline marketer. The commenters felt
these costs would be imposed on entities that are not the source of the
emissions, since refueling vapors are emitted from the vehicle, not the
retail outlet .(I--H-1U8, I-H-liy).
Response: The Agency considers owners and operators liable for
the control of significant pollutant emissions arising from the opera-
tions carried out at their facilities. These operations occasionally
involve equipment owned by other parties and not a permanent part
of the facility. (A direct parallel can be drawn between the service
station refueling operation and the loading of for-hire tank trucks
at a bulk terminal. Direct responsibility for control liny emissions
rests with the terminal owner or operator.)
With Staye II systems, it is true that service station owners
generally bear the entire burden of purchasing, installing, and maintain-
-------
iny emission control equipment. However, these costs would typically be
recoverable by passing them through to the customer in the form of
small (less than one cent per yallon) gasoline price increases or
increases in other prices at the facility. Thus, control costs would
be ultimately borne by the consumers patroniziny controlled facilities.
Comment: Une commenter stated that tPA was not thorouyh in
assessiny the costs associated with the implementation of a Stage II
reyulatory strateyy. The commenter felt there are other costs besides
the direct costs normally associated with Stage II that were not considered.
The commenter indicated that these costs include: (1) a loss of Federal
tax revenues for each gallon of gasoline not consumed due to increased
gasoline prices (estimated as $47 million/yr, NPV of $463 million), and
(2) a decline in. crude sales as gasoline demand declines with accompanying
losses in tax revenues, jobs, and income. This commenter felt that EPA
should consider the costs associated with an increase in the concentration
of large gasoline outlets as either small- or medium-size firms leave
the market as a result of Stage II (I-H-1U2).
Response: The reduction in government tax revenue cited by the
commenter represents a gain in revenue to the private sector. These
funds are not lost, but are shifted within the economy.
The loss of crude sales to the gasoline marketing sector is
accounted for in the loss of gasoline sales. The value of gasoline
sales lost is comprised of the value of the crude product plus the
value added by each stage of the production and marketing processes.
Moreover, the crude oil no longer, consumed in this market has other uses
and other markets, so that jobs and income will not necessarily be lost. .
Increased market Si.are for large stations does not necessarily
mean increased industry concentration, which depends on the number of
firms in the industry and .their share of the market. The gasoline
marketing industry includes numerous firms of varying sizes, and most
would be able to remain in business, including many in each size class.
No substantial increase in industry concentration is anticipated to
occur as a result of the implementation of Stage II requirements. .
Comment: Une commenter felt that, due to the lower flow rate of
Stage II nozzles, the labor costs to operate full service islands
would increase if Stage II were required (I-ri-12b).
3-23
-------
Response: According to the American Petroleum Institute (I-U-38),
flow rates throuyh current uncontrolled dispensiny nozzles vary widely,
depending on the number of dispensers drawiny simultaneously from the
same submersible pump, problems with nuisance shutoff, whether full
service or self-serve, and so forth. Rates throuyh these nozzles
vary from approximately t>.5 to 12 gpm. In Stage II balance systems in
California, flow rates throuyh second-yeneration systems are b to 8 ypm,
while the newly certified systems allow ID ypm. From these figures,
there appears to be no significant difference between flow rates from
conventional and Stage II nozzles. Thus, there should not be any
noticeable difference in the refueling times at uncontrolled and con-
trolled facilities.
Comment: Many service station owners and operators pointed out
that they operate with low profit margins and stiff competition, and
several stated that the cost of Stage II would ultimately be passed on
to consumers (I-H-1, I-H-1A1, I-H-1A3 to 1-H-lAb, I-H-1A8, I-H-1A9,
I-H-1A13, I-H-1A17, I-H-lAiy to I-H-1A^1, I-H-1A26, I-H-1A^7, I-H-1A29,
I-H-1A3U, I-H-13 to I-H-16, I-H-3U, I-H-36, I-H-64, I-H-66, I-H-rby,
I-H-71, I-H-76, I-H-76, I-H-77, I-H-78, I-H-84, I-H-86, I-H-88, I-H-8y,
I-H-1U2, I-H-1U8). Several commenters indicated that Stage II costs
would not be recoverable in the current market economy, and that they
might be driven out of business. Three commenters indicated that even
if capital were available, a facility would have to pass the cost of
Staye II on to the consumer or fail economically; however, in the
>
current market, intense competition would not permit the passing on of
Staye II costs (I-U-67, I-H-1U2, I-H-1U8, I-H-liy). Five commenters
stated that they- could not pass on Staye II costs without losiny market
share (I-H-IMZ, I-H-1A24, I-H-14, I-H-46, I-H-49); this is especially
true for independent operators (I-H-b4, I-H-by). Another commenter
emphasized that the competitiveness in the gasoline market leaves
little profit for independents. The commenter felt the imposition of
Stage II control, at a time when margins are low and when marketers are
being required to replace underyround tanks because of State reyulations,
could well cause an unprecedented number of bankruptcies for independent
marketers (I-H-64). Uther commenters added that the automobile industry
has been able to pass on most cost increases that were necessary to
enable them to do business (I-H-34, I-H-38). Another of the commenters
-------
said that his company had no after-tax profits for the past 3 years
laryely as a result of expenditures for Staye I installation, expenditures
to adjust meters to display full price, and automobile fleet mileage
standards, which have decreased yasoline consumption (I-H-lAy).
Kesponse: It should be noted that cornmenters (oil companies and
oil trade associations) disayree as to whether they can pass Staye II
costs on to consumers in the form of hiyher prices. In the lony run,
these costs must be recovered from the consumer. If they are not
recovered initially, some facilities will close, and closures will
place upward pressure on gasoline prices and yenerate cost recovery.
Predicted price increases for yasoline reflect the transfer of Staye II
costs to the consumer.
A close examination of these comments indicates that the commenters
are expressiny control costs as current expenses and compariny them
only to internal sources of funds. The Ayency believes it is inappro-
priate to assess costs only as current expenses. Annualization of
costs ana amortization of equipment in terms of NPv must be considered.
Furthermore, it is not sufficient to compare costs only to internally
yenerated funds; external financiny must also be considered. Many
variables affect capital availability and these factors vary from firm
to firm.
tvaluatiny the effect of controls on a particular firm is compli-
cated by the numerous factors involved in determininy investment recovery,
includiny terms of loans, total sales, yasoline throuyhput, existiny debts,
and methods of financing. For this reason, generalizations based on a
firm's ownership status are unreliable. While some facilities would be
likely to close, many others of corresponding volume would remain open.
A number of yasoline marketing firms have experienced reduced
profitability in recent years. While existing environmental and other
control expenditures miyht have contributed to pressure on marketiny
profits, they have not been solely responsible and many enterprises
continue to be profitable despite this pressure.
Comment: Several commenters asserted that neyative economic . .
effects would result from service station closures. The commenters
stated these effects include: (1) a reduction in employment (I-H-1A2,
I-H-1A7, I-H-1AB, I-H-lA^b, I-H-2b, I-H-bl), (2) an end to "Mom and
-------
Pop" operations (I-H39), (3) a reduction in automobile maintenance
service availability (I-H-lAa), (4) a reduction in fuel availability
tnat would nurt local/ State economies (I-H-^y), and (b) niyher consumer
prices resultiny from a loss of competition in the marketplace as small
independents are forced to shut down (I-H-1A3', I-H-1A4, I-H-lAy, I-H-iy,
I-H-40, I-H-44, I-H-b4, I-H-79). Another commenter felt that Staye II
would have a catastrophic effect on the financial ability of Maine
petroleum marketers to remain in business, especially since they are
currently worKiny on a proyram to monitor, replace, and/or install
underyround storaye tanks. The commenter felt closures would reduce
fuel availability, adversely affectiny commerce and industry, particularly
the very important Maine tourist trade (I-H-29).
Response: The EPA ayrees that there can be neyative economic
effects stemming from potential closures of service stations subject to
a Stage II requirement. A much more extensive economic analysis would
be needed to fully characterize these effects; nevertheless, some yeneral
responses can be made to the comments. The net employment effects of
any of the reyulatory strategies are unknown. Production and installation
of control equipment creates jobs, while closure of stations eliminates
jobs. The extent to which these effects offset each other is not estimated
in the analysis.
"Mom ana Pop" establishments are smaller facilities with hiyher
per-unit production and control costs. However, because they are
small, they are the facilities most likely to be exempted from control
requirements under small facility exemptions. Thus, not all such
operations would be expected to close.
Automobile maintenance is not necessarily tied to gasoline
marketiny. If a maintenance establishment is profitable, it is likely
to remain open as a maintenance establishment reyardless of gasoline
sales; if not, the operation miyht close reyardless of gasoline sales.
Fuel availability should not be affected by control strategies.
While there may be an impact on the number of locations of available
fuel, this number will change even in the absence of additional control.
In the lony run, throuyhput will be as hiyh as demand for gasoline at
the market price requires it to be. It is unlikely that availability
in an entire region, State, or city would be affected.
2-26
-------
Competition in the marketplace is expected to be preserved. Tne
emeryence of the convenience store that sells yasoline and the contin-
uiny survival of many small facilities should yuarantee prolonged
competition. An overall decline in the number of facilities is not
sufficient to yenerate market power amony the numerous yasoline retailers
remaininy in operation.
Comment: Two commenters noted that facility operators would be
forced to borrow money to pay for Staye II controls and that banks and
other lendiny institutions are reluctant to loan money for non-income-
produciny equipment (I-H-1U3, I-H-119). Two commenters stated that
capital is yenerally not available for nonproductive investments and,
therefore, it would be difficult for many small firms to obtain loans
to finance Staye II (I-H-lb, I-H-Bl). Une commenter quoted a recent
survey of 3y New Jersey jobbers which disclosed that the implementation
of Staye II would involve an initial expenditure of $8,712,UUU (based
on an API cost survey of oil companies and data supplied by equipment
manufacturers). The comenter stated that many small jobbers do not
internally yenerate sufficient funds to cover this expenditure, nor can
they obtain loans since such an expenditure does not contribute to the
productivity or profitability of the outlet (I-H-4U).
Response: Various marketiny firms will finance Staye II controls
with different combinations of internally and externally yenerated funds.
Uver time, market forces will push up yasoline prices, yeneratiny
internal funds. However, some firms will require some initial external
financiny. Althouyh in certain circumstances availability of adequate
financing may be a constraint, it should not be a major barrier to stable
firms. Control-equipment represents a productive investment in tne
sense that the lack of it would result in closure of the facility, and
so it is income-produciny. Profitable businesses are able to obtain
outside financiny for improvements to their capital since they are able
to repay loans. Lendiny institutions should be capable of perceiviny this.
Comment: Une commenter stated that the Clean Air Act favors an
equitable distribution of the burdens and costs of improving air
quality. The commenter felt that, since a marketer usiny Staye II
would not benefit from recovered vapors, Staye II will not provide an
equitable distribution of the costs; onboard will more successfully
achieve this yoal (I-H-1U2).
3-27
-------
Response: As discussed in a response in Chapter 7, tne gasoline
vapors recovered in a Staye II control system reduce or eliminate
evaporation in underground storaye tanks, and the subsequent loss of
product that occurs through tank venting. The gasoline not lost
through evaporation and venting can certainly be considered a benefit
of Staye II to a marketer.
The Agency agrees that onboard would provide an equitable distri-
bution of control costs, by placing the actual purchase and maintenance
of controls in the hands of consumers. This would be accomplished when
a new vehicle was purchased at a slightly higher price than it would
have been without controls. Costs of Stage II controls would be similarly
distributed, however, if a marketer passed through control costs in the
form of higher gasoline prices. Then, under either strategy, the con-
sumer would in the end absorb the costs. The equitability of cost dis-
tribution to the providers of the emission controls (auto manufacturers
or service station owners) is less clear.
3.7 STAGE II MAINTENANCE
Comment: Several commenters stated that the cost of Staye 'II
maintenance is prohibitive and it is often costly and difficult to
detect defective equipment (I-H-1, I-H-1A4, I-H-1A13, I-H-1A14,
I-H-lAlb, I-H-1A23, I-H-4, I-H-8, I-H-11, I-H-14, I-H-19, I-H-31,
l-H-52, I-H-SJ4). Two other commenters cited experience in California
to support this claim (I-H-69, I-ri-78). Une commenter from the Dis-
trict of Columbia said that the rubber boots on nozzles underyo
excessive wear by customers, the climate, and yasoline, and can "crack,
tear and rip." The commenter indicated that gasoline causes wear to
many internal parts made of rubber or plastic, and some of these plastic
parts are prone to breakage due to customer abuse. This commenter
stated that the "permanent" band on the nozzle comes apart and causes
recirculation of fuel (I-H-18). Another commenter claimed that the
vacuum assist system is more complex than a balance system and is
therefore more often subject to mechanical failure (I-H-84).
Response: The EPA has included estimates of maintenance costs for
systems such as those currently being installed in Washington, U.C. and
California in the revised cost analysis. These estimates were based on
discussions with equipment manufacturers and vendor experience in
existing Stage II areas. The estimates include a weighted averaye of
3-28
-------
appropriate maintenance costs for all the available types of systems.
The vacuum assist system, being more complex with the additional burner
and vacuum pump equipment, may require additional maintenance for these
items. However, these systems are generally easier for the consumer to
use and they are beiny used successfully in California.
The EPA is aware that Staye II equipment does require maintenance
and periodic replacement. The equipment defects referred to by the
commenters were taken into account not only in the estimate of mainte-
nance costs, but also in estimatiny the actual in-use efficiency of
Staye II equipment. In its 1983 report to the State Leyislature (I-F-78),
the California Air Resources Board pointed out that many of these equip-
ment defects were caused by sharp objects around the vicinity of the fill-
pipe, abuse, normal wear, improper system maintenance,.and improper system
installation. The CAKB has been working with equipment manufacturers
based on these findings to develop more durable and reliable components
that will allow a more moderate, acceptable level of maintenance.
Comment: Three commenters believe that facility operators making
a good faith effort to comply with Stage II regulations would be faced
with large and continuous maintenance costs on Stage II equipment that
can easily be rendered ineffective, thereby exposing them to liability
for violating the regulations (I-H-102, I-H-1U8, I-H-liy).
Response: As stated in the previous response, manufacturers claim
that improvements being made in the durability of system components
should reduce the amount of maintenance required by facility operators,
in areas where Staye II may be implemented. In California, minor
maintenance failures or defects found by an inspector are tagged "out
of service.". The station operator is given 7 days to make the necessary .
repairs. Specific major defects (bellows missing, major tears, etc.) or
continued failures with no effort at maintenance may be considered a
violation of the regulations.
As with any equipment where the untrained general public has access
(such as the regular self-service equipment now in routine use around
the country), a regular maintenance program is especially necessary.
Such a maintenance program should provide for the repair or replacement.
of Stage II components before major violations occur. The costs
reflected in the analysis are "average" costs and could vary signifi-
cantly from one location to another.
-------
3.8 ENFORCEMENT OF STAGE II REQUIREMENTS
Comment: Une commenter stated that it is easier to ensure proper
function!ny of Staye II systems in use than it is for onboard systems
because it is easier to inspect 2UU.OUU yasoline stations than 1UU
million motor vehicles (I-H-1UU).
Response: The EPA ayrees that, in terms of numbers, the inspec-
tion of service stations would appear to require less effort than the
inspection of every automobile. However, each vehicle will not
necessarily need inspection. The EPA certification program for new
vehicles yoes a lony way to ensure that vehicles and systems produced
under a certificate function properly for the vehicles' useful lives.
Moreover, there are remedies under the Act (warranty and recall) that are
intended to correct those situations where a defect may exist, even
without inspectiny every vehicle. In addition, there are many pollution-
related inspections required on current new vehicles ana it is antici-
pated that inspection of onboard systems could be incorporated into the
existing vehicle inspection proyram at no additional cost. Similarly,
onboard systems in use would require little or no maintenance,
and the effects of tamperiny could be determined within the existing
vehicle enforcement proyram.
Comment: Several commenters felt that vigorous enforcement would
be needed to maintain the high effectiveness of Staye II systems (I-U-b4,
I-D-bb, I-U-b7, I-U-68, I-H-36, I-H-bU, I-H-68, I-H-76, I-H-lUy, I-H-llb,
I-H-124). Une of them felt that the required enforcement level has
been demonstrated in California and should be readily attainable in
other areas of the country if EPA works with the States and the public
to emphasize the public health issue (I-H-llb). Another commenter
thought that effective enforcement on a nationwide basis would be very
difficult and, without large yovernment involvement, may be impossible
(I-H-122). Several others believed Staye II would require a reyulatory
ayency of tremendous size to police the continual problems of these
systems (I-H-1, I-H-1A4, I-H-1A14, I-H-4, I-H-37, I-H-b4). Une
commenter felt that EPA's in-use effectiveness estimate of 86 percent
for the balance system is reasonable only if strict enforcement is
employed; another stated that at least annual inspections would be
required (I-H-1U9, I-H-118).
3-3U
-------
A number of commenters also referred to the anticipated hiyh costs
of enforciny compliance (I-H-4, I-H-46, I-H-UU, I-H-84, I-H-87). Two
commenters mentioned that the added certification and inspection fees
required for a Stage II system can be costly. Une commenter referred
to the lb)83 API study indicatiny that permit fees could averaye $*8b
for a 3-hose outlet and $1,2U1 for an 18-hose outlet, and the cost to
perform an annual inspection would be about $13U (I-H-84, I-H-87).
Another commenter favored permit fees or penalty fines as a method to
make a Staye II enforcement proyram self-supportiny. The commenter
noted that EPA had advocated permit fees in the past in response to
retirements under Section llu of the Clean Air Act (I-H-21).
Response: Experience in California with its Staye II proyram has
indicated that an in-use efficiency of 86 percent can be maintained
through a relatively viyorous enforcement effort. This level is beiny
considered as the upper end of the in-use efficiency ranye, or best
that can reasonably be expected, in the Ayency's revised cost/benefit
analysis of the reyulatory strategies. Thus, an assumed in-use
efficiency of 86 percent reflects the annual inspection enforcement
scenario. Annual inspections of .service stations would be relatively
costly but not impossible.
Enforcement costs contribute to social costs because they repre-
sent real resource use. A system of permit fees and penalty fines
shifts the burden of payiny for the enforcement proyram from the public.
sector to the private sector. The cost itself is not removed. There-
fore, it is necessary to consider the costs of enforcement in compariny
regulatory strateyies. Enforcement costs were estimated for each reyu-
latory strategy and have been included in the final cost estimates used
to compare the strategies. These costs are described in more detail
in the Volume I RIA.
The fee structure for certifyiny and inspectiny systems has not
been established. Undoubtedly, each jurisdiction would set fees based
on its own budyet needs and the costs of administeriny the proyram.
These added costs would not be expected to be a major portion of the
total costs of purchasing, installing, and maintaining a Staye II
system.
3-31
-------
Comment: Une commenter felt EPA's analysis should have considered
the quarterly inspection scenario because this level of" enforcement
increases the actual Stage II efficiency by 4 percent (I-H-127). Two
other commenters thought that the zero-enforcement option should not be
considered by EPA as a serious option because under this strateyy most
of the total control costs would be incurred, while the amount of
captured VUC's would De drastically reduced (I-H-lUa, I-H-lly).
Response: Likely cost and emission reduction impacts for Staye II
were determined based on actual operatiny systems in California and
Washington, D.C. Uue to the very different levels of enforcement effort
that exist in these two areas, these two proyrams were considered to
represent reasonable upper and lower limits on the enforcement levels
that could be expected in future proyrams. In California, where at
least annual inspections of Stage II installations are conducted, rather
hiyh in-use control efficiencies averaging 8fa percent are observed.
In Washington, D.C., on the other hand, much lower in-use efficiencies
of about 6Z percent are observed. While the program in Washington,
D.C., does not represent a total lack of enforcement, EPA has termed it
"minimal enforcement." Thus, any Stage II program instituted in the
future would be likely to operate at least as efficiently as that in
Washington, U.C., and probably no better than the one in California.
The Agency prefers to see the most efficient control programs possible,
but most enforcement activities depend on the capabilities and interest
of State or local authorities.
Comment: Two commenters thouyht that Stage II would require an
unending series of subsequent rules and regulations to cover methods of
handling consumer complaints, development of operating instructions
and certification procedures, requirements for inspection reports', etc.
The commenters referred to the extensive rulemaking addendums that
exist in California (I-H-1, I-H-84).
Response: The extensive rulemaKiny referred to in California,
where Stage II systems have been developed and refined, indeed repre-
sents a significant quantity of material. This material has evolved
over several years. Areas becoming subject to Stage II requirements in
the future would not likely require such extensive developmental
material. First, the work performed in California (and in Washington,
3-3*
-------
U.C. and St. Louis, Missouri) could be drawn upon in forminy new proyrams.
Second, EPA would provide yuidance that would also make extensive
rulemakiny unnecessary.
Comment: Une commenter asserted that poor consumer acceptance
would lead dealers to be lax in maintaininy Staye II equipment, makiny
enforcement difficult. The commenter also believed that station size
exemptions would cause reyulated dealers to feel unfairly treated and
that the maintenance of equipment, therefore, would not carry any
"moral imperative" (I-H-12U). Another commenter stated that enforce-
ment would be easy to accomplish because the vacuum assist Staye II
system manufactured by his company would pay for the initial investment,
maintenance, and repair of the system throuyh the yasoline recovered
(U.3 percent of throuyhput) (I-H-lzy).
Response: Exemptiny facilities based on the volume of their yaso-
line business is considered desirable because the cost of controls can
represent a larye proportion of a station's profits, especially for
smaller stations. Certain exemption options have been analyzed in
response to requirements in Section 324, which applies to possible
Federal reyulation. The selection of cutoff levels for State-adopted
requirements would rest with the States. The Ayency belie./es that the
principle of reyulatory exemptions is fair and should be applied when
it is dictated by an analysis of the potential impacts.
It is not clear how consumer acceptance would affect the level of
maintenance of Staye II. It seems likely that a consistent level of
maintenance, whether hiyh or low, would be practiced throuyhout an
entire facility.- The costs to enforce Staye II requirements have been
considered in the strateyies analysis. While the Ayency prefers that
an operator maintain his control equipment voluntarily, enforcement
proyrams provide an incentive to operators to keep systems in proper
operatiny condition.
Comment: Une cornmenter felt that EPA's analytical approach did
not fully capture the subtleties of the enforcement costs issue. The
commenter remarked that yovernrnent budyetiny is not an entirely rational
process, in that resources are not always commensurate with an ayency's
responsibilities or public expectations, and that EPA must consider
likely future enforcement resources and possible alternative uses for
3-33
-------
those resources. The commenter further remarked that if other levels
of yovernment are beiny relied upon, EPA must consider whether the
political will exists to allocate resources for the particular purpose
for the time period required (I-H-12U).
Kesponse: The Ayency recoynizes the political factors involved in
the allocation of scarce yovernment resources. Because of the uncertainty
of the response from area to area, EPA's analytical approach does not
attempt to consider such political ramifications. However, these
factors and related issues not covered in the analysis were considered by
the Administrator as the reyulation was developed for proposal.
Comment: Une commenter stated that, unfortunately for the tax-
payers, a siynificant proportion of the inevitably hiyher purnp price of
yasoline attributable to the cost of installiny Staye II vapor recovery
controls will be wasted unless those same taxpayers also pay additional
taxes to finance the cost of a Staye II inspection proyram (I-H-y4).
Kesponse: The Ayency's cost analysis for a Staye II control proyram
included the costs associated with enforcement inspections needed to
maintain a reasonable control effectiveness of the Staye II equipment.
The cost for the Staye II proyram is reflected in the estimated
yasoline price increase, and so no additional taxes or costs beyond
those considered in the analysis are anticipated for this purpose.
Comment: Une commenter noted that, consideriny current State
resource constraints, it is unlikely that their State ayency would be
able to commit the level of enforcement resources necessary to maintain
a hiyhly effective Staye II vapor recovery control proyram (I-H-y2).
Two'commenters, from two control jurisdictions in California,
discussed their experience in the enforcement of local Staye II reyula-
tions. Une of these commenters indicated that the record and experiences
for the South Coast Air Quality Manayement uistrict, California, indicate
that theirs is a very effective proyram. The commenter indicated that
fourteen positions are budyeted, allowiny maintenance inspections at
least twice a year. The commenter stated that each facility that
receives frequent notices of violating equipment is placed on a
special inspection proyram. 'The commenter further stated the public is
encourayed to provide the Uistrict, over toll-free numbers, information
3-34
-------
about maintenance or misfueliny problems they have encountered. The
commenter noted that the District also uses the assistance of the
Department of weiyhts and Measures in cases such as meter leaks or
malfunctions (I-H-6tf).
The other commenter pointed out that, due to siynificant improve-
ments in equipment reliability, San Dieyo does only one installation
inspection instead of the two mentioned in the report (unless a problem
is encountered), enforcement inspections have been reduced from four to
three per year for stations that sell to the public, and private Staye
II facilities have been reduced to one inspection per year. The
commenter noted that the most effective aspect of San Dieyo's proyrarn
has been the initial installation test requirements measuring pressure
decay, vapor flow resistance, and liquid blockaye; these tests were the
primary reason a State survey found that San Dieyo had far fewer custo-
mer and dealer complaints than other districts. The commenter also
pointed out that the hiyh-hany hose loop requirement, which Keeps
liquid from accumulating in tne vapor return hoses in balance systems
and causing premature nozzle shutoff, was a contributing factor to the
low number of complaints in San Dieyo (I-H-yu).
Response: The California experience clearly indicates that budget
problems can be addressed and overcome. However, it is equally clear
that the will to do so might vary from State to State. Such factors
were considered by the Administrator in proposing the regulation.
3.9 SCUPE/CUVERAGE UF STAGE II REQUIREMENTS
Comment: Une commenter suggested that Federal action be taken to
make it'mandatory that controls be placed on all filling station pumps
(I-H-6).
Response: From a nationwide perspective, the Agency would not
require controls on all service stations since it is required to obey
the limitations set forth in Section 324(a) of the Clean Air Act. This
section of the Act states that any EPA regulations applicable to vapor.
recovery from fueling of motor vehicles at retail outlets will not
apply to an outlet owned by an independent small business marketer of
gasoline and having monthly sales of less than 5U,UUU gallons. In
3-3b
-------
addition, the regulatory options that .included exemptions for small
service stations were considered because of the smaller emissions
contribution from these stations and because the economic impacts on
such stations could be excessively severe. However, States actiny
independently on localized situations are free to adopt lower (or no)
exemptions.
Comment: A number of commenters felt that Staye II controls may
represent a viable means of reducing hydrocarbon emissions in order to
attain regional ozone taryet levels in nonattainment areas (I-U-bb, I-H-99,
I-H-1U1, I-H-128). Two of the commenters stated that such controls
should be required only if the Ayency can demonstrate that they are a
cost-effective means of achieving those ozone targets (I-U-bb, I-H-yy).
Response: As discussed in the preamble to the accompanying
proposal, EPA believes that onboard control is the most appropriate
long-term solution to the vehicle refueling problem. Moreover, although
Stage II controls could theoretically realize earlier emission reductions,
trie uncertainty associated with rapid implementation in many areas at
the same time, coupled with the duplication in costs involved and
other factors, weigh against the Agency's requiring Stage II (in addition
to onboard) as an interim measure. However, where a State considers
Stage II to be a reasonable and necessary measure, EPA will support the
State fully in its efforts to ensure that such'controls are implemented.
Comment: Une commenter pointed out that the cost per unit of hydro-
carbon controlled by Stage II systems is markedly different between urban
areas, which represent most of the areas not in ozone attainment, and
rural areas, which generally are in attainment. The commenter
stated that rural service stations have low throughputs and, conse-
quently, would have high recovery costs per unit of hydrocarbon
controlled if Stage II is required (I-H-yy).
Response; The Agency agrees that many rural stations would face
high costs if they were covered by Stage II requirements. However, as
discussed in Section 3.6, many of the smaller service stations may be
exempted due to lower business volumes, if Stage II were adopted.
Moreover, stations in rural areas outside of designated nonattainment
areas would probably not De covered under potential SIP requirements.
