United States Office of Air and Radiation EPA-450/3-84-012b
Environmental Protection (ANR-443) July 1984
A9ency Washington, DC 20460
Air
Evaluation of
Air Pollution
Regulatory
Strategies for
Gasoline
Marketing
Industry
Executive Summary
-------
EPA 450/3-84-012b
Evaluation of Air Pollution
Regulatory Stategies for
Gasoline Marketing Industry
Executive Summary
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 1984
-------
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
Title Page
1.0 EXECUTIVE SUMMARY 1-1
1.1 Operations, Emissions and Control Technology. ... 1-2
1.1.1 Bulk Terminals 1-2
1.1.2 Bulk Plants 1-4
1.1.3 Tank Trucks 1-5
1.1.4 Service Stations 1-5
1.2 Analyses of Regulatory Strategies 1-7
1.2.1 Regulatory Strategies, Model Plants, and
Projections 1-7
1.2.2 Air Pollution Emissions, Health-Risk, and
Control Cost Analyses 1-12
1.3 Results of Regulatory Strategy Analyses 1-17
1.3.1 Nonattairment Area Strategy Results .... 1-20
1.3.2 Nationwide Strategy Results 1-22
1.3.3 Cost Per Incidence Reduction 1-29
111
-------
LIST OF TABLES
Title Page
1-1 Gasoline Marketing Regulatory Strategies 1-8
1-2 Major Analytical Considerations 1-11
1-3 Unit Risk Factor Summary 1-14
l-4a Estimated Risks from Gasoline Marketing Source Categories
(Using Plausible Upper Limit Unit Risk Factor for Gasoline
Vapors) 1-18
l-4b Estimated Risks from Gasoline Marketing Source Categories
(Using Maximum Likelihood Estimate Unit Risk Factor for
Gasoline Vapors) 1-19
1-5 Vehicle Refueling Controls in Nonattainment Areas 1-21
1-6 Control of Benzene from Gasoline Marketing Sources. ..... 1-23
1-7 Summary of Theoretical Impacts for Selected NESHAP
Regulatory Strategies 1-25
1-8 Impacts Based on "In-Use" Effectiveness for Stage II Controls
and Onboard Controls (1986-2020) 1-26
1-9 Economic Impact of Regulatory Strategies Based on Theoretical
Efficiencies 1-30
1-10 Economic Considerations 1-31
1-11 Estimated Regulatory Costs and VOC Benefits 1-32
1-12 Benzene Regulatory Costs and Incidence Reduced 1-34
1-13 Benzene Regulatory Costs Per Cancer Incidence Avoided
(Assuming VOC Benefits) 1-33
1-14 Benzene and Gasoline Vapors Costs Per Cancer Incidence
Avoided (Using Rat Data Unit Risk Factor for Gas Vapors) . . . 1-35
-------
LIST OF FIGURES
Title
Page
1-1 Gasoline Marketing in the U.S. (1982 Baseline
VOC Emissions) 1_3
1-2 Effect of Onboard and Stage II Controls on Benzene Incidence 1-28
-------
Preface
This report contains an executive summary of a regulatory strategies
analysis performed by the Environmental Protection Agency. Also, this
report refers the reader to other chapters and appendices that are contained
in the complete regulatory strategies document. Interested persons should
read the regulatory strategies document (EPA-450/3-84-012a) for a more
comprehensive discussion of the methodology and assumptions used in the
analysis, and the results of the analysis. The complete regulatory
strategies document can be obtained from the EPA library and NT IS (see
page ii, for address).
VI 1
-------
1.0 EXECUTIVE SUMMARY
The purpose of the study was to evaluate the air pollution
regulatory strategies available to reduce emissions from gasoline
marketing operations of benzene (Bz), ethylene dibromide (EDB), ethylene
dichloride (EDC), and gasoline vapors (GV). Gasoline vapors or
volatile organic compound (VOC) emissions contribute to ambient ozone
concentrations and, thus, in some areas contribute to a failure to
attain the national ambient air quality standard for ozone. Benzene is a
known carcinogen, which has been listed as a hazardous air pollutant
under Section 112 of the Clean Air Act and is present in varying amounts
in gasoline. In addition, EDB, EDC and gasoline vapors each have been
shown to cause cancers in laboratory animals. EDB and EDC are generally
added to leaded gasoline, but are not present in unleaded gasoline.
The following segments of the gasoline marketing industry were considered:
bulk terminals (including storage tanks and tank trucks), bulk plants
(including storage tanks and tank trucks) and service stations (both
inloading of underground storage tanks and refueling of vehicles). The
regulatory strategies examined controls on all segments of the industry,
both with and without selected size cutoffs for small facilities, as
well as controls onboard the vehicle to reduce refueling emissions.
As noted, there are still areas of the country which have not yet
attained the national ambient air quality standard (NAAQS) for ozone.
The Clean Air Act requires that all areas achieve the NAAQS by
December 31, 1987. Some States, as part of their State implementation
plans to meet the statutory requirement, are considering control of
gasoline marketing sources, especially the refueling of motor vehicles.
Thus, an analysis of gasoline marketing regulatory strategies must
address the need to attain the ozone NAAQS in selected areas. However,
the emissions from gasoline marketing sources may induce public health
risks which require control on a national basis. The analysis evaluated
regulatory strategies which address both the more limited nonattainment
issue in part of the country and the broader question of the need for a
national control program to limit potential hazardous exposure.
1-1
-------
1.1 OPERATIONS, EMISSIONS AND CONTROL TECHNOLOGY
This section briefly outlines the operations and emissions of each
source category and major associated type of facility in the gasoline
marketing industry, as well as the commonly used control techniques.
The segments of the gasoline marketing industry analyzed in this study
include all elements and facilities that move gasoline starting from
the bulk terminal to its end consumption. Gasoline produced by refiner-
ies is distributed by a complex system comprised of wholesale and
retail outlets. Figure 1-1 depicts the main elements in the marketing
network. The flow of gasoline through the marketing system is shown
from the refinery, through bulk terminals, and sometimes bulk plants,
to retail service stations or commercial or rural dispensing facilities,
primarily via pipeline and tank truck. The wholesale operations storing
and transporting gasoline including delivery and storage in a service
station underground tank are commonly called Stage I operations.
Retail-level vehicle refueling operations are commonly termed Stage II.
The baseline nationwide VOC emission estimates are also given for
the various source categories on Figure 1-1. VOC emission factors for
the individual point source operations at each source category were
estimated. Emissions at baseline were calculated based on the gasoline
throughput and current regulations for each source category in each
county in the nation. Emission estimates for the other pollutants (Bz,
EDB, EDC) were calculated using a ratio of the vapor pressures and
thus, vapor emission rates.
1.1.1 Bulk Terminals
Bulk gasoline terminals serve as the major distribution point for
the gasoline produced at refineries. Gasoline is most commonly delivered
to terminal storage tanks by pipeline with no emissions. Gasoline is
stored in large aboveground tanks and later pumped through metered
loading areas, called loading racks, and into delivery tank trucks,
which service various wholesale and retail accounts in the marketing
network.
Most tanks in gasoline service at terminals have an external or,
less commonly, an internal floating roof to prevent the loss of product
through evaporation and working losses. Floating roofs rise and fall
with the liquid level preventing formation of a large vapor space
1-2
-------
Imported
Gasoline
Barge
Pipeline
Tanker
200,000 MG/YR
222,000 i%/YR
Servlca
Station
7,000 MG/YR
Commercial,
Rural
Consumer
Imoorted
or
Domestic
Cruae
Wholesale
Distribution
Level
208,000 MG/YR
I I s Storage
= Transoort
Figure 1-1. Gasoline Marketing in the U.S
(1982 Baseline VOC Emissions)
1-3
-------
and resulting emissions. Fixed-roof tanks, which are still used for
gasoline in some areas, use pressure-vacuum (P-V) vents to control the
smaller breathing losses and may use processing equipment to control
the much greater working (filling and emptying) losses. Breathing
losses result from volume variation due to daily changes in temperature
and barometric pressure. Emptying losses occur when air drawn into the
tank during liquid removal saturates with hydrocarbon vapor and expands
beyond the vapor space of a fixed-roof tank. Filling losses occur when
the vapors in the fixed-roof tank are displaced by the incoming liquid
and forced to the atmosphere. The largest potential source of losses
from external floating-roof tanks is an improper fit between the seal
and the tank shell. Withdrawal loss from exposed wet tank walls is
another source of emissions from floating-roof tanks.
