WASTE OIL STORAGE
FINAL DRAFT REPORT
by
Suzanne C. Metzler, Nicholas S. Artz,
Jacob E. Beachey, and Robert G. Hunt
Franklin Associates, Ltd.
Prairie Village, Kansas 66206
and
PEDCo Environmental, Inc.
Cincinnati, Ohio 45246
FRANKLIN ASSOCIATES, LTD.
Engineering/Environmental/Management Consultants
8340 Mission Road, Suite 101
Prairie Village, Kansas 66206
-------
WASTE OIL STORAGE
FINAL DRAFT REPORT
by
Suzanne C. Metzler, Nicholas S. Artz,
Jacob E. Beachey, and Robert G. Hunt
Franklin Associates, Ltd.
Prairie Village, Kansas 66206
and
PEDCo Environmental, Inc.
Cincinnati, Ohio 45246
Project Officer: Michael J. Petruska
Hazardous and Industrial Waste Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF SOLID WASTE (WH-565)
WASHINGTON, D.C. 20460
January 1984
-------
CONTENTS
Figures iii
Tables iv
Acknowledgment vi
Executive Summary vii
1. Introduction 1-1
2. Technological Characterization of Waste Oil Storage 2-1
2.1 Characterization of waste oil storage 2-2
2.2 Characterization of waste oil losses 2-13
2.3 Summary 2-38
References for Section 2 2-42
3. Environmental Fate of Waste Oil Lost From Oil Storage
Sites 3-1
3.1 Mechanisms of waste oil movement 3-1
3.2 Above-ground tanks 3-8
3.3 Below-ground tanks 3-15
3.4 Spills from containers and drums 3-17
3.5 Summary 3-38
References for Section 3 3-42
Appendix A Derivation of Estimated Failure Probabili-
ties in Below-Ground Waste Oil Storage
Tanks A-l
Appendix B Oil Infiltration Into Soil Using Green-
Amp t Model B-l
Appendix C Derivation of Spill Penetration Depth
Equation C-l
11
-------
FIGURES
Number Paqe
I Breakdown of Waste Oil Stored by Estimated
Volume of Oil at Each Source ix
2-1 Estimated Quantity of Stored Waste Oil by Type
of Storage 2-5
2-2 Storage Practices by Source in Millions of
Gallons Stored 2-9
3-1 Spill Area Versus Spill Volume for Oil Spills
Without Secondary Containment 3-19
3-2 Volume of Saturated Soil Versus Spill Volume
for a Soil Porosity of 50 Percent 3-22
3-3 Volume of Contaminated Soil Versus Soil Porosity
for a Spill Volume of 220 Gallons 3-23
3-4 Depth of Spill Penetration Versus Soil Porosity
for a Catastrophic Spill of 220 Gallons 3-26
3-5 Geometry of Cone Assumed for Calculating Pene-
tration Depths 3-28
3-6 Depth of Spill Penetration Versus Soil Porosity
for Three Cone Angles for Residual Saturation
of 0.1 3-29
3-7 Depth of Spill Penetration Versus Soil Porosity
for Periodic Spills of One-Half Gallon Each 3-32
3-8 Depth of Spill Penetration of Soil Porosity for
Three Cone Angles for Residual Saturation of
0.1 (Light Oil) 3-33
3-9 Depth of Spill Penetration Versus Soil Porosity
for Periodic Spills of One Pint Each in Four
Spill Areas 3-35
3-10 Depth of Spill Penetration Versus Soil Porosity
for Three Cone Angles for a Residual Saturation
of 0.1 (Typical for Light Oil) 3-36
111
-------
TABLES
Number Page
I Waste Oil Storage Summary xi
II Annual Waste Oil Losses xiv
III Maximum Expected Waste Oil Losses From Different
Size Tanks xv
IV Time Required for Spilled Oil to Contaminate
30.5 Centimeters (12 Inches) of Soil xviii
V Oil Migration Time From an Above-Ground Tank
Water Table 100 Centimeters (39.4 Inches) Deep xviii
VI Oil Migration Time From a Below-Ground Tank to
a Water Table 100 Centimeters (39.4 Inches)
Deep xx
2-1 Estimated Quantities of Stored Waste Oil by
Source and Type of Storage 2-3
2-2 Probability of Losses From Above-Ground Tank
Facilities for Hazardous Liquid Storage 2-22
2-3 Storage Spill Percentages From EPA and U.S.
Coast Guard Data Bases 2-25
2-4 Storage Spill Incidents From EPA and U.S.
Coast Guard Data Bases 2-35
2-5 Size Distribution of Spills Reported to EPA 2-36
3-1 Time Required for Penetration of Spilled Oil
to a Depth of 30.5 cm (12 Inches) for Various
Soil Types 3-10
3-2 Above-Ground Tank Sizes and Typical Oil Levels 3-12
3-3 Oil Migration Time From an Above-Ground Tank
to a Water Table 100 Centimeters Deep 3-13
3-4 Oil Migration Time From an Above-Ground Tank
to a Water Table 1,000 Centimeters Deep 3-14
IV
-------
TABLES (continued)
Number Page
3-5 Oil Migration From a Below-Ground Tank to a
Water Table 100 Centimeters Deep 3-16
3-6 Range of Values of Porosity 3-24
v
-------
ACKNOWLEDGMENT
This report was prepared for the U.S. Environmental Protec-
tion Agency Office of Solid Waste under Contract No. 68-02-3173.
The EPA Project Officer was Michael J. Petruska. PEDCo Environ-
mental was the prime contractor; Franklin Associates Limited was
the major subcontractor and author of this report.
Franklin Associates gratefully acknowledges the assistance
in many forms rendered by Michael J. Petruska, Project Officer,
and by Penelope Hansen, Branch Chief. We also gratefully acknowl-
edge the assistance given by Catherine Jarvis, Project Manager at
PEDCo Environmental, Inc., and Marty Phillips, technical editor
at PEDCo.
VI
-------
EXECUTIVE SUMMARY
The purpose of this study was to evaluate the potential for
environmental contamination from waste oil storage systems. The
report findings and methodologies used are summarized herein.