3-36
-------
Comment: One commenter remarked that Stage II nas the advantage
of flexibility of application, and thus could be applied on a limited
or provisional basis without makiny a premature commitment to a
particular control technology. The commenter provided an example
where Staye II could be implemented only in selected areas where the
need was greatest, or only at certain service stations or service
station operations where the exposure was likely to be the highest
(enhancing cost effectiveness). The commenter concluded that with
Stage II, the lead-time necessary to make modifications would be
greatly reduced (as compared to onboard controls)(I-H-93).
Response: Given the nature of the ozone problem, and the poten-
tial for hazardous exposure, EPA believes there is a clear need for
control of vehicle refueling. Because refueling technologies are well
understood, it is doubtful there would be need for repeated modification
of the regulatory requirements whether onboard or Stage II is required.
However, due to the smaller number of entities controlled under Stage
II (as compared to the number of motor vehicles controlled under require-
ments for onboard controls), and the .fact that it is easier to communicate
with operators of, and effect changes on, stationary businesses than
owners of motor vehicles, in concept it could be easier to modify Stage
II systems in-use or apply them in a limited manner. This issue is
being considered in the decisionmaking process.
3.1U CONSUMER REACTION TO STAGE II
Comment: One commenter remarked that he had used Stage II systems
in California and considered them superior to onboard canisters in that
they not only prevent loss of vapors, but also prevent overfilling gas
tanks and the resultant spillage. The commenter thought the very act
of using Stage II systems would create a safety awareness in people
that a carbon canister would not (I-H-3).
Response: Both Stage II and onboard type systems are designed to
prevent or reduce overfilling and spillage when used properly. The
Administrator has considered the advantages and disadvantages of the
available control approaches in developing the proposal.
Comment: Two commenters thought that EPA should reject any
suggestion that California's experience with Stage II is readily adapt-
able to other States or local jurisdictions, because California's Air
3-37
-------
Quality Control Regions yenerally are valleys or basins surrounded Dy
mountains. Tne commenters felt motorists residing in AQCR's currently
covered by Stage II have little choice as to whether they purchase
gasoline from a facility with Stage II equipment (I-H-lub, I-H-119).
These commenters (and one other) stated that the migration of customers
to retail gasoline outlets without Stage II would be counterproductive
to environmental objectives and would lead to a loss in tax revenues
for the controlled jurisdiction (estimated as high as $47 million)
(I-H-1U2, I-H-1U8, I-H-liy).
Response: There would almost certainly be some migration of custo-
mers seeking to avoid using self-service dispensing facilities equipped
with Stage II. However, such migration would occur only under a partial
coverage scenario, and not under a scenario in which Stage II was
mandated nationwide. Further, although California may have additional
geographical restraints to migration, migration would occur primarily
in fringe or boundary areas where unregulated stations were conveniently
accessible to motorists.
Althouyh improved Stage II equipment and increasing customer
familiarity and acceptance would tend to reduce this problem, it is
a factor being considered in the decision-making process.
Comment: Several commenters described present Stage II equipment
as inconvenient, cumbersome, and unpopular (I-H-1, I-H-1A3, I-H-1A4,
I-H-1A8, I-H-1A12, I-H-1A27, I-H-4, I-H-63, I-H-67, I-H-137). Une com-
menter explained that the Stage II nozzles fill slowly, click off
continuously, and spill gasoline (I-H-bj. Several other commenters
described the nozzles as heavy, bulky, and at best unwieldy, leading to
bad consumer response (especially in the District of Columbia) (I-H-11,
I-H-13, I-H-lb, I-H-19, I-H-33, I-H-3b, I-H-37, I-H-b9, I-H-61, I-H-71,
I-H-76). One commenter, especially noting L).C. experience, said that
consumers find Stage II equipment awkward and unwieldy, and nave a
problem with spills on their shoes and clothing (I-H-12U). Two com-
menters felt that the complexity, size, and weight of the Stage II
nozzles could end self-service altogether (I-H-47, I-H-78). Some
commenters said that the nozzles are difficult to use, unpopular witn
customers, and often abused (I-H-4U, I-H-50, I-H-69). Two commenters
noted that Stage II equipment handling problems are particularly acute
»
3-38
-------
for older and handicapped Americans, who miyht have to foreyo the lower
cost of self-service refueling if Staye II is mandated (I-U-b4,
I-U-b7, I-U-68, I-H-12U).
Seven commenters quoted a survey completed in U.C. following
implementation of Staye II that showed 68 percent of the- respondents
bought gasoline elsewhere to avoid cumDersome Stage II equipment, 6u
percent felt that further research should have been done before imple-
menting Stage II, and only 4 percent supported Stage II control (I-H-78,
I-H-84, I-h-86, I-H-1U2, I-H-1U8, I-H-119, I-H-13U). Eight commenters
in the U.C. area noted an actual shift or anticipated a shift in consumer
buying patterns to uncontrolled areas (I-H-4, I-H-lb, I-H-18, I-H-3b,
I-H-37, I-H-4U, I-H-76, I-H-89). Une commenter stated that in both
California and the District of Columbia the evidence to date indicates
that consumers do not want to operate Stage II equipment and, given the
choice, will purchase gasoline at facilities with conventional nozzles;
i.e., at exempt facilities. The commenter cited a consumer survey at
20 service stations in Washington, U.C., indicating widespread dissatis-
faction (I-H-12b). Une commenter felt that air quality would not be
improved if a system is installed that is so difficult for the average
consumer to use that he or she devises methods to sabotage or defeat
the proper operation of the equipment, or if unhappy consumers find
they can drive to areas without Stage II to avoid the inconvenience
of buying gasoline through Stage II equipment (I-H-IUZ).
Kesponse: Many of the problems described by these commenters can
be attributed to older, first-generation Stage II equipment at service
stations. Recognizing the importance of reducing consumer complaints,
manufacturers have taken action to improve system components. They
state they are developing more durable, reliable, and easily used
components for Stage II vapor recovery systems. Among the changes
described are improvements to nozzle bellows and faceplates, vapor hoses,
swivels, and nozzle latching bands and springs. In addition, nozzle
manufacturers have introduced a new tear-resistant urethane-type
bellows material,that will be used on new nozzles and as replacements.
High-retractor twin and coaxial hose and high-hang coaxial nose
configurations suspend hoses off the ground. Suspended hoses are less
likely to be run over and flattened. New, lighter and I'ess cumbersome
3-39
-------
nc^zles are beiny introduced, which should make refueling easier for
more people.
The Agency believes that such recent improvements may reduce
consumer attempts to sabotage, or avoid the use of, Stage II dispensing
equipment. Nevertheless, Stage II is still more complicated and cumbersome
than standard equipment, and some customer inconvenience can be expected
where it is used.
Comment: Four commenters consider the 1982 California complaint
rate (2,b33 in 11 months) on Stage II to be very high, particularly
when complaints were only accepted by CAKb in writing (I-H-1U2, I-H-1U8,
I-H-119, I-H-130). Two of these commenters cited consumer apathy and
customer "avoidance" as reasons why the level of dissatisfaction in
California and Washington, D.C. is potentially much higher than indi-
cated by the numbers of complaints. The commenters felt that some
motorists have learned to bypass the proper operation of the system,
making use of the equipment easier, and will therefore no longer Have
reason to complain (I-H-1U8, I-H-119).
Kesponse: The California AKB, in its March 1983 report to-the
State Legislature (I-F-78), noted that there are approximately one
billion fuelings per year made with Stage II equipment in California.
Therefore, the complaint rate amounted to about one complaint in 3bU,UUU
refueling operations. As discussed previously, improvements in equipment
should reduce many of the difficulties experienced previously by users.
Comment: Une commenter noted that, while consumers should be
willing to pay a yrice in personal inconvenience for the sake of clean
air, the Agency's imposition of burdens on the public (as with btage II)
is a resource "that should be expended carefully. The commenters
suggested that onboard would be less likely to draw down tPA's stock of
public support (I-H-12U).
Response: The EPA has attempted to consider all pertinent issues
and aspects of the regulatory options under evaluation. Some intangible
elements, such as burdening the public with part of the responsibility
for maintaining air quality, are difficult to quantify so that they can
be weighed against other costs. The Agency agrees that the burden -to
the public at-large of using Stage II equipment could be greater than
that of purchasing an onboard system, and this has been considered in
developing the rulemaking being proposed at this time.
3-4U
-------
3.11 REFERENCES (Comment letters are not repeated here. See Chapter 1,
Table 1-1, for a complete list of comment letters.)
I-A-22 Cost and Cost-Effective Study of Onboard and Staye II Vapor
Recovery Systems. Prepared by R.A. Luken for U.S. EPA, Office
of Air Quality Planning and Standards, Research Trianyle ParK,
NC. August 1978.
I-U-38 Letter and enclosures from Crockett, P.E., American Petroleum
Institute, to Garay, C.L., U.S. EPA. August 8, 1984.
Information on gasoline dispensing rates.
I-E-19 Telecon. Eldridye, K., Pacific Environmental Services, with
Stray, I., b.F. Goodrich. October 16, 1984. Cost of hose
assembly.
I-E-2U Telecon. Levine, A., W.M. Wilson's Sons, with Eldridye, K.,
Pacific Environmental Services. October 18, 1984. Costs of
nose confiyurations and dispenser.
I-E-22 Telecon. Eldridge, K., Pacific Environmental Services, with
Simon, J., Petro Vending. October 2b, 1984. Costs and life
expectancy of components.
I-E-23 Telecon. Purcell, R., Pacific Environmental Services, with
Taylor, B., Hint Combustion Enyineers. October 26, 1984.
Costs of Stage II components.
I-E-25 Telecon. Purcell, R., Pacific Environmental Services, with
Healy, J., Cambridge Enyineeriny. October 29, 1984. Costs of
VR system.
I-E-26 Telecon. Eldridye, K., Pacific Environmental Services, with
Madden, M., Pomeco. October 29, 1984. Costs of equipment.
I-E-27 Telecon. Tayyart, L)., SMP Company, with Eldridye, K., Pacific
Environmental Services. October 29, 1984. Costs of equipment.
I-E-37 Telecon. Eldridye, K., Pacific Environmental Services, with
Van cleave, R., Ameron. January 7, I98b. Service life of
, fiberglass pipe.
I-E-38 Telecon. Oswald, K., A.O. Smith, Incorporated, with Eldridye, K.,
Pacific Environmental Services. January 9, I98b. Service
life of fiberglass pipe.
I-E-47 Telecon. Norton, B., Pacific Environmental Services, Inc.,
with Simeroth, U., California Air Resources Board. January 30,
198b. Stage II system in-use efficiencies.
I-E-53 Telecon-. LaFlam, G., Pacific Environmental Services, with Tpdd,
D., California Air Resources Board. September 24, 198b.
Resolution of spit-back problem in early Stage II systems.
3-41
-------
I-E-b4 Telecon. Sirneroth, u., California Air Resources Board, with
LaFlam, G., Pacific Environmental Services. September Zb,
198b. Fire safety of Staye II systems.
I-F-78 Air Resources Board, State of California. A Report to the
Legislature on Gasoline Vapor Recovery Systems for Vehicle
Fueliny at Service Stations (Staye II systems). March 1983.
I-F-98 Cost Comparison for Staye II and Un-Board Control of Refueliny
Emissions. American Petroleum Institute. Washinyton, U.C.
January 1984.
3-42
-------
4.0 TRADEOFFS BETWEEN STAGE II AND ONBOARD*
4.1 GENERAL ISSUES
Comment: One commenter expressed concern over the capability of
both Stage II and onboard systems to control hydrocarbon emissions
from alternative fuels such as methanol and ethanol blends. They noted
that the evaporative emissions of such fuels may contain oxygenated
hydrocarbons in addition to the light hydrocarbons. Some of these
oxygenated hydrocarbons may affect control efficiencies; for example,
adsorption and desorption efficiencies of methanol on activated carbon
may not be identical, which may cause saturation of the activated carbon
for methanol and a resulting drop in control efficiency. The commenter
concluded that while it is not clear if EPA has studied the effects of
alternative fuels on Stage II and onboard controls, it is a topic that
should be addressed in light of the anticipated increased use of such
fuels (I-H-126).
Response: It is not anticipated that methanol or ethanol blends
would affect the performance of Stage II equipment. The balance and
hybrid system equipment is only a capture system and so any hydrocarbon
vapors, regardless of composition, should be captured equally and piped
to the storage tank. The alcohol fuel vapors also should not reduce
the efficiency of the incinerators associated with vacuum assisted
systems.
As discussed in Section 2.1.4(b), the Agency has conducted two
separate test programs to evaluate the effects of alcohols on onboard
canister working.capacity. The findings were that alcohols have little
or no effect in reducing the performance of the activated carbon used
in onboard systems.
Comment: One commenter stated that if control of VOC emissions
from motor vehicle fueling is determined to be necessary, then enhanced
carbon canisters on new vehicles is the preferred approach. The
commenter pointed out that the evaluation report indicates onboard to
be more effective, and cited these advantages: (1) the control effi-
ciency of onboard is higher;-(2) onboard would also control excess
*1984 Federal Register topic.
4-1
-------
evaporative emissions; (3) an onboard program would extend to all areas;
(4) the existence of exempted facilities would not reduce the effectiveness
of an onboard program; (5) the effectiveness of onboard is.not dependent
upon extensive enforcement actions; and (6) onboard is not susceptible to
circumvention resulting from low consumer acceptance (I-H-92).
Another commenter ayreed that onboard controls will capture not
only refueling emissions, but also evaporative emissions that escape
existing evaporative emission controls, and that, as the Agency pointed
out, when evaporative emissions are included in the overall analysis,
the emission reduction potential for onboard controls is "roughly
double" that of Stage II (I-H-94).
Response: All of the issues cited by the first commenter, control
efficiencies, evaporative emission controls, program coverage, exemp-
tions, enforcement, and public acceptance, have been thoroughly considered
by the Agency in evaluating control options for the gasoline marketing
industry. In reanalyzing the impacts of controls, both enlarged carbon
canisters and gasoline RVP limitations in combination with Stage II
controls have now been considered (see Vol. I KIA).
Comment: One commenter noted that, if a consumer ruined a fill-
pipe seal on a car equipped with onboard controls, only that one car
would be affected. However, a disabled Stage II nozzle could refuel
hundreds of cars within days (I-H-1U2). Another commenter felt that
Stage II systems are relatively fragile compared to onboard, and esti-
mated that the typical expected emission loss for a single car with a
ruined onboard seal would be about 1.6UU grams per year, contrasted
with 1,60U grams-per day for a defective Stage II nozzle (I-H-20).
Response; The Agency agrees that a disabled Stage II nozzle, if
used to refuel a large number of vehicles before maintenance or enforce-
ment action, could potentially allow relatively large quantities of
emissions to escape. Under either control approach, proper design,
installation, maintenance, and inspection are important to ensuring that
the systems continue to control emissions at their optimum efficiencies.
Comment: One commenter stated that Stage II controls would save'
almost 100 million gallons of gasoline per year that would be lost if
onboard controls are required. This advantage, the commenter pointed
4-2
-------
out, could be important beyond the current dollar value of that gaso-
line in a future petroleum scarcity and should not be ignored
(I-H-93).
Response: The fuel savings resulting from the implementation
of Stage II systems (with exemptions) was shown in Table 5-12 of the
July 1984 analysis to be 8.1 billion liters of gasoline over the
35-year period of the analysis, or about 61 million gallons per
year. For the no-exernption scenario, this figure increased to
86 million gallons per year. Zero fuel savings were indicated for
onboard systems, because the canister weight added to the vehicle was
presumed to offset any enhanced fuel economy.
These figures have been revised in the new analysis to reflect
updated emission factors and a more detailed examination of onboard
system designs. The Draft Volume I RIA annual fuel savings of 23U to
33U million liters of gasoline per year or about 60 to 9U million
gallons per year for Stage II applied nationwide. Onboard systems are
now estimated to produce a savings of 61U million liters or about 16U
million gallons of gasoline per year. The Volume I RIA also shows that
Stage II-nationwide plus evap controls can have an associated fuel
savings as high as 650 million liters per year (17U million gallons per
year). Section 2.7 discussed in detail the fuel consumption aspects of
vehicles with onboard systems.
The recovery of valuable motor fuel otherwise lost is considered
an important factor in the analysis of control strategies, and has
been considered by the Administrator in developing the rulemaking
proposal.
4.2 FACILITY EXEMPTIONS FROM STAGE II
Comment: Three commenters felt that the effectiveness of Stage II
would be reduced (in contrast to the total coverage of onboard controls)
because of the exemptions planned for smaller service stations, which
outnumber the large facilities (I-D-54, I-H-1A14, I-H-40). Une of the
commenters also pointed out that governmental facilities, car rental
agencies, and other commercial installations would, through exemption,
continue to contribute to the emissions problem (I-H-40).
Two commenters agreed that size exemptions should be considered
(I-H-42, I-H-43).
4-3
-------
Response: Upon full implementation, Stage II would be somewhat
reduced in effectiveness because of uncontrolled facilities. The
magnitude of the reduction would depend upon the exemption levels
selected and the intensity of enforcement programs. It is true that,
depending on the exemption level, a large number of smaller facilities,
including the types mentioned by the commenter, could remain uncontrolled.
Comment: A number of commenters remarked on the appropriateness
of exemption size cutoffs for service stations. One commenter noted
that it would be difficult to determine the gallonage (throughput) of
service stations (I-H-1A14). Another commenter agreed that the true
sales volume could be difficult to determine and added that seasonal
fluctuations can change a dealer from exempt to regulated. In addition,
they remarked that proposing a cutoff invites a battle over fairness
and equity (I-H-21). Another commenter felt the Agency's suggested
exemption was too large, noting that the Bay Area Air Quality Management
District of California exempts tanks smaller than 26U gal from Stage I
and stations with throughputs less than 180,DUO gal/yr from Stage II.
The experience in the Bay Area is' that proposing exemptions will give
rise to claims of giving an unfair competitive advantage (l-H-82). One
commenter objected to the unequal treatment of independents and non-
independents regarding cutoffs (I-H-24).
One commenter believed that the provisions of Section 325 apply
to Section 112 standards; however, they did not think that EPA was
prohibited from establishing exemption levels greater than 50,DUO
gal Ions/month. The commenter felt that a size standard of 85,000
gallons/month would be adequate to protect the same type of small
marketer Congress had in mind when it amended the Clean Air Act in'1977
(I-H-102).
Response: In the 1984 analysis, EPA examined the impacts of
several regulatory strategies, including Stage II at service stations.
Specific options included the exemption of independent stations with
gasoline throughputs less than 50,000 gallons per month, and all other
stations with less than 10,000 gallons per month, and a "no-exemption" '
4-4
-------
option. As stated in the 1984 evaluation report (paye 1-9), these cut-
offs were based on the relatively higher costs of control for small
facilities, existing size cutoffs under State and local regulations,
and statutory requirements for small throughput independent service
stations under Section 324 of the Act. Exemption levels were again
examined in the new analysis to investigate the impact on nationwide
regulatory strategies. Although other levels were examined, the assumed.
exemption level cutoff of 10,000 gal/month for all stations and 50,000
gal/month for independents was retained. Other responses in this section
and in Section 3.6 discuss the fairness of exemptions and the effects
on competition. Section 10.1 discusses the legality of the 50,000 gal/
month cutoff for independent gasoline marketers.
Similar exemptions are used in other emission categories to avoid
imposing costly controls on low-emitting entities. The Agency
believes that such an approach is fair, and comports with its objective
of reducing emissions while considering impacts.
Comment: One commenter suggested that EPA exempt al1 stations
with a throughput under 50,000 gallons per month in order to avoid
competitive discrimination. The commenter pointed out that, under this
approach, independent stations retain the Section 324 protection agai.'St
infeasible capital expenditures, but not at the expense of their non-
independent counterparts. The same commenter stated that EPA's example
exemption option would put nonindependent service stations with through-
puts between 10,000 and 50,000 gallons per month at a competitive dis-
advantage with respect to independent stations of like size. The
commenter pointed out that any prudent business entity, large or small,
will analyze the economic impact of a proposed capital expenditure c.t
any one of its operating units on a "stand-alone" basis; that is,
significant capital expenditures at a service station will be evaluated
in light of the revenue produced by that particular station. The
commenter added that if the size of the proposed expenditure is dispro-
portionately large relative to the revenue generated by the station,
the "payback period" (the time required for the revenue generated by
virtue of the expenditure to'recoup the expenditure) will be unacceptably
long, and the expenditure will not be made. The commenter pointed out
4-5
-------
that, assuming independent and nonindependent service station owners
are equally prudent, they would decline to make such a capital expendi-
ture at roughly the same point. In the case of a capital expenditure
for mandatory Stage II controls, declining to make the expenditure
would mean that the station could not legally operate and would have to
close (I-H-94).
Response: Under the original 1984 exemption scenario, noninde-
pendent facilities in the ID,QUO to 50.0UU gallon per month throughput
size class would be at a competitive disadvantage. The reanalysis
examines two additional exemption options that posit uniform size
cutoffs for independent and nonindependent facilities.
Under "stand-alone" economics, the same closure decision would be
made regardless of station ownership. However, Section 324 does not
address this process. Rather, it is motivated by a concern about
access to financing and cost impacts for independent firms. Further
discussion on the statutory basis for facility exemptions is contained
in Section 10.1.
Comment: Une commenter granted that under a Federal program, EHA
would be required by statute to grant certain exemptions, but noted
that States, acting under the aegis of SIP revisions, could ah>ply
Stage II without exemptions. This commenter feels, therefore, that
ultimately Stage II programs would likely cover more facilities than
the current "with exemption" estimates indicate. Thus, both costs
and emission reductions appear to be underestimated in the forecast .
(I-H-12U).
Response; The commenter is correct in that States may set specific
exemption cutoff levels for service stations within their jurisdictions,
provided that these levels are stricter, i.e., equal to or lower, than any
levels specified under Agency requirements (see Section 116 of the Clean
Air Act). The reanalysis of regulatory strategies, discussed in the
Draft Volume I RIA, includes the evaluation of additional exemption
considerations.
4.3 IN-USE CONTROL EFFICIENCIES
Comment: Two commenters felt that EPA's assumptions concerning
the in-use efficiencies of the two control approaches skewed the
4-6
-------
evaluation in favor of onboard (I-H-57, I-H-82). One of them felt that
tampering with onboard controls would be considerable, causing EPA's
worst-case onboard control efficiency of 92 percent to be extremely
optimistic. Also, the worst-case figure for Stage II of b6 percent
(minimal enforcement) is much too low, considering the commenter's
experience with Stage I controls (I-H-57). These commenters felt, that
using correct assumptions about in-use efficiencies would show that
Stage II is the superior control approach. Another commenter agreed
with EPA's estimates of 92 percent for onboard and 66 to 86 percent for
Stage II (I-H-94).
Response: Due to the latest design concept under consideration
for the onboard system (J-tube liquid seal in fuel tank instead of
nozzle/filIneck vapor seal) and the phase-out of leaded gasoline,
tampering with the nozzle restrictor/fi 1 Ineck should occur only rarely.
The Agency now estimates an in-use efficiency of 93 percent for onboard
controls. The "worst-case" estimate for Stage II in-use efficiency
is now 62 percent, based on extensive observations made at controlled
service stations in Washington, D.C. (I-A-61).
Comment: Several commenters thought the in-use control efficiency
of onboard systems would be much higher than the efficiency of Stage
II, and should be selected as the control approach that would achieve
the greatest emission reductions (I-H-94, I-H-1U2, I-H-1U8, I-H-119,
I-H-122, I-H-125).
Response: As discussed in the previous response and in the sec-
tions on onboard controls, Agency estimates of in-use control effi-
ciencies have been re-evaluated. For Stage II, the range is now
estimated at 62 to 86 percent, based on the latest surveys made in
areas currently employing Stage II controls. For onboard, the re-
analysis is based on an in-use control efficiency of 93 percent, due
to negligible effects from tampering and system deterioration (see
Section 2.1.2.8).
4-7
-------
4.4 REFERENCES (Comment letters are not repea :ed here. See Chapter 1,
Table 1-1, for a complete lis*. of comment letters.)
I-A-61 D.C. Gasoline Station Inspections to Assure Compliance with
Stage II VOC Vapor Recovery Requirements. U.S. EPA Region
III. Philadelphia, PA. 9271.UU/12b-N. January Iy8b.
4-8
-------
5.0 EPA's 1984 CONTROL STRATEGY EVALUATION
5.1 GENERAL METHODOLOGY
Comment: One commenter remarked that EPA had attempted to find a
single solution to the two problems of ozone in nonattainment areas
and gasoline vapors with their associated cancer incidences. Thus,
EPA compared tried-and-true Stage II systems, effective in reducing
very real ozone nonattainment problems, with experimental onboard
systems said to be effective in reducing speculative cancer incidences.
In the process, EPA stretches its predictive capabilities to an
"amazing" 35 years (I-H-21).
Another commenter felt that it is misleading to compare on common
terms Stage II data from full scale public use programs with extra-
polated onboard data from controlled test programs (I-H-118).
A number of other commenters thought that a period of 35 years is
too long to make confident projections in the analysis of control options.
One commenter felt that the cost estimates for onboard systems had been
made and projected 35 years into the future without knowing the size,
construction, or location of the unit. This commenter maintained that
extrapolating beyond the year 2006 serves no purpose, since no one can
predict whether automobiles will even be powered by gasoline engines at
that time (I-H-101). Another commenter did not see the 35-year
analysis as reasonable, given the uncertainties about the motor vehicle
fuel of the future. The commenter felt there was no justification for
a time horizon beyond 25 years (I-H-127). Finally, one commenter felt
that, while -the -35-year analysis period is relevant, it dramatically
distorts the cost comparisons, and EPA need not make a decision in 1984
for the entire period through 2020. The commenter stated that the 35-
year period inflates the apparent cost of a "Stage II only" program in
which EPA mistakenly assumes that Stage II equipment must be replaced
in 2002 and 2017, which more than triples the apparent capital cost of
Stage II (I-H-93).
Response: The commenters in general appear to be questioning EPA'-s
selection of 35 years (1985-2020) as the period for estimating impacts
from implementing onboard and Stage II control strategies. The selection
of this projection period is based on confidence that there will be
5-1
-------
a complete turnover of the vehicle fleet within 35 years and that
onboard is anticipated to be fully implemented by then (most likely
well before 2010). The analysis was carried out to 2020 to allow
amortization of the latest cycle of Stage II equipment. Thus, the two
control approaches can be compared after they both are at their full
coverage. Two analysis approaches were evaluated in the comparison of
the regulatory strategies: (1) a net present value (NPV) analysis that.
evaluated all future costs and analyzed the future cost stream in 1987
dollars, and (2) an analysis for the year 2010 that compared the costs
of all strategies after full implementation was reached. These analyses
do not distort the cost comparisons but instead provide a reasonable
solution to evaluating regulatory strategies that have significantly
different implementation rates.
The £PA agrees with one of the commenters that there are some
uncertainties about motor vehicle fuel in the future and the effect
this might have on gasoline marketing controls. In spite of these
uncertainties, however, EPA must use the best data available in estimating
impacts of these controls.
The EPA has looked not only at single strategies to control the
complex emissions associated with gasoline marketing, but also at com-
binations of strategies. For example, a regulatory strategy was eval-
uated that combined onboard (which leads to the greatest long-term VOC
and incidence reduction) with Stage II in nonattainment areas (which
could achieve faster control of ozone in the areas where it is most
needed.)
Section.2.l_contains a response to the concern that onboard
technology has been evaluated only during controlled test programs.
Section 3.4 discusses the changes made in the analysis to better
estimate the replacement cycle and replacement costs associated with
Stage II.
Comment: Two commenters felt that EPA's analysis should not have
included the benefits of reduced hot soak evaporative emissions in the
determination of the cost effectiveness of onboard controls (I-D-58.,
I-H-95). One of them said the Agency should analyze other alternatives
for its control (such as enlarging the existing carbon canister or
limiting the RVP of gasoline in summer months), and should evaluate the
5-2
-------
cost effectiveness of measures that address refueling and hot soak
evaporative emissions as separate issues constituting the complete
problem. Three potential measures would be: (1) onboard controls,
(2) Stage II vapor recovery combined with enlarged carbon canisters
required on new cars, and (3) Stage II vapor recovery combined with a
limitation on the RVP of gasoline sold in the summer months. If addi-
tional control alternatives for evaporative emissions are not evaluated
along with onboard controls, evaporative emission reduction credits
should not be included in the evaluation of regulatory strategies
(I-H-95).