Emissions from the tank truck loading operations at terminals occur
when the product being loaded displaces the vapors in the delivery
truck tank and forces the vapors to the atmosphere. In order to con-
trol these loading emissions, the displaced vapors can be ducted to a
vapor processor such as a carbon adsorber, thermal oxidizer, or refrig-
erated condenser for recovery or destruction. The quantity of emissions
generated during loading a tank truck are dependent on the type of
loading. Splash loading from the top of the truck creates considerable
turbulence during loading and can create a vapor mist resulting in
higher emissions. Top submerged loading, which uses an extended fill
pipe, or bottom loading admit gasoline below the liquid level in the
tank and can be used to reduce turbulence and emissions (about a 60
percent reduction). The recently promulgated bulk terminal new source
performance standards (NSPS), as well as a large number of State regula-
tions, currently require the use of vapor processors and submerged
loading at bulk terminals. Most State regulations limit truck loading
emissions to 80 mg/liter transferred (equivalent to about 90 percent
reduction) and require the tank truck to be vapor tight. The NSPS is
more stringent and requires a lower emission limit of 35 mg/liter.
1.1.2 Bulk Plants
Bulk gasoline plants are secondary distribution facilities that
typically receive gasoline from bulk terminals via truck transports,
1-4
-------
store it in aboveground storage tanks, and subsequently dispense it
via smaller account trucks to local firms, businesses and service
stations. As discussed in the previous section, vapors can escape from
fixed-roof storage tanks at bulk plants due to breathing losses
even when there is no transfer activity. The majority of bulk plants
already use top or bottom submerged loading, largely in response to
State regulations. Vapor balancing is required by many State regulations,
primarily for incoming loads, but also for outgoing loads in some
instances. Vapor balancing enables the vapors from the tank being
filled to be transferred via piping to the tank being emptied. Thus, the
vapors are not forced to the atmosphere, and as a result, working losses
are greatly reduced (by 90 percent or greater).
1.1.3 Tank Trucks
Gasoline tank trucks are normally divided into compartments with a
hatchway at the top of each compartment. Loading can be accomplished by
top splash or submerged fill through the hatch, or by bottom filling.
The majority of trucks have dual capability. Either top or bottom
loading can be adapted for vapor collection. However, the trend is
toward bottom loading because of State vapor recovery regulations and
operating and safety advantages. The vapor collection equipment is
basically composed of vapor domes enclosing each top hatch along with
various connectors and pipes (some removable) that enable the vapors
from the tank being filled to be transferred to the tank being emptied
of liquid. Tank trucks with vapor collection equipment can become a
separate source of emissions when leakage occurs (estimated to average
about 30 percent of potentially captured emissions). Many States
require gasoline tank trucks equipped for vapor collection to pass an
annual test of tank vapor tightness and pressure limits for the tanks
and vapor collection equipment (reducing average leakage to about
10 percent).
1.1.4 Service Stations
Gasoline handling operations, emissions, and controls at service
stations are basically divided into two steps: the filling (or inloading)
of the underground storage tank, commonly called Stage I, and vehicle
refueling, commonly called Stage II. The filling of underground tanks
1-5
-------
at service stations ends the wholesale gasoline marketing chain. The
automobile refueling operation at service stations is the part of the
marketing chain that interacts directly with the general public.
Emissions from underground tank filling operations at service
stations can be reduced significantly (by about 95 percent) by the use
of a vapor balance system. Instead of being vented to the atmosphere,
the vapors are transferred into the tank truck unloading at the service
station and, ultimately, to the terminal vapor processor for recovery
or destruction. Such controls have been incorporated into many State
regulations.
Vehicle refueling emissions are another major source of emissions,
attributable to spillage and to vapor displaced from the automobile
tank by dispensed gasoline. The two basic vehicle refueling regulatory
strategies are: (!) control systems on service station equipment
{termed Stage II controls), and (2) control systems on vehicles and
trucks (termed onboard controls). Stage II controls consist of either
vapor balance systems or assisted systems. Assisted systems use a
variety of means to generate a more favorable (negative or zero) pressure
differential at the nozzle-vehicle interface so that a tight seal is
not necessary between the vehicle and the nozzle "boot" (a flexible
covering over the nozzle which captures the vapor for return to the
underground tank via a vapor hose). Stage II controls are currently
being used in 26 counties in California and the District of Columbia and
are being considered for other ozone nonattainment areas. Onboard vapor
controls consist of a fillpipe seal and a carbon canister that adsorbs
the vapors displaced from the vehicle fuel tank by the incoming gasoline.
The onboard system has undergone only limited testing to date. It is
unclear what design problems could be encountered if onboard were
required for the entire vehicle fleet; however, the technology is an
extension of a system already installed on light-duty cars and trucks.
Since 1971, new cars have been equipped with similar carbon canister
systems for collecting evaporative emissions (breathing losses caused
by temperature changes in the vehicle tank and carburetor).
Both Stage II and onboard controls can be highly effective (as
high as 95 and 98 percent, respectively). However, these high theoretical
1-6
-------
efficiencies are likely to be somewhat reduced in-use (to as low as
56 percent for Stage II programs under a minimal enforcement scenario
considering a 20 percent rate of noncompliance, and to about 92 percent
for onboard controls with the expected level of tampering).
It should be noted that vehicle refueling controls not only reduce
ambient concentrations of VOC and hazardous emissions dispersed from
the service station, but also reduce the much higher exposures to
hazardous pollutants during self-service refueling (as discussed later).
In addition, it has been found that the present canisters for controlling
evaporative emissions on many models of new vehicles are undersized.
The expansion of the onboard system to control refueling emissions
could achieve additional evaporative emission reductions roughly equal
to the emission reductions achieved through refueling control. The
estimate of excess evaporative emissions is preliminary because EPA
testing is not yet complete.
1.2 ANALYSES OF REGULATORY STRATEGIES
The regulatory strategies selected for this evaluation were assessed
with regard to their air pollution emissions, health-risk and cost
impacts. Impacts analyses were conducted using a model plant approach
for most industry segments and source categories, along with certain
key assumptions. Economic impacts were also assessed, as were the
effects of various enforcement levels on in-use effectiveness of the
vehicle refueling control systems. The following sections summarize
the regulatory strategies and analytical methods and assumptions used.
1.2.1 Regulatory Strategies, Model Plants, and Projections
A total of 14 industry-wide regulatory strategies were selected for
evaluation. These strategies, which are presented in
Table 1-1, are composed of a mixture of control options for the individual
source categories. For the strategies calling for additional controls
assessment was made of the relative emissions, risks, costs, and cost
effectiveness of:
(1) nationwide control of Stage I sources,
1-7
-------
TABLE 1-1. GASOLINE MARKETING REGULATORY STRATEGIES3
No Additional Controls (Baseline)
Stage II - Selected Nonattainment Areas (NA*)b
Stage II - All Nonattainment Areas (NA)C
Stage I - Nationwide
Stage II - Nationwide
Stage I and Stage II - Nationwide
Onboard - Nationwide
Stage II - Selected Nonattainment Areas & Onboard - Nationwide
Stage II - All Nonattainment Areas & Onboard - Nationwide
Stage I & Onboard - Nationwide
Stage II - All Nonattainment Areas and Stage I & Onboard - Nationwide
Stage II & Onboard - Nationwide
Stage I & Stage II & Onboard - Nationwide
Benzene Reduction in Gasolined
facility Size Cutoffs:
Stage I:
(1) bulk plants with throughputs <4000 gal/d from balance controls on
outgoing loads; and
(2) service stations with throughputs <10,000 gal/mon.
Stage II:
(1) all service stations with throughputs <10,000 gal/mon; and
(2) all independent service stations with throughputs <50,000 gal/mon.
b
Ozone nonattainment areas needing vehicle refueling controls to help meet
their ozone attainment goals by 1987.
c
Areas predicted by State or EPA to be nonattainment for ozone in 1982.
d
Benzene reduction:
A. removal of 94.5 percent of Bz from reformate fraction for total
reduction of 62.4 percent;
B. removal of 94.5 percent of Bz from reformate and fluid catalytic
cracker (FCC) fractions for total reduction of 81.3 percent.
1-8
-------
(2) nationwide control of vehicle refueling emissions
(Stage II controls, onboard controls, or both),
(3) ozone nonattainment area control of vehicle refueling
emissions (selected or all ozone nonattainment areas), and
(4) combinations of the above.
The regulatory strategies consider these approaches either singly or in
combination, both for controlling all facilities and for including size
cutoffs for some facilities. The facility size cutoffs were assumed
based on the relatively higher costs of control for small facili-
ties, existing size cutoffs under State and local regulations, and
statutory requirements for small and medium throughput independent
service stations under Section 325 of the Clean Air Act (section titled
"Vapor Recovery for Small Business Marketers of Petroleum Products")..