Generally, stored waste oil falls into one of two categories:
automotive/diesel or industrial. Automotive/diesel waste oils
consist primarily of crankcase oils generated by cars, trucks,
and other vehicles. Because these oils are usually consistent in
composition and levels of contamination, increased contamination
as a result of mixing the oils from different sources is not
likely. The contaminants that are common in these oils are the
metals barium, chromium, and lead. Lead is still the contaminant
of greatest concern, despite the fact that the decrease in the
use of leaded gasoline has lessened its significance. These
waste oils also contain some potentially hazardous polynuclear
aromatic compounds (PNA's).
Industrial waste oils, as the name implies, are generated by
industry. They include metal working, hydraulic process, elec-
trical, refrigeration, and turbine oils. These waste oils can
contain a wide range of potentially hazardous constituents,
including halogenated solvents, aromatic solvents, polychlori-
nated biphenyls (PCB's), and heavy metals (cadmium, chromium, and
VII
-------
zinc). The levels of these contaminants range from very high to
essentially zero.
Waste oil is stored in below-ground tanks, above-ground
tanks, and 55-gallon drums. Most of the tanks now in use are
made of unprotected steel, but this practice is changing, par-
ticularly for below-ground tanks. For example, to avoid cor-
rosion problems, the major oil companies are replacing most of
their below-ground steel tanks that fail with fiberglass units.
Tank sizes vary widely, but the vast majority of them (both
below-ground and above-ground) hold 500 gallons or less. Some
facilities, however, have 5,000- to 10,000-gallon tanks, and
*
collector-processors of waste oil occasionally have tanks that
hold a few hundred thousand gallons.
WASTE OIL STORAGE FACILITIES
Waste oil is stored by both generators and collector-
processors (Figure I). Automotive/diesel oil is generated by
service stations, automotive repair shops, automotive dealers,
fleet maintenance garages, and a miscellaneous group classified
as "others." This combined group of generators stores an esti-
mated 64 million gallons of waste automotive/diesel oil. Indus-
trial generators store an estimated 41.7 million gallons of waste
industrial oil, and collectors and collector-processors store an
estimated 67.8 million gallons of automotive/diesel and indus-
trial waste oil.
Collector-processors both collect and process waste oil.
Collectors collect waste oil, but they do not process it.
vni
-------
Figure I . Breakdown of waste oil stored by estimated volume of oil
(In millions of gallons) at each source.
ix
-------
Because the waste oil stored by collectors and collector-
processors includes both industrial oils and automotive/diesel
oils, various types of industrial oils may be mixed together and
automotive/diesel oils may be mixed with industrial oils. Such
mixing increases the potential for a greater variety of contami-
nants in a given source of collected waste oil. Some collectors
and collector-processors segregate their waste oils by source,
but most practice some mixing. Cross-contamination can also
occur as a result of storing one type of oil in a tank that
previously contained a different type.
The proportional relationships between the quantities of
stored waste oil and the number of facilities and storage methods
vary greatly. Whereas more than 65 percent of the facilities use
below-ground tanks, only 49.4 percent of total waste oil is
stored in below-ground tanks. On the other hand, only an esti-
mated 7.1 percent of the facilities store waste oil in above-
ground tanks, but this group accounts for 43.9 percent of the
waste oil stored. For drum storage, the situation is reversed;
27.5 percent of the storage facilities use drums, but drum storage
accounts for only 6.7 percent of the waste oil stored. The total
amount of waste oil stored is estimated to be 173 million gallons
(Table I) .
x
-------
TABLE I . WASTE OIL STORAGE SUMMARY6
Type of storage
Below-ground tanks
Subtotal
Above-ground tanks
Subtotal
Drums
Total
Average
tank sizes,
gallons
500
600
5,000
10,000
250
500
5,000
10,000
50,000
55
Number of
facilities
269,000
42,000
25,000
70
336,070
1,500
33,400
800
570
255
36,525
141,490
514,085
Number of
tanks and drums
269,000
42,000
25,000
70
336,070
1,500
33,400
1,600
1,070
2,550
40,120
424,470
800,660
Storage quantity,
106 gallons
40.5
7.6
37.5
0.2
85.8
0.1
5.0
4.0
3.2
63.8
76.1
: 11.6
173.5
Storage quantity,
% of total
23.4
4.4
21.6
0.0
49.4
0.0
2.9
2.3
1.9
36.8
43.9
6.7
100.0
a Based on numerous contacts with waste oil processors and information developed by Development
Planning and Research Associates, Inc., of Manhattan, Kansas.
-------
DETERMINATION OF WASTE OIL LOSSES
Frequency of Losses
Because insufficient data were available to serve as a basis
for a direct assessment of the frequency of waste oil losses,
alternative methods were used to estimate loss probabilities.
For determination of losses from above-ground tanks and
drums, a previously developed "fault-tree" analysis proved to be
useful. This analysis provided estimated failure probabilities
for the various components in an above-ground storage system.
These probabilities were used to estimate above-ground spills
from typical above-ground waste oil storage systems.
The results of recently performed research into leakage from
below-ground gasoline tanks were used to estimate the frequency
of losses in below-ground waste oil tanks. Leaks in below-ground
tanks can result from external or internal corrosion, piping
failure, tank design and fabrication faults, or improper tank
installation, but external corrosion is by far the most common
cause. Over three-fourths of all unprotected steel tanks will
experience localized external corrosion, which leads to an accel-
erated rate of failure.
Two approaches were used to estimate the probability of
leaks in below-ground waste oil tanks: 1) the use of a mathe-
matical model developed in an American Petroleum Institute (API)
study to estimate tank age failure under assumed soil conditions,
and 2) the use of recent data compiled by Warren Rogers Associ-
ates on the expected age of underground gasoline tanks at time of
XII
-------
failure under various soil conditions encountered at automotive
service stations. Both approaches were useful only in predicting
leaks caused by external corrosion. In each case, a uniform tank
age distribution of 0 to 20 years was assumed.
Estimated probabilities and frequencies of waste oil losses
from the three modes of storage (below-ground tanks, above-ground
tanks, and drums) indicate that the incidence of losses is far
greater from below-ground waste oil tanks than from either of the
other two storage modes (see Table II). The probability of
leakage in a below-ground tank has been conservatively estimated
to be between 12 and 14 percent, based on an assumed uniform tank
distribution of 0 to 20 years. This percentage translates into
43,500 leaks per year from below-ground tanks.