Another commenter stated that the inclusion of excess evaporative
controls distorts the onboard analysis. The commenter feels this
should be excluded because: 1) evaporative emissions do not have the
same health implications as refueling vapors because they are not
emitted in the breathing zone, 2) onboard controls will not assure
the elimination of excess emissions that exist due to fuel volatility,
3) the fuel volatility difference between certification fuel and commer-
cial gasoline must be eliminated or reduced, and 4) the analysis did
not analyze other approaches to the control of fuel volatility (I-H-114).
Another commenter further contended that the consideration of
excess evaporative emissions and their control should not be a part of
the analysis, noting that EPA had recently announced its intent to
eliminate the current volatility difference between commercial and
certification test fuels (I-H-127).
Response: The EPA's revised analysis considered the three alter-
native measures posed by the first commenters. The first alternative,
onboard controls without controlling excess evaporative emissions, is
not reasonable because the control of excess evaporative emissions is
an integral part of the current onboard system concept. Therefore, it
is not technologically feasible to implement the onboard program without
controlling excess evaporative emissions. However, EPA did evaluate
the impacts of an onboard strategy incremental to evaporative controls
(thus assessing 'onboard on the basis of refueling only). . .
The EPA has evaluated the second and third alternatives, enlarged
carbon canister controls in conjunction with Stage II vapor recovery
b-3
-------
systems and the combination of Stage II systems with RVP controls. The
EPA's Office of Mobile Sources (QMS) is studying the elimination of the
current RVP difference between certification fuel and commercial gasolines,
and placing limitations on gasoline RVP during the summer months, which
includes options to increase and change the evaporative canister system
(I-A-66). These alternatives have been considered in the revised
analysis.
In response to the second commenter's first point, that excess
evaporative emissions should not be included in the onboard analysis
because of differences in health implications, EPA in its reanalysis
examined the effects of using an onboard system to control refueling
vapors in conjunction with excess evaporative vapors. (Both of these
sources of emissions would be controlled by an onboard system.) Further
discussion on approaches to control of excess evaporative emissions is
contained in Sections 2.1.7 and 2.6.4.
Comment: One commenter remarked that it was difficult to verify
EPA's estimated bulk terminal and bulk plant emission reductions and
corresponding recovery credits. The commenter suggested that EPA should
clarify the assumptions used and provide sample calculations (at the
bottom of tht tables in a way similar to Table 7-lU of the 1984 report).
The commenter cited some specific examples of recovery credit calculations
requiring clarification: (1) The gasoline throughput rate and units
for Q are not provided in Table 7-5 to verify calculations based on
bulk terminal internal floating-roof losses of (7.3285 x 10'3 Q) +
2.4 Mg/yr. They noted that on pages 2-12 and B-22, the floatiny-roof
tank withdrawaT -losses are given as 0.46 x 10'7 Q where Q is in barrels/yr.
(2) The bulk plant emission reduction changes with size of plant and
type of control as expected; however, the recovery credit does not
change with type of control. They considered it illogical that with an
exemption, the smallest plant would have a higher recovery credit.
They noted that the recovery credits in Tables 7-6 and 7-7 could not be
verified (I-H-126).
Response: The units for Q in the internal floating-roof emis-
sion factor equation are in -barrels of product per year. A weighted Q
was used to calculate the recovery credits. This weighted Q was based
5-4
-------
on tf.-: bulk terminal model plant throughputs, number of tanks assumed
for each.model plant, and the distribution of each model plant. The
weighted annual throughput used for the internal floating-roof recovery
credit calculations was approximately 2,700,000 barrels per year.
An error was discovered in the recovery credit calculations for
bulk plants under the no-exemption option. The original 1984 analysis
included only the emission reductions attributed to the storage tank
draining losses and not the emission reductions associated with the
tank truck loading operations. This error nas been corrected in the
revised analysis.
Comment; One commenter felt that Stage II capital costs should be
calculated using actual data from California and that onboard should be
credited with MOBILES credits for evaporative emissions capture. If
these changes in the analysis are made, onboard then becomes three
times more cost effective than Stage II (I-H-120).
Response: In EPA's reanalysis, Stage II costs were revised based
on California-certified systems. These revised costs are presented in
Appendix B of the Volume I RIA. In the revised analysis, EPA displays
results for onboard both with and without excess credit for evaporative
emissions capture.
Comment: One commenter claimed that the three types of Stage II
systems described in paragraph 3.7.1 of the evaluation report do not
include the system manufactured by his company. The commenter noted
differences between his system and a system described in the report,
and stated that these differences form the basis for his system's
superior performance (I-H-129).
Response: The EPA's July 1984 analysis discussed three types of
Stage II systems being, used currently in California and the District
of Columbia: the vapor balance, the hybrid, and the vacuum assist sys-
tem. The commenter's system is apparently a vacuum assist type system
with some differences in detail from the general description EPA pro-
vided. It was not feasible to describe fully in the report the various
specific systems.currently available. Moreover, it is highly unlikely
that the Agency would base its regulation on a particular system of a
given technology for control of refueling.
Comment: One commenter felt that EPA did not make a balanced
use of information from the oil and auto industries in its analysis,
5-5
-------
since the list of references in the EPA report includes many references
from the oil industry (I-H-118).
Another commenter felt that EPA's almost sole reliance on a
single API report on the effectiveness of onboard controls is question-
able given the oil industry's economic interest in promoting onboard
controls. The commenter recommended that EPA use an independent means
of verifying any data from any industry with an economic interest
associated with the result (I-H-90).
Response: Throughout the development and analysis of regulatory
strategies for controlling emissions from gasoline marketing, EPA has
requested, received, and considered input from all interested parties.
On June 27, 1978, the Agency solicited (43 FR 27892) information on the
costs and effectiveness of possible controls on motor vehicle refueling
(I-G-5). Respondents to the Federal Register notice included API, General
Motors, Ford, and AMC. With regard to Stage II, the recent practical
experience and recommendations of authorities in California and the
District of Columbia have been closely studied. Moreover, in 1984 EPA
solicited comments on its draft evaluation of gasoline marketing strategies.
As discussed in Appendix C of the EPA July 1984 analysis report, the API,
GM, and Ford information contained data from tests with onboard control
hardware. All of these data have been reviewed and analyzed by EPA.
Since the publication of the July 1984 analysis, EPA's Office of
Mobile Sources (OMS) has developed and tested a liquid seal configura-
tion for onboard control systems that addresses many of the concerns
commenters had with regard to the mechanical seal (nozzle/fillpipe
interface) (Section 2.1.2 describes this design concept in more detail).
The 1978 studies of API and its contractors, the technical input and
comments of the auto makers, as well as comments solicited on the
material in the July 1984 analysis (I-G-15), and subsequent experimental
work and analysis by OMS have been carefully considered by EPA in
arriving at the regulatory proposal.
Comment: One commenter felt that the calculations using equipment
useful lives are appropriate only for equipment placed in service prior
to 1980. Recovery of capital costs for tangible depreciable property
placed in service after 1980 is accomplished through a method called
Accelerated Cost Recovery System (ACRS). Under this method, the cost
5-6
-------
of an asset is recovered over a predetermined period that is generally
shorter than its useful or income-producing life, while salvage value
is disregarded. Under ACRS, an automobile or light-duty truck is
considered to have a life of 3 years, while bulk terminal, bulk plant,
and petroleum storage equipment is considered to have a life of b
years. This commenter suggested that since most businesses have to use
the ACKS method for capital cost recovery, EPA should use a similar
method that more reasonably reflects the tax consequences of control
equipment. The commenter added that the true cost of the control systems
depends on the type of financing, debt/equity ratio, depreciation
schedule, tax effects, and other factors (I-H-126).
Response: The EPA's cost analysis is a before-income tax
analysis and, consequently, it does not address the tax implications of
control equipment. The before-tax approach is chosen for several
reasons. First, the main purpose of the cost analysis is to estimate
the social or real resource costs of regulatory strategies. From an
aggregate or societal viewpoint, a tax on net earnings is best viewed
as a transfer of resources from the private to the public sector,
rather than as a real resource cost. Of course, individual firms would
consider the tax consequences of control equipment, and the market
would reflect this consideration. However, computation of likely tax
impacts is confounded by the wide variety of firms, firm conditions,
and firm accounting practices in the industry, and by possible competi-
tive pressures created by the regulatory strategies. Furthermore, in
other analyses it has been observed that before- and after-tax cost
impacts ,do not.differ markedly. Depending on tax rates, credits,
depreciation, other taxable income, etc., the after-tax costs can be
higher or lower than the before-tax costs, so no prior belief about
the direction of the difference between the two methodologies can be
held. In this case, EPA used before-tax calculations of costs for
the preliminary economic analysis (excepting the local tax discussed
above). Questions concerning the appropriate depreciation schedule in
this context are'moot. .
In order to better respond to the comment, however, EPA has
computed before-tax and after-tax costs per unit of throughput by
model plant for Stage II control. For all computations, EPA assumed
5-7
-------
that the cost of capital was 10 percent. For the after-tax case, a
10 percent investment tax credit was claimed, the 95 percent basis for
depreciation of environmental control equipment was chosen, and accele-
rated cost recovery was applied to capital equipment. The results
displayed in Table 5-1 show that the per-liter average total cost of
control for each model plant is slightly less before taxes than after
taxes. Differentials between before-tax and after-tax estimates
range from less than 0.03<£/liter to 0.076^/liter. The percentage
differences between before-tax and after-tax estimates are within
the error bounds of the cost calculations and the limits of accuracy of
the impacts methodology. Larger percentage differences are observed
only when the before-tax cost of control is close to zero.
Whether income tax payments represent real resource costs or
whether they represent, all or in part, a transfer of resources from
the private to the public sector, depends on what assumptions the
analyst is willing to make about the value of the services provided in
exchange for these tax payments and the relative tax burden associated
with these investments relative to the average investment made in the
economy. In this analysis, EPA chose to treat taxes on net income as a
transfer, and thus not a real resource cost from an economic perspective.
Comment: One commenter noted that the net present value (NPV)
rate of time preference approach used in the evaluation normally applies
to income or other positive values. For health risks associated with
emissions, the time preference would be negative; given a choice between
emissions now or later, later will usually be chosen. The commenter
felt that this, creates differences from the traditional incremental
steps involved in a yearly discount rate, and raises questions about the
methodology (I-H-24).
Response: While the time preference with regard to emissions (and
associated health risk) is negative, the time- preference with respect
to reductions in emissions is positive. Thus, the application of NPV
analysis to emission reductions is consistent with the traditional time
preference approach. Moreover, the NPV's included in the analysis.are
calculated in the traditional manner: emission reductions are deter-
mined incrementally, discounted, and summed to arrive at the present
value of emission reductions.
5-8
-------
Table 5-1. STAGE II AVERAGE TOTAL COST OF CONTROL BY
MODEL PLANT (1984 {/liter)
Tax basis3 MP1 MP2 MP3 MP4 MP5
Before tax
After tax
1.756
1.956
0.526
0.602
0.552
0.627
0.367
0.421
0.105
0.133
aAll calculations assume a 10 percent cost of capital. After-tax esti-
mates also assume a Federal tax rate of 46 percent, a State tax rate of
6 percent, a 10 percent investment tax credit, the 95 percent basis
for depreciation of fixed capital after the tax credit, and accelerated
capital recovery based on 5 years on an accelerated depreciation
schedule.
5-9
-------
Comment: Two commenters believed EPA's cost analysis was faulty
and a significant departure from previous Agency analyses in that it
used a discount rate to discount emissions, rather than discounting
the monetized benefits of emissions reduction (I-H-102, I-H-120).
One commenter (I-H-1U2) argued that, just as society has a preference
about the timing of emission reductions, it must also have preferences
concerning various levels of emission reductions. In order to determine
the best strategy, the analysis should calculate the costs and benefits
of reducing VOC each year under the various strategies, express them in
monetary terms, and then discount them. However, the evaluation document
did not furnish the necessary information to do this. Since this
information is not available, one can instead discount emissions as EPA
has done, and select different discount rates based on the VOC reduction
potential of the various strategies. This would mean that onboard
would be assigned a higher rate of discount than would any of the various
Stage II options. Using EPA's sensitivity analysis, it would be appropriate
to assign onboard the rate of 1U percent and Stage II the rate of
5 percent. Under the use of this discounting method, Stage II costs
double, while onboard costs remain the same (I-H-102).
The other commenter (I-H-120) stated that a more traditional
approach would assign a monetary value and take the net present value
of the monetized stream of benefits. Calculating the number of years
required to achieve a certain level of cumulative VOC reductions is
another alternative. This commenter felt that further consideration
should be given to the most appropriate method of comparing streams of
costs with streams of benefits.
Response: It is not necessary to discount the monetary equivalent
of emission reductions if all units of reduction are assigned the same
value, as they are in this analysis. The monetary value is needed as
a common basis for evaluation only if, for example, the value of the
last unit of emissions reduction is different from the first. Placing
the same value on each unit of reduced emissions, regardless of the
current emission level, is consistent with a linear dose-response
function constructed independently of baseline emissions. Since the
risk analysis is based on the linear dose-response function, the
discounting of emissions in this manner is appropriate.
5-10
-------
Discounting is used to remove the time element from investments
with diverse time streams so that they can be compared. The discounting
procedure captures the effect of time without the use of different
discount rates for various projects. Adopting different discount rates
to reflect diverse time streams would defeat the purpose of discounting,
which is intended to account for variations in timing by applying the
same calculation methodology to all alternatives.
Comment: One commenter suggested that a more accurate cost
comparison can be achieved using the ratio of vehicles that would be
equipped with onboard controls to each service station that would be
required to have Stage II. The commenter felt that costs in both cases
should include initial installation, maintenance and repairs, total
enforcement costs, and level of effectiveness (I-H-90).
Response: Using ratios of vehicles and service stations compli-
cates the analysis and introduces uncertainty, due to varying distribu-
tions of vehicles and service stations, ratios of controlled to
uncontrolled sources, exempted categories of sources, and other factors.
The Agency believes that the best nationwide total cost estimates can
be made by first determining the most accurate component and system
costs, and then multiplying these values by the total number of emitting
sources likely to have the controls installed. All of the cost cate-
gories mentioned by the commenter were considered in the cost analysis.
Comment: One commenter differed with the assumption that onboard
is an all-or-nothing choice, i.e., that it would be impossible to
have differently equipped cars going to different geographic areas. '
They felt this truncated the analysis artificially since respected
academicians have suggested the appropriateness of a "two-car" strategy
and California already receives different cars from the rest of the country.
The commenter suggested it might be possible to impose special require-
ments on cars shipped for sale in ozone nonattainment areas (I-H-120).
Response: In the Agency's reanalysis, both a 49-State and a 5U-State
analysis was reviewed. California was excluded from coverage by vehicles
equipped with onboard controls under the 49-State scenario. Stage II is
prevalent throughout the more populated areas of the State, and vehicles
destined for California have different emission requirements, so the pre-
sumption of different refueling controls is logical and its implementation
-------
would be feasible. Un the other hand, the advisability of extendiny
this type of strategy to other areas is hiyhly doubtful. The EPA
considers it nearly impossible to taryet vehicles sufficiently over
such a large number of individual areas, and the travel of most vehicles
outside of their relatively limited "designated" areas would cause the
effectiveness of such an approach to be lost quickly.
Comment: One commenter felt that the approximate $1UU million
cost per cancer incidence reduction calculated by EPA (Table 1-1? of
the 19b4 EPA analysis report) for the nationwide regulatory strategies
was applicable only to ozone attainment areas. In EPA's analysis,
various values for the concurrent VUC reductions for ozone control are
subtracted until, at $2,000/Mg for VOC control, the benzene cancer
incidence reduction cost is zero. The commenter thought this method-
ology was flawed because VOC reduction has significant value only in
ozone nonattainment areas (I-H-127).
Response: The EPA agrees that VOC reduction can be considered to
have its greatest "value" in ozone nonattainment areas. Nonetheless,
VOC reduction in other areas also has considerable value. It is diffi-
cult to place relative quantitative values on environmental improvements
because many diverse, often conflicting value systems and priorities
can be applied in making such evaluations. The EPA has performed a
new analysis (presented in Section 3.4 of the Draft Volume I RIA) that
assumes various benefit values for VOC reductions. In the revised
analysis, most of the regulatory strategies show a net cost benefit.at
a value of $1,000 per megagram of VOC (nationwide average).
5.2 EMISSION ESTIMATES
Comment: Several commenters questioned the validity of the results
from EPA's original test program to estimate in-use evaporative
emissions. Two of the commenters felt that the figure of 0.13 gram per
mile was probably incorrect, and may be twice as high (I-H-109, I-H-114).
Another said that the reliability of the preliminary emissions data was
questionable, due primarily to the nonstandard test procedures used
(I-H-125). Finally, one commenter thought that the Agency should not
use its original test results for any regulatory action until repeat
tests have been completed, and gave the following reasons: (1) contra-
dictions between the results from tests conducted by EPA and the
5-12
-------
results of California's in-use surveillance program, and (2) the
serious flaws pointed out to EPA by API and MVMA in EPA's tests on the
failure rate of recent model passenger cars with respect to evaporative
emission standards (I-H-99).
Response: As was stated in Appendix C to the 1984 Gasoline Market-
ing Study, EPA's preliminary assessment that a problem exists with
excess evaporative emissions in-use was based on a vehicle evaluation
and testing program which was in progress at that time. The problem of
excess evaporative emissions and increased in-use fuel RVP has subse-
quently been characterized much more thoroughly and confirmed, and has
become the subject of a separate EPA study. Much more information on
those issues can be found in that study (I-A-66). Nevertheless, the
question of the validity of EPA's test results is not an issue for
analyses regarding the control of refueling emissions, since refueling
emissions control is being evaluated incremental to any measure to
address problems with excess evaporative emissions.
b.3 ENERGY IMPACT ANALYSIS
Comment: One commenter felt EPA's discussion of energy impacts
was too brief and did not allow verification of the calculations, since
assumptions and methodologies were not described. This commenter
believed that the following information is required in order for the
analysis to be evaluated: (1) assumed yearly throughput and uncon-
trolled emissions for each of the industry segments, and (2} assumed
energy requirements for operation of control equipment and the effect
on energy sayings of the various control strategies (I-H-126).
Response: A simplified energy impacts analysis wa? included in the
1984 strategies analysis to determine the amount of vapors recovered as
product under each of the regulatory strategies. It is true that a
more detailed energy analysis could be performed; however, most of the
control systems evaluated, with the exception of vapor processors at
bulk terminals and incinerators in some (about b percent) of the
Stage II systems; are passive or displacement type systems and use or. .
consume no energy. Therefore, in most cases, the recovery credits are
equivalent to the energy savings. A revised energy analysis, using the
same methodology as used previously, is presented in Section 2.4.2 of
the Draft RIA Volume I.
-------
5.4 GASOLINE CONSUMPTION PROJECTIONS*
Comment: One commenter stated that EPA's assumption of declining
yasoline consumption was inconsistent with its own recent projections
contained in the lead phasedown proposal of August 2, 1984 (I-H-127).
Another commenter summarized gasoline consumption projections from
various sources, noting that they varied from 80 to 97 billion gallons
in 1990 and that the DOE forecasts for 1990 have been increasing over
the past few years. Apparent trends in consumer preferences toward
larger, less efficient caj~s have resulted in higher demand efficiencies.
Therefore, given the importance of gasoline consumption projections to
the overall assessment, the commenter suggested that the analysis
should reflect an understanding of the variability and uncertainty of
recent projections, even through the midterm (I-H-126).
Response: Projections of future gasoline consumption are subject
to as much speculation and uncertainty as forecasts of future gasoline
prices and, in fact, these two parameters are closely related. The
difficulty in making these estimations lies in their dependence on in-
fluences that are traditionally hard to predict, such as the preferences
and tastes of consumers, the world economic situation, and so forth.
Estimates from all available sources are regulaily updated to reflect
these and many other variables. The projections made in "Regulation of
Fuels and Fuel Additives; Lead Phase Down," 49 FR 31032 (August 2,
1984), were not yet completed when the final drafts of the strategies
evaluation report were being prepared. Furthermore, the projection
methods used in the two forecasts relied on different types of informa-
tion, i.-e., world supply estimates versus consumer usage habits, to
calculate the results. Even though the results are somewhat different,
both sets of figures are considered by the Agency to be valid for the
purposes for which they were intended.
In summary, EPA is aware of the uncertainty inherent in all pro-
jections of this type. To the extent that the EPA analysis underesti-
mates fuel consumption, then the analysis also underestimates risk,
emissions capture, incidence reduction, and recovery credits, and .
overestimates the several cos't parameters. For this reanalysis, up-
dated projections of future gasoline consumption have been prepared.
*1984 Federal Register topic.
5-14
-------
b.b SIZES AND DISTRIBUTION UF FACILITIES*
Comment: Two commenters felt that EPA has iynored trends toward a
smaller number of service station facilities and laryer throughputs for
the remaining facilities, and that correcting for this would reduce
Stage II cost estimates by 30 percent and increase emission reductions
by 30 percent (I-H-114, I-H-127).
Response: A new set of facility projections has been developed
that incorporates a change in size distribution based on a shift toward
larger service stations (see Appendix D of the Draft Volume I RIA). The
throughputs of all model plant sizes were assumed to decrease uniformly
in proportion to the decrease in nationwide consumption. A percentage
of the throughput decrease of smaller model plant stations was projected
for closure in each year and a number of new, laryer stations were
added. The number of new stations added was estimated such that the
throughput of the larger new stations equalled the throughput of the
small stations that closed. In this manner the number of stations,
when multiplied by their respective model plant throughputs, would
approximate the total gasoline consumption projections.
The facility projection methodology developed by EPA does not
produce major effects on the emission reduction calculations, since the
emission reductions are based on total gasoline consumption and not on
the number of facilities. (Minor effects are noticed because the shift
from small to large facilities affects the number of facilities and the
amount of throughput exempted each year due to size cutoffs). The faci-
lity cutoffs had a major effect on reducing the cost estimates for the
Stage II strategies. Fewer numbers of stations results in lower costs.
Of course, new data on unit costs produced upward pressure on overall
Stage II costs. The net effect is shown in the Draft Volume I RIA.
Comment: The same commenters stated that an average throughput of
2,000 gallons/month better represents exempt stations less than 10,000
gallons/month (i.e., represented by Model Plant 1) than the EPA assumed
average value of ,5,000 gal Ions/month. The commenters suggested that
using the alternate value reduces the throughput exempted from controls
by 7 percent (I-H-114, I-H-127).
*1984 Federal Register topic.
b-15
-------
Response: The representative throughput selected for Model
Plant 1 was reconsidered. The 5,UOU yal/mo average throughput for all
Model Plant 1 facilities was changed to an average throughput of
2,QUO gal/mo for "private" facilities and an average of 6,UUU gal/mo
for "public" facilities.
Comment: One commenter thought that EPA's assumed total number of
service stations appeared high, pointing out that the lU3rd edition
of the Statistical Abstract of the United States (December 1982) reports '
this figure for 1981: 153,500 ret&il gasoline stations, deriving at
least 50 percent of their gross revenues from the sale of gasoline. No
estimates for 1982 are available, but they are expected to be even lower.
In addition, the commenter stated that there were 17,300 franchised and
an unknown number of nonfranchised convenience stores in the U.S. in 1982.
Assuming all franchised convenience stores dispense gasoline, the total
number of "public" service stations would be 170,800, nearly 20 percent
lower than the 210,875 assumed by EPA. The commenter went on to say
that while EPA's estimates for some of the "private" outlets (e.g.,
trucking and local service, taxis, and school buses) appear reasonable,
the government and miscellaneous categories seem high since they correspond
to averages of 1,709 and 1,900 facilities, respectively, per State (I-H-126).
A second commenter pointed out that, while EPA had estimated that
the number of stations would not increase appreciably, the number has
steadily decreased in recent years (I-H-127).
Response: Several new publications were researched for current
data on the number of service stations nationwide. As indicated by
this commenter, the 103rd edition of the Statistical Abstract reports
153,500 retail gasoline stations 'or 1981. Also, this report states
that there were 17,300 franchised convenience stores, which results in
170,800 (153,500 + 17,300) total "public" service stations.
The December 1984 issue of National Petroleum News (NPN) reports
the preliminary and incomplete findings of the U.S. Census Bureau's
1982 count of service stations as 116,154 (for service stations that
derive at least 50 percent of their gross revenues from sale of gaso-
line) (I-F-126, I-F-127). The-1984 NPN Factbook estimated the total
number of all service stations in 1982 to be 144,690 (I-F-124, I-F-125)..
As reported in the January 1985 issue of NPN, the Department of Commerce,
5-16
-------
Bureau of Industrial Economics, estimated the number of service stations
that derived at least 50 percent of their gross revenues from the sale
of gasoline as 144,691) in 1982, 136,570 in 1983, and 132,000 in 1984.
In addition, their estimates of franchised convenience stores were
14,683 in 1983 and 15,331 in 1984 (I-F-128). The Lundberg estimate for
mid-1984 of 190,000 "public" retail outlets, including convenience
stores, is also discussed in the January 1985 issue of NPN.
In light of these recent estimates and the continuing decline in
the number of service stations, the Agency's original estimate of
210,875 "public" service stations was revised to reflect the Lundberg
mid-1984 estimate of 190,000 public stations. The estimated number
of "private" outlets was not changed, since no new information was
available for this category.
Comment: One commenter felt that EPA's assumptions concerning the
distribution of bulk terminals led to a serious underestimation of the
costs to implement Stage I vapor controls. The commenter (an oil company)
stated that 67 percent of its terminals subject to potential Stage I
controls use top loading, and would require conversion to bottom loading
in order to be compatible with the Stage I system (EPA had assumed only
10 percent would need t.'.is conversion). Also, the commenter indicated
that all of its terminals subject to controls use submerged loading,
whereas EPA had assumed that 10 percent use splash loading. This assumption
inflates the apparent benefits of Stage I controls by leading to a larger
gasoline recovery credit due to the controls. The commenter recommended
that EPA re-examine its bulk terminal model plants and reassess its
calculation of control costs at terminals (I-H-125).
Response: The assumptions relating to the distribution of bulk
terminals used in the Stage I cost analysis were based on information
obtained in developing the 1980 proposal of the recently promulgated
(August 1983) New Source Performance Standards for bulk gasoline
terminals (I-A-34). In re-evaluating these assumptions, the Agency
determined that approximately 60 percent of the facilities in attain-
ment areas (currently uncontrolled), are practicing top loading of
tank trucks. This assumption is based on inputs from throughout the
industry, including many smaller oil companies. Whereas it was
assumed in the original analysis that 10 percent of all uncontrolled
5-17
-------
facilities use splash loading, it is now estimated that 10 percent of
the top loading facilities (6 percent of all uncontrolled facilities)
practice splash loading. These new assumptions have reduced recovery
credits slightly and increased estimates of the net cost for incorpor-
ating vapor recovery controls at bulk terminals.
5-18
-------
5.6 REI- IRENCES (Comment letters are not repeated here. See Chapter 1,
Table 1-1, for a complete list of comment letters).
I-A-34 Bulk Gasoline Terminals - Background Information for Proposed
Standards. U.S. EPA. Research Trianyle Park, NC. EPA-450/3-
8Q-U38a. December 1980.
I-A-66 Study of Volatility and Hydrocarbon Emissions from Motor
Vehicles. U.S. EPA, Office of Mobile Sources. EPA-AA-SDSB-
85-5. November
I-F-124 Service Station Shakeout Continues. 1984 National Petroleum
News (NPN) Factbook Annual Issue, 76_(6a):lU3.
I-F-125 Service Station Population by States - 1984 vs. 1977. 1984
NPN Factbook Annual Issue, 7_6(6a):104.
I-F-126 Latest Station Count 21.4% Below 1977;Fuel Oil Dealer Population
Falls 17.5%. National Petroleum News, 7_6(12) :34. December 1984.