If a Section 112 standard is pursued requiring Staged or Stage II
controls, the actual size cutoffs could vary from these assumptions
based upon a more thorough economic analysis for small businesses and
an assessment of whether Section 325 applies to Section 112 standards.
For the purposes of the analysis, initial installation of Stage I and
II control equipment was assumed to occur in 1986 for the nonattainment
area strategies, in 1987 for nationwide strategies, and on new vehicles
beginning with the 1988 model year for onboard controls. All of the
strategies were compared with a baseline reflecting 1982 Federal, State,
and local regulations. The base year of 1982 was selected because this
represented the final implementation year for many State regulations
affecting gasoline marketing sources, and because at the beginning of
the analysis the most recent complete data reflected 1982 totals.
A number of model plants were developed to represent the entire
spectrum of facilities in the analyses. Four model plants were used
for both bulk terminals and bulk plants while five model plants were
used for service stations. The model plants for each source category
were differentiated on the basis of size, in terms of gasoline throughput.
Estimates of typical costs, emissions, and resultant health-risks could
then be generated for each model plant and, thus, for the entire population
spectrum of each facility type.
1-9
-------
Because onboard controls would be installed only on new vehicles
(this analysis did not consider a retrofit option), the onboard regulatory
strategies take a number of years after initial implementation to
control the entire vehicle fleet. Therefore, in order to evaluate the
comparison of onboard with other controls during both phase-in and full
implementation, the analyses examined the time period from 1986 (when
the first controls would begin to be implemented) through 2020. Thus,
the projection of certain basic parameters was necessitated.
Total and leaded gasoline consumption were extrapolated to the
year 2000, based on the projections by EPA through 1990 for the phasedown
of lead in gasoline (47 FR 49329). Due to a lack of confidence in
extrapolating beyond 15 years, gasoline consumption was assumed to
remain constant from the year 2000 to the year 2020. The number and
fuel consumption of onboard controlled vehicles in a given year were
estimated based on projections of new vehicles, retirement rates, fuel
economy, and mileage accrual rates through the year 2000. These parameters
were also assumed constant from 2000 through 2020. Although the number
of bulk plants and service stations have been decreasing, no quantitative
data was readily available with which to project facility populations.
Therefore, the numbers and size distributions of facilities were assumed
to remain constant at the values estimated for the base year of 1982.
However, the throughputs per facility and corresponding recovery credits
were decreased in proportion to the projected decline in nationwide
gasoline consumption. The economic effects of alternative assumptions
(e.g., constant gasoline consumption, declining marketing facilities,
etc.) are examined in Chapter 8.
A summary of major analytical considerations that should be noted
in assessing the results is given in Table 1-2. Both costs and emissions
were summed to a cumulative value, and also were discounted (at 10 percent)
to a net present value in 1986 (by summing the equivalent worth in 1986
of each annual amount). Discounting was necessary because impacts are
not uniform over the time period analyzed due to: 1) the slower phase-in
period of onboard controls compared to Stage I and II equipment; 2) the
respective useful lives of service station and onboard control equipment;
and 3) the declining gasoline consumption which directly influences the
1-10
-------
TABLE 1-2
MAJOR ANALYTICAL CONSIDERATIONS
o A standard exposure lifetime of 70 years was used for exposure to ambient
concentrations (i.e., those away from the immediate vehicle fueling area).
A period of 50 years was used for self-service fueling exposure because
one would not be expected to operate a vehicle for a complete lifetime.
o Because of the different phase-in times for Stage II and onboard, and to
analyze the impact of the strategies after onboard was fully implemented,
the analysis covered a period from 1986 to 2020.
o In this analysis, it was assumed that a national Stage II program would be
in place in 1989, and that 1988 model year vehicles would be the first to
incorporate onboard controls. It is expected that it will take about 20
years to convert the entire vehicle fleet to onboard control.
o Both costs and emissions capture over the 35-year analysis period were
discounted at 10 percent to calculate cost effectiveness. This was done
to address the difference in phasing-in Stage II and onboard controls.
o Due to the number of sources and lack of siting data, simplifying assumptions
were required. These included:
— Model Plants
-- Typical Locations of Plants
-- Number of Facilities Within Source Categories
-- Distribution of all Sources
o Sources were distributed by best estimate considering size and population
densities.
o Risk is assumed to be linearly related to dose (concentration x duration
of exposure) and combinations of concentration and duration yielding the
same dose are assumed to be equivalent for risk estimation purposes.
o The impacts of exposure to other substances beyond benzene, EDC, EDB, and
gasoline vapor are not addressed. Health Effects other than cancer are
not explicitly addressed.
o Total gasoline and leaded gasoline consumption is based on EPA's lead
phase-down projection extrapolated to the year 2000. The consumption
estimate for the year 2000 is assumed for all years from 2001 to 2020.
o Fleet average cost estimate of onboard systems used in this analysis was
approximately $15 per vehicle. Recent studies by API and Ford Motor
Company estimate average onboard costs of $13 and $53 per vehicle,
respectively.
o The average capital costs (equipment and installation costs) per station
for Stage II control systems used in this analysis are $5,700, $6,100
$6,600, $9,800 and $14,800 for 5, 20, 35, 65, and 185 thousand gallon
per month throughput stations, respectively.
1-11
-------
recovery cost credits. Cost effectiveness was calculated using the
discounted costs and emissions values.
1.2.2 Air Pollution Emissions, Health-Risk, and Control Cost Analyses
Several underlying methods and assumptions were made in all of the
emissions, health-risk, and control cost analyses, in addition to the
projections noted in the previous section. Generally, emissions and
health-risk impacts of the various regulatory strategies (including
baseline, which reflects current controls) were estimated for a base
year (1982) and then extrapolated to the years 1986 through 2020 in
proportion to the total (for benzene and gasoline vapors) or leaded
(for EDB and EDC) gasoline throughput for each source category. In
addition, the phasing-in of control installations with time were
considered in accordance with statutory requirements. All affected
facilities were assumed to install controls (linearly with time) within
one year for a CTG and within 2 years for a NESHAP except for independent
service stations, which may be allowed up to 3 years in accordance with
Section 325 of the Clean Air Act. Capital costs were attributed to the
year of installation.
Estimates of annualized costs, emissions, and subsequent health-risks
during phase-in periods were based on the number of facilities and
corresponding gasoline throughput controlled for an entire year. The
impacts of vehicles with onboard controls in each year were calculated
based on the vehicle fleet projections noted previously. After nearly
the entire vehicle fleet was projected to be equipped with onboard
controls (in about 2002-2003), Stage II controls were not replaced, but
instead gradually phased out after the completion of useful equipment
lives for those strategies combining Stage II and onboard.
The health-risk analysis estimated both annual cancer incidences
nationwide and lifetime risk from high exposure, assuming a linear
dose response relationship with no threshold. The term "lifetime risk
from high exposure" is conceptually similar to the term "maximum lifetime
risk" which has been presented in other EPA documents, including those
on benzene sources regulated or considered for regulation under Section
112 of the Clean Air Act. The term "lifetime risk from high exposure"
rather than "maximum lifetime risk" is used in presenting the risk
1-12
-------
calculations for the gasoline marketing study because EPA is less
certain in this case that the assumptions used result in the maximum
exposure to any single person or group. For example, high exposure
from self-service was assumed to occur to a person pumping 40 gallons
per week. To the extent that some people may pump more gas than that,
risks may be underestimated.
The estimates of risk, in terms of individual lifetime risk from
high exposure and aggregate incidence, are applicable to the public in
the vicinity of gasoline marketing sources and those persons who refuel
their vehicles at self-service pumps. This analysis did not examine
the risk to workers from occupational exposure (e.g., terminal operators
and service station attendants). The lifetime risk from high exposure
for these workers is probably substantially higher than for the general
public. In addition, the estimates of aggregate incidence would be
higher if such worker populations were included in the analysis. Of
course, any controls to reduce gasoline marketing emissions would reduce
exposure for workers as well as for the general public.
The unit risk factors used in the analysis for the four pollutants,
presented in Table 1-3, were developed by the EPA's Carcinogen Assessment
Group based on available health studies. Two values of unit risk are
shown for gas vapors - a "maximum likelihood estimate" and a "plausible
upper limit." Both values are used in the analysis to provide a broader
base for evaluating the impacts of exposure to gasoline vapors. For a
detailed description of the derivation of the gasoline vapors unit risk
numbers, see the EPA staff paper "Estimation of the Public Health Risk
from Exposure to Gasoline Vapor Via the Gasoline Marketing System",
June 1984. The risk factor for benzene is based on studies of humans
occupational^ exposed to benzene. The risk factors for gasoline
vapors, EDC, and EDB are based on animal studies only. Because of the
significance of the gasoline vapor animal studies, conducted for the
American Petroleum Institute, they are examined in detail in the EPA
staff paper cited above. The staff paper was submitted on June 22,
1984, to the EPA Science Advisory Board for review.