Considerable evidence indicates that the assumed age range
(0 to 20 years) is conservative and that a significant number of
waste oil tanks in current use are more than 20 years old. The
probability of failure in these older tanks is believed to exceed
50 percent. In areas of the country where below-ground tanks are
consistently exposed to moisture-saturated soil, the probability
of leakage is much higher than the 12 to 14 percent estimate.
Based on engineering judgment, about 25 to 35 percent of the
underground waste oil tanks are believed to be leaking in some
areas, especially where a large number of tanks over 20 years old
are still in service.
Xlll
-------
TABLE II. ANNUAL WASTE OIL LOSSES
Storage mode
Large above-ground tanks
Small above-ground tanks
Total above-ground tanks
Total drums
Total below-ground tanks
Annual probability
of loss, percent
2.9
1.6
1.7
1.1
13.0b
Annual number
of loss incidents
150
550
700
4,500
43,500b
3 Reflects all estimated incidents of uncontained losses from tanks and drums
indicated in Table I. The losses shown for below-ground tanks include
ongoing or continuous leaks.
These numbers are the averages of an estimated range.
xiv
-------
Magnitude of Losses
Table III presents a summary of losses typically expected
from various size tanks. Losses from above-ground tanks are
based on the maximum quantities expected to be stored in these
tanks. Losses from below-ground tanks are based on the expected
maximum quantities received in these tanks over a period of time.
It should be noted the maximum expected loss values on this table
represent worst-case scenarios.
TABLE III. MAXIMUM EXPECTED WASTE OIL LOSSES FROM DIFFERENT SIZE TANKS
Storage mode
Below-ground tanks0
Above-ground tanks
Tank size,
gallons
500
5,000
250
500
5,000
10,000
Average loss
if tank is
emptied, gallons
75/incident
150/incident
2,500/incident
5,000/incident
Maximum .
expected loss,
gallons
375/month
3,750/month
188/ incident
375/incident
3, 750/ incident
7, 500/ incident
Assumes smaller tanks are 30 percent full, on the average, and larger
tanks are 50 percent full.
Worst-case scenarios, and losses from widely used tanks without secondary
containment.
C Losses from below-ground tanks are shown in gallons/month because they
usually emanate from slow, continuous leaks.
Nontransportation storage spills reported to the EPA and the
Coast Guard between 1974 and 1980 primarily involved losses of
less than 250 gallons. Nearly 30 percent involved less than 50
gallons. Only 21 percent of the estimated spills from above-
ground waste oil tanks involve the larger tanks (capacities
xv
-------
greater than 1000 gallons). The great majority of above-ground
waste oil tanks have capacities around 500 gallons and are esti-
mated to be only 30 percent full on the average. Thus, the
average total loss from one of these tanks would not be more than
150 gallons. Because many tank spills are probably stopped
before the tank is emptied, spills from above-ground waste oil
tanks are typically less than the average quantity contained in
the tanks.
Based on reported underground gasoline losses, the probability
of a waste oil loss of 3750 gallons per month from a below-ground
waste oil tank appears to be remote. Some of the largest reported
gasoline losses are known to be less than 3750 gallons per month,
and these were from larger-capacity tanks. For example, a 30,000-
gallon underground gasoline loss in New York is believed to have
averaged no more than 2500 gallons per month from two 4000-gallon
tanks. Inasmuch as total gasoline storage in service stations is
greater than total waste oil storage and tank sizes are generally
much larger, the few documented gasoline losses of 3750 gallons
or more per month suggest that a waste oil loss of this size
would be very unlikely.
ENVIRONMENTAL IMPACT OF WASTE OIL LOSSES
Waste oil that is lost as a result of spills or leaks may
contaminate the land, groundwaters, surface waters, and even the
air. This report focuses on an evaluation of soil contamination.
xvi
-------
Evaluation was limited to the rate or depth of penetration of
waste oil into the soil.
A worst-case scenario approach was selected for determina-
tion of the environmental impact of losses from waste oil storage
systems. This scenario describes the worst conditions for envi-
ronmental contamination that reasonably can be expected to occur.
If environmental contamination is low under these conditions,
more typical situations are likely to result in little or no
contamination.
Above-Ground Tanks
Environmental contamination from waste oil loss from above-
ground tanks can result in seepage of spilled oil from the im-
pounded area around the storage tank or from leaks in the tank
bottom. The time required for spilled oil to contaminate a depth
of 30.5 centimeters (12 inches) of soil depends on the type of
soil present within the secondary containment (impoundment) area,
as shown in Table IV. It is predicted that a spill with an
average depth of 30.5 centimeters within the secondary contain-
ment area will penetrate a typical sandy soil to a depth of 30.5
centimeters in only a few minutes. The amount of oil lost depends
on soil porosity, but it will certainly be more than 25 percent.
Because cleanup times range from an hour to several days, much of
the oil will be lost before it can be cleaned up; thus, a secondary
containment system with a sandy soil bottom is virtually useless.
The rate of oil seepage is much slower if the soils in the sec-
ondary containment area are silt or clay, and expeditious cleanup
of spilled oil lessens the loss considerably.
xvii
-------
TABLE IV. TIME REQUIRED FOR SPILLED OIL
TO CONTAMINATE 30.5 CENTIMETERS (12 INCHES) OF SOIL9
Soil type
Time
Clay
Silt
Sand
37.6 to 772 years
30.2 to 16.9 days
9.03 to 13.1 minutes
a Calculations assume average soil conditions and an
average spill depth of 30.5 centimeters.
TABLE V. OIL MIGRATION TIME FROM AN ABOVE-GROUND TANK
WATER TABLE 100 CENTIMETERS (39.4 INCHES) DEEPa
Soil type
Clay
Silt
Sand
Time
286 to 877 years
12.0 to 17.1 days
12.3 to 12.8 minutes
Calculations assume average soil conditions and an
average oil depth in the tank of 500 centimeters.
xvi 11
-------
Leaks from the bottom of above-ground tanks (Table V) also
pose severe problems if the soil under and around the tank is
sandy. In the event of a major rupture, oil may reach a shallow
(100 centimeters or 39 inches deep) water table in a matter of
*
minutes. It would take several days for an oil to reach the
groundwater table if the tank were placed on a silty soil.
Regular monitoring of oil levels within the tank is necessary to
assure that a failure does not go undetected.