I-F-127 Census Bureau Data - Service Stations/1982-1977. National
Petroieum News, .76(12) :35. December 1984.
I-F-128 Slide in Station Count Ebbing; Survivors Start to Get Fitter.
National Petroleum News, _77_(1) :28-29. January 1985.
I-G-5 On-Board Hydrocarbon Technology. Study of the Feasibility and
Desirability. U.S. EPA. 43 FR 27892. June 27, 1978.
I-G-15 Regulatory Strategies for the Gasoline Marketing Industry.
Notice of availability of a regulatory strategies analysis
document for public comment. U.S. EPA. 49 FR 317U6. August
8, 1984.
5-19
-------
6.U REASONABLENESS UF CONTROL CUSTS VERSUS HEALTH RISK REDUCTION
6.1 NEED FUR STANDARDS
Comment: One commenter felt that refueliny emissions represent
only a small part of VOC emissions and their recovery by either
Staye II or onboard controls will make only a minor contribution to
ozone attainment goals, and at considerable cost per unit of VOC con-
trolled (I-H-102).
Several commenters pointed out that previous measures (i.e.,
Staye I, consumption decrease, Stage II in California and Washington,
D.C.) had minimal or no measured effect on ambient ozone levels (I-H-1,
I-H-1A11, I-H-1A27). One commenter stated that Staye II controls have
"not been proven to improve the quality of our air" (I-H-1A11), while
another thought that implementation of Stage II recovery would not
result in a significant improvement in air quality (I-H-1A2U).
Response: Both Stage II and onboard systems have the capability
of reducing the release of hydrocarbon vapors to the atmosphere. In-use
efficiencies for Stage II and onboard have been determined to be 62 to
86 percent and 93 percent, respectively. The volatile organic compounds
(VUC's) in these vapors are known to participate in atmospheric chemical
reactions that produce ozone. While refueling emissions are not a major
proportion of the VOC emissions in all areas, they constitute a signifi-
cant fraction of the uncontrolled emissions in most urban ozone non-
attainment areas..
California's Air Resources Board has noted that, in 1981, Stage II
systems'in California reduced 48,000 tons of hydrocarbon emissions
(130 tons per day), and prevented the waste of Ib million gallons of
gasoline (based on 10 billion gallons dispensed) (I-F-78). The EPA
believes that proportional- reductions associated with refueling control
will serve to improve the air quality in the many areas with current
ozone problems.
Comment: Une commenter stated that society has limited resources
to devote to its problems. This commenter felt that, when billions of
dollars are to be devoted to a problem that even EPA admits will show
no measurable improvement for the effort, there are dozens of other
environmental health problems that would respond more favorably (I-H-lul).
6-1
-------
Two other commenters felt that the available data on the health effects
of yasoline emissions and/or ozone do not justify the expenditures
required under the control strategies (I-H-24, I-H-116). Another com-
menter felt that unless the facts pointed to significant human health
effects, then the cost of $15 per car on 10 million cars (per year)
could be better spent in other health areas (I-H-28).
Une commenter remarked that gasoline fumes are an important contri-
butor to ozone problems and must be curbed, especially in areas where
Stage I and Stage II controls are not yet in place. This commenter
expressed concurrence with the view that gasoline vapors are carcino-
genic, and added that recent evidence now links benzene with additional
forms of human cancer besides leukemia. The commenter felt there could
be no valid basis for postponing regulatory action in light of this
evidence as well as the potential ozone reductions. Finally, the
commenter cited the following measures as being necessary parts of any
regulation: installation of Stage I controls by 1987 on all sources not
yet having them, installation of onboard on new vehicles beginning
model year 1987, and installation of Stage II in all ozone nonattain-
ment areas by 1987 (I-H-115).
Another commenter was supportive of controls on the gasoline
marketing industry, since these controls (Stage I and Stage II) have
proven to be very cost-effective and have saved millions of gallons
of fuel annually in his State (I-H-118).
One commenter felt that a nationwide program is warranted if
either the cancer risk is shown to be significant or if such a program
is shown to be a cost-effective way to help areas attain the ozone
standard (I-H-98).
Response: The comments illustrate the variety of opinion surround-
ing this controversial issue. EPA believes that the control costs are
warranted. The basis for this decision is thoroughly discussed in the
preamble to the proposed rulemaking.
Comment: Une commenter felt that implementation of both the on-
board and Stage II nationwide control approaches is not warranted, .and
EPA should act to discourage'Stage II at the State level (I-H-113).
Two other commenters expressed strong opposition to the option of
onboard nationwide plus Stage II in ozone nonattainment areas. The
6-2
-------
backlash from haviny two different systems with the same purpose being
implemented at the same time would be considerable (I-D-70, I-H-124).
Response: Although the simultaneous implementation of both
onboard and Stage II would involve certain inherent difficulties, the
Agency does not consider them insurmountable. These difficulties were
considered in developing the regulatory proposal. Although EPA is n.ot
requiring the implementation of Stage II as an interim measure in
nonattainment areas where it is not now in place or in the process of
being installed, it may be effective in particular situations. The
EPA will support State implementation where the State determines that
the strategy is appropriate and desirable.
Comment: Two commenters saw no advantage in adding new service
station controls in areas where there is no danger of exceeding the
national ambient ozone standard (I-H-42, I-H-43). One commenter sup-
ported the need for control in certain communities, but did not favor
Stage II on a statewide or nationwide basis (I-H-1A2U).
Another commenter stated that EPA should delay the decision on
this "expensive and controversial" control measure until a conclusion
is reached on the health effects issue. The commenter stated that he
understood the need for controls in nonattainment areas, but did not
see why nationwide control was needed (I-H-1U4).
Response: The rationale for the proposal of onboard controls is
thoroughly discussed in the preamble to the proposed rulemaking. In
part, onboard was selected for proposal because of greater long-term
VOC (and therefore ozone) reduction benefits.
Comment: Three commenters claimed that the levels of benzene in
gasoline vapors are below those posing a human health risk, and thus
standards are not needed. They pointed out that EPA has decided not to
regulate certain other sources of benzene because the Agency had deter-
mined the maximum lifetime risks posed by those sources (about 1 x
10"4) to be insignificant. Since EPA's analysis predicts a lifetime
risk from high exposure to be 1.1 x 10"s at uncontrolled self-service
stations and 1.2 x 10~4 at the boundary of a complex of uncontrolled
bulk terminals, a similar finding of no Federal action necessary should
6-3
-------
De made in this case (I-H-91, I-H-94, I-H-1UU). Another commenter
stated that the estimate of lifetime risk from benzene of 2.4 x 1U'6
"baseline" for service stations is well below the level defined by EPA's
own Carcinoyen Assessment Group as "negligible" (I-0-b8, I-H-117).
Response: The Agency is proposing to regulate vehicle refueling
emissions for a number of reasons, including the benefits of reducing
VUC emissions in nonattainment (and attainment) areas, and the benefits
from reducing exposure to known or probable carcinogens in gasoline
vapors (including benzene, a known human carcinogen, and the mixture of
gasoline vapors, which is considered a probable human carcinogen). The
decision to proceed with a proposal is not based primarily on the
reduction of hazardous emissions but rather an overall assessment of
total benefits.
Comment: Several commenters thought that the human health risk
posed by EDB, EUC, and benzene in gasoline vapors is unproven or too
small to warrant regulation. Others stated that refueling vapors in
general do not present a significant risk that would justify controls.
These commenters generally felt that the existing scientific data, as
revealed in EPA's analysis, as well as the literature on epidemiological
studies, are insufficient to justify any program to control these emis-
sions (I-D-55, I-D-63, I-H-11, I-H-28, I-H-.91, I-H-94, I-H-99, I-H-lUU,
I-H-1U2, I-H-1U8, I-H-1U9, I-H-113, I-H-114, I-H-117, I-H-119, I-H-12U,
I-H-127). Two of them claimed that EPA's own estimates demonstrate the
risks to be too small to warrant Federal regulation (I-H-1U2, I-H-113).
One commenter said that epidemiological studies of workers indicate no
major health problems (I-H-11). Another expressed the view that, if
EPA determines refueling controls to be a cost-effective way to achieve
ozone goals, the rationale for this conclusion should be distributed in
a separate document for public comment (I-H-99). Four commenters felt
that additional data on health risk should be collected, and that the
Health Effects Institute should examine these data for soundness (I-H-1U1,
I-H-114, I-H-116, I-H-127). Another thought that the nationwide control
strategies should be set aside until a human health hazard has been
6-4
-------
determined, subjected to peer review, and accepted by EPA's Science
Advisory Board (I-H-1U7). A final comrnenter suggested that EPA expedite
its cancer research in the area of gasoline vapor and resolve the risk
analysis, which is now only preliminary (I-H-98).
A number of commenters also concurred that controls to address EDB
and EDC are unwarranted in light of the small present exposure levels
and the certainty that even these small exposures will rapidly decrease
as leaded gasoline is phased out of the market (I-H-83, I-H-91, I-H-94,
I-H-99, I-H-1U1). Une of these commenters stated that since EUB and
EDC are not listed as hazardous air pollutants under Section 112, the
basis for imposing nationwide vapor recovery to reduce exposure to
these compounds is even weaker than for benzene. The commenter stated
that assuming the Agency's new lead phasedown strategy is successful,
annual sales of leaded gasoline should fall more rapidly than EPA's
original analysis assumed, further reducing EDB and EDC emissions from
the marketing of leaded gasoline. The commenter further pointed out
that the concentrations of these compounds are so minute as to render
detection in leaded gasoline vapors virtually impossible, and the
nonoccupational exposures are intermittent and brief (I-H-94).
Response: It is true that the small risk associated with exposure
to EDB and EDC will decline as leaded gas is phased out of the market
place. In the revised analysis, EDB and EDC risk calculations have
been dropped. However, as was pointed out in the evaluation report
(I-A-55) and Federal Register notice (I-G-15), the Agency believes that
there is sufficient evidence derived from human epidemiological studies
to support the existence of a causal association between exposure to
benzene and the onset of cancer. On June 8, 1977, Denzene was listed
as a hazardous air pollutant under Section 112. The listing was based
on the evidence referred to above, and EPA's finding that ambient
exposures to benzene may constitute a cancer risk and should be reduced.
Uther evidence also implicates gasoline vapors as a possible serious
health hazard for humans. The unit risk factor developed for gasoline
vapors is based on laboratory studies with mice and rats.
Given the pervasive ozone nonattainment problem and the potential
hazardous exposure of millions of gasoline consumers and those living
6-5
-------
in the vicinity of service stations, the Agency believes that regula-
tion of refueling emissions is warranted. The rationale for the regula-
tory proposal is contained in the Federal Register notice accompanying
this document.
Comment: Une commenter felt that new EPA requirements would repre-
sent regulatory duplication because USHA already controls exposure of
employees at gas stations (I-H-3).
Response: It is EPA's understanding that USHA has not taken, and
has no immediate plans to take, steps to limit the exposure of service
station employees specifically to the vapors generated during refueling
operations. Controls on refueling would have multiple benefits,
including health risk reduction for employees, self-service (and
other) customers, and the community, as well as reductions in ambient
ozone concentrations. The extent to which these controls provided
beneficial effects for service station employees would not duplicate
current benefits conferred by USHA's occupational standards.
Comment: A commenter felt there is no need to regulate small,
rural stations since they do not contribute significantly to air pollu-
tion, and they provide a tremendous service to rural communities (I-H-1A6).
Response: It is true that smaller stations located in rural areas
contribute a relatively small percentage of overall VUC emissions compared
to large, urban stations, and hence are smaller contributors to the
ozone problem. In addition, these facilities are generally situated in
attainment areas, where control of ozone precursers may not be needed
specifically to attain the ambient standard.
Since it is'likely that facilities below some particular through-
put cutoff would not have to install refueling vapor .controls, many of
the smaller facilities referred to by the commenter would not be regu-
lated even under possible State initiations. However, uncontrolled
facilities in rural areas would be contributing to health impacts,
especially during self-service refueling, and so could be controlled
just as the stations in urban areas. Individual States may decide to
impose a more stringent distribution of Stage II regulatory coverage
than that permitted to EPA by the Clean Air Act, based on the environ-
mental , health, and economic impacts of the controls.
6-6
-------
Comment: Two Ohio-based petroleum marketers felt that the hydrocarbon
emissions from natural sources (i.e., trees, leaves, and shrubbery) are
equal to, if not yreater than, hydrocarbons emitted by dispensing
petroleum products (I-H-1, I-H-1A20). One commenter argued that the
current state of knowledge about the cause and effect of ozone levels .
is insufficient to justify control measures (I-H-1).
Response: The EPA has specific statutory obligations to limit
certain sources of air pollution that affect human health or welfare.
The fact that plants also produce hydrocarbons that may play a part in
pollution in no way changes EPA's responsibility or authority.
While all of the ozone formation mechanisms are not fully understood
with certainty, prevailing scientific evidence and opinion has established
the link between VOC emissions (such as those from petroleum sources) and
the formation of ozone in the atmosphere. Since emissions from the
segments of the gasoline marketing industry form a significant portion
of total uncontrolled VOC emissions, especially in more populated
areas, the Agency decided to evaluate strategies for reducing the
adverse environmental and health impacts from this emission category.
The commenter did not suggest any control techniques that would
reduce the hydrocarbon emissions from plants. The locations of hydro-
carbon emissions from vegetation are generally concentrated in rural
areas that are attaining the ambient ozone standard. In addition,
natural hydrocarbon emissions generally do not contain compounds that
are likely to be hazardous, such as benzene.
Comment: One commenter stated that in light of the uncertainties
associated wit-h the API studies, the Agency should move forward on the
basis of valid and reliable information currently available (which ex-
cludes the API studies), with the provision that its regulatory approach
may require supplementation or amendment as additional information is
developed (I-H-94).
Response: The Health Effects Institute has recently evaluated
the scientific information regarding the human health risk of gasoline
vapors (I-A-65). The Institute's conclusion was that although the API
animal study was well conducted and demonstrates that some components
of gasoline increase cancer rates in rats and mice, its relevance to
6-7
-------
human risk assessment is uncertain. The Institute also concluded that
the available epidemiologic evidence neither negates nor confirms the
interpretation that gasoline vapors are a potential human carcinogen.
Although the Institute suggested that the lower bound of gasoline vapor
risk was equal to zero, the Institute noted that a decision to implement
mandatory controls based on available estimates, despite the uncertainty,
is a policy question.
The EPA believes that it is appropriate to consider the API studies,
as well as other available information, in estimating the potential
health risk from gasoline vapors, despite the uncertainties involved.
These uncertainties were fully considered when interpreting the results
of the risk assessment in order to reach a decision on appropriate
regulatory strategies.
6.2 COST/BENEFITS OF CONTROLS
Comment: One commenter stated the nationwide onboard cost and
benefits analyses were in error because: 1) the per-vehicle costs
were greatly underestimated, 2) the analysis included benefits where
Stage II systems are already in place, 3) the analysis ignored the fact
that onboard will increase emissions at stations with Stage 11,4) the
analysis underestimated spillage, and 5) a large portion ot the in-use
effectiveness is due to evaporative emission controls that may not
occur (I-H-114).
Response:
1) The EPA has reanalyzed its onboard cost estimates in detail.
A discussion of these costs by cost category is contained in Section
2.6. '
2) An error was discovered in the original analysis of emission
reductions, whereby additional Stage II emission reductions were
calculated for areas where Stage II was already in place. This error
has been corrected in the revised analysis. The cost and incidence
calculations in the original analyses correctly considered the areas
where Stage II controls are already in place.
3) The impacts of onboard controls at stations already having
Stage II controls were considered in the 1984 analysis. These impacts
included the lack of control of underground storage tank emptying
losses and reduced recovery credits.
6-8
-------
4) No data on spillage could be found that were significantly
better than the basis of the AP-42 emission factor of 84 mg/1. There-
fore, this estimate was used in the 1984 analysis, as well as the new
analysis. Table 2-2 of the Draft Volume I RIA summarizes all the
emission factors used in the analysis.
5) In the reanalysis, excess evaporative emissions have been
considered separately; thus, Stage II and onboard have been compared
on an equal basis.
Comment; One commenter felt that EPA's cost-benefit analysis
approach seemed reasonable. However, they thought the analysis
overlooked or did not address the newer technology Stage II recovery
systems and the economic cost recovery mechanisms that exist (I-H-65).
Response: The EPA has re-evaluated Stage II costs and developed
cost estimates that reflect currently available systems (see Appen-
dices B and C of the Draft Volume I RIA). No specific economic cost
recovery mechanisms were mentioned by the commenter apart from fuel
recovery, tax advantage, and credits. These mechanisms were considered
for inclusion in the Stage II cost reanalysis.
6.3 ALTERNATIVE CONTROLS
Comment: One commenter opposed to Stage II controls uryed the
investigation of other control alternatives because there are other
solutions that would be technically better, less costly, and have their
costs borne more equally (I-H-88).
Response: This commenter did not specify any particular alterna-
tives. The Agency selected for evaluation all feasible regulatory strate-
gies and options it considered appropriate based on initial estimations
of costs, technical feasibility, emission reductions, and other factors.
Although there are other sources of VOC emissions that are candidates for
control, they are not substitutes for refueling control, but complementary
parts of the overall strategy to address the pervasive ozone problem.
Comment: One' commenter felt that, since onboard controls would
collect both refueling and excess evaporative emissions, any comparisons
made with Stage II should include the impacts of volatility, or Reid
6-9
-------
Vapor Pressure (RVP), limitations on commer:ial gasoline. (RVP limita-
tion is an alternative to control the evaporative emissions uncontrolled
by Stage II.) These added measures should be accounted for in comparing
both the costs and the time required to implement onboard and Stage II
controls. The commenter noted that the costs of KVP control would
include the capital costs of reconfiguring refineries to maintain
gasoline quality at lower KVP and the higher cost of lower vapor pressure
high octane blending components (I-H-94).
Another commenter suggested that a strategy including Stage II
plus extra evaporative emission control (employing an enlarged carbon
canister) would be more cost effective than any of the vehicle
refueling emission control approaches considered by EPA (I-H-118).
Response: The reanalysis of regulatory strategies performed since
the evaluation report was issued in 1984 takes into account the additional
strategies of Stage II combined with gasoline RVP limitations, and
Stage II combined with enlarged canisters for control of excess evaporative
emissions. The impacts of these and the other strategies are discussed
in the Draft Volume I KIA and in another EPA report (I-A-66).
Comment: One commenter felt that EPA failed to evaluate the
reduction of the volatility of gasoline as a control strategy, despite
the substantial impact this measure would have in reducing vapors at
every step in the gasoline marketing system (I-H-101). Another com-
menter stated that controls on commercial fuel volatility are needed to
reduce in-use evaporative emissions (I-H-1UU).
One commenter felt that EPA (through its powers under Section 211(c)
of the Act) should encourage all States to adopt and enforce the ASTM
Standard D 439. setting volatility limits on gasoline. Some gasolines
fall outside of these "good industry practices," and some States (such
as New York, New Jersey, and Texas) have not adopted the ASTM limits.
The commenter also saw no benefit in EPA's considering any restrictions
on gasoline RVP below the limits in the ASTM Standard, because of the
resulting disruption in the butane market and the gasoline price
increase of l.b cents per gallon that would occur during the control
period.. The installation of Stage II or onboard systems would be
preferable as measures to control vehicle refueling or evaporative
emissions (I-H-99).
6-lu
-------
Response: The Agency has evaluated RVP limitations and has proposed
a decision along with the onboard proposal. The rationale for these
proposals is contained in the respective Federal Register notices.
6.4 REACTION TO GASOLINE PRICE INCREASE (STAGE II)*
The Agency received only four comment letters from the general
public concerning price increase; i.e., these commenters were not repre-
senting any company, agency, or organizational affiliation. One of
these commenters expressed the viev: that higher gasoline or automobile
prices would not be justified unless a more positive link was established
between cancer deaths and gasoline dispensing by the public (I-H-28).
Several companies and organizations referred to gasoline price
increases in terms of a facility's ability to compete in light of its
control costs, and discussed whether these costs could be passed through
to consumers. These comments on the economic impacts of Stage II con-
trols are summarized in Section 3.6.
6.5 REACTION TO INCREASE IN PRICE OF A NEW VEHICLE (ONBOARD)*
Comment: Three commenters felt EPA's estimate of an onboard system
cost of $lb/vehicle to be reasonable and supported by API and ARCO studies.
In addition, the commenters believed there was no reliable technical basis
for the onboard costs estimated by the automobile manufacturers that
are far in excess of those made by API and ARCO (I-H-1U2, I-H-108, I-H-119).
One commenter felt that the impact of onboard controls was over-
stated, and that a $15 increase in the. purchase price of vehicles whose
price tags average upwards of $10,000 will not have any discernible
effect on new .car sales (I-H-115).
Response: Comparisons of abiolute dollar values associated with
the cost of control technology and the price of vehicles are not
sufficient to determine the effect on buyer behavior of small increases
in vehicle price. Price elasticities, which measure the percentage
change in quantity demanded due to a 1 percent change in the price of
a good, more accurately reflect buyer response to small price changes.
*1984 Federal Register topic.
6-11
-------
Since elasticities are intended to capture consumer responses to
small price changes, elasticities are used to estimate consumption
declines associated with price increases attributed to controls.
Since there is disagreement in estimates of vehicle price elasticities,
the analysis examines the impacts of vehicle cost increases assuming
several different price elasticities. (See pp. 8-22 through 8-26 of the
July 1984 Analysis Report.) The analysis reflects market history, which
indicates that some behavioral response on the part of buyers will be
observed, even when the price increase is small.
Comment: One commenter considered the onboard control system to
be an "anti-pollution gizmo" designed to raise the prices of American
cars (I-H-7).
Response: Both onboard and Stage II controls are designed to
reduce emissions of ozone precursors and hazardous pollutants, and
thereby protect the health of the public. As discussed in detail in
Section 2.6, the extra cost for onboard controls is estimated to be
about $2U for each passenger car (see Table 2-b). This cost to the
consumer seems reasonable in light of the environmental and health
benefits that would result from the controls. Moreover, an onboard
requirement would apply tc all cars, both foreign and domestic.
Comment: One commenter thought that, given the option of paying
for VOC controls through an increased' gasoline price of approximately
1 cent per gallon or $lb per vehicle, the consumer who drives more than
30,000 miles would generally be better off paying an extra $15 per new
vehicle. The commenter calculated that, assuming constant dollars and
20 mpg, the $15 increase for onboard would be equivalent to Stage II
costs associated with 30,000 vehicle miles. The same commenter noted
that, today, most automobiles are purchased on credit. Consequently, if
vehicle financing rates and time value of money are taken into account,
the one-time onboard cost would be significantly higher than $lb and
might not provide any benefit (I-H-126).
Response: As indicated by the commenter's analysis, many vari-
ables are involved in determining the relative attractiveness of
these regulatory strategies to an individual consumer. These vari-
ables include gas mileage, miles driven, length of vehicle ownership,
6-12
-------
and the credit terms involved in the vehicle purchase. Variations in.
one or more of these factors may result in greatly differing results.
In the reanalysis, such factors play an even larger role. The
latest onboard control analysis projects a per-vehicle onboard control
price that varies with vehicle vintage (model year) and posits that
captured vapors will improve fuel efficiency and generate recovery
credits for vehicle owners. Vehicle vintage, miles driven, length of
vehicle ownership, gasoline price, and the rate of time preference
(discount rate) determine the credit received by the individual vehicle
owner; vehicle vintage and credit terms determine the gross (initial)
cost of the control to the purchaser. Accordingly, the net cost of
onboard control to the individual consumer will vary with all of
these factors.
6-13
-------
Since elasticities are intended to capture consumer responses to
small price changes, elasticities are used to estimate consumption
declines associated witn price increases attributed to controls.
Since there is disagreement in estimates of vehicle price elasticities,
the analysis examines the impacts of vehicle cost increases assuming
several different price elasticities. (See pp. 8-22 through 8-26 of the
July 1984 Analysis Report.) The analysis reflects market history, which
indicates that some behavioral response on the part of buyers will be
observed, even when the price increase is small.
Comment: One commenter considered the onboard control system to
be an "anti-pollution gizmo" designed to raise the prices of American
cars (I-H-7).
Response: Both onboard and Stage II controls are designed to
reduce emissions of ozone precursors and hazardous pollutants, and
thereby protect the health of the public. As discussed in detail in
Section 2.6, the extra cost for onboard controls is estimated to be
about $2U for each passenger car (see Table 2-5). This cost to the
consumer seems reasonable in light of the environmental and health
benefits that would result from the controls. Moreover, an onboard
requirement would apply tc all cars, both foreign and domestic.
Comment: Une commenter thought that, given the option of paying
for VUC controls through an increased" gasoline price of approximately
1 cent per gallon or $15 per vehicle, the consumer who drives more than
30,UUO miles would generally be better off paying an extra $15 per new
vehicle. The commenter calculated that, assuming constant dollars and
20 mpy, the $15 increase for onboard would be equivalent to Stage II
costs associated with 3C,UUO vehicle miles. The same commenter noted
that, today, most automobiles are purchased on credit. Consequently, if
vehicle financing rates and time value of money are taken into account,
the one-time onboard cost would be significantly higher than $15 and
might not provide any benefit (I-H-126).
Response: As indicated by the commenter's analysis, many vari-
ables are involved in determining the relative attractiveness of
these regulatory strategies "to an individual consumer. These vari-
ables include gas mileage, miles driven, length of vehicle ownership,
6-12
-------
and the credit terms involved in the vehicle purchase. Variations in
one or more of these factors may result in yreatly differing results.
In the reanalysis, such factors play an even larger role. The
latest onboard control analysis projects a per-vehicle onboard control
price that varies with vehicle vintaye (model year) and posits that
captured vapors will improve fuel efficiency and generate recovery
credits for vehicle owners. Vehicle vintage, miles driven, length of
vehicle ownership, gasoline price, and the rate of time preference
(discount rate) determine the credit received by the individual vehicle
owner; vehicle vintage and credit terms determine the gross (initial)
cost of the control to the purchaser. Accordingly, the net cost of
onboard control to the individual consumer will vary with all of
these factors.
6-13
-------
6.6 REFERICES (Comment letters are not repeated here. See Chapter 1,
Table 1-1, for a complete list of comment letters.)
I-A-b'b Evaluation of Air Pollution Regulatory Strategies for Gasoline
Marketing Industry. U.S. EPA. Research Triangle park, N.C.
EPA-45U/3-84-U12a. July 1984.
I-A-6b Gasoline Vapor Exposure and Human Cancer: Evaluation of
Existing Scientific Information and Recommendations for Future
Research. U.S. EPA. Prepared by the Health Effects Institute.
September 1985.
I-A-66 Study of Volatility and Hydrocarbon Emissions from Motor
Vehicles. U.S. EPA, Office of Mobile Sources. EPA-AA-SUSB-
8b-b. November 1985.
I-F-78 Air Resources board, State of California. A Report to the
Legislature on Gasoline Vapor Recovery Systems for Vehicle
Fueling at Service Stations (Stage II systems). March 1983.
I-G-15 Regulatory Strategies for the Gasoline Marketing Industry.
Notice of Availability of a regulatory strategies analysis
document for public comment. U.S. EPA. 49 FR 31706. August
8, 1984.
6-14
-------
7.0 STAGE I CONTROLS*
Comment: One commenter stated that the 1984 cost of a carbon
adsorption unit for vapor control at bulk terminals was about 80 percent
of the 1981 cost (based on a maximum emission from the unit of 30 mg/1).
This change in cost would affect the Agency's bulk terminal control
cost estimates (I-H-27).
Response: The reduced cost suggested by this manufacturer was
factored into the updated bulk terminal cost analysis. The net effect
is a reduction of about 10 percent in the unit purchase cost for carbon
adsorption (CA) systems. Since the CA system costs are averaged with
thermal oxidizer and refrigeration system costs to determine average
bulk terminal control costs, the net effect is a reduction of only
about 4 percent in the average unit purchase cost for all vapor recovery
systems at terminals.
Comment: One commenter noted that, in Tables 7-8, 7-9, and 7-10
of the July 1984 Analysis Document, "NA" (not applicable) was shown for
the product recovery credit for for-hire tank trucks at terminals, for
bulk plants, and for service stations using Stage I control. The
ctwnenter felt that either he had misunderstood the costing material or
that cost credits should be allowed for the vapors saved through con-
trols in each of these cases (I-H-132).