There can be substantial uncertainty in unit risk factors. Reasons
for this uncertainty include extrapolations which must be made from
1-13
-------
TABLE 1-3. UNIT RISK FACTOR SUMMARY
Pollutant
Unit Risk3
(probability of cancer
given lifetime exposure
to 1 ppm)
Health Effects
Summary
Continents
Gasoline Vapor
Plausible Upper Limit:
Rat Studies
Mice Studies
3.5 x ID'3
2.1 x 10-3
Maximum Likelihood Estimates:
Rat Studies
Mice Studies
Benzene
Ethylene
Dibronide
Ethylene
Dichloride
2.0 x 10-3
1.4 x ID'3
2.2 x 10-2
4.2 x 10'1
2.3 x 10-2
Kidney tumors in
rats, liver tumors
in mice.
Human evidence of
leukemogenicity.
Zymbal gland
tumors in rats,
lymphoid and other
cancers in mice.
Evidence of carci-
nogenicity in
animals by inhalation
and gavage. Rats:
nasal tumors; Mice:
liver tumors.
Evidence of carci-
nogenic ity 1n
animals. Circulatory
system, forestomach,
and glands; Mice:
Liver, lung, glands,
and uterus.
Gasoline test samples in
the animal studies were
completely volatilized,
therefore may not be
completely representative
of ambient gasoline vapor
exposures.
EPA: listed as a hazardous
air pollutant, emission
standards proposed.
IARCC: sufficient evidence
to support a causal associ-
ation between exposure and
cancer.
EPA: suspect human carci-
nogen; recent restrictions
on pesticidal uses.
EPA: Suspect human
carcinogen. Draft
health assessment
document released
for review March 1984.
Unit Risk Factor 1s in terms of the probability of a cancer incidence (occurrence)
a single individual for a 70-year lifetime of exposure to 1 ppm of pollutant.
in
The plausible upper limit is calculated as the 95 percent upper-confidence limit of the
incremental risk due to exposure to 1 ppm of gasoline vapor, using the multistage dose-
response model for low-dose extrapolation.
IARC: International Agency for Research on Cancer.
1-14
-------
workers or animals to the general population and from the higher concen-
trations found in studies to the lower concentrations found in the
ambient air.
Estimates of risk due to exposures from bulk terminals, bulk
plants, service stations, and self-service vehicle refueling were
generated for each of the four pollutants. In order to calculate
community exposure to emissions (and the resultant risk) from bulk
terminals and plants, assumptions were made concerning their geographical
distribution. The fundamental assumption was that facilities were
located in proportion to the gasoline throughput for an area—for
example, the largest model plants would be located in large urban areas
where throughput (and population density) were highest. Further, each
model plant type in each source category (bulk terminals and bulk
plants) was distributed over a range of ten urban area sizes. The
largest terminals, for instance, were assumed to be located in cities
ranging in size from New York City to Des Moines, Iowa; the smallest
terminals were assumed to be located in cities ranging in size from
Spokane to Effingham, Illinois. Estimates were also made of the extent
of existing control at these terminals. Most of those in the large
cities (likely to be ozone nonattainment areas) were considered controlled,
based upon existing regulations, with proportionately fewer facilities
controlled in the smaller areas.
Thus, for both terminals and bulk plants, there were 40 model plant
locations (four model plant sizes each distributed to 10 representative
areas) for which estimates of ambient concentrations, population exposure
and incidence were made. Total national incidence was calculated by
multiplying the model plant incidence by the number of facilities
represented by each model plant. In somewhat similar fashion, model
service stations were allocated to 35 localities (multi-county metropol-
itan areas or single counties) and grouped by seven population size
ranges. The model plants were selected to be representative of total
national service station distribution. The localities and seven popula-
tion size ranges were selected to be representative of the total national
population distribution.
1-15
-------
Ambient concentrations, exposure, and incidence for bulk terminals,
bulk plants and service stations were calculated using the SHEAR version
of the EPA Human Exposure Model (HEM). The HEM is a model capable of
estimating ambient concentrations and population exposure due to
emissions from sources located at any specific point in the contiguous
United States.
Annual incidence due to self-service vehicle refueling was
estimated based on benzene and VOC concentrations in the region of
the face of a person filling the tank, as measured in a study for the
American Petroleum Institute (API).* API selected thirteen gas stations
in 6 cities in which samples were collected to characterize typical
exposures to total hydrocarbons, benzene and eight other compounds.
Samples were collected using MSA-type battery operated pumps operated at
a one liter per minute flow rate and analyzed using a gas chromatograph.
Results were expressed as mg/nP air and ppm (vol.).
The lifetime risk analysis was designed to estimate high exposures
of the four pollutants. The Industrial Source Complex (ISC) dispersion
model was used to calculate annual concentrations in selected years at
a number of receptors in the vicinity of a bulk terminal complex, a
bulk plant complex, and a service station complex. Meteorological data
for several cities expected to produce high concentrations were used.
The highest concentration at any receptor under a given regulatory
strategy was used to estimate the risk over a 70-year lifetime. The
lifetime risk due to self-service exposure was estimated based on the
API measurements and an assumed lifetime exposure pattern of an individual
using a relatively high amount of gasoline (i.e. traveling salesman):
the risk was based upon an assumed exposure to four ID-gallon self-service
refuel ings per week for a working lifetime (estimated as 50 years).
The control cost estimates were developed using a method similar to
that for emissions. Capital and operating cost data were obtained
(largely from previous EPA studies) and developed on a per facility
*CTayton"TrivTronmental Consultants, Inc. Gasoline Exposure Study for
the American Petroleum Institute. Job Mo. 18629-15. Southfield,
Michigan. August 1983.
1-16
-------
basis for each model plant size of each source category. These per
facility costs were then combined with data on the number of facilities
requiring controls within each source category. Capital costs over the
35 years of the analysis were incorporated during the initial phase-in
years and then repeated in the years in which the economic life of the
equipment ended if replacement equipment was required. Annualized
costs reflected the capital costs and also were adjusted each year, as
appropriate, to reflect reduced recovery credit due to the assumed
decreases in gasoline throughput.
1.3 RESULTS OF REGULATORY STRATEGY ANALYSES
Although only results of strategies involving size cutoffs are
given in this summary, all strategies are discussed in detail in Chapters 2
through 9. The estimated risks from the various source categories are
given in Table 1-4 for baseline (no additional controls) and when
controlled. Table l-4a contains estimated risks using the plausible
upper limit unit risk factor for gasoline vapors and Table l-4b contains
the estimates using the maximum likelihood estimate unit risk factor
for gasoline vapors.
The lifetime risk from high exposure estimates the probability
that exposure by an individual to a relatively high ambient concentration
throughout his lifetime would result in a cancer incidence. The lifetime
risk from high exposure to bulk terminal emissions is higher than the
lifetime risk from high exposure to uncontrolled emissions from any of
the other source categories.
The average annual incidence for each regulatory strategy is the sum
of the estimated average annual incidences from each industry segment
expected to result from exposures during a given year. (The estimated
annual incidences decrease during the study period in proportion to a
projected decrease in gasoline consumption.) The average annual inci-
dences given are estimates of cancer incidence due to benzene and
gasoline vapors; for the latter, results are shown based on both mice
and rat health studies. Subsequent incidence numbers in this Executive
Summary will be given in terms of rat data only (for both plausible
upper limit and maximum likelihood estimates) as the rat numbers are
higher and the tables can be simplified. The incidences due to EDB and
1-17
-------
TABLE l-4a. ESTIMATED RISKS FROM GASOLINE MARKETING
SOURCE CATEGORIES
(USING PLAUSIBLE UPPER LIMIT UNIT RISK FACTOR FOR GASOLINE VAPORS)
A. BASELINE
Source Category
Bulk Terminals
Bulk Plants
Service Stations
Sel f-service
Source Category
Bulk Terminals
Bulk Plants
Lifetime Risk from
High Exposure
(probability of effect)
1.2 x 10'4(2.4 or 3.9
6.4 x 10-6(1.2 or 2.0
2.4 x 10-6(4.4 or 7.2
1.1 x 10-5(5.5 or 9.0
B. CONTROLLED WITH SIZE
x 10-3)
x 10-4)
x 10-5)
x 10-5)
CUTOFFS5
Lifetime Risk from
High Exposure
(probability of effect)
Bz(GVa)
2.0 x 10-5(3.9 or 6.4
1.7 x 10-6(3.2 or 5.3
x lO'4)
x 10-5)
Average Annual Incidence
Over 35 years (1986-2020)
Bz(GVa)
0.07(1.3 or 2.2)
0.04(0.68 or 1.1)
0.19(3.3 or 5.5)
3.2(19 or 31)
Average Annual Incidence
Over 35 years (1986-2020)
0.05(0.92 or 1.5)
0.02(0.32 or 0.53)
Service Stationsc
Stage I controls
only
Stage II controls
only
Onboard controls
only
1.6 x 10-6(2.9 or 4.7 x 1Q-5) 0.17(3.0 or 4.9)
1.3 x 10-6(2.5 or 4.2 x 1Q-5) 0.11(1.9 or 3.2)
1.6 x 10-6(2.9 or 4.8 x lO'5) 0.10(1.3 or 3.0)
Sel f-servicec
Stage II controls 5.1 x 1Q-7(2.S or 4.1 x 1Q-6) i 2(6 3 or 12)
only
Onboard controls 7.5 x 10-3(4.4 or 7.2 x 10'7) 1.1(5.9 or 9 3)
only
j— - • . . _ ,
Bz = benzene, GV = gasoline vapors (mice or rats studies). Incidences and
lifetime risk due to gasoline vapors are presented to reflect the two unit
risk factors (for liver cancer in mice or for kidney cancer in rats).
b
Based on theoretical control efficiencies.