Below-Ground Tanks
Failure of an underground tank will result in seepage of oil
into the surrounding soils. Because leaks are not visible from
the surface, they are likely to go undetected for a much longer
period than those from above-ground tanks. Failure of a tank
placed in an average sandy soil may result in oil migration to a
water table 100 centimeters (39 inches) deep in less than an hour
(Table VI). An average silty soil may lengthen migration times
to 1 or 2 months. Because of the long periods of time that may
elapse before detection of oil loss from a below-ground tank, the
potential for environmental contamination from a below-ground
tank in a silty soil is still significant. Clay is the only type
of soil that is believed to be safe for burying below-ground
tanks, and this belief may be overly optimistic. Recent research
Oil that is leaking because of tank bottom failure migrates
much more rapidly than spilled oil because of the head exerted
by the oil within the storage tank.
xix
-------
indicates that interaction of some organics with clay can greatly
increase its permeability.
TABLE VI. OIL MIGRATION TIME FROM A BELOW-GROUND TANK TO
A WATER TABLE 100 CENTIMETERS (39.4 INCHES) DEEP3
Soil type
Clay
Silt
Sand
Time
365 to 2598 years
22.4 to 52.1 days
34.9 to 39.3 minutes
a Calculations assume average soil conditions and an
average oil depth in the tank of 120 centimeters.
Containers and Drums
Spills from containers and drums will result in some seepage
of oil into soils. Depth of oil penetration was evaluated for
both catastrophic spills and sequential small spills. Catastrophic
spills tend to spread over a large surface area. Soil penetration
varies with soil type and the type of oil spilled. A light oil
spilled on a gravel surface results in the deepest oil migration.
Sequential small spills do not spread over such a large area, but
the repeated spillage usually occurs in the same location. The
result is a deeper localized penetration of oil, even though the
total volume of oil may be small.
In general, groundwater contamination due to spills from
containers and drums should be minimal. Because cleanup of these
spills is typically minimized, however, some soil contamination
can be expected, and leaching of some oil components from oil-
contaminated soil may occur.
XX
-------
SECTION 1
INTRODUCTION
The U.S. Environmental Protection Agency's Office of Solid
Waste is funding a study to assess the environmental impact of
waste oil as a fuel, waste oil as a dust suppressant, and the
storage of waste oil. Three separate reports completed as part
of this study characterize each of the practices. This report
presents an evaluation of waste oil storage practices in the
United States.
Approximately 4.3 billion liters (1.1 billion gallons) of
waste oil are generated each year. Regardless of its end use,
virtually all waste oil is stored at some time. The composition
of waste oil is highly variable, and much of it contains poten-
tially hazardous contaminants. The contaminants in waste oil are
highly dependent on its source. Some of those found in waste oil
include heavy metals, particularly lead; organic solvents such as
benzene, xylene, and toluene; and chlorinated organics such as
trichloroethane, trichloroethylene, and polychlorinated biphenyls
(PCB's).
The potential for losing waste oil from storage sites (e.g.,
through leaks, spills, and evaporation) presents the possibility
of the release of hazardous materials into the environment. This
1-1
-------
study is designed to assess the environmental impact of such
releases from waste oil storage sites.
The report is divided into two primary parts. The first
part (Section 2) presents a technological characterization of
waste oil storage in the United States, and the second part
(Section 3) examines and summarizes the fate of waste oil that is
released into the environment as a result of leaks and spills at
storage sites.
The technological characterization includes discussions on
the quantity and sources of waste oil, the various types of
storage facilities, the composition of stored waste oil, the
frequencies and volume of waste oil losses, and the failure
mechanisms and their relative importance.
The section addressing environmental impact examines the
rate and degree of contamination from typical and worst-case
waste oil releases from storage sites. The three mechanisms of
oil movement from the storage sites (i.e., evaporation, surface
runoff, and seepage into the soil) are discussed. The primary
mechanism of spilled oil movement is seepage, which is examined
in detail. Mathematical models used to estimate environmental
pollution consider the spill conditions, including oil and soil
types, distance to the groundwater, and time from spill to detec-
tion.
1-2
-------
SECTION 2
TECHNOLOGICAL CHARACTERIZATION OF WASTE OIL STORAGE
Included in this section are discussions on the quantity and
sources of waste oil, the various types of storage facilities,
the composition of stored waste oil, the frequencies and volume
of waste oil losses, and the mechanisms and relative importance
of storage failures.
This technological characterization of waste oil storage is
based on the limited amount of data available. Data on waste oil
leaks from underground storage tanks are especially sparse because
these leaks usually are not reported to any central agency. The
data that do exist generally belong to the private sector and are
not available to the public. Also, underground leaks often go
undetected (and thus undocumented) for long periods of time. Of
necessity, some of the data presented here have been derived from
the input of a combination of information sources. We believe
these derived data are reasonable, however, and in combination
with the other data in the report, will provide a foundation for
determining potential environmental effects of waste oil storage.
Generally, stored waste oil falls into one of two categories:
automotive/diesel or industrial. Automotive/diesel waste oils
consist primarily of crankcase oils generated by cars, trucks,
2-1
-------
and other vehicles. Industrial waste oils, as the name implies,
are generated by industry.
Because automotive/diesel waste oils are usually consistent
in composition and levels of contamination, there is little
probability of increased contamination as a result of mixing the
oils from different sources. These oils are likely to be con-
taminated with the heavy metals barium, chromium, and lead. Lead
is still the contaminant of greatest concern, despite the fact
that the decrease in the use of leaded gasoline has lessened its
significance. These waste oils also contain some polynuclear
aromatic compounds (PNA's), which are potentially hazardous.
Industrial waste oils include metal working, hydraulic
process, electrical, refrigeration, and turbine oils. These
waste oils can contain a wide range of potentially hazardous
constituents, including halogenated solvents, aromatic solvents,
polychlorinated biphenyls (PCB's), and heavy metals (cadmium,
chromium, and zinc). The levels of these contaminants range from
very high to essentially none.
2.1 CHARACTERIZATION OF WASTE OIL STORAGE
Waste oil is stored in below-ground tanks, above-ground
tanks, drums, and some surface impoundments. Since the use of
surface impoundments (once a major factor in waste oil storage)
has been declining rapidly, discussions in this section concern
only tank and drum storage.