Response: Recovery cost credits are applied in calculating net
control costs in two situations: (1) when vapors are condensed into
usable liquid product, and (2) when, through the application of controls,
product, is preve-nted from evaporating and being lost. The former
situation applies at a bulk terminal using a vapor processor to recover
'the gasoline vapors displaced from tank trucks during the loading
operation. Recovered product is piped as a liquid to storage tanks for
sale with the regular product. Tables 7-1, 7-2, and 7-4 of the July
1984 Analysis Report show recovery credits for bulk terminals using vapor
recovery controls. The latter situation applies in the case of storage
tanks and at bul'k plants using a vapor balance system. In the case of
storage tanks, the recovery credit represents the difference between
*1984 Federal Register topic.
7-1
-------
the emissions from a fixed-roof tank and an internal floating-roof
tank. Recovery credits for a bulk plant usiny vapor balance arise due
to the prevention of storage tank emptying losses following the loading
of account trucks. Without vapor balance, fresh air is drawn into the
storage tank as the tank is drained, and the product subsequently
evaporating into this air space is lost through tank venting. With
vapor balance, a vapor saturated air space is maintained above the
product to suppress evaporation and loss. Tables 7-6 and 7-7 of the
July 1984 Analysis Report show recovery credits for bulk plants controlled
by vapor balance.
As pointed out on page 7-15 of the report, a recovery credit does
not accrue to for-hire tank trucks. Vapors drawn into these tanks
during vapor balance do not suppress product evaporation and loss
(as occurs at a bulk plant using vapor balance) because the truck tank
is emptied of product during the vapor balance operation. For this
reason, product recovery credits do not apply in the cost calculations
shown for for-hire tank trucks in Tables 7-8 and 7-9 of the July 1984
Analysis Report. In the case of service stations, the displaced vapors
are lost from the underground tank regardless of whether Stage I is in
use or not. Therefore, Table 7-.1U does not apply a recovery credit in
presenting the costs associated with Stage I control at service stations.
Finally, credits are allowed for Stage II for reasons similar to those
given above for a bulk plant using vapor balance.
7-2
-------
8.0 EFFECTS ON STATE IMPLEMENTATION PLANS (SIP's)*
The EPA received numerous comments on the effect of gasoline
marketing controls on State regulations, particularly the feasibility
of the States' incorporating Stage II or onboard controls into their
respective SIP's. This section summarizes their comments and EPA's
responses to the following issues:
o Ozone NAAQS Attainment Deadline
o Emission Credits
o Effect of State Adoption of Controls
o EPA's Role in'Selecting Controls
o Miscellaneous Other Issues
8.1 OZONE NAAQS ATTAINMENT DEADLINE
Many commenters expressed concern about the States' meeting the
December 1987 deadline for implementing controls. These commenters were
also concerned with EPA's schedule for deciding whether to regulate VOC
emissions from refueling operations.
Comment: A number of commenters stated that EPA should decide
promptly whether further gasoline marketing controls are required
because States are actively searching for options to meet the
approaching 1987 NAAQS deadline. In this regard, some commenters sug-
gested that EPA select onboard and stem any move by the States toward
Stage II (I-H-99, I-H-102, I-H-108, I-H-109, I-H-119). One of these
commenters stated that it is important for EPA to define promptly the
role of hydrocarbons in oxidant formation before States with nonattain-
ment areas revise their implementation plans by the December 31, 1987,
deadline. The commenter pointed out that if EPA fails to act in a
timely manner, industry could be faced with the needless implementa-
tion of Stage II vapor recovery systems, possibly even duplicating
other systems for control of vehicle refueling emissions (I-H-99).
In contrast, other commenters felt EPA should not delay in issuing
guidance on implementing Stage II controls (I-D-56, I-H-21, I-H-100).
One felt that a Stage II regulation would bring the nation closer to
achieving the ozone NAAQS by the 1987 deadline (I-D-56). Another
*1984 Federal Register topic.
8-1
-------
commenter stated that States are ready to implement Stage II, noting
that seven States have already proposed to use Stage II controls to
meet ozone standards in certain areas and that their initiative should
not be discouraged (I-H-100). One State supporting a national onboard
control program said that it is proceeding with implementati.on of
Stage II as the only option available to control refueling emissions
on a short-term basis (1-0-65).
According to one commenter (I-H-92), a decision is needed soon
on whether to regulate VOC emissions from motor vehicle refueling in
order to: (1) provide guidance to States that have been directed by
EPA to revise ozone SIP's to adopt additional VOC control measures by
the end of 1987; and (2) give States time to determine what, if any,
regulatory approach they should adopt to control benzene emissions from
motor vehicle refueling in areas other than those that are nonattain-
ment for ozone, if a national regulation requiring Stage II controls in
urban ozone nonattainment areas is adopted by EPA.
Response: The Agency is proposing to regulate vehicle refueling
emissions by means of onboard.controls. This will result in effective
long-term reduction of refueling emissions, and nationwide protection
from exposure to benzene and other potential carcinogenic constituents
in gasoline. Because of the time necessary to phase in onboard con-
trols, however, the Agency has evaluated other Federal requirements
that could provide near-term VOC reductions. As a result, EPA also
is proposing to limit in-use fuel volatility which will produce sub-
stantial reductions relatively quickly. The EPA also considered
requiring Stage II controls in some ozone nonattainment areas as an
interim measure until onboard controls become effective. However, EPA
proposes not to impose Stage II as a requirement in those areas where
it is not now being implemented or is not contained as a commitment in
a State Implementation Plan (SIP). Nevertheless, States may choose to
adopt Stage II controls as an appropriate VOC control measure in such
areas where it is not now implemented or committed to, depending on a
case-by-case assessment of local ozone nonattainment problems and
available VOC control alternatives.
Comment: One commenter stated that EPA should defer any action
requiring States to adopt controls on VOC emissions from motor vehicle
8-2
-------
refueling in ozone nonattainment areas until a final national policy
has been established, and specifically, Stage II controls should not be
required until this form of control has been defined as reasonably
available control technology (RACT). The commenter felt that a require-
ment for action prior to such time would be unwarranted (I-H-92).
Response: As discussed in the accompanying Federal Register
announcement concerning proposed onboard controls and in other
responses, the Agency does not intend to require Stage II controls in
all nonattainment areas. However, Stage II can be effective in parti-
cular situations, and is an available measure for States to consider in
revising their implementation plans applicable to ozone nonattainment
areas. The EPA will support such efforts made by the States where
analysis indicates it is a desirable alternative for ozone control.
Comment: Two commenters suggested that EPA continue to focus on
the 16 original target AQCR's (all of which already have Stage II).
They felt no nationwide Stage I and Stage II measures should be con-
sidered until meaningful, reasonable procedures for the 16 AQCR's are
developed (I-H-42, I-H-43).
Response: The problem of ozone nonattainment clearly extends
beyond some "16 original target AQCRs" in which Si age II is now in
place. There are more than 6U major urban areas outside of California,
containing over 8U million people, which have ozone levels significantly
above the national ambient air quality standard. The Agency considers
the problem of ozone nonattainment to be a serious national concern
requiring a broad-based solution, of which the proposals for onboard
controls and reduction of fuel volatility are only a part.
Comment: Some commenters expressed concern about EPA's imple-
menting Stage II or onboard controls in. time to meet the 1987 deadline
(I-D-b4). Une commenter estimated that the phase-in period for Stage
II implementation would be 7 years, rather than 2 years; therefore,
implementation would not be faster than onboard (I-H-83). Uther com-
menters stated that Stage II vapor recovery systems cannot reasonably
be expected to be implemented in most nonattainment areas by the 1987
deadline (I-H-99, I-H-1U2, I-H-119, I-H-124). One of the commenters
stated that 6 years would be required to complete installation in non-
attainment areas (I-H-y9). According to another commenter, a minimum
8-3
-------
of 2 years would be required for the State (Pennsylvania) to develop
and adopt Stage II requirements, and another 2 years or more would be
needed to actually install the equipment (I-H-122). Another pointed
out that it would take his State (Virginia) close to a year to adopt
and implement a Stage II regulation (I-D-b3).
Response: To determine the phase-in period for Stage II implemen-
tation, EPA considered in its reanalysis a range of periods from 3 years
to 7 years. A number of commenters believed that 3 years would not be
enough time for all regulated businesses to install Staye II controls,
maintaining that there would not be a sufficient number of installation
contractors to perform the necessary work in 3 years. For that and other
reasons, EPA included an evaluation of a 7-year phase-in period in
the examination of a nationwide Stage II program. However, since Stage
II controls would be implemented on a smaller scale when considering
controls only in nonattainment areas, EPA believes that a somewhat
shorter period could be achieved. Therefore, EPA used a range of
phase-in periods in evaluating Stage II controls in nonattainment areas
in the reanalysis. A more complete discussion of control strategy
phase-in issues is contained in Section 2.5.
Comment: Two commenters favoring onboard urged that EPA delay the
1987 ozone attainment deadline or grant SIP extensions for reduction
shortfalls, since neither Staye II nor onboard would allow timely
attainment of the NAAQS for ozone (I-H-37, I-H-6U).
Response: The EPA has no authority to delay the 1987 ozone
attainment schedule, which is provided for in the Clean Air Act. How-
ever, the Agerrcy recognizes that a number of areas will not attain the
standard by the specified deadline and plans to develop a comprehensive
post-1987 ozone policy to deal with the need for revised SIP's and
adoption of additional and more aggressive VUC control measures.
8.2 EMISSION CREDITS
A number of commenters felt that emission credits (i.e., an
allowance for future reductions) should be given for Stage II and
onboard controls since implementation may not occur by the 1987 deadline.
Comment: Several commenters favoring the onboard control approach
suggested that, since neither onboard nor Stage II controls could be
8-4
-------
implemented in sufficient time for States to achieve the ozone stan-
dard by the December 31, 1987 deadline, EPA should grant advance SIP
credits for onboard controls, recognizing the delayed impact of this
control strategy (I-D-55, I-H-44, I-H-84, I-H-85, I-H-87, I-H-97,
I-H-102, I-H-105, I-H-106, I-H-109, I-H-110, I-H-121, I-H-122, I-H-123).
According to several of these commenters (I-H-85, I-H-87, I-H-102),
such a policy would give flexibility to the States, and would eliminate
the need for temporarily implementing Stage II controls.
Four commenters who recommended onboard controls stated that it
is not necessary for the selected control measure to be fully imple-
mented by the attainment deadline because EPA has the authority to
approve revised SIP's for ozone nonattainment areas which anticipate
controls that will be fully implemented by December 31, 1987 (I-H-98,
I-H-108, I-H-109, I-H-119). Two of these commenters further stated
that EPA should not create the impression that the controls necessary
to achieve the desired VOC reductions must be in place by the end of
1987 but rather the reductions must be "provided for". The Agency's
attainment policy and guidelines adopted for its implementation allow
EPA to approve a commitment to onboard without requiring additional
Stage II measures in the interim to offset the effects of any delayed
capture of emissions (I-H-108, I-H-119).
One commenter (I-H-94) added that if EPA required onboard controls,
the Agency should then conduct analyses comparing other available short-
term, interim control strategies to be used as guidance by individual
States needing additional VOC emission reductions to attain the NAAQS.
One commenter stated that, for judging nonattainment status,
credits should be given for hydrocarbon emission reductions that will
occur after 1987 through programs implemented prior to that date
(I-H-99).
Response: It is apparent that many areas will not attain the
ozone standard by the 1987 deadline. Thus, the commenters1 concerns
about whether measures have been "implemented" or "provided for" in
SIP's for such areas can be better addressed by considering whether the
SIP's contain measures sufficient for expeditious attainment of the
standard. As discussed earlier, the Agency is developing a post-1987
ozone policy which addresses the need for calls for revised SIP's,
8-5
-------
the ranye of activities to be covered in SIP revisions, and the need
for adequate demonstrations of progress toward attainment. As part of
the strateyy development process, the Agency has evaluated a broad
ranye of potential VUC control measures and will work with the State
and local agencies in the development of reasonable, effective proyrams
for expeditious ozone standard attainment.
Comment: Une commenter believed, that implementation of controls
would not achieve emission reductions until after the iytf7 deadline and,
therefore, the reyulations could be considered unnecessary (I-H-122).
Response: The EPA agrees that neither Stage II nor onboard
controls can be implemented in time to meet this deadline. However,
nonattainment area control strategies still will be needed even after the
deadline passes. The statutory requirement to attain the ozone standard
will still exist after December iy«7.
Comment: Une commenter stated that onboard provides significantly
more emission reductions than Stage II at rouyhly the same costs, and
emphasized that the most cost-effective control measure should be
implemented first (1-H-yo).
Kesponse: As described in the preamble, there were a variety of
factors that led to the proposal of onboard as the most desirable
refueling control technology. One of those factors was the projection
that'onboard, when fully implemented, would likely achieve greater
emission reductions, and cover more refueling events than would Staye II.
Comment: Une commenter felt that EPA's calculation of net present
value estimates of costs and benefits over a 3b-year period obscures
the fact that Stage II controls would reduce hydrocarbon emissions in
time to improve air quality by the lyb? deadline (I-H-114).
Kesponse: The analysis period of 3b years was used to compare
strategies with different phase-in times more objectively. It was
noted that Stage II could possibly get in place earlier than onboard,
but not likely on a national basis before the 1987 deadline. Careful
consideration was given to targeted Staye II versus onboard. Un the
basis of the broad range of benefits associated with onboard, and the
pervasive, long-term nature of the ozone problem, the Agency decided
8-b
-------
to propose onboard as the most desirable refueling control option (see
the Federal Register notice for additional discussion).
8.3 EFFECT OF STATE ADOPTION OF CONTROLS
Comment: Five commenters indicated that EPA should discourage the
piecemeal promulgation of Stage II by the States in their SIP's since a
State-by-State "creep" adoption of Stage II would eventually result in
a nationwide Stage II program. Pennsylvania was cited by one commenter
as an example of Stage II "creep," where State regulations would require
Stage II controls essentially statewide rather than in only the non-
attainment portion of the State (I-H-102, I-H-108, I-H-109, I-H-119,
I-H-120). All of these commenters suggested that EPA select onboard
controls instead of Stage II if further gasoline marketing controls are
required. One of them believed that States would consider Stage II
controls as their only available alternative (I-U-54, I-H-120).
Another commenter preferred onboard because its broader coverage
would make it more effective than area-targeted Stage II controls.
Also, area-targeted Stage II would almost certainly expand into addi-
tional areas as attainment areas were reclassified to nonattainment,
or as States applied the regulations statewide in the interest of equity
or administrative simplicity. This would undermine the apparent cost
effectiveness of a Stage II program that was originally targeted to non-
attainment areas (I-H-83). A second commenter also pointed out that,
due to the shifting nature of ozone attainment, States often effectively
turn a nonattainment area strategy into a statewide requirement. In
this way, nonattainment area strategies can tend to become equivalent
to nationwide proqrams. This commenter felt the use of onboard
canisters would effectively address the issue of "shifting or spreading"
nonattainment areas (I-H-24).
Response: States have the authority, in most cases, to prescribe
controls that are more stringent than Federal regulations require. If
a State chooses to require a control strategy statewide that EPA would
require in nonattainment areas only, or that EPA has not required at all,
that State effectively is enacting more stringent controls than those
required by EPA. States may implement such control strategies if they
deem them necessary to reduce ozone and protect the public health and
welfare.
8-7
-------
However, yiven tnat EPA has proposed to require onboard technoloyy
on new vehicles, which will eventually make Staye II unnecesssary, it
is doubtful that the "creep" issue will become a major problem. Even
those States that adopt Staye II as an interim strateyy are likely
to taryet the requirement to existiny problem nonattainment areas.
Comment: Several commenters from State control ayencies noted
their commitment or lack of commitment to adopt or investiyate Staye II .
controls. One commenter thought EPA's statement, that several States
had "made commitments to adopt Stage II" as part of their strateyy to
attain ozone yoals by 1987, was not entirely correct. This State ayency
(and others) had committed to study various alternatives and adopt
the best choice. The commenter believes such State efforts would be
compromised by an EPA decision to adopt onboard nationwide, unless the
attainment deadline is changed (I-H-21). Another State ayency felt EPA
should hold States to their original SIP commitments where necessary to
achieve the ozone standards (I-H-6b).
One commenter indicated that they had committed to investiyate
Staye II controls in the Pennsylvania 1982 SIP revisions for ozone.
However, until EPA makes a decision on whether to require onboard
canisters, the commenter believed it is unlikely that Staye II reyula-
tions will be adopted (I-H-122).
Response: As noted in earlier comments, EPA has evaluated Staye
II as an interim VUC control measure, and expects areas with Staye II
in place (or in the process of beiny installed) to continue to maintain
the system until such time as it is effectively made unnecessary by
onboard-equipped- cars and the appropriate SIP revisions are approved.
In addition, EPA expects States with existiny commitments in their
SIP's to implement Staye II to do so unless they can identify adequate
substitute measures and include such measures in approved SIP revisions.
The EPA recognizes that not all references to Staye II in an implemen-
tation plan constitute a "commitment." However, it must be realized
that even though a given plan may have committed only to study the
measure as a possibility if further VOC reductions proved necessary,
many areas will probably need such reductions. If Staye II is determined
to be impractical in areas where it has not yet been implemented in
terms of timing, cost-effectiveness, or other selection criteria, the
8-8
-------
necessary reductions nevertheless will nave to be made up by available
but perhaps equally difficult choices.
Comment: One commenter stated that if EPA demonstrates that
control of refueling vapors is necessary and a cost-effective way for
States to meet oxidant standards, then controls should be required only
in areas projected to be nonattainment at the 1987 deadline (I-H-99).
Response: Onboard is being proposed as the refueling control
strategy because, among other things, of greater overall benefit potential,
simplicity of administration and use, and cost effectiveness. The proposal
of onboard also reflects concern for the future of marginal attainment
areas (i.e., a need to maintain satisfactory air quality). Also, onboard
achieves emission reductions in areas surrounding nonattainment areas,
which contribute to violations of the ozone standard.
8.4 ERA'S HOLE IN SELECTING CONTROLS
The EPA received several comments about EPA's role in selecting
gasoline marketing controls.
Comment: Two commenters questioned EPA's role in selecting gasoline
marketing controls and recommended instead that the decision to select
controls should be delegated to the States. One of these commenters
(I-H-91) felt that EPA should not issue a Control Techniques Guideline
covering Stage II, but should leave the flexibility of choosing the
most cost-effective controls with the States. The second commenter
(I-H-94) stated that it is more appropriate for each State, rather than
EPA, to conduct an analysis of the unique problems of each nonattainment
area and then -to develop a specific control strategy for that area
based on a range of available control techniques. The commenter added
that EPA's role should be to analyze the pros and cons of all possible
ozone control techniques and provide guidance to the States in their
individual SIP revisions.
Another commenter stated that EPA has both the legal and moral
authority to guide States in choosing the most appropriate control
strategy so that States do not mistakenly implement short-term, less
cost-effective controls that would be obsolete in several years
(I-H-98).
8-9
-------
Response: As stated before, the Agency sees the ozone problem as
pervasive and long-term and there is much to be done by both the EPA and
the States. The EPA believes that vehicle refueling should be controlled
nationally, using onboard controls, because of the broad basis of the
ozone nonattainment problem, and because of the potential for exposure to
hazardous emissions. The Agency is also proposing to regulate volatility,
which will have national ozone benefits. The Agency feels that it is
best left to the States to adopt, as needed, a range of other measures
based on local needs.
8.5 MISCELLANEOUS
Comment: Une commenter noted that the July 1984 analysis listed
only 10 counties in Ohio as potentially subject to Stage II, while 26
counties are currently listed as nonattainment for ozone (I-H-24).
Response; The July 1984 EPA analysis evaluated two nonattainment
area groupings. One contained all nonattainment areas given an exten-
sion to 19H7 for achieving the standard. This grouping contained 1U
Ohio counties. The other grouping, called "selected nonattainment
areas," was a subset of the extension areas and consisted of areas
where a commitment was contained in their SIP to evaluate Stage II as
an additional means of achieving the ozone standard. No Ohio counties
were included in this grouping.
In the new analysis, the nonattainment area evaluations centered upon
three area groupings (11, 27, and 61 areas). The 11-area grouping con-
sisted of 11 metropolitan areas most likely to need refueling control,
based on design levels and estimated ozone shortfalls. The 27-area
grouping consisted of areas that were predicted to continue to be non-
attainment in 1987 and currently do not require Stage II. Neither of
these two groupings contain any Ohio counties. The-61 area grouping
consisted of non-California areas that recorded design values greater
than the ambient ozone standard (0.12 ppm) in 1982-1984. This grouping
contained 19 Ohio counties. Therefore, the number of Ohio counties
considered in the analysis is dependent on the definition of nonattain-
ment areas used. [When EPA considered all designated nonattainment
areas, the grouping included 27 Ohio counties.]
8-10
-------
9.U EXPOSUKE/KISK ANALYSIS
9.1 UNIT RISK FACTORS
Comment: One commenter noted that, using the information
developed in the evaluation, one cannot conclude that the control
options discussed are reasonably available or cost effective. Howeve'r,
if one were to take into account the exposure of gasoline service
station workers and also to factor in new information on benzene expo-
sure coming out of California, this conclusion may change. For example,
cancers from benzene exposure may be as much as three times higher than
EPA has originally reported (reference: Health Effects of Benzene,
Part B, California Department of Health Services, Epidemiology Section,
July 1984) (I-H-65). Another commenter asserted that EPA's 1981
assessment of benzene risk did not take into account a number of
important studies showing that benzene causes other forms of cancer in
addition to leukemia and that the unit risk factor should be increased
by as much as 15-fold (I-H-115).
Response: The unit risk factor for benzene has been re-evaluated
and increased in light of new information, in part contained in the July
1984 report from California (I-F-103). The new benzene risk factor,
which is 17 percent higher, was used in the reanalysis. Occupational
exposures of service station workers to benzene and gasoline vapors
were approximated in the revised analysis from time-weighted averages
taken from a Shell Oil industrial hygiene study (I-F-13). These new.
exposure situations were considered in the reanalysis of occupational and
lifetime risk.-
Comment: One commenter opposed the listing of benzene as a
hazardous air pollutant under Section 112 and argued that the EPA
assessment of unit risk is inflated (I-H-120). Another commenter noted
that EPA has calculated a unit risk factor for benzene based on informa-
tion gathered during a separate rulemaking on benzene (49 FR 23478).
This benzene rulemaking was controversial, and was challenged by many
commenters as significantly overestimating cancer risks (I-H-101).
Response: The Clean Air Act defines hazardous air pollutants under
Section 112 as those substances judged to cause or contribute to air
pollution "which may reasonably be anticipated to result in an increase
9-1
-------
in mortality or an increase in serious irreversible, or incapacitatiny
reversible, illness."1 EPA based the decision to list benzene on a
growiny consensus in the scientific and reyulatory community, as evidenced
by reports by the National Academy of Sciences,'^ the National Institute
for Occupational Safety and Health,3 and proposed reyulations issued by
the Occupational Safety and Health Administration,4 that benzene was
causally linked to the occurrence of leukemia in occupationally exposed .
populations. In EPA's judgment, leukemia clearly fits the criteria
described in Section 112 as a "serious irreversible, or incapacitatiny
reversible, illness."
The EPA's judyment that benzene present in the ambient air could
"reasonably be anticipated" to pose a significant health hazard to the
general population relied on two arguments advanced in the listing
notice: first, that benzene is released to the air at a rate of as
much as 2bU million pounds annually to which "large numbers of people
are routinely exposed" and, second, that EPA had "adopted a reyulatory
policy which recognizes that some risk exists at any level of exposure
to carcinogenic chemicals."5
Based on the above, EPA believes that the decision to list benzene
was appropriate. The subsequent assessments of low-level exposure
and carcinogenic risk were intended, as indicated in the listing notice,
for use in "determining which sources of benzene emissions must be
controlled, and the extent of control needed."6
2, U.S.C., 7412(a)(l).
^National Academy of Sciences, Health Effects of Benzene: A Review.
Washington, U.C., June 1976.
^National Institute for Occupational Safety and Health, "Update
Criteria and Recommendations for a Revised Benzene Standard,"
September 1976.
^Occupational Safety and Health Administration, "Occupational Exposure
to Benzene, Emergency Temporary Standards," 42 FR 22bl6 (May 3, iy77).
5U.S. EPA, "Addition of Benzene to List of Hazardous Air Pollutants,"
42 FR 29332 (June 8, 1977).
642 FR 29333 (June 8, 1977).
9-2
-------
Comment: .Four commenters pointed out that the Science Advisory
Board's (SAB) review of the June 1984 EPA Staff Report entitled "Estima-
tion of the Public Health Risk Exposure to Gasoline Vapor Via the
Gasoline Marketing System" concluded that the reported unit risk factor
for yasoline vapors was not suitable for quantification of human risk
and that the EPA Staff Report should not be used as a basis for regula-
tion. The commenters asserted that the SAb was critical of the Staff
Report for its selective approach to inclusion of epidemiological
studies and its failure to reveal the uncertainties inherent in the
available data (I-H-1U1, I-H-1U2, I-H-117, I-H-127).
Two other commenters reiterated three arguments as to why the
existing gasoline vapor health effects data base does not provide an
adequate basis for developing a quantitative risk assessment suitable
for regulatory decisionmaking: (1) there are fundamental biological
uncertainties as to the relevance for humans of the animal models in
the API studies; (2) both the chemical mixture and type of exposure in
the underlying study differ significantly from real world human
experience; and (3) the EPA risk estimates have not been supported by
the available epidemiological studies. Thus, they claimed that extensive
modifications in the development of the unit risk factor should be made
to take into account the purportedly substantial, fundamental, uncertainties.
At a minimum, they felt EPA should clearly characterize the uncertainties
in the unit risk factor. They considered their opinions to have been
confirmed by the SAB (I-H-12U, I-H-yi).
Response: There were two reasons for the reservations expressed by
the SAB. The first reason was that the vapor and whole gasoline may-
differ in the concentration of carcinogenically active components. This
issue will be discussed in a later response. The second reason concerns
the biological relevance of the male rat kidney response. However, as
discussed earlier in this document, the Agency believes that it is
appropriate to consider this evidence of animal carcinogenicity as indicative
of a potential risk to human health unless convincing evidence becomes
available showing that both the male rat kianey and female mouse liver .
9-3
-------
carcinogenic responses are anomalous. This issue is discussed in more
detail later in this document.
Comment: Several commenters expressed concern that EPA recoynize
and acknowledge the uncertainties associated witn the API exposure
studies (1-0-54, I-U-63, I-H-94, I-H-101). These commenters referred
to API's ongoing research to resolve the questions raised by the
original studies. They felt that the current risk factors should be
considered preliminary, and that further research using representative
gasoline vapors, as well as additional epidemiological studies, may
allow an acceptable unit risk factor to be established in the future.
The API briefly described its current ongoing healtn effects
research as having three parts: (1) to determine whether the rat is
an appropriate model for assessing effects in humans, (2) to observe
the nephrotoxic potential of various components or fractions of unleaded
gasoline, and (3) to acquire exposure and health data on occupationally
exposed populations. Results to date, according to API, suggest that
a significant human health risk is unlikely; however, a decision for
the onboard canister would represent a prudent action in light of
present uncertainties, and would also control the known benzene healtn
hazard associated with gasoline operations (I-U-54).
Response: The uncertainty in the risk assessment based on the API
study is discussed in considerable detail in a later response. Included
in the discussion are differences in composition between a liquid
aerosol of unleaded gasoline and a saturated vapor, the likely effects
of this difference upon the results, the relative contribution of
benzene-to the- carcinogenic response, and the validity of rat kidney
tumors and mouse liver tumors for extrapolation to humans.
The commenters imply that EPA should wait for an extended period
of time before making the judgment as to whether gasoline vapors present
a serious risk of cancer. However, the Agency believes that human
exposure to these vapors can have serious health consequences. The
precise quantitative measure of health risk calculated in the analysis
is secondary to the possibility that these health effects may be actually
occurring.