Indicates reduction of lifetime risk from high exposure for controlled sources.
1-18
-------
TABLE l-4b. ESTIMATED RISKS FROM GASOLINE MARKETING
SOURCE CATEGORIES
;USING MAXIMUM LIKELIHOOD ESTIMATE UNIT RISK FACTOR FOR GASOLINE VAPORS)
A. BASELINE
Source Category
Sulk Terminals
Bulk Plants
Service Stations
Self-service
Lifetime Risk from Average Annual Incidence
High Exposure Over 35 years (1986-2020)
(probaoility of effect)
8z(GV3) BzlGV*)
1.2 x 1Q-4(1.5 or 2.2 x 10'3)
6.4 x 10-6(8.2 x 10-5 or 1.1 x lO"4)
2.4 x 10-6(2.9 or 4.1 x 10'5 )
1.1 x 10-5(3.7 or 5.1 x 10"5)
0.07(0.90 or 1.3)
0.04(0.46 or 0.64)
0.19(2.2 or 3.1)
3.2(13 or 18)
3. CONTROLLED WITH SIZE CUTOFFS6
Source Category
Bulk Terminals'3
Bulk Plants
Lifetime Risk from
High Exposure
(probability of effect)
Bz(GVa)
2.0 x 10-5(2.6 or 3.7 x 10'4)
1.7 x 10-6(2.2 or 3.0 x 10'5)
Average Annual Incidence
Over 35 years (1986-2020)
Bz(GVa)
0.05(0.62 or 0.86)
0.02(0.22 or 0.30)
Service Stations0
Stage I controls
only
Stage II controls
only
Onboard controls
only
1.6 x 10-6(1.9 or 2.7 x 1Q-5)
1.3 x 0-6(1.7 or 2.4 x 10'5)
1.6 x 10-6(2.0 or 2.3 x 10'5)
0.17(2.0 or 2.8)
0.11(1.3 or 1.8)
0.10(1.2 or 1.7)
Self-service0
Stage II controls 5.1 x 1Q-7(1.7 or 2.3 x 10"6)
only
Onboard controls 7 .5 x 1Q-8(3.0 or 4.1 x 10"7)
only
1.2(4.8 or 6.6)
1.1(4.0 or 5.6)
Bz = benzene, GV = gasoline vapors. Incidences and lifetime risk due to
gasoline vapors are presented to reflect the two unit risk factors (for
liver cancer in mice or kidney cancer in rats.
Based on theoretical control efficiencies.
Indicates reduction of lifetime risk from high exposure for controlled sources.
1-19
-------
EDC only increase the incidences due to benzene by less than 3 percent
in most cases and by 5 percent or less in all cases. Because the
estimated incidences due to EDB and EDC are relatively small, they were
omitted from the summary tables. The average annual incidence from
self-service refuel ings at service stations contributes about 80
percent of the total incidence from all source categories. The annual
incidences due to service stations without any additional controls are
approximately 3 for benzene and from about 15 to 36 for gasoline vapors,
considering both community exposures to ambient concentrations and
individual exposures to self-service refueling concentrations.
1.3.1 Nonattainment Area Strategy Results
The effects of vehicle refueling controls in nonattainment areas
are shown in Table 1-5. The primary focus of the nonattainment area
strategies is to reduce VOC emissions in order to attain the national
ambient air quality standard (NAAQS) for ozone; reduction of risk due
to hazardous pollutants is an added benefit. If onboard controls
nationwide under Section 202(a)(6) of the Clean Air Act are used to
replace or supplement nonattainment area regulatory strategies, in
addition to the primary aim of reducing VOC emissions in some or all
nonattainment areas, VOC and hazardous emissions are also reduced
nationwide. It should be noted that the strategy of Stage II controls
in all nonattainment areas also includes the costs, emissions, and
risk reductions for Stage I controls at service stations in two non-
attainment areas (Atlanta and Phoenix) where they currently are not
instal led.
The average annual baseline level of YOC emissions (and zero addi-
tional control cost) from service stations of 91,300 Mg/yr is given in
parentheses in Table 1-5. The average annual VOC emission reduction
from the baseline, net present value of control costs, and discounted
cost effectiveness due to Stage II refueling controls were estimated to
be 20,600 Mg/yr, $210 million, and $940/Mg VOC, respectively, for the
"selected nonattainment areas" (NA*) strategy and 60,900 Mg/yr, $570
million, and $870/Mg VOC, respectively, for the "all nonattainment
areas" (NA) strategy. The costs and cost effectiveness values for
strategies involving onboard controls are much higher (costs about $2
1-20
-------
TABLE 1-5. VEHICLE REFUELING CONTROLS IN NONATTAINMENT AREAS
Regulatory Strategy
BASELINE
Service Station
Emi ssions
Average Annual VOC
Emission
Reduction3, In
Nonattainment
Areas, Mg
(1986 - 2020)
(91,300)d
Net Present
Value of
Costs, SMilllon
(1986 - 2020)
Discounted
Cost
Effectiveness
$/Mg VOC
Nationwide
Service Station
Average Annual
Incidence"
(1986 - 2020),
Bz (GV)C
3.4(21 or 36)
REGjJUa^JT_RATEGIES
Stage II In Selected Non-
attainment areas (NA* with
size cutoffs)6
Stage II In all Nonattainment
areas (NA with size cutoffs)^
Onboard Nationwide?
Stage II In Selected Nonattain-
ment areas (with size cutoffs),6
Onboard Nationwide
Stage II in All Nonattalnwent
areas (with size cutoffsF,
Onboard Nationwide
20,600
60,900
56.3009
60,9009
69,1009
210
570
1,900
2,100
2,400
940
870
4,760"
4,130"
3,440"
3.1(19 or 33)
2.6(16 or 28)
1.2( 7.3 or 13)
l.H 6.8 or 12)
1.0( 6.1 or 11)
Emission reductions from vehicle refueling only assuming theoretical efficiencies.
b
Fran community exposure to ambient concentrations from service stations and from exposure of Individuals during self-service
vehicle refueling considering theoretical efficiencies.
c
GV = Gasoline vapors (rat data only), two numbers given represent estimates using maximum likelihood estimate unit risk
factor and plausible upper limit unit risk factor, respectively.
rj
Baseline anission level for vehicle refueling in nonattairanent areas.
e
Selected Nonattaimient Areas (NA*) - areas needing Stage II to meet their attainment goals by 1987.
All Nonattainment Areas (NA) - areas predicted by State or by EPA to be nonattainment for ozone in 1982. Includes Stage I
for two metropolitan areas currently without Stage I in place.
g
Onboard emission reductions in nonattainment areas only. Total nationwide average annual emission reductions (Mg/yr):
Onboard - 197,000, Stage II (NA*) *- Onboard - 203,000, Stage II (NA) »• Onboard - 211,000.
h
Cost effectiveness based on total nationwide emission reductions are: Onboard - 1,360, Stage II (NA*) * Onboard - 1,370,
Stage II (NA) * Onboard - 1,410.
f
h
-------
billion and cost effectivenesses about $5,000/Mg VOC) when considering
only the emission reductions in the affected nonattainment areas and
nationwide costs for onboard controls. The annual average incidences
presented are the cumulative nationwide cancer incidences expected to
result from service station community and self-service exposures during
the 35-year period under the given regulatory strategy averaged over the
35 years. As can be seen from Table 1-5, the incidence reduction due to
Stage II controls in nonattainment areas are relatively small (less than
one incidence per year due to benzene) compared with that associated with
onboard controls nationwide (more than two incidences per year due to ben-
zene reduced). The results shown in Table 1-5 are based on theoretical
control efficiencies and do not account for reduced efficiency in-use.