Estimates have been made of the quantities of waste oil
stored in tanks and drums at various facilities (Table 2-1 and
2-2
-------
TABLE 2-1. ESTIMATED QUANTITIES OF STORED WASTE OIL
BY SOURCE AND TYPE OF STORAGE9
Waste storage establishments
With below-ground tank storage
Service stations
Automotive repair shops
Automotive dealers
Fleet maintenance garages
Industrial generators
Railroads
Subtotal
With above-ground tank storage
Airplane service facilities
Fleet maintenance garages
Collectors
Col 1 ectros/processors
Marine service facilities
Automotive repair shops
Railroads
Subtotal
With drum storage
Service stations
Automotive repair shops
Automotive dealers
Fleet maintenance garages
Collection centers
Airplane service facilities
Industrial generators
Railroads
Subtotal
TOTAL - Tanks and Drums
Number of
facilities
113,000
93,000
63,000
42,000
25,000
70
336,070
1,500
2,600
800
255
500
30,800
70
36,525
6,000
31,900
21,000
27,000
300
4,000
51,000
290
141,490
514,085
Assumed
average
tank size,
gallons
500
500
500
600
5,000
10,000
250
500
5,000
50,000
10,000
500
10,000
55
55
55
55
55
55
55
55
Assumed
average
number
of units
1
1
1
1
1
1
1
1
2
10
2
1
1
3
3
3
3
3
3
3
3
Quantity
of waste
oil stored,"
106 gallons
17.0
14.0
9.5
7.6
37.5
0.2
85.8
0.1
0.4
4.0
63.8
3.0
4.6
0.2
76.1
0.5
2.6
1.7
2.2
<0.1
0.3
4.2
<0.1
11.6
173.5
Based on numerous contacts with waste oil processors and information devel-
oped by Development Planning and Research Associates, Inc., of Manhattan,
Kansas.1"19
Assumes tanks 30 percent full on average; for collectors and collectors/
processors, tanks assumed to be 50 percent full. Drums assumed to be 50
percent full.
2-3
-------
Figure 2-1). Much of this information was compiled by Devel-
opment Planning and Research Associates, Inc., of Manhattan,
1 2
Kansas. ' Some was gathered through personal communications
3-19
with various waste oil processors.
An estimated 173 million gallons of waste oil is stored in
tanks and drums. Although storage capacity is more than twice
this amount, the competitive nature of the waste oil business
normally results in the waste oil being collected long before it
11 1 VI
reaches storage capacity. Reported inventories of waste oil
processors vary from near capacity to almost none, depending on
the processed product and the season of the year.
Less than 7 percent of the estimated total quantity of waste
oil is stored in drums. Although many facilities use drums to
store waste oil, the quantities produced by these facilities are
usually small. Industrial generators tend to use drums only if
waste volumes are less than 500 gallons per month.
About half of the total waste oil is stored in below-ground
tanks, and nearly two-thirds of the 500,000 facilities that store
waste oil use this storage method. Although only 7 percent of
these facilities use above-ground tanks from waste oil storage,
these tanks account for more than 40 percent of the stored oil.
Collector-processors of waste oil use above-ground storage almost
exclusively, and the quantities they store are enormous compared
with that stored by other establishments.
Stored waste oil generally falls into one of two categories:
automotive/diesel oil or industrial oil. About 40 percent of
2-4
-------
IUU
o
S. 80
B
0
E 70
^
o* 60
2
0
"> 50
5 40
LLJ
| 30
u.
0
>_ 20
P
1 10
_
-
^
-
"
-
-
-
85.8
£:-S%S&^££
?S:S^SS?^
BELOW-
GROUND
STORAGE
76.1
ABOVE-
GROUND
STORAGE
_
-
-
—
-
-
_
11.6
"
DRUMS
Figure 2-1. Estimated quantity of stored waste oil by type of storage,
2-5
-------
the total stored waste oil (excluding that stored by collectors
and collector-processors) comes from industrial oil sources; the
remaining 60 percent is from automotive/diesel oil establishments.
The waste oil stored by collectors and collector-processors
represents both categories, and after collection, oils from the
two categories are sometimes mixed.
Collector-processors store far more waste oil (63.8 million
gallons) than any of the other facilities. Storage is typically
in above-ground tanks ranging from 10,000 to more than 250,000
gallons in capacity. Based on the information obtained from
these references, the average storage capacity per facility is
believed to be at least 500,000 gallons.
Industrial generators rank second in the amount of waste oil
stored. Most of the total quantity of this industrial waste oil
is stored in below-ground tanks, even though a larger proportion
of industrial facilities uses drums. Tank sizes vary considerably,
but the average capacity is estimated to be 5000 gallons.
Automotive repair shops outnumber any other type of waste
oil storage facilities. Most store in below-ground tanks, but
some use above-ground tanks or drums. The typical tank size is
500 gallons. Service stations have more below-ground tanks than
any other type of waste oil storage facilities, but a few use
drums for storage.
With the exception of collectors, collector-processors, and
marine service facilities, most sources using tanks to store
waste oil have only one tank.
2-6
-------
Below-ground waste oil storage tanks with capacities greater
than 1000 gallons are used primarily by industrial generators and
railroads. Above-ground waste oil storage tanks with capacities
greater than 1000 gallons are used primarily by waste oil collec-
tors, collector-processors, marine service facilities, and rail-
roads.
Data gathered on storage of waste oil at marine service
facilities are somewhat conflicting. The information presented
in Table 2-1 reflects that reported by Development Planning and
1 2
Research Associates, Inc. ' Discussions with a few identified
marine service facilities, however, suggest that they often store
20 — 23
waste oil together with large quantities of water. Total
storage quantities of mixtures of oil and water may be in mil-
lions of gallons.
Although the storage quantities of waste oil shown in Table
2-1 are not intended to include substantial amounts of other sub-
stances mixed with oil, the quantities for marine service facili-
ties seem questionable, since it is not clear what is included
and the storage capacities reported in recent discussions are
20-23
much higher than indicated.
2.1.1 Composition of Stored Waste Oil
The comprehensive data base established as part of this
study characterizes the composition of various types of waste
oil. Ideally, this data base would be used to predict the most
likely composition of the waste oil in each of the storage units
for each facility type, and this is done (to some extent) in a
2-7
-------
supplementary report that deals strictly with waste oil composi-
tion issues. This report, however, presents only a general
assessment of waste oil composition as it relates to the various
sources of oil.