9-4
-------
Comment: Une commenter felt that data are insufficient to conclude
whether benzene levels below 1U ppm result in leukemia (I-H-101). Two
other commenters stated that several studies indicate a nonlinear rela-
tionship between risk and exposure duration, rather than a linear
relationship for benzene. This error could inflate the risk analysis
by 2 to 5 orders of magnitude (I-H-114, I-H-127).
Response: Because a specific environmental carcinogen is likely
to be responsible for at most a small fraction of a community's overall
cancer incidence and because the general population is exposed to a
complex mixture of potentially toxic agents in their daily lives, it is
currently not possible to directly link actual human cancers with
ambient air exposure to chemicals such as benzene. Today's epidemic-
logic techniques are not sensitive enough to measure a direct associa-
tion. Therefore, EPA must rely largely upon mathematical modeling
techniques to estimate human health risks. These techniques, collect-
ively termed "quantitative risk assessment," are the means whereby the
risk of adverse health effects from exposure to benzene in the ambient
environment can be estimated mathematically; for example, effects found
at higher occupational exposure levels can be extrapolated to lower
concentrations characteristic of human exposure in the vicinity of
industrial sources of benzene. The analysis estimated the risk of
cancer at various levels of exposure. A unit risk factor for benzene
is derived from the dose-response relationship observed in the occupa-
tional studies. The unit risk factor represents the cancer risk for an
individual exposed to a unit concentration of a carcinogen (e.g., one
part per million (ppm)) for a lifetime.
In the evaluation of benzene emissions from the gasoline market-
ing system, EPA has given great weight to the nature and relative
magnitude of potential public health risks. In the absence of scientific
certainty, regulatory decisions must be made on the basis of the best
information available. For benzene, this is represented by the epidemi-
ologic studies of the occupationally exposed population. The association
between benzene exposure and human leukemia is given strength by the . .
fact that leukemia mortality rates were observed among independent
studies in different occupational settings by independent investigators.
The epidemio.logical studies showed a 3-fold to 2L)-fold increase in
9-5
-------
risk of leukemia above that of individuals not exposed to benzene.
These findings present unequivocal evidence that chronic inhalation of
benzene causes leukemia in humans.
Comment: Several comments concerning the linear, non-threshold
dose-response model were received. One commenter felt that the gasoline
vapor unit risk factors may not be valid for dose levels outside the
range of 6U to 3UO ppm, based on the male rat data and a monotonically
decreasing gradient with incremental dose level. This suggests dose
saturation and questions a linear estimation of unit risk. In addition,
the commenter felt that both the male rat and female mice data are
inadequate for a valid estimation of true dose-response for low level
exposure (I-H-114).
Another commenter stated that the unit risk factors are based on a
nonthreshold, linear dose-response model which assumes that even one
molecule of a carcinogen presents a finite risk of a cancer. This model
has not been validated biologically for benzene or gasoline vapors and
is quite "conservative" since it always predicts cancer no matter how
small the exposure (I-H-1U1). One commenter did not question the
established risk assessment assumptions, e.g., linearity at low dose
levels, nonthreshold for carcinogenicity, and plausible uppe, limits,
since they are conservative estimates of risk. The commenter felt
protection of the public health requires the application of these
methodologies (I-H-126).
Response: While EPA agrees that the linear, nonthreshold model
is conservative in nature and would te'nd to provide a reasonable upper
bound to the statistical risk range, the Agency does not believe that
the assumptions upon which it is based are unreasonable noi that the
results of its use are exaggerated. The dose-response model with
linearity at low dose was adopted for low-dose extrapolation by EPA
because at the time of its introduction, it had the best, albeit
limited, scientific basis of any current mathematical extrapolation
model.
The risk estimate for gasoline vapor represents an extrapolation
below the dose range of experimental data. There is currently no
experimental basis for any mathematical extrapolation model that relates
exposure to cancer risk at the extremely low concentrations that must
9-6
-------
be dealt with in evaluating environmental exposures. For practical
reasons the correspondinyly low levels of risk cannot be measured
directly either by animal experiments or by epidemiologic studies.
Low-dose extrapolation must, therefore, be based on current scientific
understanding of the complex mechanisms of carcinoyenesis. At the
present time the dominant view of the carcinogenic process involves the
concept that most, but not all, cancer-causing agents also cause
irreversible damage to DNA. This position is based in part on the fact
that a very large proportion of chemicals that cause cancer are also
mutagenic. There is reason to expect that the quantal response that is
characteristic of mutagenesis correlates with a linear nonthreshold
dose-response relationship. Indeed, there is substantial evidence
from studies with both ionizing radiation and a wide variety of carcino-
genic chemicals that this type of dose-response model is the appro-
priate one to use. This is particularly true at the lower end of the
dose-response curve; at high doses, there can be an upward curvature,,
probably reflecting the effects of multi-stage processes on the bioloyic
response. ...
The linear nonthreshold dose-response relationship is also con-
sistent with the relatively few epidemiologic studies of cancer responses
to specific agentst that contain enough information to make the evalua-
tion possible. Examples of such agents include radiation-induced
leukemia, breast and thyroid cancer, skin cancer induced by arsenic
in drinking water, leukemia induced by benzene, and liver cancer induced
by aflatoxins in the diet. Some supporting evidence also exists from
animals experiments, such as the initiation stage of the two-stage
carcini genes is model in rat liver and mouse skin.
Because its scientific basis, although limited in some respects,
is the best of any of the current mathematical extrapolation models,
the nonthreshold model, which is linear at low doses, has been adopted
by EPA as the primary basis for risk extrapolation to a low levels of the
dose-response relationship. The cancer risk estimated with the non-
threshold model should best be regarded as conservative and representing
a plausible upper limit for the cancer risk (i.e., the true cancer risk is
not likely to be higher than that estimated, but it could be lower) when
extrapolating from animal studies.
y-7
-------
The quantitative aspect of carcinogen risk assessment for yasoline
vapors is best used in the regulatory decisionmaking process in deciding
whether a need exists for Federal regulation, or in evaluating the adequacy
of technology-based controls. Risk assessment is best construed as a
relative indication of likely risk. The linear extrapolation model used
provides an approximate but plausible estimate of the upper limit of
risk from exposure to a unit concentration of gasoline vapor (i.e., with
this model it is not likely that the true risk would be much more than
the estimated risk, but it could be considerably lower).
Comment: Four commenters criticized EPA's two unit risk factors
for gasoline vapors, which were based on the mouse and rat data from
the API gasoline inhalation study. The API work used wholly volatil-
ized unleaded gasoline, not gasoline vapors. Thus, the test animals
were exposed to all components of the gasoline, and not just the
lighter ends that would evaporate during refueling. The API approach
seriously overrepresented the heavier fuel components, which are more
likely to be carcinogenic (I-D-51, I-D-63, I-H-1U1, I-H-116, I-H-117).
One commenter questioned the relevance of administering a totally
evaporated gasoline vapor mixture continuously for hours as compared to
an ambient gasoline vapor associated with intermittent exposure for a
short time, followed by a long period without exposure, as is the case
with self-service gasoline distribution. This commenter remarked
that an error in the exposure estimate or health effect will influence
the cancer risk factor to a considerable degree. This commenter also
noted that h.igJi boiling point gasoline fractions are the most mutagenic
and carcinogenic in short-term bioassays anl animal studies. Thus,
the carcinogenic potency of the actual mixtures to which the population
is exposed may be entirely different from that calculated from the
animal study. Another animal study cited in the EPA Staff Paper indi-
cated high carcinogenic potency in the lighter fractions. The commenter
stated that contradictory health impacts lend a greater uncertainty to
the animal study results. Therefore, actual carcinogenic potency may
be overstated or understated- (I-H-126).
9-8
-------
One commenter felt that the lifetime risk from high exposure for
self-service ("traveling salesmen" case) is only 0.0005 times that of
a petroleum worker. The commenter suggested that EPA's estimate of unit
risk may be overstated if the transient nature of the self-service
exposure is taken into account (I-H-114).
Another commenter pointed out that EPA used the results from an
API study employing wholly vaporized gasoline and applied them to
exposure data generated during typical self-service refueling opera-
tions (referred to as the Clayton project). The commenter stated that
the Clayton project does not properly identify the composition of the
gasoline vapors. The commenter also pointed to a joint study conducted
by Amoco and Shell that showed the composition of gasoline vapor emitted
in vehicle refueling to be substantially different from the totally
vaporized gasoline administered in the API study. For example, the
totally vaporized gasoline administered in the API study contained 151
identifiable compounds, of which 42 accounted for 75 percent of the
mixture. In contrast, the Amoco and Shell data reveal that workplace
vapor exposures comprise about 20 components (of greater than 0.5 per-
cent by weight or volume) that account for at least 85 percent of the
sample. Of these, four C4/C5 hydrocarbons: n-butane, isobutane, n-
pentane, and isopentane, comprise 67 percent of the gasoline vapor.
Also, the Clayton data reveal that the proportion of various constitu-
ents measured were very similar and, in some cases, almost identical to
the proportions Amoco and Shell found. .The commenter further pointed
out that these four hydrocarbons are not known to be either nephrotoxic
or carcinogenic.. For these reasons, the commenter contended that EPA's
use of total hydrocarbon exposure ir. the risk analysis is faulty (I-H-99)
One commenter stated that the use of the API gasoline studies is
not appropriate as a basis for regulation of gasoline vapors because:
(1) the test animals were exposed to wholly vaporized gasoline - an
exposure totally unrepresentative of anything experienced by humans;
(2) potentially the most significant outcome of the studies - an
elevated incidence of kidney tumors found in male rats - is questionable
due to the apparently unique susceptibility of the male rat kidney to
hydrocarbon vapors; and (3) the other significant outcome - an elevated
incidence of liver tumors in female mice - is also open to question,
9-9
-------
since this type of tumor is known to occur spontaneously in laboratory
mice at a highly variable, and therefore unpredictable, rate. The
commenter concluded that since occupational exposures are shown to
involve only a mildly increased cancer risk, the risk to the general
population is likely to be negligible (I-H-94).
Response: The fraction of gasoline shown to be responsible for
subchronic pathological effects in the kidneys of rats included pri-
marily branched chain aliphatic hydrocarbons having 6 to y carbon
atoms ("C6-C9"). The relative percentages of these compounds in
gasoline vapors versus whole gasoline varies depending upon temperature
and other conditions, but on the average, vapors contain about 2t>
percent as much of these compounds as aerosolized whole gasoline. If
it is assumed that the C6-C9 fraction is also responsible for tumorigenic
activity, then the risk from tha vapors could be as low as about 25
percent of the risk previously estimated.
The claim that the C6-C9 branched chain hydrocarbons are respon-
sible for cancer induction is based upon the hypothesis that tumor
induction is a result of renal nephropathy. However, this hypothesis
is unproven. In fact, the presence of liver and renal tumors in mice
without accompanying pathology, and renal tumors in some rats without
mineralization of the renal pelvis, suggested that tumorigenesis is not
necessarily preceded by pathology. Therefore, to assume that gasoline
vapor is less carcinogenic than an aerosol of whole gasoline, and to
make a quantitative adjustment based upon this assumption, would result
in the possibility of underestimating risk.
The quantitative estimate of risk was based upon experiments from
animals enclosed intermittently, i.e., during the work day instead of
24 hours/day. Uhile it is true that humans are normally exposed for
much shorter periods, except for occupational conditions, it cannot be
assumed that short exposure periods are necessarily less harmful. For
example, exposure to ozone for 8 hours/day results in almost as much
lung damage as exposure to the same concentration for 24 hours/day.
In this case, extropolation of the risk from shorter exposure using
estimates derived_from continuous exposure (based on the product of time
and concentration) could greatly underestimate risk from shorter exposures.
9-1U
-------
Under EPA's Proposed Guidelines for the Assessment of Carcinogenic
Risk, the induction of liver tumors in mice is considered to be as valid
as any other tumor response in animals as evidence for carcinogenicity,
unless the response which occurred is weak in one or more of several
ways specified in the guidelines. The spontaneous rates in mice were
not high enough to throw the results into question. For gasoline, the
API study showed carcinogenic effects in both rats and mice, so that a
categorization of sufficient evidence in animals is warranted.
Comment: One commenter criticized several of the research methods
used in the API study, as follows: (a) The low air exchange rate
increased the ammonia, ($2, and CH4 concentrations in the exposure booth;
since the exposure atmosphere was controlled via the total hydrocarbons
only, and no background measurements were conducted in the control
booth, one cannot differentiate between the methane emitted by the test
animals and the hydrocarbons in the evaporating gasoline; (b) temperature
and relative humidity in the booths were not maintained (test conditions
were 21° to 29°C and 15 to 92 percent RH, while the OECD-GLP guidelines
recommend 22 +_ 2°C and 30 to 70% RH); (c) it cannot be judged whether
the produced gasoline vapor is comparable to the emissions that occur
during fueling and/or operation of vehicles; (d) the amount of aromatics
and the benzene concentration in the exposure atmosphere must be measured,
not just the total hydrocarbons; and (e) 2U percent of the mice died
during the quarantine when the room temperature dropped to 10°C.
Although it cannot be excluded that the health of the surviving animals
was affected by this drop in temperature, these animals were still used
in the test (.I-rH-,116).
Response: The flow rate through the chambers varied from 900 to
1,900 liters/minute. The maximum number of animals in a chamber was
200 mice and 200 rats. This number of rats and mice will require an
estimated 40 liters/minute. If the expired air contains about 5 percent
CO^, the COg production will equal 2 liters per minute. Based upon the
minimum air flow of 900 liters per minute in the chamber, the maximum
COg level should 'not exceed one-quarter of one percent. This is .
insignificant physiologically. Past experience has shown that if the
animal cages are changed regularly and the chambers are washed daily,
as was assumed to have been done under GLP procedures, ammonia buildup
9-11
-------
with three or more air exchanges per hour will be insignificant. Since
the air flow through the chamber was 45 to 95 times the estimated
expiratory minute volume of the entire animal population, and since the
expired air itself is likely to contain only a few ppm of methane, it
is unlikely that endoyenously produced methane would influence the
analysis of chamber vapor content to any detectable degree.
According to published results, temperatures in the four chambers
ranged from 24 +_ 1.4 to 26 +_ 1.3°C and relative humidities from b2 _+
9.5 to 56 _+ 7.2 percent. These are within the normal range for room
temperature conditions.
Also, according to the published report, only animals that were
apparently healthy were used. Since the animals were observed daily
during the study for any detectable abnormalities, without any unusual
findings in controls, it appears that the animals were normal. No
mention of any animal deaths during the quarantine period was made in
the published article.
Comment: One commenter noted that the lack of supporting weight
of evidence from short-term bioassays (e.g., mutagenic responses), as
well as other unexplained inconsistencies, may weaken the validity of
the animal sw.udy results used for the health risk component of the
cancer risk factor (I-H-126).
Response: While positive short-term bioassays would provide
supportive evidence for carcinogenesis, the induction of tumors in two
species, along with suggestive evidence from epidemiological studies,
provides strong enough evidence to justify a quantitative estimate of
risk.
Comment: Une commenter pointed out that for gasoline vapors, EUB,
and EDC, there is no definitive evidence of human cancer and the unit
risk factor is based entirely on animal studies. The commenter suggested
that attempting to quantitatively extrapolate animal data to humans is
extremely risky (I-H-1U1).
Response: In this uncertain field, it is appropriate to make the
conservative assumption that animal toxicology experiments are indicative
of potential human health effects unless it can be shown that animal
results are irrelevant, which is a difficult fact to establish.
9-12
-------
The primary uncertainty in the use of animal data to assess cancer risk
in humans is due to differences in sensitivity of the test species in
comparison with humans. This could be due to differences in absorption
efficiency, excretion rates, metabolic pathways, levels and inducibility
of xenobiotic metabolizing enzmyes, DNA repair mechanisms, etc. The
EPA/GAG is aware of these uncertainties and they are accounted for in
the quantitative estimate of risk.
The compounds ethylene dibromide (EDB) and ethylene dichloride
(EUC) are used as lead scavengers in additives to leaded gasoline. As
leaded gasoline continues to become a smaller percentage of the gaso-
line sold in the U.S., emissions of these compounds are becoming
negligible. Therefore, as discussed elsewhere in this document,
although EDB and EDC have been shown to cause adverse health effects in
experimental animals, these two substances are no longer being considered
in this analysis.
Comment: Une commenter noted the Agency's implicit assumption
that the concentration at the location of the receptor is equal to the
dose administered to the target organ (kidneys or liver) where the
health impact will occur. The commenter did not think it was clear
frorr, the animal data whether this is the case for human exposure. The
pharmacokinetics may not be the same for humans as it is for the test
animals (I-H-126).
Response: While differences in pharmacokinetics may exist
for many chemicals, such as cadmium compounds, the solubility of the
compound appears to be a greater factor in concentration in the kidneys
and liver than pharmacokinetic differences. Unless it can be shown that
phannacokinetic factors are responsible for a lesser susceptibility to
gasoline vapors in humans than in animals, then any adjustment of the
quantitative risk based on pharmacokinetics is not justified.
Comment: One commenter conceded that based upon historic data
regarding spontaneous kidney carcinomas in rats, the incidence for the
animals exposed ,in the API study must also be attributed to external
influences. This statement must be qualified, however, by the fact
that in this case of exposure, the tumors could only develop due to an
endogenous neph.ropathy of the male rats (I-H-116).
9-13
-------
'Response: The male rat has been shown to be more susceptible to
renal nephropathy from inhalation of gasoline than most other laboratory
species tested. However, there is no direct evidence indicating that
humans are less susceptible in this regard than male rats. Moreover,
even if the general population is less susceptible, sensitive sub-
groups may exist. For example, Kazantzis, et_ aK, 1962. Quart. J. Med.
31:403-419, reported data indicating the presence of a subpopulation of
humans extremely sensitive to the nephrotoxic effects of mercury. This
population resembled that of a genetically susceptible strain of rats
in this respect.
The most significant evidence linking the development of renal
tumors with renal nephropathy is a positive correlation between the
two endpoints. While it is possible that tumorigenesis may be related
to a promoting mechanism, such as continuous cell damage, it is also
possible that the positive correlation may be completely fortuitous.
In fact, several bits of evidence suggest that a tumorigenic response
may not be dependent upon kidney pathology. First of all, mineraliza-
tion of the renal pelvis was one of the common damaging effects of
gasoline exposure in male rats, but there was no correlation between
rats with mineralization and animals with kidney tumors. Secondly, two
renal tumors were found in the kidneys of exposed female mice, a
species which generally did not exhibit kidney pathology. Although the
presence of only two tumors does not constitute a statistically signif-
icant increase, nevertheless, due to a low historic incidence the
numbers are suggestive. Finally, a significant increase in liver
tumors was detected in female mice, again with little accompanying
liver pathology.
In summary, in order to discount the rat data it would be necessary
to assume that male rats are both uniquely sensitive to gasoline induced
renal nephropathy and that renal tumors arise as a result of the acute
nephrotoxic phenomena being currently explored by API. Since neither
hypothesis has been proven, the rat kidney cancer data are considered
as indicating possible human carcinogenicity.
Comment: One commenter pointed out that in the API study some of
the female mice contracted liver cancer and some of the male rats con-
tracted kidney cancer, yielding two different cancer rates.. The fact
9-14
-------
that the mouse data cannot even predict the rat data (or vice versa)
reflects on the reliability of the numbers for humans (I-H-lul).
Response: The rates of tumor induction in different strains,
species, or target organs are seldom identical. The dose at the target
organ, rates of activation or deactivation, UNA repair mechanisms, and
many other factors will alter the response. This type of variability
is inherent in any toxicology study, whether tumors or a nonocogenic
response is the endpoint, and does not invalidate the results. In
addition, the revised range of risk factors based upon rat data, which
incorporate a data correction by API as presented to the SAB, encompasses
the range of maximum likelihood to plausible upper limit estimates of
risk based on mouse data.
Comment: One commenter did not understand why EPA used the
results of only three of the available epidemiological studies in
estimating risk and urged EPA to draw on all sound epidemiological
studies, both negative and positive, in assessing potential hazards to
human health (I-H-99).
Response: In the current analysis, bb studies concerning exposure
to gasoline vapors have been analyzed. Many of these studies did not
prove to be relevant because of multiple exposures to compounds other
than gasoline vapors.
9.2 RISK ASSESSMENT METHODOLOGY
9.2.1 General
Comment: One comrnenter felt that the health risk assessment
methods, were s-tate-of-the-art and comprehensive. However, even though
the approach was sound, the number of assumptions used in the analysis
made it difficult to place a reasonable degree of confidence in the
conclusions drawn in the report. (However, none of the assumptions
appeared to be extremely unrealistic.) Also, they felt that since
there appears to be some risk to persons in gasoline handling occupa-
tions, service station workers should have been included in the
occupational group risk assessments (I-H-66). Two other commenters '
thought that risk estimates should be presented for all the employees
at a service station (I-H-llb, I-H-12U).
9-lb
-------
Response: Since publication of the EPA analysis, the exposure
assessment associated with benzene and gasoline vapors has been modi-
fied to now include occupational exposure of service station attendants.
Shell Oil conducted personnel monitoring of service station attendants'
exposure to gasoline vapors and benzene at seven service stations
(I-F-13). Based on data presented in this study, a lifetime risk and
oase year incidence were estimated for these employees. The base year
baseline occupational incidences were projected to other years by
assuming that baseline incidence is proportional to total national yas-
oline throughput. The incidence reduction for occupational exposures
was assumed to be 75 percent of the total percentage reduction calcu-
lated for a refueling strategy, since station attendants are exposed to
sources of gas vapors and benzene other than refueling operations.
Comment: One commenter felt that the EPA analysis invariably
made "worst-case" assumptions, thus maximizing estimates of potential
risks and, in effect, factoring a hugely exaggerated safety margin into
the results. This commenter saw little useful purpose in a calculation
for an individual based on a worst-case unit risk factor multiplied
by a worst-case exposure time (i.e., the case of a traveling salesman
using 4U gal/week for 50 years). The end result is neither realistic
nor objective (I-H-101).
Response: The EPA risk analysis is conservative in order to ensure
that the public health and welfare are adequately protected in spite of
uncertainty. However, the assumptions were not truly "worst-case" but
are the middle to high end of the reasonably expected range. Thus, the
resulting risk estimates may reflect relatively high exposure, but
should be realistic and a proper basis for a prudent public health
analysis. For example, in the case of lifetime risk, emission factors
based on national average conditions (RVP and temperature) were used,
while emission factors for the area with the most extreme conditions
could be significantly higher.
Comment: One commenter suggested that the following factors should
be resolved before a decision is made. Are the exposure estimates
based on the most appropriate exposure models, modes of exposure, and
calculation of exposed population? The methods used may be the best
presently available, but are they grossly overestimating (or under-
estimating) the numbers of exposed population, concentration levels at
9-16
-------
the receptor, and the duration of exposure? The commenter asserted
that these factors have a direct influence on the estimation of the
cancer risk factor (I-H-125).
Response: The methodologies used in the original analysis were
reassessed and determined to still be the best available for the purpose
of reasonable estimates. The exposure models are the best available for
such numerous and widespread sources. The modes of exposure considered
both community exposure from dispersed emissions and individual expo-
sure at service stations. The entire population of the country was
considered at some exposure concentration in the various incidence
analyses. The estimates in the risk analysis are of a reasonable mag-
nitude. The assumptions and strategies were refined and revised as
seemed warranted. For example, the unit risk factors, gasoline consump-
tion, in-use efficiencies, and nonattainment coverages were revised.
Also, additional exemption levels, onboard coverage (HDGV), and service
station projections were reconsidered.
Comment: One commenter noted in particular that the benzene inci-
dences are actually part of the gasoline vapor incidences, not an addi-
tion to them. The commenter stated that benzene is present only at
refueling as a component of gasoline vapors, and should not be counted
twice as EPA has done (I-H-1U1).
Response: The association between benzene exposure and leukemia
has been well documented in several epidemiological studies. The ben-
zene unit risk factor estimated from three occupational studies was
used to calculate the benzene cancer incidence. The gasoline vapor
unit risk factor .was estimated from the API chronic inhalation study of
unleaded gasoline vapors in rats and mice, and was used to calculate
the gasoline vapor incidence. It is possible tnat some of the response
to gasoline shown by the mice and rats is due to the benzene content.
However, there is no conclusive evidence to support or deny that this
response was due to the benzene component of gasoline. Further research
is needed to identify which compounds or fractions of compounds are
responsible for the carcinogenic effect. However, since the cancers
caused by the gasoline vapors were tumors in the kidneys of rats and
the livers of mice, rather than being leukemia (as would be expected
from exposure to benzene only), the effects were assumed to be additive.
9-17
-------
Comment: One commenter compared cumulative residual benzene inci-
dence with onboard controls at 44 incidences over 35 years to Stage II
control (with only 70 percent of self-service refuelings controlled)
at 50 incidences. The commenter felt that this difference is probably
not statistically significant, considering the many uncertainties
involved (I-H-101).
Response: The revised analysis shows larger differences in the
cumulative residual between incidences (as much as 85 for onboard vs.
106 for Stage II * Evap over 33 years). It is true that from a statis-
tical viewpoint, confidence intervals of the two estimates may overlap.
However, in light of limitations in knowledge and understanding, the
estimates must be used as the best available basis for regulatory
decisionmaking.
9.2.2 Exposure Measurements
Comment: One commenter noted that the EPA strategies evaluation
report refers to "average" exposures being calculated from the exposure
data for persons refueling their vehicles in an API study. However,
the EPA report does not explain how this calculation was done (I-H-101).
Response: Data presented in the API/Clayton self-service
refueling exposure study were used as the basis for calculating
"average" exposures (I-D-17). During this study, samples were collected
to characterize typical exposures to total hydrocarbons (measured as
n-hexane), benzene, and eight other compounds at 13 service stations.
API/Clayton calculated the geometric mean for each type of gasoline
refueled (leaded, unleaded, and/or (unleaded) premium)) at each station.
The geometric means of the mass concentrations reported by Clayton/API
were used, along with the station temperature and the molecular weight
of benzene or gasoline vapors, to derive a value of volume concentration
in parts per million. An arithmetic mean volume concentration of ben-
zene and gasoline vapors for each type of gasoline was calculated from
these adjusted geometric means for each station.
9.2.3 Incidence
Comment: One commenter felt that EPA's assumption that the loca-
tion of source sizes should be based on population densities (Section
4.1.2.1 of the July 1984 Analysis Report), and the estimate that bulk
terminals have the highest emissions, would greatly overstate the
9-18
-------
exposed population and directly affect the calculated incidence. The
commenter stated that bulk terminals are not usually located in the
center of a high population-density area, and not all of the population
in the vicinity of the facility would be equally exposed. Consequently,
they felt that selecting the highest exposure level and applying it to
the greatest number of people skews the cancer risk factor considerably.
The commenter remarked that the bias may be somewhat modified by the
terminal location method used. Exposure levels may be biased upward
due to the assumption that the SHEAR version of the Human Exposure
Model (HEM) is most appropriate. Use of a surrogate (benzene) to
predict gasoline vapor concentration may be a source of error, but its
use may provide a more accurate prediction of gasoline vapor concentra-
tions. They also felt that use of the ISC dispersion model for the
smaller gasoline vapor generators may modify the uncertainties associ-
ated with the HEM model. They considered this methodology to be the
most conservative one to use for estimating exposure levels of a particular
pollutant from a point source at the receptor locations (I-H-126).
Response: Bulk terminals and bulk plants could not be modeled on
an individual basis since for both cases there are too many facilities
nationwide (i.e., 1,500 bulk terminals, 15,000 bulk plants). A limited
amount of data was available on the locations and throughputs of these
types of facilities. In the case of bulk terminals, the model plant
configurations, characteristics, and size distribution were taken from
available data established during the development of the bulk terminal
new source performance standards (NSPS) (I-A-34). Bulk plant model
plant characteristics were based on previous EPA studies of this
industry sector (I-A-9). The bulk terminal and bulk plant model
plants were assumed to be distributed among locality sizes in which it
was most likely that each size of model plant would be located. Based
on industry data and experience, larger model plant sizes were generally
assumed to be more likely to be located in more populated localities.