1.3.2 Nationwide Strategy Results
The issues involved in gasoline marketing operations do not relate
to ozone attainment only -- there is concern for hazardous exposure also.
This section presents summaries of the results of analyses of alternative
nationwide regulatory strategies and combined nationwide/nonattainment
strategies. Table 1-6 shows the estimated baseline annual average
incidence for benzene and gasoline vapors, and the residual incidence
and cost for each of the three primary nationwide strategies: Stage I,
Stage II and onboard. The baseline benzene incidence attributable to
vehicle operations emissions (tailpipe and evaporative) is also shown;
this incidence is more than twice as large as that from gasoline marketing
operations. In addition, the baseline incidence from possible additional
evaporative emissions is shown separate from vehicle operations. This
value was separated because of its uncertainty, being based on preliminary
test results. Controls on gasoline marketing sources will not reduce
incidences associated with vehicle operations.
For gasoline marketing sources, the average annual reduction in
benzene incidence is estimated to be 0.1 with Stage I controls, 2.1
with Stage II controls and 2.3 with onboard. For gasoline vapors, the
comparable numbers are 1.0 or 1.8, 12 or 22, and 14 or 24. Only
limited benzene incidence reduction is achieved by removing benzene
from gasoline (2.4-3.2 per year). This reduction primarily results
from reduced exposure from gasoline marketing sources. Benzene tailpipe
1-22
-------
TABLE 1-6. CONTROL OF BENZENE AND GASOLINE VAPORS
FROM GASOLINE MARKETING SOURCES
Regulatory
Strategy
Average Annual
Incidence
Of Cancers Expected
From Exposures
Over 35 years
(1986-2020)
3z (GVa)
Cost Impacts
of Strategies
(SBillion)
BASELINE
Gasoline Marketing^
Vehicle Operations0
Evap. Emissions Mot Captured6
Total
IMPACTS AFTER SELECTEDf
NATIONWIDE STRATEGIES
Stage I - Nationwide
(with size cutoffs)
Stage II - Nationwide
(with size cutoffs)
Onboard Controls
Mationwide
.- w/o Evap
- w/ Evap
Benzene Reduction
in Gasoline
Gasoline Marketing
Vehicle Operations0
Evap. Emissions Not Captured
3.5
9.7
0.2
13
(23 or 40)
(NAd)
(4.0 or 7.0)
(NA)
AVERAGE ANNUAL
INCIDENCE REDUCTION
0.1 (1.0 or 1.8)
2.1 (12 or 22)
2.3 (14 or 24)
2.5 (18 or 31)
2.2-2.9 (0 or 0)
0.18-0.23 (NA)
0.09-0.12 (1.5 or 3.3)
1986 NPVb Total 1982
of Costs Dollars Spent
(1986 - 2020) (1986 - 2020)
0.9
1.6
1.9
1.9
7.4-22
3.2
6.3
9.7
9.7
30-90
GV = gasoline vapors (rat data only), two numbers given represent estimates using
maximum likelihood estimate unit risk factor and plausible upper limit unit risk
factor, respectively.
3NPV = Net Present Value.
"incidences due to exhaust and evaporative benzene emissions during vehicle operations
were estimated using an area source approach similar to that used for service stations.
Unanticipated evaporative emissions were not considered here (see Footnote e).
Not applicable.
a
~B.ased on preliminary estimate of possible emissions not captured by the existing
evaporative emissions control system on vehicles. Deduction in such incidence is account
for under "Onboard Controls, Mationwide, w/Evap."
Impacts based on theoretical control efficiencies.
1-23
-------
emissions are not affected substantially by benzene removal, since
benzene is formed in the combustion process.*
Costs of additional controls beyond baseline are presented both as
the net present value of cost (discounted at 10 percent to 1986) and the
cumulative value of the estimated expenditures from 1986 through 2020
(all in 1982 dollars). The costs of all available nationwide strategies
are greater than $800 million net present value of costs or $3 billion
cumulative costs. The cost of benzene reduction in gasoline is a
factor of 2 to 10 greater than for the next most costly strategy.
Table 1-7 summarizes the estimates of average annual incidence,
emission reductions, cost, and cost effectiveness for the nationwide
control strategies (with size cutoffs) evaluated. These impacts also
are based on theoretical control efficiencies. This table presents
more detailed results: average annual incidence under the strategy,
cumulative VOC and benzene emission reductions, net present value of
costs, and discounted cost effectiveness for both basic regulatory
strategies and combinations of strategies. Although Stage I controls
result in large emission reductions at relatively low cost, they result
in substantially less incidence reduction than Stage II or onboard.
Strategies with either Stage II or onboard refueling controls achieve
greater incidence reductions because of their effect on self-service
emissions.
Table 1-8 presents costs, emission reductions, incidence reduction
and cost effectiveness with theoretical and in-use efficiencies for
Stage II and onboard, and gives two levels of enforcement for Stage II.
The in-use efficiency of Stage II programs is highly dependent on the
level of enforcement used, varying from 56 percent with no inspections
to 86 percent with annual inspections. Enforcement costs are not
included in the cost-effectiveness figures given (their impact is addressed
in Chapter 8). Although average annual enforcement costs for Stage II
nationwide with annual inspections are about $7.7 million, including
* Black, P.M., I.E. High, and J.M. Lang. Composition of Automotive
Evaporative and Tailpipe Hydrocarbon Emissions. Journal of the Air
Pollution Control Association. 30:1216-1221. Movember 1980.
1-24
-------
Table 1-7. SUMMARY OF THEORETICAL IMPACTS FOR SELECTED REGULATORY STRATEGIES
Regulatory
Strategies
(with size cutoffs)
Dasel1ne( No Additional
Controls)
Stage 11 - Nationwide
Onboard -
Nationwide
Stage II NA* Areas,
Onboard Nationwide
Stage II NA Areas,
Onboard Nationwide
i Stage I - Nationwide
rx>
c_n
Stage 11 and
Onboard Nationwide
Stage I and Stage II -
Nationwide
Stage I and Onboard-
Nationwide
Stage I, Stage II,
Onboard - Nationwide
Average Annual Average Annual Average Annual
Incidence VOC Emission Bz Emission
(1986 - 2020) Reduction, Mg Reduction, Mg
Bz (GVa) (1986 - 2020) (1986 - 2020)
3.5(23 or 40) [710,000]d [4,400]d
1.4(10 or 18) 163,000 1,100
1.3(9.2 or 16) 197,000 1,300
1.2(8.7 or 15) 202,000 1,300
1.1(8.0 or 14) 210,000 1,400
3.4(22 or 38) 217.000 1,300
I
0.8(6.7 or 12) 228,000 1,500
1.4(9.3 or 16) 380,000 2,300
1.2(8.1 or 14) 414,000 2,500
0.8(5.7 or 9.9) 446,000 2.700
Net Present Discounted
Value of Cost
Costs5. JMIlllon Effectiveness0
(1986 - 2020) $/Mg VOC
1 ,600 1 .040
1,900 1,360
2,100 1,380
2,400 1.390
860 410
3,500 1,700
2,500 680
2,780 790
4,350 1,040
GV = gasoline vapors (rat data only), two numbers given represent estimates using maximum
likelihood estimate unit risk factor and plausible upper limit unit risk factor, respectively.
b
Net present value of costs Is the sum of the present worth of each year's annual I zed cost
(discounted at 10 percent).
c
Discounted cost effectiveness Is the net present value of costs divided by the net present
value of emissions, both discounted at 10 percent.
d
Average annual emissions at baseline control levels.
-------
TABLE 1-8. IMPACTS BASED ON "IN-USE" EFFECTIVENESS FOR STAGE II AND ONBOARD CONTROLS (1986 - 2020)
CTl
Regulatory Strategies Average Annual Emission Met Present Value Discounted Cost Average Annual Average Annual
(% Control Efficiency) Reduction of Benzene of Annual Costs, Effectiveness, Incidence Enforcement
(Gasoline Vapors) $ Billion $/Mg of VOC Reduction Resources
(103Mg) Oz(GVa) Person-years/$ Mil lion
SI AGE II NATIONWIDE
(Ho size cutoffs)
o Theoretical (95%)
o In-use
. Annual
Inspections (86%)
. No Inspections (56%)c
STAGE 11 NATIONWIDE
(w/size cutoffs)
o Theoretical (95%)
o In-use
. Annual
Inspections (86%)
. No Inspections (56%)c
ONBOARD
o Theoretical (98%)
. w/o Add! tional
Evaporative Control
. w/Add) tional
Evaporative Control
o In-use (92%)
. w/o Additional
Evaporative Control
. w/ Additional
Evaporative Control
1.6(237) 4.9 2,130 2.9(18 or 31)
1.4(214) 5.0 2,380 2. 6(16 or 28) 815/24.3
0.9(140) 4.0 2,950 1.6(9.8 or 17)
1.1(163) 1.6 1,040 2.1(12 or 22)
1.0(146) 1.7 1.200 1.9(11 or 20) 260/7. /
0.6( 94) 1.4 1,570 1.1(6.7 or 12)
1.3(197) 1-9 1,360 2. 2(14 or 24)
2.5(374)b 1.9 720 2.5(18 or 31)
1.2(183) 1-9 1.460 1.8(11 or 20) 4.9/0.1
2.3(346)b 1.9 770 2. 1(15 or 27) 4.9/0.1
dGV = gasoline vapors (rat data only), two numbers given represent estimates using maximum
likelihood estimate unit risk factor and plausible upper limit unit risk factor, respectively.
b
Based on preliminary data.