As in the case of its storage, the composition of waste oil
falls into two basic categories: automotive/diesel oils and
industrial oils. The composition of automotive/diesel oils is
fairly consistent; the major variable contaminant is lead, which
is directly related to whether the vehicles generating the oil
used leaded or unleaded fuel. The composition of industrial
oils, on the other hand, varies considerably. Among the several
types of industrial oils are metal working, hydraulic process,
electric, refrigeration, and turbine oils. Each is used in a
unique environment that contributes its own contaminants to the
oil. The levels and types of these contaminants differ from
industry to industry. All waste oils contain some polynuclear
aromatic compounds (PNA's) as part of their basic hydrocarbon
makeup. These PNA's also present a hazard potential of some
concern.
Both automotive/diesel and industrial oils are stored in one
of the three types of storage units already discussed: below-
ground tanks, above-ground tanks, or drums. The flow diagram in
Figure 2-2 shows the estimated storage for both categories of
waste oil. The following subsections discuss qualitatively the
composition of the waste oil most likely to be stored in each of
these storage units.
2-8
-------
ONSITE STORAGE
10
vo
WASTE OIL SOURCES
AUTOMOTIVE/DIESEL,
64 0
INDUSTRIAL,
41.7
-
*~
ABOVE-GROUND TANKS
AUTOMOTIVE/DIESEL, 8.3
DRUMS
AUTOMOTIVE/DIESEL, 7.4
INDUSTRIAL, 4.2
BELOW-GROUND TANKS
AUTOMOTIVE/DIESEL, 48.3
INDUSTRIAL, 37.5
rni i FETOR'S' rni i mop-
PROCESSORS STORAGE
ABOVE-GROUND TANKS
AUTOMOTIVE/DIESEL
** AND 1NUUSIK1AL,
67.8
Figure 2-2. Storage practices by source in millions of gallons stored (derived from Table 2-4),
-------
2.1.1.1 Below-Ground Tanks—
This type of storage unit accounts for greater overall
quantities of waste oil storage than the other units combined.
Most (more than 90 percent) of these units are small-capacity
tanks (about 500 to 600 gallons) used to store waste crankcase
oil generated at service stations and other automotive repair and
service shops. Most of the remaining underground tanks represent
the much larger tanks used to store industrial waste oil. Despite
their smaller number, these industrial tanks account for nearly
half of the waste oil stored underground.
Most of the waste oil stored in below-ground tanks is segre-
gated; that is, each tank usually contains oil generated by a
single source. This is particularly true for the crankcase oils,
which are relatively consistent with respect to composition and
contamination levels. Although crankcase oils stored in below-
ground tanks contain some contaminants of concern (primarily
heavy metals), they are not likely to be contaminated with other
unknown materials (e.g., solvents). Because all waste crankcase
oils are usually similar in composition, there also is little
probability of contamination resulting from mixing oil from one
tank with that from another tank.
Industrial waste oils that are stored underground are far
more subject to variability within a single storage tank, from
plant to plant, and from industry to industry. Consistency can
vary at a single site because more than one oil or oily waste is
generated there. Also, the variability in the nature of indus-
2-10
-------
trial processes among different industries, or even from plant to
plant within a single industry, can produce significant differ-
ences in waste oil composition.
Based on an evaluation of approximately 400 composition
analyses from various sources, automotive/diesel oils are likely
to be contaminated by some heavy metals with potentially hazard-
ous characteristics. Lead is still the primary metal of concern,
despite the decrease in its significance as a result of the trend
away from leaded gasoline. Other heavy metals of concern are
barium and chromium.
Industrial oils can contain a much wider range of poten-
tially hazardous constituents, including heavy metals and organic
compounds (such as halogenated solvents, aromatic solvents, and
polychlorinated biphenyls). The quality of industrial waste oils
ranges from very clean to highly contaminated.
2.1.1.2 Above-Ground Tanks—
Most above-ground storage tanks are located at generator
sites, but most of the oil stored in this manner is held by
collectors or collector-processors, whose tanks are much larger
than those at generator sites. The general discussion of indus-
trial versus automotive/diesel oil quality just presented applies
also to segregated oils stored in above-ground tanks at generator
or collector-processor sites. One additional type of waste oil
is that generated at airplane service facilities. This oil is
similar to automotive waste oils, but it is much more likely to
2-11
-------
be contaminated with chlorinated cleaning solvents used at these
sites.
Because of mixing, the composition of waste oil stored by
collectors and collector-processors is less predictable than the
composition of waste oil stored at the generator sites. Automo-
tive oils may be mixed with industrial oils, and various types of
industrial oils may be mixed together. Some collectors and
processors segregate their oils according to source, but most of
them mix the oils and thereby create a potential for contamina-
tion by a wide variety of substances, including heavy metals,
halogenated solvents, aromatics, and PCB's.
2.1.1.3 Drums—
Fifty-five gallon drums are used to store waste oils at
generator sites. Because of the small capacity of these storage
units, the oil in a given drum is likely to be generated from a
specific source. Virtually every type of oil is stored in drums,
but these units are used somewhat more frequently for crankcase
oil storage than for industrial oil storage. The contaminants
likely to be present in waste oil stored in drums include the
entire range of contaminants identified in all waste oils.
2.1.2 Changes in Composition Resulting From Storage Practices
Switching the material stored in a given tank for inventory
purposes or because of product demand is a common industry prac-
tice. Such changes affect the composition of the waste oil
stored in any type of unit. For example, if a relatively clean
automotive waste oil is placed in a storage tank that previously
2-12
-------
contained oil contaminated with PCB's or some other material, the
clean oil will become contaminated if the tank has not been thor-
oughly cleaned beforehand. This well-documented phenomenon has
caused significant problems with regard to misrepresented oil.
Such incidents are most likely to occur in the tanks of a col-
lector or a collector-processor, but they also could occur in
industrial generator storage tanks.
2.2 CHARACTERIZATION OF WASTE OIL LOSSES
Because waste oil losses may have a significant impact on
both health and the environment, it is important to address the
frequency and mechanisms of waste oil releases and the magnitude
of the resulting losses.