Ten localities of varying sizes were chosen to represent the demographic
and meteorological characteristics of all facilities within each model
plant size range. The geographical distribution of localities repre-
senting each model plant size was as widespread as possible so that a
cross-section of nationwide climatological conditions was also
9-19
-------
represented. The facility location method used did not locate >u1k
terminals or bulk plants within high population-density areas. On
the contrary, most of the selected sites were in industrial areas or on
the outskirts of a metropolitan area.
The SHEAR version of HEM is most appropriate when: (1) multiple
pollutants from one source must be analyzed, (2) there are a large
number of sources within an area that cannot be located individually,
or (3) when the sources are so widespread that they act as an area
source rather than several point sources. Thus, the SHEAR version was
considered appropriate to model incidence from bulk terminals, bulk
plants, and gasoline service stations. For bulk terminals and bulk
plants, the point source routine of HEM was used. In this method, the
population in the vicinity of the facility is not considered to be
equally exposed. The HEM point source routine considers both the
dispersion of pollutants around a source and the number of people
residing (in the various census block groups) around the source in its
calculation of the number of people exposed to various concentrations,
and the resulting incidence. For service stations, the area source
routine of HEM was used to calculate a single exposure concentration
for the entire population of an area, assuming a uniform distribution
of emissions.
The ISC model was used to estimate lifetime risk to an individual
living around a complex of facilities, whether bulk terminals, bulk
plants, or service stations. This model was used for lifetime risk
because the contribution of each emission source to the ambient
concentration at each receptor is calculated separately, considering
the location and characteristics of each source. This level of
detail was considered unwieldy and infeasible for the number of sources
nationwide considered in the incidence analysis. This level of detail,
however, was needed to assess the local scenario that was used to
extrapolate to a nationwide analysis of the lifetime risk from high
exposure.
Comment: One commenter felt that the total population exposure
due to self-service refueling was erroneously obtained by dividing
the annual self-service throughput by the average pumping rate.
The commenter felt that the Agency's procedure is faulty because it
does not define the size of the population exposed (I-H-127).
9-20
-------
Response: The self-service vehicle refueling analysis assesses
the risk to individuals other than station attendants from exposure to
the high concentrations present near the tank fillneck duriny refueling.
Thus, the computation of self-service incidence is based on the
exposure incurred by one person during each self-service refueling.
Since it is assumed that it will always be a single person pumping
fuel, the total annual incidence is directly proportional to the annual
self-service gasoline throughput. This relationship holds true because
the total exposure time for all refueling is equal to the self-service
throughput divided by the assumed average gasoline pumping rate of 8
gal Ions per minute.
Thus, the incidence can be calculated by taking the product of the
exposure time for all self-service refuelings, the refueling concentra-
tion, and the number of people exposed per refueling (i.e., one).
Under this approach, it can be seen that the impact of self-service
refueling is independent of the size of the population exposed. In
effect, while the entire self-service refueling population is con-
sidered, tenuous assumptions regarding the gasoline usage of specific
individuals need not be made. Table 6-5 in the July 1984 Analysis Report
outlines the calculation procedure used.
Comment: Une commenter noted purported discrepancies in the July
1984 Analysis Report between EPA's stated method of risk analysis
(Tables 6-5 and 6-8), the data base EPA states is used (Table 4-6), and
the results presented (Table F-8). The commenter cited alleged errors,
including: 1) inconsistencies in benzene levels used, 2) use of total
gasoline consumption rather than only the self-service fraction, and
3) double counting of the population at risk-(I-H-127*.
Response: 1) For the self-service incidence analysis, average
benzene exposure concentrations were calculated for both leaded and
unleaded gasoline. In the case of unleaded gasoline, a weighted average
benzene exposure for premium and regular unleaded gasoline was calcu-
lated. The average benzene exposures during self-service refueling
were calculated to be Q.96 ppm for unleaded and 1.46 ppm for leaded. . .
For the self-service incidence analysis, some consumers will be
exposed to regular, and others to premium, unleaded gasoline. Thus,
it was important to calculate a weighted average benzene exposure
9-21
-------
associated with the unleaded yasoline throughput, which was not differen-
tiated between reyular and premium. However, for the selfservice lifetime
risk analysis, an averaye benzene exposure concentration for only reyular
unleaded yasoline was used, rather than a weighted average for both
premium and regular (as in the self-service incidence analysis). This
resulted in benzene concentrations of 0.98 ppm for unleaded gasoline and
1.46 ppm for leaded gasoline. The self-service lifetime risk analysis
estimates the risk associated with hiyh exposure to any one individual.
Since leaded yasoline is being phased-out of the marketplace and was ex-
pected to be gone by the time frame of this analysis, the benzene concentra-
tion for reyular unleaded yasoline was used in the lifetime risk analysis.
2) The EPA used the procedure outlined in Table b-b of the July
1984 analysis for the calculation of self-service incidence. (The
procedure in Table 6-8 was used for self-service lifetime risk to an
individual with hiyh exposure.) This procedure does use only the
self-service yasoline fraction. As suyyested by the commenter, only
the nonayricultural throuyhput was considered for service stations and
the throughput already controlled by Stage II was considered. However,
the existing Stage II throuyhput was calculated to be about 9 percent,
rather than the 1U percent assumed by the commenter, and the associated
incidence was calculated at controlled levels ratner than zero.
3) The method used by EPA does not double count the population at
risk. Although it may not have been clear in the abbreviated derivation
given in Table 6-b, the projections of incidence in any given year were
calculated using the ratio of leaded or unleaded throuyhput in the
given year to the base year total national gasoline throughput. Thus,
no correction is needed for the fraction of time each week spent pumping
leaded or unleaded gasoline. In fact, the weekly pumping rates are
assumed only for lifetime risk calculations, but are not used in the
incidence calculations.
Comment: One commenter stated several alleged reasons for believing
that the risk analysis underestimates the public health risks. The EPA
has excluded passengers in cars at service stations who are exposed during
refueling. Furthermore, even though OSHA has the authority to protect
employees, it is indefensible that the analysis ignore the cancer inci-
dences -associated with employees and the resulting employee incidence
reduction resulting from the control strategies (I-H-llb).
-------
Response: The EPA considered the inclusion of passengers in cars
at service stations duriny refueling. However, the exposure concentra-
tion is proportional to the inverse of the cube of the distance from
the source; windows are closed about one-half of the year, even on
cars without air conditioning; and the fraction of cars with air condi-
tioning and, thus, with windows closed nearly year-round is increasing.
Therefore, the incidence due to passenger exposure was considered in
the analysis to be negligible. For lifetime risk due to high exposure,
a high estimate of passenger exposure was considered in developing the
assumption of 50 years of equivalent 4U gal/week usage. Even so, the
passenger exposure contribution to total lifetime risk was nearly
negligible. Although OSHA has the authority to protect employees,
estimates of the incidence resulting from occupational exposure of
service station employees were included in the reanalysis. The inclu-
sion of service station occupational incidence provides a more complete
assessment of the baseline incidence due to vehicle refueling and the
benefits of tne various control strategies.
9.2.4 Lifetime Risk ...
Comment: One commenter stated that EPA declined to sum the
category-by-category lifetime risk on The assumption that it is
unlikely that any one individual would be exposed to high exposures
from any two source categories. The commenter claimed that this
assumption is flawed because it is highly probable that some of the
most exposed persons near various types of wholesale operations also
frequently visit, or may even work at, service stations (I-H-llb).
Response:. .The EPA still deems it appropriate to calculate the life-
time risks by individual source category and consider them separately.
In particular, the individual categories are controlled by different
control options, so presenting the individual categories presents the
resulting risk reductions more clearly. Admittedly, some individuals
may be exposed to several source categories at the same time. However,
it is less likely that they would be subject to high exposures from all
categories, because the high exposure scenarios assume that several of
one type of facility are close together. To examine the possible impacts,
the lifetime risks could be added together, but the resulting sums
would not differ significantly from the risk for bulk terminals alone.
9-23
-------
9.3 EXPOSURES DURING SELF-SERVICE REFUELING*
9.3.1 Gasoline Pumping Rate*
Comment: Two commenters felt that EPA's use of 8 gal/min to
represent the range of pumping rates of 8 to 12 gal/min inflates the
apparent risk from self-service (I-H-114, I-H-127). Another commenter
also disagreed with the EPA assumption that gasoline pumps operate at
an average of 8 gal/min. The EPA had justified this value on the basis
that filling rates are frequently extended by persons "topping off", or
by automatically reduced pumping rates at the end of prepaid fill-ups.
Although the commenter agreed that this does increase exposure time, it
seems that gasoline vapor concentrations would be proportionally reduced
due to a slower pumping rate. The end result of EPA's assumption is a
50 percent higher prediction of cancer among people who fuel their own
vehicles (I-H-101).
Response: Data received from API (I-D-38) and other background
data on gasoline pumping rates were re-examined. These data showed
that gasoline pumps typically operate within a range of 6.5 to 12 gal-
lons per minute. The API stated that roughly one-half of the service
station pumping systems use suction pumps, which normally can operate
at 8-9 gpm, while the other half use submersible pumps, which typically
can operate at 10 gpm or more, but vary in rate depending upon the number
of nozzles operating simultaneously and the pumping distance. The API
also stated that some cars must be filled at low rates to avoid nuisance
shutoff, and reported the statement of one nozzle manufacturer's engi-
neer that the average automobile fill configuration would not accept
more than about 8 gpm without nuisance shutoff. The Agency concluded
that the 8 gallon per minute level was a reasonable assumption. A
Stage II system may generally require a slower pumping rate than an
onboard system; however, this could not be clearly determined from the
data. The API reported, for example, that "second-generation" balance
systems operate at 5-8 gpm, while newly certified systems allow 10 gpm.
*1984 Federal Register Topic.
9-24
-------
9.3.2 Number of Tank Fillings*
Comment: One commenter cited recent studies showing average annual
miles traveled per household vehicle as ranging between 10,100 and
11,600. Average fuel efficiency of household vehicles was given as
14.7 mpg in 1982, and projections of average fleet fuel economy of
17.5 mpg in 1984, 23.8 mpg in 1990, and 27.9 mpg in 1995 were reported.
Assuming that the average vehicle miles traveled (VMT) in the 1995 to
2000 time period will be 11,000 miles per year at a fuel efficiency of
27.9 mpg, the commenter calculated the average fuel consumption per
vehicle as 394 gal/yr. Assuming a weekly fill, this would be equivalent
to a consumption of approximately 7.5 gal/week. If a high exposure is
represented by a salesperson traveling 3 to 5 times the average VMT,
the fuel consumed would range between 23 and 38 gal/week, so that the
EPA assumption of 40 gal/week seems very high. Furthermore, they noted
that the assumed high exposure period of 50 years of life implies that
such people would be driving the same number of miles, even when 70
years old (assuming the average salesperson begins employment at age
20). Since most people retire at around 65 years of age, they thought
that driving 3 to 5 times the average VMT for 50 working years over-
states the exposure risk. The commenter suggested that an estimate of
30 to 35 working years and an average use of about 30 gal/week may be
more reasonable for the high exposure scenario (I-H-126).
Response: As the commenter estimated, 38, or about the 40 gal/
week used in the analysis, is a high, but reasonable, usage rate for a
high exposure scenario. Although a correspondingly high working life
might be 45 years (from 20 to 65 years of age), the 50-year exposure
life was assumed in ordjr to account for lower usage rates before
employment and after retirement. The 30 to 35 working years and average
use of about 30 gal/week suggested by the commenter are, indeed, reason-
able estimates of average usage or high typical usage. The EPA estimate,
however, although not a maximum exposure scenario, is designed to
reflect the high end of exposures that could reasonably be expected.
9.3.3 Exposure Concentrations*
Comment: One commenter stated that his company had collected
14 air samples to measure face level customer exposure to benzene and
*1984 Federal Register Topic.
9-25
-------
other hydrocarbons during self-service filling, and found the average
total hydrocarbon concentration to be similar to that used in the EPA
evaluation. However, benzene concentrations were measured at U.U4 to
0.15 ppm, and averaged U.U6 ppm (compared to EPA's figures of 0.96 ppm
for unleaded and 1.46 ppm for leaded gasoline) (I-H-24).
Two other commenters stated that the API/Clayton self-service
exposure concentrations are most likely unrepresentative in terms of
both population distribution and seasonal variation. In addition,
these commenters felt that the concentrations were questionable because
the benzene-to-gasoline vapor ratio seemed too low and because the ben-
zene vapor content was found to be higher for leaded gasoline than for
unleaded gasoline (I-H-114, I-H-127). One of the commenters concluded
that the results of the API/Clayton study are not relevant for deter-
mining health effects from refueling vapors (I-H-127). The commenters
further concluded that the uncertainties cast doubt on whether self-
service is dominant in nationwide risk.
One ccmmenter requested further information about the methodology
used in the API exposure study, stating that ambient and tank tempera-
tures, wind speed and direction, humidity, position of sampler, fill
rate, and sampling pump flow rate can all affect air concentration
measurements (I-H-72). Another said that there is some question as
to the validity of the API/Clayton exposure data, with regard to the
sample flow rates used, the statistical manipulation of data, and the
high benzene concentrations found in the gasoline bulk samples
(I-D-bl).
One commenter stated that the variability in the actual measured
exposure data could not be determined from the evaluation document,
remarking that use of the maximum observed vapor concentration may over-
state the exposure, while average observed concentrations may understate
the same exposure (I-H-126).
Response: The Clayton work is, in EPA's assessment, the most
comprehensive refueling study available. In addition, the results of
the Clayton study are basically similar to the results of other studies.
It is true that the mean results of studies assessing time-weighted
averages (TWA) of service station employees vary from 0.08 to O.bb ppm
for benzene, with a total 'range of individual measurements'of <0.oi to
9-26
-------
2.08 ppm (I-F-13, 1-0-17). The only study assessing total gasoline
vapors reported a mean of 8.3 ppm and a range of 0.42 to 114 ppm (I-F-13).
These results for time-weighted averages of occupational exposure are
in general agreement with those reported by the commenter (I-H-24).
The Clayton study, however, assessed exposure concentrations during
refueling time only, which is the relevant concentration for self-
service refueling. A statistical analysis was performed of the station
geometric means of concentrations by volume, calculated from the Clayton
concentrations by mass and station temperatures. The results are as
follows:
BENZENE
Unleaded Regular U.98
Leaded Regular 1.4U
Premium Unleaded U.82
All Unfeaded 0.96
(Ann. Avg.)
All Gasoline 1.03
(Ann. Avg.)
GASOLINE VAPORS
Lower 9b%
Number of Confidence
Stations Minimum Maximum Limit
13
12
8
NA
NA
0.41
0.54
0.10
0.38
0.40
1.79
4.11
1.61
1.77
2.1b
0.70
0.78
0.32
0.65
0.67
Upper 9b%
Confidence
Limit
1.26
2.02
1.32
1.26
1.39
Unleaded Regular
Leaded Regular
Premium Unleaded
All Unleaded
(Ann. Avg.)
Al 1 Gasoline
(Ann. '-U'-j.)
61.1
76.2
69.2
62.0
64.3
13
12
9
NA
NA
19.1
18.0
26. 4
19.8
19. b
126
174
IbO
129
136
41.8
43.0
40.9
41.6
41.9
80.4
109.4
97.6
82.4
86.8
As can be seen, the annual' average values for benzene from all gasoline
is 1.0 ppm with 9b percent confidence limits of 0.67 and 1.39 ppm. The
range of station geometric means was 0.40 to 2.15 ppm weighted for all
gasoline types, or 0.10 to 4.11 ppm for the most extreme values for any
fuel type. Other studies reported values for refueling time only with
means of 0.1 to 1.2 ppm and a total range of <0.0l to 3.2 ppm (I-F-13,
9-27
-------
I-F-141). As shown above, the Clayton-based annual average values for
yasoline vapors from all gasoline is 64.3 ppm with 95 percent confidence
limits of 42 and 82 ppm. The range of station geometric means was 21)
to 136 ppm weighted for all gasoline types, or 18 to 174 ppm for the
most extreme values for any fuel type. The one other available study
reported values for refueling time only with a mean of 41.7 ppm and a
range of 1.8 to yy ppm.
The EPA's Environmental Monitoring Systems Lab (EMSL) also con-
ducted a brief refueling exposure study (I-A-67). The results of this
study agreed well with the data used as a basis for the risk analysis.
y-28
-------
9.4 REFERENCES (Comment letters are not rep.eated here. See Chapter 1,
Table 1-1, for a complete list of comment letters.)
I-A-9 Study of Gasoline Vapor Emission Controls at Small Bulk Plants.
U.S. EPA, Region VIII. Prepared by Pacific Environmental
Services, Inc. October 1976.
I-A-34 Bulk Gasoline Terminals - Background Information for Proposed
Standards. U.S. EPA. Research Trianyle Park, NC. EPA-45U/3-
8U-U38a. December 1980.
I-A-67 Self-Service Station Vehicle Refueliny Exposure Study.
Environmental Monitoring Systems Laboratory, U.S. EPA. Research
Triangle Park, NC. Undated.
I-D-17 Letter and enclosure from O'Keefe, W.F., American Petroleum
Institute, to Newburg-Rin, S., U.S. EPA, Office of Toxic
Substances. September 20, 1983. Transmittal of final report:
"Gasoline Exposure Study for the American Petroleum Institute
(API), Washington, D.C.," prepared by Clayton Environmental
Consultants, Inc., dated August 2b, 1983.
I-D-38 Letter and enclosures from Crockett, P.E., American Petroleum
Institute, to Gray, C.L., U.S. EPA. August 8, 1984. Information
on gasoline dispensing rates.
I-F-13 Quest for a Gasoline TLV. McDermott, H.J., et al., Shell Uil
Company, San Ramon, California. American Industrial Hygiene
Association Journal (39)-.110-117. February 1978.
I-F-103 Health Effects of Benzene. Part B. State of California
Department of Health Services/Epidemiological Studies Section.
July 1984.
I-F-141 "Service Station Attendant's Exposure to Benzene and Gasoline
Vapors." American Industrial Hygiene Association Journal
(4U): 315-321. April 1979. H.J. McDermott and G.A. Vos,
Shell Oil Company, San Ramon, California.
9-29
-------
1U.U OTHER METHODOLOGIES AND CONSIDERATIONS
10.1 LEGAL AND POLICY CONSIDERATIONS
Comment: One commenter felt that EPA should adopt Stage I controls
for all sources that currently do not have these controls (including
bulk plants, terminals, and service stations). Further, the Agency
should require achievement of the lowest achievable emission level,
rather than RACT limits, which are considerably higher than the limits
that can be reached with the same technology using good operational and
maintenance practices. An example of this difference was provided by
the commenter for bulk terminals, where the RACT limit for vapor pro-
cessors is 80 mg/liter, and the LAER limit is 30 mg/liter (I-H-115).
Response; The Agency is currently considering the role of Stage I
controls in the gasoline marketing regulatory program. No final deci-
sions have been reached on the application and stringency of such
controls.
Comment: One commenter felt that the comprehensiveness of EPA's
July 1984 analysis makes for an unwieldy document in which important
issues are lost in trivia. For example, since Stage I was shown to
have a negligible effect on annual incidences of cancer and all ozone
nonattainment areas should have adopted Stage I, they felt there was
little sense in carrying on with the analysis of Stage I on a nation-
wide basis (I-H-21).
Response: The Agency considers it important to include an evaluation
of Stage I impacts because Stage I is part of the gasoline marketing
chain. 'Moreover, Stage I annual incidence estimates in the analysis
provide a basis for comparison of the relative impacts of the several
exposure pathways and sources.
Comment: One commenter felt that attention should be given to the
strategies that will meet established needs, such as: nationwide
Stage I control to remove more than half of the gasoline vapors at low
cost, Stage II control in areas that must attain ozone air quality
standards by the 1987 deadline, and reduction of the vapor pressure of
gasoline to restore current evaporative controls to a better working
order (I-H-101).
10-1
-------
Response: The question of whether to require adoption of Staye I
controls beyond nonattainment areas, where it is already required, is
still being considered by the Agency. The rulemaking actions taken by
the Agency include a proposal to reduce in-use fuel volatility, as the
commenter suggests. As described in the proposal for onboard control
of vehicle refueling, the Agency currently does not intend to require
Stage II controls in areas where it is not presently installed (or being
installed) or where the applicable SIP does not contain a commitment to
install Stage II. However, it is a potential measure for States to
consider in developing revised nonattainment plans (see the onboard
Federal Register announcement for further discussion of Stage II issues).
Comment: One commenter felt that the July 1984 EPA analysis repre-
sents an important departure from custom at EPA, in that health effects
and control alternatives are analyzed simultaneously. In the past,
they noted that EPA has established that there is a problem and then
independently endeavored to solve it. They felt the evaluation estab-
lishes that there are essentially two problems, i.e., ozone nonattain-
ment and cancer incidences, that may not have the same solution. They
felt the effort to simultaneously analyze health effects and control
alternatives fails to assign a monetary value to reductions in morbi-
dity or mortality associated with the control strategies. Furthermore,
they noted that the-precedent is hereby established for the selection
of control alternatives based on "dollars per cancer incidence prevented."
They felt Congress clearly did not intend such compromises when it
promulgated the Clean Air Act (I-H-21).
Response:- -The EPA does not agree with the commenter1 s view that
it is Agency policy to "gnore potential solutions when analyzing
environmental problems. The Agency assesses possible courses of action
in its analytical work; to do otherwise would be irresponsible. The
gasoline marketing analysis on which comment was sought presented the
regulatory alternatives in some detail for the express purpose of
stimulating comment from the public. Concerning the nature of the
problem, the commenter suggests that there are two aspectsozone non-
attainment and hazardous exposure. In fact, there are more, including
transport of ozone and ozone precursors from attainment to nonattainment
1U-2
-------
areas, maintenance of air quality in ozone attainment areas, and
potential benefits derived from reducing ozone levels in attainment
areas. As explained in the Federal Register proposal, EPA believes that
vehicle onboard controls is the solution that, overall, deals best with
all of these problems.
It is true that a specific monetary value for reduction in morbidity
was not included in the analysis (except for the implied benefit for
reducing VOC in nonattainment areas, which is primarily concerned with
morbidity effects). The ability to monetize morbidity effects (as
related to levels of exposure) for pollutants such as benzene and gas
vapors is limited, because such linkages are not well defined. To the
extent that such effects (and benefits) exist, the analysis understates
the advantages of controlling refueling emissions.
Comment: Several commenters questioned the interpretation of
Sections 112 and 324 of the Clean Air Act applied in the analysis. Two
commenters examined Sections 112 and 324 and the relevant legislative
history, and believe it is clear that Congress intended to allow the
control of a hazardous air pollutant to take precedence over the exemp-
tion for "independent small gasoline marketers" (I-H-114, I-H-127).
Another commenter felt the legislative history and a legal construction
of Sections 324 and 112 indicate that Section 324 applies to any regu-
lations or statutory authority that require the installation of vapor
recovery equipment. Therefore, a reasonable interpretation would
conclude that the exemption of small marketers allowed in Section 324
is appropriate with regard to any refueling strategies mandated under
Section. 112 .(L-H.-1U2). One commenter pointed out that neither Section
324 of the Act noi anything in its legislative history affirmatively
requires EPA-to regulate nonindependent stations below 50,DUO gallons
per month and that Congress was not trying to give independent stations
a competitive advantage, but was seeking to protect them from economi-
cally infeasible capital expenditures (I-H-94).
Response: Section 112 of the Act authorizes the Administrator to
set emission standards for hazardous air pollutants. The strategies . .
evaluation was conducted, in part, to examine the feasibility of setting
Section 112 standards requiring vapor recovery on gasoline refueling
operations.
10-3
-------
Sectic is 323 and 324 provide:
§323 ... (a) The regulations under this chapter
applicable to vapor recovery with respect to mobile
source fuels at retail outlets of such fuels shall
provide that the cost of procurement and installation
of such vapor recovery shall be borne by the owner
of such outlet ... .
§324 ... (a) The regulations under this chapter
applicable to vapor recovery from fueling of motor
vehicles at retail outlets of gasoline shall not
apply to any outlet owned by an independent small
business marketer of gasoline having monthly sales
of less than 50,UUU gallons ... .
(b) Nothing in subjection (a) shall be
construed to prohibit any State from adopting or
enforcing, with respec.t to independent small business
marketers of gasoline having monthly sales of less
than 50,UUU gallons, any vapor recovery requirements
for mobile source fuels at rental outlets. . . .
Any vapor recovery requirement which is adopted by
a State and submitted to the Administrator as part of
its implementation plan may be approved and enforced
by the Administrator as part of the applicable
implementation plan for that State ....
The Conference Report on these provisions states:
Under the conference agreement, no station
v?hich is owned by an independent marketer and which
has monthly throughput of less than 5U,UUU gallons
of gasoline may be required, directly or indirectly,
by the Administrator under this Act to install and
use vapor recovery equipment.
H.K. Rep. No. 564, 95th Cong., 1st Sess. 182 (1977).
The House Report explains the origin of the provisions:
This approach [vapor recovery] is regarded as a
necessary, but not sufficient, strategy in many areas
for attaining the national primary oxidant standard.
However, one major problem has become evident
in the implementation of these [vapor recovery]
controls. That problem[ ] is the capital costs of
control which currently must be borne by the owner
or operator of the station.
H.R. Rep. No. 294, 95th Cong., 1st Sess. 299 (1977). Rep. Whalen
had introduced Section 324 as a floor amendment with these words:
1U-4
-------
It would be a serious blow to our economy it
the price we had to pay for applying the Staye II
L vapor recovery] regulations to independent marketers
were the loss of, or serious damage to, the inde-
pendent segment of the yasoline marketing industry.
Conyressional Research Service, A Leyislative History of the Clean Air
Act Amendments of 1977 at 6483 (1978).
Absent persuasive reasons to the contrary, statutes are inter-
preted according to their plain meaning. E.g., CPSC v. UTE Sylvania,
447 U.S. 102 (198U). Sections 323 and 324 on their face apply to any
regulations under the Clean Air Act includiny standards promulgated
under Section 112. Their broad application is also supported by the
legislative history; the Conference Report states that they forbid that
small gas stations "... be required directly or indirectly, by the
Administrator under this Act to install and use vapor recovery equipment."
(emphasis added).
It is true that Congress did not specifically address the applica-
tion of Sections 323 and 324 to Section 112 standards, but "the
absence of congressional focus is immaterial wnere the plain lanyuaye
applies." Jefferson City Pharmaceutical Ass'n v. Abbot Laboratories,
46U U.S. IbU, 1U3 S.Ct. lull, 1U17 n. 18 (1983). "[I]t is no bar to
interpreting a statute as applicable that the question which is raised
ori the statute never occurred to the legislature." Eastern Air Lines,
Inc. v. C.A.B., 364 F.2d 5U7, 511 (D.C. Cir. 1966), citing Cardozo,
The Nature of the Judicial Process, 16 (1921). Accord, Montana Power
Co. v. F.P.C., 446 F.2d 739, 746 (U.C. Cir. 197U), cert den. 4UU U.S.
1U13 (1971); Portland Cement Ass'n v. Ruckelshaus, 48b F.2d 376, 38U-
(U.C. Cir. 1973), cert. den. 417 U.S. 921 (1974).
Moreover," statutes should be interpreted consistent with Congress'
intent. E.g., U.S. v. Braverman, 373 U.S. 4U6 (1963). Congress' clear .
intent in Sections 323 and 324 was to relieve at least some gasoline
retailers of certain economic burdens which might be imposed by EPA
regulations. Construing these sections to apply to Section 112 standards
is consistent with that intent.
Comment: Une commenter expressed concern over the omission in
the EPA report of the statement by Vice President Bush in April 1981,
that onboard control of refueling emissions would not be imposed on the
automotive industry (I-H-117).
lu-6
-------
Response: The Vice-President's statement of April 1981 and EPA's
own public announcement of its decision (I-G-7), were made at a time
period when the domestic automobile manufacturiny industry was experiencing
massive layoffs and corporate losses. Since that time, the economic
health of this industry has improved considerably. In light of that
and more relevant information on the potential need for refueling controls,
the Agency believed that an up-to-date evaluation of the potential
impacts and desirability of onboard controls was warranted.