Cprogram efficiency reduces to 56 percent when the rate of noncoinpllance (20 percent) Is considered.
Actual average control efficiency of Installed recovery systems Is estimated at 70 percent.
-------
them would cause only a slight increase in cost effectiveness. The
in-use efficiency of onboard controls is expected to be about 92 percent
(based on current levels of tampering to use leaded gasoline, disregarding
the phase-out of leaded gasoline). The enforcement costs for onboard are
lower than Stage II (average annual cost of $0.1 million) since they
are for the incremental cost above the current certification and in-use
testing program, and the incremental costs of inspecting onboard rather
than evaporative control systems on selected vehicles at the assembly
line.
Figure 1-2 graphically depicts annual incidences due to benzene
with either Stage II or onboard regulatory strategies and with no
additional controls (baseline). Estimated incidence with Stage II is
shown both for the assumed statutory phase-in requirements and for an
alternative phase-in schedule suggested by the American Petroleum
Institute (API). The assumed statutory phase-in requirements used in
this analysis are installation within 2 years for non-independents under
a NESHAP (Section 112) and 3 years for independents assuming Section 325
of the Clean Air Act applies to CTG's and Section 112 standards. The
API phase-in schedule assumes 3 years for nonindependents and 7 years
for independents. Installation of Stage II controls was assumed to
begin in 1987 for both the statutory and API phase-in scenarios.
Onboard controls were assumed to be installed on new vehicles beginning
with the 1988 model year. Therefore, all of the strategies shown begin
at baseline levels of about 4.8 incidences expected from 1986 benzene
emissions from the entire gasoline marketing system. The baseline (no
further controls) levels of annual incidence also decrease with time in
proportion to the projected decrease in gasoline consumption. The
Stage II strategies reduce incidence more rapidly than the onboard
strategy. The numbers in parentheses on the graph indicate the differences
in cumulative incidence before or after 1995 when the onboard strategy
is projected to reduce incidence to below the level reached by the
Stage II strategies. Thus, although Stage II can achieve incidence
reduction sooner than onboard, by 2020 the cumulative incidence reduction
with onboard controls has surpassed the cumulative reduction with
Stage II controls, since the steady-state levels of annual benzene
incidence are about 1.2 for Stage II versus 0.5 for onboard.
1-27
-------
Figure 1-2. Effect of Onboard and Stage II Controls on
Benzene Incidence
(Based on Theoretical Efficiencies)
5.0 ~-
4.5
4.0 ~-
3.;
3.0
3.5
2.0
1.5 --
1.0 --
0.5 -~
STAGE II (WITH SIZE CUT-OFFS)
STATUTORY PHASE-IN (2 YRS. NON-IND., 3 YRS. IND.)
STAGE II (WITH SIZE CUT-OFFS)
API PHASE-IN (3 YRS. NON-IND., 7 YRS. INO.)
ONBOARD (BEGINS IN 1988)
BASELINE INCIDENCE -
A
A
A
A
A
( ) DIFFERENCE IN INCIDENCE DUE TO EFFECT OF CONTROL OPTION
A
A
4.9 FROM 1986 THROUGH 1995
(4.0 FROM 1986 THROUGH 1994 )
14.7 FROM 1996 THROUGH 2020
36 38
YEAR
1-28
-------
The control costs associated with each of the regulatory strategies
are assumed in this study to be passed on by producers to consumers of
gasoline and vehicles in the form of higher prices. The magnitude of
these price increases for components of nationwide regulatory strategies
with size cutoffs are presented in Table 1-9. Most show gasoline price
increases of less than half a cent per gallon of gasoline. Price
increases for benzene reduction strategies range from 1.5 to nearly
5 cents per gallon. These are average figures; in practice they would
vary both with location and over time.
Consumer resistance to these price increases can reduce the sale
of vehicles and gasoline. Estimates of the reductions in the rate of
consumption are displayed in percentage terms in Table 1-9. The
estimated reduction in gasoline consumption ranges from 38 million
gallons a year for Stage I to 128 million gallons a year for a combination
of Stage I, Stage II, and onboard controls, and to over 1,200 million
gallons a year for the most costly benzene reduction option. For
regulations involving onboard controls, annual LDV and LOT rate of sales
are estimated to decline by 17.7 and 5.3 thousand vehicles, respectively.
Other economic impacts of the regulatory strategies were also
examined. These included an analysis of the sensitivity of cost
calculations to underlying assumptions and consideration of distributional
impacts of the regulatory strategies by firm size. Results are summarized
in Table 1-10.
1.3.3 Cost Per Incidence Reduction
An analysis was performed to determine the residual costs expended
per cancer incidence avoided for selected nationwide and nonattainment
area regulatory strategies. The residual costs were determined by
obtaining the annualized costs of the controls associated with the
regulatory strategy and subtracting a range of assumed benefit values
per megagram of VOC emissions reduced. The assumed VOC benefits are
those in addition to cancer prevention, such as reduction in non-cancer
health effects and agricultural damage due to ozone. The residual cost
per incidence was then calculated by dividing residual costs by the
appropriate amount of cancer incidences avoided.
1-29
-------
TABLE 1-9. ECONOMIC IMPACT OF REGULATORY STRATEGIES BASED ON
THEORETICAL EFFICIENCIES3
I
CO
Average Unit
Cost Increase
Regulatory Strategies
(With Size Cutoffs)
Stage I
Stage II
Stage I and Stage II
Onboard controls
Stage I and Onboard
Stage II and Onboard
Stage I , Stage II and
Onboard
Benzene reduction In
re formate gasoline
Benzene reduction in
reformate and FCC gasoline
Gasoline
cents/liter
(cents/gallon)
0.034 (0.129)
0.065 (0.246)
0.099 (0.375)
0.034* (0.129JC
0.082* (0.311)*
0.116* (0.440)C
0.40 - 0.43
(1.51 - 1.64)
1.01 - 1.17
(3.81 - 4.44)
Light
Motor
Vehicles
^/vehicle
LDV LDTC
13 22
13 22
13 22
13 22
Gasol1ned
(Percent)
0.055
0.104
0.158
0.055
0.131*
0.186*
0.64-0.76
1.61-1.86
Average National Reduction In
Consumption/Production"
(Percent)
Light
Motor
Vehicles
LOV LOT
(106 Gal.) (%} (103) (%) (103)
37.7
71.5
109.2
0.16 17.7 0.18 5.3
37.7 0.16 17.7 0.18 5.3
90.5 0.16 17.7 0.18 5.3
128.2 0.16 17.7 0.18 5.3
439-473
1110-1280
Based on the declining recovery credit assumption with constant number of facilities.
b
Reductions attributable to the Imposition of new regulations.
c
Weighted average costs of trucks with single (80 percent) and dual tanks (20 percent)
at $18.19/tank.
d
Percent* based on average annual gasoline consumption over the projection period
(69 x 109 gallons/year; 261.5 x 109 liters/year).
-------
TABLE 1-10
ECONOMIC CONSIDERATIONS
o The costs of nationwide strategies without facility exemptions are 80
to 100 percent higher than for the comparable strategies with exemp-
tions. For strategies covering only nonattainment areas, costs with-
out facility exemptions are 15 to 35 percent higher than for comparable
strategies with exemptions.
o Stage I and Stage II controls will cost 1/2 to 2 1/2 cents more per
gallon of throughouput at small gasoline marketing facilities than at
large ones. Facility exemptions reduce this cost differential, but
do not el iminate it.
o Facility exemptions improve the competitive position of the smallest
facilities, but small facilities tend to be less efficient than large
facilities.
o This analysis assumes the per vehicle cost for onboard controls is
$13 per tank. If, however, the cost really is $25, the net present
value of nationwide onboard control costs would increase by more than
50 percent, and would then exceed those of nationwide Stage I and
Stage II controls.
o This analysis assumes gasoline consumption will decline in the years
ahead. If, however, consumption holds at current levels, then costs
for Stage I and Stage II controls would be less because there would be
more recovery credits.
o This analysis assumes the number of gasoline marketing facilities re-
mains constant in the years ahead. If, however, the number declines,
then costs for Stage I and Stage II controls would be less because
there would be less control equipment needed.