2.2.1 Frequency and Mechanisms of Waste Oil Releases
2.2.1.1 Below-Ground Tanks—
Until very recently almost no data were available on the
magnitude of leaks from below-ground storage tanks. A recent
study sponsored by the American Petroleum Institute (API), how-
ever, focuses on the causes and predictability of leaks in below-
21
ground unprotected steel tanks containing gasoline. This study
provides a mathematical model and basic data that are also appli-
cable to failure predictions for other petroleum-filled, below-
ground steel tanks. The potential sources of leaks in below-
ground tanks are listed as external corrosion, internal corro-
sion, piping failure, tank design and fabrication, and tank
24
installation.
2-13
-------
Although examples of failure resulting from each of these
causes have been documented, leaks from external corrosion occur
24
most frequently. External corrosion that is essentially uni-
form over the entire surface of a steel tank is not usually a
concern; tanks with such corrosion are expected to last as long
as the rest of a facility. In most (±77 percent) below-ground
steel tanks, however, localized external corrosion occurs, which
is likely to produce failure in a much shorter time. One or
several localized anodes established on the tank surface during
installation eventually lead to corrosion perforation(s). The
localized anodes may be caused by impurities in the backfill
adjacent to the tank surface, physical damage to the tank surface
(e.g., scrapes), etc. A mathematical model for predicting the
mean age to the outset of leakage from external corrosion was
developed in the API study:
Age = 5.75 x R°'05 x S'0'018 x e(0'13 ?H ' °'41M ' °'26 Su)
where
R = resistivity of soil in ohm-centimeters
S = tank capacity in gallons
pH = soil acidity
e = 2.72
M = 1 if soil saturated with water, 0 otherwise
Su = 1 if sulfides present in soil, 0 otherwise
The model is reported to be very accurate in predicting
f\ » O C *) £
failures where point-source external corrosion occurs. ' '
2-14
-------
Whereas the average age of a tank to the time of failure can be
calculated with the model, the actual age at the time of failure
may be less or more than the calculated age. The standard de-
viation from the calculated mean age is estimated to be 2.5
years.
An effort has been made to estimate the probability and
number of failures in below-ground waste oil storage tanks. The
probability estimates are based on data from the API study and
subsequent data developed and provided by the author of the API
study, Warren Rogers Associates. ' These data are considered
applicable because the vast majority of below-ground waste oil
tanks are unprotected steel tanks subject to essentially the same
25
external soil conditions as gasoline tanks.
Two approaches were used in estimating the probability of
leaks in below-ground waste oil tanks. They first utilized a
mathematical model from the API study to estimate tank age at
failure. The values used in the model were based on two sets of
assumed soil conditions and a tank size of 500 gallons. (The
assumed soil conditions have a dramatic impact on predicted ages
at failure, whereas the tank size is relatively unimportant.)
The failure predictions from the mathematical model were applied
to an assumed uniform age distribution of 0 to 20 years in below-
ground tanks. This permitted determination of the cumulative
probability of leaks in tanks equally distributed in this age
grouping.
2-15
-------
The second approach utilized recent data on the predicted
ages at which underground gasoline tanks will fail. These pre-
dictions were based on numerous tests of soil conditions at
automotive service stations throughout the United States and were
applied to an assumed uniform age distribution of 0 to 20 years
for below-ground tanks.
Both of these approaches are described in detail in Appendix
A. Both are useful in predicting failures caused by external
corrosion.
Based on the results of these two approaches, a probability
of leaks in below-ground waste oil storage tanks of 12 to 14
percent was calculated. These figures are believed to be overly
conservative, however because they consider only external corro-
sion failures and because evidence indicates that the assumed
tank age distribution is probably conservative. Although most
of the larger oil companies have begun tank replacement programs
o c 0*7 p ft
in the service stations they control, ' ' at least half of
the service stations in the United States are not within their
control and generally have no replacement programs. It is also
doubtful that below-ground waste oil tanks receive any attention
from the numerous other establishments that use them until an
obvious problem arises. For these reasons, it seems likely that
many of the waste oil tanks in service have been buried for well
over 20 years. Indeed, discussions with representatives of oil
companies suggest that many of the below-ground tanks currently
27 29
used in service stations may be 15 to 25 years old. ' One
2-16
-------
recent study indicates that the age of close to one-third of the
1.2 million below-ground gasoline and fuel oil tanks is 16 years
or older. Correspondence received from one major oil company
indicated that the age of an estimated 20 percent of their waste
26
oil tanks is 21 years or older.
An estimated 25 to 35 percent of the below-ground waste oil
tanks in some areas of the country are believed to be leaking,
and nearly all of these tanks are buried in moisture-saturated
soil. The higher figure would be expected where the tanks are
not only buried in moisture-saturated soil, but many are more
than 20 years old.
The use of a 12 to 15 percent leak probability in below-
ground waste oil tanks results in an estimate of 40,000 to 47,000
leaking tanks nationwide. Although these figures seem alarmingly
high, they may well be conservative. In any case, the number of
leaking below-ground waste oil tanks far exceeds the annual
number of spills from above-ground waste oil storage in tanks and
drums. Also, many of the below-ground tanks have been leaking
undetected over a period of many years.
2.2.1.2 Above-Ground Tanks—
Insufficient empirical data are available for direct deter-
mination of the frequency and probability of waste oil losses
from above-ground tanks. Regulations require that spills of oil
and/or hazardous substances affecting U.S. inland surface waters
be reported to the EPA. Similarly, spills that could affect
2-17
-------
navigable waters must be reported to the U.S. Coast Guard.
Unfortunately, the spill reports received are not considered a
reliable source by which to determine spill probability because
the belief is that many spills are not reported. The reported
spills are believed to represent only those that could affect
surface water.
An attempt was made to estimate how many of these reported
spills from 1975 through 1980 were waste oil. It was determined
that over 80 percent of the spills reported to the U.S. Coast
32
Guard were petroleum products. Although a similar determina-
tion could not be made for spills reported to EPA, 73 percent of
the spills reported in Region IV (approximately one-third of the
national total) were petroleum products. The Coast Guard data
suggested that waste oils were involved in about 11 to 12 percent
of the petroleum product spills. These figures lead to an esti-
mate that approximately 360 of the nontransportation storage
spills reported to EPA and U.S. Coast Guard were waste oil. This
is equivalent to 60 spills of waste oil each year from 1975
through 1980.