Comment: One commenter stated that EPA has a legal requirement to
make three decisions concerning emissions of hydrocarbons in vehicle
refueling: (1) Is regulation needed for direct protection of public
health? (2) Is regulation necessary and cost effective for the reduc-
tion of atmospheric oxidants in nonattainment areas? and (3) If regula-
tion for either reason is needed, should it be accomplished by control
onboard the vehicle or at service stations (I-H-99)? This commenter
also stated that EPA must decide whether vehicles in-use are complying
adequately with the evaporative emission standard and, if not, how best
to assure the adequate compliance of these and future vehicles. Further-
more, this commenter stated that it is essential that the option chosen
be one that will achieve the desired result at a minimum cost to the
motoring public and, thus, EPA must avoid redundant actions. Specifi-
cally, the commenter felt EPA must avoid requiring both onboard and
service station controls, both RVP restrictions and service station
controls, or both RVP restrictions and onboard controls. For the
public good, EPA must identify the one most cost-effective control
option (I-H-99-).-
Response: The Agency clearly is not limited to identifying "the
one most cost-effective control option." Based on an extensive analysis
of a broad range of options, the Agency has proposed to regulate refuel-
ing emissions with onboard controls and to limit the volatility of in-
use motor fuel. The rationale for these actions is thoroughly explained
in the rulemaking proposals and accompanying support documents.
Comment: One commenter felt that the Clean Air Act does not permit
EPA to reject available measures for controlling carcinogenic pollutants
on the basis of cost considerations (I-H-115).
10-6
-------
Response: Neitner Section 112; nor Its legislative history states
that costs may not be considered. On the contrary, the legislative
history indicates that cost factors may be taken into account.
Section 112 was enacted in 1970. All of the bills considered by
Congress provided that EPA could consider feasibility in setting stan-
dards for hazardous air pollutants. For example, the Senate Report
notes:
The committee recognizes that some of these
hazardous pollutants . . . are present in nearly
all raw materials ... . Recognizing that complete
control . . , may not be ... practicable, the
Committee has provided the Secretary with authority
to differentiate amony categories ... .
S. Rep. No. 1196, 91st Cong., 2d Sess. 2U (197u), reprinted in
Congressional Research Service, A Legislative History of theCleanAir
Act Amendments of 197U at 42U.
There are sound reasons for interpreting Section 112 as permitting
consideration of cost and feasibility when carcinogens are concerned.
As EPA has consistently observed In Section 112 rulemakings, there is
no direct evidence that air pollutants cause cancer. That is, there
are no studies showing an increased incidence of cancer in humans due
to exposure to particular pollutants in the ambient air. Instead,
regulation of airborne carcinogens is based on studies of workers
exposed in the workplace and animals exposed in laboratory experiments.
However, the workplace and laboratory exposures are much higher, gen-
erally many orders of magnitude higher, than the levels of these carci-
nogens found in the ambient air. It is only by extrapolation from the
higher exposures to ambient levels that the Agency concludes that the
air pollution due to these carcinogens threatens public health within
the meaning of Section 112. This extrapolation is based on the "no
threshold" assumption, i.e., that any exposure to a carcinogen presents
some risk, although the risk becomes vanishingly small as the exposure
does. Under the no threshold assumption, emission standards for carci-
nogens under Section 112 could prevent all risks only by preventing all
emissions and, hence, all exposures.
But EPA does not believe that Congress intended Section 112 stan-
dards to prevent all risks. Section 112(b)(l)(B) simply requires that
standards be "at the level which in [EPA's] judgment provides an ample
10-7
-------
maryin of safety to protect the public health from such a hazardous air
pollutant." First, the requirement for an ample maryin of "safety"
does not imply eliminatiny all risk. For example, the Occupational
Safety and Health Act requires standards that "provide safe or healthful
employment and places of employment." As the Supreme Court held in
reviewiny an OSHA benzene standard, "... safe is not the equivalent
of 'risk-free.'" Industrial Union Dept., AFL-CIU v. American Petro-
leum Institute. 448 U.S. 6U7, 642 (1980).
Second, Congress did not contemplate that Section 112 would require
the elimination of all risk from, and all emissions of, carcinogens.
Virtually every basic industry in our society - chemicals, petroleum,
electric power, metals - emits one or more of the six carcinogens listed
under Section 112 (asbestos, vinyl chloride, benzene, inorganic arsenic,
radionuclides, and coke oven emissions). In most cases, these emissions
cannot be completely eliminated. Therefore, standards prohibiting all
emissions (and eliminating all risk) would effectively shut down the
Nation's basic industry. The legislative history of Section 112 makes
clear that Congress had no such draconian results in mind. See Proposed
Policy for Airborne Carcinogens, 44 Fed. Reg. 58642, b8659 - 58661
(October 10, 1979).
In these circumstances, EPA has established Section 112 standards
for carcinogens at levels that reflect demonstrated, effective control
systems. This reduces, but does not eliminate, emissions, exposure,
and risk. It necessarily involves considering feasibility and cost.
First, a control system cannot be considered demonstrated or effective
unless it is feasible. Second, whether a system is feasible depends
in part on the reasonableness of its cost.
The EPA has consistently interpreted Section 112 to permit the
Agency to consider cost and feasibility, at least when setting standards
for carcinogens. The EPA began implementing Section 112 on March 31,
1971 (only 3 months after its enactment), when it listed 3 hazardous
air pollutants, including the carcinogen asbestos. 36 Fed. Reg. 5931.
The EPA thereupon proposed asbestos emission standards which, in every
case, were based on identified and feasible control techniques. 36
Fed. Reg. 23239 (December 7, 1971). In promulgating those rules EPA
stated that:
10-8
-------
The EPA considered the possibility of banning
production, processing, and use of asbestos or
banning all emissions for asbestos into the atmosphere,
but rejected these approaches ... . Either approach
would result in the prohibition of many activities
which are extremely important ... . For example,
demolition of any building containing asbestos fire
proofing or insulating materials would have to be
prohibited ... 38 Fed. Reg. 8820 (col. 2)
(April 6, 1973).
The EPA has continued consistently to base Section 112 standards
for carcinogens on consideration of, among other things, cost and
feasibility. Amendments to standards for asbestos, 39 Fed. Rey. 38U64
(October 25, 1974); 40 Fed. Reg. 48229 (October 14, 1975); standards
for vinyl chloride, 40 Fed. Reg. 59532 (December 24, 1975); 41 Fed. Reg.
46560 (October 21, 1976); proposed amendments to standards for vinyl
chloride, 42 Fed. Rey. 29005 (June 7, 1977); amendments to standards
Tor asbestos, 43 Fed. Reg. 26372 (June 19, 1978); standards for benzene,
46 Fed. Reg. 1165 (January 5, 1981); 49 Fed. Reg. 23498 (June 6, 1984);
proposed standards for benzene, 45 Fed. Reg. 26660 (April 18, 1980); 45
Fed. Reg. 83448 (December 18, 1980); 45 Fed. Rey. 83962 (December 19,
1980); standards for radionuclides, 48 Fed. Reg. 15076 (April 6, 1983);
49 Fed. Reg. 43906 (October 31, 1984); standards for inorganic
arsenic, 51 Fed. Reg. 27956 (August 4, 1986)
This consistent and long-standing view of the Agency charged
with implementing Section 112 is entitled to substantial deference.
E.I, du Pont de Nemours & Co. v. Collins. 432 U.S. 46, 54-55 (1977).
See Chevron USA, Inc. v. NRDC, 467 U.S. 837 (1984).
Moreover, when Congress re-enacted the Act in 1977, it was well
aware of EPA's settled interpretation of Section 112. By that time,
the standards for asbestos and vinyl chloride had been proposed,
promulgated, and litigated. Indeed, the principal amendment to Section
112 was the explicit authorization of design, equipment, work practice,
and operational standards (Section 112(e)), which were the heart of the
asbestos standards. The legislative history states that "[t]his limited
provision would fully authorize the present EPA regulations governing
asbestos." S. Rep. No. 127, 9bth Cong., 1st Sess. 44 (1977), (enacted
Amendments adopted the Senate bill without significant change, See
10-9
-------
H.K. Kep. No. b64, ybth Cony., 1st Sess. 131-132 (ly77) (Conference
Keport)). Conyress' re-enactment of Section 112 with full knowledye of
EPA's interpretation further shows that interpretation to be reasonable.
North Haven Board of Education v. Bell, 4b6 U.S. 512, 1U2 S. Ct. iyi2,
iy2b.(1982); NLKb v. Bell Aerospace Co.. 416 U.S. 267, 274-27b (1974),
and cases cited therein.
The EPA disayrees with the commenter's claim that there is a
yeneral leyal principle forbiddiny consideration of cost under any
statute usiny the phrase "margin of safety." For example, in National
Ass'n of Demolition Contractors (NADC) v. Costle, b6b F.2d 748 (D.C. Cir.
1977), the Court upheld EPA's action in settiny asbestos standards
based on feasible control systems:
NADC argues that the Administrator's statutory
mandate to protect the public health with "an ample
maryin of safety" is inconsistent with his decision
to use the "best available control methods .... We
disayree.
Protection of the public with "an ample maryin
of safety" may necessitate use of different control
measures .... 56b Fed. Key. F.2d at 7b3.
The cases cited by tne commenter, Lead Industries Ass'n v. EPA, 647
F.2d 113U (D.C. Cir. 1980), cert, den. 449 U.S. 1U.42 (198U); American
Petroleum Institute v. Costle, 66b F.2d 1176 (U.C. Cir. 1981), cert.
deru 102 S.Ct. 1737 (1982); Hercules, Inc. v. EPA, b98 F.2d 91
(U.C. Cir. 1978), interpret only two statutory provisions, Section 109
of the Clean Air Act, and Section 3U7 of the Clean Water Act. Those
cases do not set forth a yeneral principle that a statute usiny the
phrase "margin, of safety" forbids consideration of feasibility and
cost. On the contrary, those cases were based on detailed examination
of Sections Iu9 and 307, each taken as a whole, alony with their
leyislative history, Ayency interpretation, and the facts involved. In
all these respects, the cases are entirely different from the reyulation
of carcinoyens under Section 112.
_ Comment: Three commenters believed Section 202(a)(6) of the Clean
Air Act requires EPA to perform an analysis of Staye II and onboard,
and the Ayenc.y must issue onboard regulations if they are desirable and
feasible. The analysis indicates that onboard is superior to Staye II;
10-10
-------
therefore, EPA must issue onboard reyulations if it is to comply with
the statutory directive (I-D-54, I-H-108, I-H-119). A second commenter
also believed that the application of Section 202(a)(6) to the findings
contained in the Strategies Document effectively prohibits the selec-
tion of Stage II vapor recovery as a control strategy (I-H-102).
Another commenter pointed out that Section 202(a)(6) of the Act requires
that onboard be studied as a substitute for gasoline vapor recovery
(Stage II), and not as a complement to it. Thus, EPA's analysis appears
to be contrary to the intent of Congress. The commenter concluded that
no such legal questions would impede the expeditious implementation of
a Stage II program (I-H-93). One commenter pointed out that in enacting
the Clean Air Act Amendments of 1977, Congress recognized the potential
advantages of onboard controls in Section 202(a)(6), which requires EPA
to prescribe onboard controls in preference to Stage II vapor recovery
if the EPA Administrator finds employment of onboard controls to be
feasible and desirable (I-H-94). One commenter stated that, if onboard
controls are found to be the preferred strategy, the Agency should man-
date these controls under Section 202(a) to "lock in" their long-term
benefits (I-H-94).
Response: The EPA agrees that Section 202(a)(6) requires promul-
gation of onboard controls if they are ultimately determined by EPA to
be "feasible and desirable." As fully discussed in the proposed onboard
rulemaking, EPA believes, at this time, that onboard controls are "fea-
sible and desirable" and has proposed such controls. Of course, EPA
welcomes comments on its current belief as to both the feasibility and
desirability issues. If EPA finally concludes that they are feasible
and desirable, onboard controls will be required.
As to the second commenter, EPA is not sure of the exact meaning
of the somewhat vague suggestion that the "findings" in the Strategies
Document (July 1984 analysis) preclude selection of Stage II controls.
However, as discussed in the proposed onboard rule, EPA is proposing
not to require Stage II controls as a reasonably available control
strategy for all, ozone nonattainment areas under Section 172 of the Act.
Regarding the comment that Section 2U2(a)(6) allows onboard only
as a "substitute" and not as a "complement" to Stage II, EPA does not
necessarily agree with the commenter's conclusions. Although Section
10-11
-------
202(a)(6) requires EPA to determine the "feasibility and desirability"
of onboard controls which would avoid the necessity of Staye II, it
does not expressly preclude EPA from requiring both onboard and
Stage II, if appropriate, after making such a determination. In any
event, as discussed in the onboard proposal, EPA currently believes
onboard is an effective substitute for Stage II and proposes not to
require Stage II for all ozone nonattainment areas under Section 172
of the Act. (This response also applies to the third commenter dis-
cussed above.) Of course, as stated in the onboard proposal, EPA
expects States which have installed (or are installing) Stage II to
continue to implement that measure unless and until they are replaced
by onboard controls, and those States which have SIP commitments to
install Stage 11 to do so unless and until adequate substitute measures
are submitted and approved.
As to the final comment discussed above, EPA has proposed to
require onboard controls and will promulgate such controls -- thus
"locking in" their benefits if, after evaluating the public comments
on the proposal, EPA finally determines that they are feasible and
desirable.
1U.2 SUGGESTED REGULATORY APPROACHES
Comment: One commenter felt that there is no reason to require
Stage II in areas now attaining the ozone ambient standard. Implemen-
tation of Stage II systems in nonattainment areas would require that
EPA take three main actions: (1) set a VOC recovery (emission1 reduc-
tion) standard, (2) develop a procedure for certifying compliance with
the standard', "and (3) define an implementation schedule.
This commenter suggested an emission standard no lower than
0.4 gram per gallon of gasoline dispensed under average conditions. A
certification test, developed in conjunction with outside technical
consultation, should be required only on a prototype system, and not at
every service station utilizing the system. With regard to the schedule
for implementation, the commenter felt all marketers should be required
to meet the same installation schedule. Further, exemptions given'to
small or independent marketers would often be unjustified because many
small facilities (non-retail trucking, automotive, or government) are
10-12
-------
financially able to afford controls, and a facility's, emissions contri-
bution does not depend on whether it is run by a major or an independent
marketer. The commenter expressed strony opposition to exemptions
based on company affiliation (I-H-99).
Response: As explained in the proposed rulemakiny, EPA does not
intend to require adoption of Staye II in areas where it is not now in
place or beiny installed or contained in a SIP commitment. However,
States with SIP commitments to adopt Stage II may avoid that requirement
if they submit adequate substitute measures that are approved by EPA.
Staye II is a viable option for other States that need to develop
strategies that will provide sufficient VUC reductions to attain the
ozone standard expeditiously. The States that choose to adopt Stage II
are free to set efficiency standards, exemption levels, and schedules
that meet their VOC reduction needs, based on case-by-case analysis of
local conditions. Such specifications would be subject to review and
approval by EPA under the SIP processing procedures.
Comment: Several commenters provided recommendations for further
studies. Two commenters suggested the field testing of new equipment
before mandating its use (I-H-42, I-H-43). One commenter suggested
that a comprehensive study be undertaken by a neutral party to determine
actual i n-use efficiency and cost effectiveness (I-H-82). Another
commenter urged EPA to research vapor recovery more fully and-, working
together with the petroleum industry, to come up with a solution that
is acceptable to all parties (I-H-63). One commenter felt EPA should
consider other strategies that combine onboard or Stage II .with controls
on commercial and certification fuels (I-H-114).
Response: As explained in the onboard proposal, EPA believes that
there are no problems with the onboard concept that cannot be resolved
within the lead-time provi-ded. The principles of vapor recovery are
well understood, and onboard is similar to the evaporative control systems
that have been on cars for years. Therefore, no long-term testing
program, nor a "comprehensive study ... by a neutral third party" is
necessary.
As noted earlier, the Agency has evaluated combining control of
refueling and controls on commercial and certification fuel volatility,
and proposes to regulate both VOC emissions problems.
10-13
-------
Comment: A number of commenters made recommendations reyardiny
possible standards. Two commenters expressed concern that the great
complexities of VOC emission control may lead to regulations that are
either unnecessary (as in the case of areas where there is no threat to
the national ambient ozone standard) or so complicated as to be diffi-
cult or impossible to implement and enforce (I-H-42, I-H-43). One
commenter suggested that EPA proceed with nationwide implementation of
Stage I controls because the findings suggest a high risk of cancer
through occupational exposure by gasoline handlers (I-H-6b). Un the
other hand, another commenter acknowledged that, while Stage I controls
.are relatively cost effective in reducing VUC emissions, a nationwide
strategy to retrofit Stage I controls would not be cost effective in
terms of environmental benefit. (Stage I controls are already largely
in place in those areas where the NAAQS is not being attained or in
question, and a nationwide program would not be cost effective in
attaining national ambient standards or reducing human exposure to gaso-
line vapors.) (I-H-83).
One commenter stated that they continue to support Stage I vapor
recovery as the most cost effective measure for controlling VOC
emissions from fuel dispensing facilities (I-H-94). One commenter
stated that the installation of additional Stage I equipment is not
justified because equipment to control hydrocarbon vapors emitted
during loading of truck transports at gasoline terminals and during
gasoline deliveries to service-stations already is installed in most
areas not in attainment with the atmospheric oxidant standard (I-H-99).
Response: As pointed out in Section 10.1, the Agency has not made
a determination as to the need to extend coverage of Stage I controls.
As some commenters noted, Stage I is presently required in ozone non-
attainment areas and is a cost-effective VOC control strategy. The
Agency is presently evaluating whether also to require Stage I controls
on gasoline marketing facilities in attainment areas, and will announce
any proposed decision in the Federal Register.
Comment: One commenter felt the only quick, effective method of
addressing a human health hazard from refueling vapors, if one is found
to exist, is to equip all service stations with Stage II controls,
remove benzene or other hazardous chemicals from commercial gasoline,
10-14
-------
and reduce the RVP of gasoline to levels that can be more readily
handled in the marketing stages and by the vehicle (I-H-1U1).
Response: As described in the proposed rulemaking, the Agency has
decided that onboard control is the most effective solution overall
to the refueling emissions problem. The Agency has examined the option
of removing benzene from gasoline (I-A-16), but that option is very
expensive compared to other alternatives, and would not eliminate
benzene tailpipe emissions completely, since benzene is also formed in
the combustion process. The Agency has, however, proposed to reduce the
RVP levels of commercial gasoline, as suggested by the commenter.
Comment: Une commenter believed that more stringent limitations
for existing bulk gasoline terminals are obtainable and cost effective.
An emission rate of 35 milligrams per liter is obtainable through the
use of "polishing" units. This commenter suggested that EPA recommend
a 3b mg/liter limit in a revised Control Techniques Guideline (CTG) for
existing bulk terminals (I-H-118).
Response: The EPA evaluated add-on or "polishing" units while
developing a new source performance standard for bulk terminals. Add-
on control systems were found to be unreasonably expensive in this
earlier cost analysis [Appendix B in Bulk Gasoline Terminals - Background
Information for Promulgated Standards (I-A-44)].
Comment: Une commenter pointed out that EPA is considering three
facets of the need to control refueling vapors: very uncertain health
effects, the known nonattainment of ozone standards, and an evaporative
issue related to commercial fuel characteristics. The commenter stated
that EPA should be most interested in those solutions that directly
attack the known problems, under the conditions of lowest uncertainty.
The commenter stated that if EPA feels compelled to control refueling
vapors, then the solution is Stage II control in nonattainment areas
and control of the volatility of commercial fuel. The commenter felt
that this answer deals most directly with the known, understood issues,
while avoiding the controversy over uncertain issues (I-H-1UU).
Another commenter strongly urged the Agency to limit any vapor
recovery regulation to the "selected nonattainment areas" as discussed
in the July 1984 Analysis Report, Chapter 4 and Tab.le 7-22, which
1U-15
-------
emphasizes the dramatic cost effectiveness advantage of so limiting the
requirement (I-H-94).
Respo'nse: The Agency has not taken the narrow views espoused by
the commenters, but rather has examined the full range of known and
potential problems associated with the refueling issue and the potentia-l
benefits to be realized from control. This examination included a
careful consideration of the uncertainties involved. As a result, EPA
proposes control of refueling emissions with onboard technology, and
proposes to reduce the volatility of commercial gasoline. The basis
for these decisions, including reduction of ozone levels in nonattain-
ment areas, reduction in VUC/ozone transport, maintenance of acceptable
ozone air quality levels, and reduction in nationwide cancer risks, is
thoroughly explained in the applicable Federal Register notices.
Comment: One commenter suggested that EPA establish a national
program using onboard controls as soon as possible rather than continue
with a piecemeal program based on Stage II controls. The commenter
claimed that 53 percent of all gasoline sold would be exempt from Stage
II controls since it would be sold at stations selling less than 50,000
gallons per month, which are exempted by the Clean Air Act (I-H-130).
Response: The rationale for adopting onboard controls is thoroughly
described in the proposed rulemaking. The following addresses an
apparent misinterpretation by the commenter. The EPA analysis shows
that 54 percent of all service stations (both independents and noninde-
pendents) have a throughput of less than 50,000 gallons/month; however,
they are not all exempted by the Clean Air Act. Only independent
service, stat.io.ns. with a throughput of less than 50,000 gallons/month
are exempted under Section 324 of the Act. The analysis indicates that
approximately 80 percent- of independent stations have a throughput of
less than 5U,UOO gallons/month and would, therefore, be exempt from
Stage II controls. Section 324 does not provide for specific exemptions
of non-independent service stations.
Comment: One commenter strongly urged the Agency to limit any
vapor recovery regulation to the "selected nonattainment areas" as
discussed in the July 1984 analysis, Chapter 4 and Table 7-22, which
emphasizes the dramatic cost effectiveness advantage of so limiting the
requirement (I-H-94).
1U-16
-------
Response: The nonattainment areas described in the tables in
Chapter 4, Table 1-22, and in the revised analysis are only examples
constructed so that the impacts of controls on various groupings of
nonattainment areas could be examined. The purpose of these groupings
was not to select a particular grouping as the target for refueling
controls.
Comment: One commenter stated that EPA should not require implemen-
tation of a dual program. This commenter stated that Stage II vapor
recovery at fuel dispensing facilities will not provide sufficient VOC
emission reductions to warrant use as an interim measure during the
gradual phase-in of enhanced onboard controls through vehicle fleet
turnover. The commenter stated that the total cost of a control program
to reduce VOC emissions from motor vehicle fueling using a dual program
approach would be unreasonable when compared to the costs and benefits
of implementing either of the two program options independently. This
commenter stated that the cost effectiveness of Stage II will be
significantly less if considered an interim measure since the full cost
of capital investments must be allocated to the reduced VOC emission
reductions realized prior to the obsolescence of the control program.
The commenter pointed out that to avoid undue economic hardship, the
Clean Air Act provides for control requirements for Stage II to be
phased in over at least a 3-year period and provides exemptions for
small volume marketers. The commenter felt that this would delay VOC
emission reduction benefits and would reduce the effectiveness of
Stage II controls as an interim measure during enhanced onboard control
phase-in (I-H--92-).
Response: The Agency, as noted, has decided not to propose a
dual program of onboard with Stage II as an interim measure. However,
States with SIP commitments will be required to fulfill those commitments,
either by carrying out Stage II or developing acceptable substitute
measures consistent with VOC reduction needs. As the commenter points
out, the efficacy and reasonableness of Stage II as an interim measure
depend in substantial part on how long it will take to complete install-
ation of the system in a given area.
10-17
-------
Comment: One commenter felt that there is no leyal basis in the
relevant sections of the Clean Air Act for the "significant risk" hurdle
that EPA has purportedly asserted must be cleared before regulatory action
may be taken against cancer-causing pollutants. In the commenter's
view, EPA has amassed sufficient evidence of a serious public health
hazard that it must act now. This commenter stated that the risks to
the individual and the level of incidence pass any conceivable
"significance" hurdles EPA might wish to fashion and, accordingly, EPA
must establish control requirements for all source categories in the
gasoline marketing system (I-H-115).
Response: This commenter argues that the legislative history of
Section 112 does not support EPA's "significant risk" test. The EPA
disagrees.
The definition of "hazardous air pollutant" was amended in 1977.
P.L. 95-95, 91 Stat. 791 §401(c). The legislative history of the
amendment shows that it permits EPA to use a "significant risk" test.
The purpose of the amendment was to codify the approach to health risks
in Ethyl Corp. v. EPA, 541 F.2d 1 (D.C. Cir. 1976), cert, denied 426
U.S. 941 (1976). As the Committee Report states:
In summary, the committee's action is intended to
support the views expressed in ... the Ethyl
case . . . The committee's bill would also apply
this interpretation to all other sections of the
act relating to public health protection [including
section 112].
H.R. Rep. No. 95-294, 95th Cong., 1st Sess. 49 (1977) (emphasis added).
The House Report reflects Congress' intent. The Conference Committee
adopted the House bill without any relevant change or comment; there
was no comparable provision in the Senate bill. H.R. Rep. No. 95-564,
95th Cong., 1st Sess. 183-184 (1977) (Conference Report).
The Ethyl case plainly states that a finding of "significant risk"
is an appropriate test for regulating:
The Administrator . . . interpreted] "will
endanger" to mean "presents a significant risk of harm."
541 F. 2d at 13;
. . . the threatened harm must be sufficiently
significant to justify health-based regulation of
national impact.
jd_. at 18 n. 32;
1U-18
-------
... a "danger" . . . can be regulated when the
harm to be avoided is widespread lead poisoning and
the risk of that occurrence is "significant."
id. at 20;
. . . regulation was justified because the aggregate
was dangerous, and because leaded gasoline was a
significant source that was particularly suited to
ready reduction.
id. at 30 (emphasis in original);
[U]nder the cumulative impact theory emissions must
make more than a minimal contribution to total expo-
sure in order to justify regulation ... We accept
the Administrator's determination that the contribution
must be "significant" before regulation is proper.
id. at 31 n. 62;
Ue believe the Administrator may regulate . . .
when he determines, based on his assessment of the
risks ... as guided by the policy judgment inherent
in the statute, that [the] emissions . . . cause a
significant risk of harm to the public health.
id. at 31-32 (Footnotes deleted throughout).
Moreover, the Committee Report makes clear that Congress was specifically
adopting the significant risk test upheld in Ethyl :
In Ethyl . . . [the court] was called upon to review
regulations promulgated by the Administrator . . .
to protect the public health from what he found was
"a significant risk."
H.R. Rep. No. yb-294, Supra, at 43;
After reviewing in detail for 2b pages the rationale
for the Administrator's judgment that lead in
gasoline did present a significant risk of harm,
the Court concluded that the Administrator had
"handled an extraordinarily complicated problem
with great care and candor."
jd_. at 47;
The committee's purposes . . . may be summarized
as follows: ... to authorize the Administrator
to weigh risks- . . .
id_. at 49;
In upholding . . . Ethyl, the committee is moving
in a direction which is consistent with most judicial
interpretations of the act. Most other courts have
held'that a substantial element of judgment, including
making comparative assessment of risks, . . . are
necessary and permissible under the act . . .
jd_. at 5U;
1U-19
-------
. . . [T]he committee language Is intended to emphasize
the necessarily judgmental element In the task of
predicting future health risks of present action
and to confer on the Administrator the requisite
authority to exercise such judgement.
id. at 51.
10-20
-------
1U.3 REFERENCES (Comment letters are not repeated here. See Chapter 1,
Table 1-1, for a complete list of comment letters.}
I-A-16 Cost of Benzene deduction in Gasoline to the Petroleum Refininy
Industry. U.S. EPA. Research Trianyle Park, N.C. EPA-45U/3-78-
021. April 1978.
I-A-44 Bulk Gasoline Terminals - Background Information for Promulgated
Standards. U.S. EPA. Research Trianyle Park, N.C. EPA-45U/
3-8U-U38D. August 1983.
I-G-7 Control of Air Pollution from New Motor Vehicles and New Motor
Vehicle Engines: Certification and Test Procedures. U.S. EPA.
46 FR 21628. April 13, 1981.
1U-21
------- |