1-31
-------
In Table 1-11, several of the regulatory strategies are presented
with their corresponding emission reductions. The emission reductions
are presented as the net present value of all the annual emission
reductions over the study period and as an annualized value representing
equal emission reductions for each year of the study period. The
residual costs were determined assuming several different dollar values
for the benefit of reducing each megagram of YOC emissions. For example,
in Table 1-11 the annual ized emission reduction associated with Stage I
is 0.218 million Mg. Multiplying this emission reduction by each of
the assumed VOC benefit values yields the annualized VOC benefit in
dollars.
Table 1-12 presents annualized cost (control equipment and
enforcement costs) and annual ized incidence reduction due to benzene
exposure associated with several of the regulatory strategies. The
cost per cancer incidence avoided, assuming no additional benefits, is
calculated by dividing the annualized costs by the annualized incidence
reduction. Table 1-13 takes this one step further by incorporating the
annualized VOC benefits into the analysis. The values presented represent
the residual cost, assuming varying benefits for reducing VOC emissions,
of reducing cancer incidences due to benzene exposure.
Table 1-14 contains a similar analysis to that used in Table 1-13,
except that Table 1-14 was developed using the sum of the incidences
due to benzene and gasoline vapors. It is assumed that the incidences
due to benzene exposure and the incidences due to gasoline vapor exposure
are additive since the respective exposure results in different types
of cancer incidences (leukemia in the case of benzene exposure and
liver or kidney tumors in the case of gasoline vapor exposure). Met
costs per annual incidence avoided are given using both plausible upper
limit and maximum likelihood estimate risk factors for gasoline vapors.
1-32
-------
TABLE 1-11. ESTIMATED REGULATORY COSTS AND VOC BENEFITS
CO
Regulatory Strategy
(In-use efficiency)
(with size cutoffs)
Stage I
Stage II-NA(87%)
Stage Il-NA(56t)
Stage II-Nation{86%)
Stage 1 I-Nation(56%)
Onboard (92%)
w/o evaporative
w/ evaporative
Net Present
Value of
Emissions
(Million Mg)
2.1
0.6
0.4
1.4
0.9
1.3
2.5
Annual! zed
Emission
Reduction
(Mil lion Mg per year)
0.22
0.06
0.04
0.15
0.10
0.14
0.26
$250/Mg
54
15
9
37
24
34
65
Annual ized
Assuming
$500/Mg
109
30
19
74
48
68
130
VOC Benefits ($Mill1ons)
VOC Benefit Value of:
$1 ,000/Mg
218
59
37
147
95
137
259
$l,500/Mg
327
89
56
221
143
205
389
$2,000/Mg
436
118
75
295
191
2/4
519
-------
TABLE 1-12. BENZENE REGULATORY COSTS AND INCIDENCE REDUCED
Regulatory Strategy
(with size cutoffs)
(In-use efficiency)
Stage I
Stage II-NA (87%)
Stage II-NA (56%)
Stage II-Nation (86%)
Stage II-Nation (56%)
Onboard (92%):
w/o evaporative
w/ evaporative
Annual ized
Costs
($ Millions)3
91
62
52
183
146
199
199
Annual ized
Benzene
Incidence
Reduction15
0.06
0.8
0.4
1.9
1.1
1.4
1.7
Costs
($ Mil lions
per Benzene Cancer
Incidence Avoided)
1,564
75
126
95
128
138
120
Includes control equipment and annual enforcement costs.
>
Incidence reduction after controls. Before-control annualized incidence:
Stage I = 0.18, Vehicle Refueling = 4.09.
1-34
-------
TABLE 1-13. BENZENE REGULATORY COSTS PER CANCER INCIDENCE
AVOIDED (ASSUMING VOC BENEFITS)a
I
OJ
OT
Regulatory Strategy
(In-use efficiency)
Stage I
Stage II-NA (87%)
Stage II-NA (56%)
Stage 1 1 -Nation (86%)
Stage 1 1 -Nation (56%)
Onboard (92%):
w/o evaporative
w/ evaporative
Annual i zed
Incidence
DpHurl" i nn
(Benzene)
0.06
0.8
0.4
1.9
1.1
1.4
1.7
Costs ($
$250/Mg
629
57
104
76
107
120
83
Mil lions
Assumi ng
$500/Mg
Ob
39
81
57
86
93
43
per Bz Cancer
VOC Benefit
$1000/Mg
0
4
35
19
44
45
0
Incidence
Value of:
$1 500/Mg
0
0
0
0
2
0
0
Avoided)
$2000/Mg
0
0
0
0
0
0
0
aCost per cancer incidence avoided =
Annualized Costs (from Table 1-12) - Annualized VOC Benefits (from Table 1-11)
Annualized Incidence Reduced
bZeros indicate that the VOC benefits outweigh the costs.
-------
Table 1-14. BENZENE AND GASOLINE VAPORS COST PER CANCER INCIDENCE AVOIDED3
(USING RAT DATA UNIT RISK FACTOR FOR GAS VAPORS)
i
C*J
CTl
•trategy Annual (zed Annuallzed
Incidence Incidence
Reduction Reduction
clency) (Bz+GV)b Jflz+GV)b
Maximum Plausible
Likelihood0 Upper
Estimate L1ra1tc
(M.L.E.) (P.U.L.)
1-1 1.9
(87%) 6.1 10
(56%) 3.1 5.1
1on (86%) 13 22
.(on (56%) 7.8 13
\
poratlve 10 17
porative 14 23
Costs ($ Million Per Bz + GVb Cancer Case Avoided)
Assuming VOC Benefit Value of:
$250/Mg
M.L.E. P.U.L.
33 19
8 5
14 8
11 7
16 9
16 10
10 6
$500/Mg
M.L.E. P.U.L.
Od 0
5 3
11 6
8 5
13 8
13 8
5 3
1 ,000/Mg
M.L.E. P.U.L.
0 0
1 0
5 3
3 2
6 4
6 4
0 0
$l,500/Mg
M.L.E. P.U.L.
0 0
0 0
0 0
0 0
0 0
0 0
0 0
$2 ,000/Mg
M.L.E. P.U.L.
0 0
0 0
0 0
0 0
0 0
0 0
0 0
Stage 1
Stage II
Stage II
Stage II
Stage II
Onboard (92'i)
Cost per cancer Incidence avoided =
Annual Lzed Costs (from Table 1-12) - Annuallzed VOC Benef Us^ j_f rom Table 1-11)
Annuali zed Tncldence Re3uce3
b
Hz • GV = Benzene plus gas vapors.
Calculated using the maximum likelihood estimate unit risk factor and plausible upper limit unit risk factor
Zeros indicate that the VOC benefits outweigh the costs.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO.
EPA-450/3-84-012b
I. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of Air Pollution Regulatory Strategies
for the Gasoline Marketing Industry - Executive
Summary
5. REPORT DATE
July 1984
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND AOORESS
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. CONTRACT/GRANT NO.
68-02-3060
12. SPONSORING AGENCY NAME AND AOORESS
Assistant Administrator for Air and Radiation
U.S. Environmental Protection Agency
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
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 atmos-
phere 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 ethylene
dichloride (EDC), ehtylene dibromide (EDB), and gasoline vapors, for which there is
animal evidence of carcinogencity. This report contains a summary of the analysis
conducted concerning the health, emission, cost, and economic impacts of several
regulatory strategies for addressing organic compound emissions from gasoline
marketing sources. (The full report is contained in EPA Document EPA-450/3-84-12a.)
The regulatory strategies evaluated are: (1) service station controls (Stage II) for
vehicle refueling emissions only in areas requiring additional VOC control to attain
the national ozone ambient standard; (2) service station controls (Stage II) for
vehicle refuleing emissions on a nationwide basis; (3) Onboard vehicle controls for
vehicle refuleing emissions on a nationwide basis; (4) bulk terminal, bulk plant,
and service station storage tank controls on a nationwide basis; and (5) various
combinations of these alternatives.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Gasoline
Air Pollution
Pollution Control
Stationary Sources
Mobile Sources
Volatile Organic Compounds Emissions
Benzene Emissions
Air Pollution Control
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (Tins Report/
Unclassified
21. NO. OF PAGES
50
Unlimited
L
20. SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
-------
------- |