Because the amount of directly applicable data is insuffi-
cient to serve as the basis for above-ground tank spill probabil-
ity, JRB Associates (in a study for EPA) developed a "fault-tree"
analysis for a reportedly "typical" storage system to examine
storage failures. This analysis considers the failure proba-
bilities of the various components of a storage system to arrive
at the failure probability of the total system. The analysis is
2-18
-------
based on combined data from the following sources: a nuclear
reactor safety study performed for the U.S. Nuclear Regulatory
Commission (published in 1975), a safety study of U.S. deepwater
port oil transport systems for the U.S. Department of Transpor-
tation (published in 1978), a Sandia Laboratory report on human
error probability (published in 1964), an American Petroleum
Institute report on fire incidents at bulk plants between 1971
and 1974, and an estimate of the probability of a hose leak
(based on the engineering judgment of JRB Associates, the devel-
opers) .
It should be noted that a fault-tree analysis is only as
good as the data used in its development and assumptions relative
to a "typical" storage facility. Reported spills have indi-
cated more identified failure causes than those provided in the
fault-tree analysis, but the analysis accounts for the major
failures identified. Natural disasters (windstorms, floods,
earthquakes, etc.) are some of the failure causes not included in
the analysis; however, these account for a small percentage of
the nontransportation storage spills reported to the EPA and the
Coast Guard. Another point to be considered with regard to the
analysis is the reliance on the nuclear energy industry for much
of the failure probability data. This industry is expected to
have higher quality standards than most others. Because the typ-
ical storage facility used in the analysis is based on standards
developed by the American Petroleum Institute and other associa-
tions, the use of the fault-tree analysis probably will lead to
conservatively low estimates of failure probability of above-
ground storage tanks used for waste oil.
2-19
-------
Consideration was given to supplementing the use of the
fault-tree analysis in this study with a separate analysis of
above-ground tank failures due to external corrosion. Many
above-ground waste oil tanks are in contact with the ground
surface, and the API study mathematical model used to predict
external corrosion in below-ground tanks was judged to have
potential here, also. Use of the model for analysis of above-
ground tank corrosion was ruled out, however, for several rea-
sons. First, major parameters in the model include the presence
(or absence) of moisture-saturated soil in contact with the tank
surface and the presence of sulfides in the soil. Although some
data are available on which to base estimates of these parameter
values for below-ground tanks, similar data for above-ground
tanks have not been found. Second, adjustments would have to be
made in the model for above-ground tanks to reflect that only a
fraction of the tank surface is in contact with the ground sur-
face. Again, data are insufficient for making these adjustments.
A final reason for excluding the mathematical model from the
above-ground tank analysis is that tank corrosion accounts for a
very small percentage of the nontransportation storage spills
reported to the EPA and the Coast Guard.
A review of both the fault-tree analysis and the nontrans-
portation storage spill reports submitted to the EPA and Coast
Guard provided some insight into failure causes at storage facil-
ities using above-ground tanks. Failure of ancillary equipment
(pipes, pumps, valves, etc.) causes the most above-ground tank
2-20
-------
losses. Failure of the tank itself, vandalism, fire, explosions,
and natural disasters are lesser causes. The fact that most
above-ground tanks used for waste oil storage are small and
simply constructed, however, eliminates much of the probability
of failure due to ancillary equipment. Failure in these above-
ground tanks is due largely to tank overflow (resulting primarily
from operational error).
Surrounding an above-ground storage facility with a second-
ary containment system (dikes and/or curbing) considerably less-
ens the probability of pollution from spills. If the surface
within such a secondary containment is relatively impermeable or
if the potential for damage from percolation of oil into the
ground is small, losses within the secondary containment may not
be significant. Although the fault-tree analysis includes fail-
ure probability of a storage facility with secondary containment,
this can be excluded if the storage facility has no secondary
containment.
The fault-tree analysis further divides the storage facility
failure probability between the storage tank system and the tank
filling and discharge system. As shown in Table 2-2, secondary
containment has a dramatic impact on the probability of a spill
leaving the boundaries of a facility. With secondary containment
around both the storage tank system and the fill-discharge system,
the probability of facility failure is estimated to be 1.3 x 10
percent per year, or roughly just over one chance in 100,000.
Conversely, at facilities without secondary containment the
failure probability increases by several orders of magnitude, to
2-21
-------
an estimated 11.6 percent annually. These numbers were arrived
at by adding the expected failure probabilities with/without
secondary containment from both the storage tank system and the
fill-discharge system.
TABLE 2-2. PROBABILITY OF LOSSES FROM ABOVE-GROUND TANK FACILITIES
FOR HAZARDOUS LIQUID STORAGE8
Category
Failure
probability/year,
percent
Loss from storage tank system:
Loss from tank
Loss from dike (around tank)
Combined loss probability
Loss from fill-discharge system:
Loss from plumbing
Loss from curbing (around plumbing)
Combined loss probability
Loss from either the storage tank system
or fill-discharge system:
Without secondary containment
With secondary containment
5.0
2.5 x 10
1.3 x 10
-2
-3
6.6
1.0 x 10
6.6 x 10
-4
-6
11.6
1.3 x 10
-3
The figures presented are considered applicable to waste oil storage tanks,
based upon the sources of supporting data used in developing the JRB
Associates fault-tree analysis.31
The failure probabilities from the fault-tree analysis were
used in this study to estimate the frequency of waste oil losses
from above-ground tank storage facilities. (The figures are
believed to be applicable, based on the supporting data used in
developing the analysis.) The first task was to estimate the
number of above-ground storage facilities with secondary contain-
ment versus those without such containment. Discussions with
collectors, processors, and others in the waste oil business
2-22
-------
revealed that most facilities with tanks that hold several thou-
sand gallons of waste oil have secondary containment, whereas
12 — 19
those with smaller tanks generally do not. In some cases,
secondary containment is simply natural drainage to a depression
(natural or manmade) capable of containing a spill until it can
be recovered.
An estimated 5200 above-ground waste oil tanks have capaci-
ties larger than 1000 gallons. Assuming 25 percent of these
tanks are at facilities that have no secondary containment, and
based on the 11.6 percent failure probability from the fault-tree
analysis, an estimated 150 noncontained spills could occur annu-
ally at these facilities. At storage facilities with secondary
containment, the number of noncontained spills each year would be
virtually zero.
The smaller above-ground waste oil tanks are believed to
number nearly 35,000. The capacities of these tanks are gen-
erally 500 gallons or less, and they are located primarily in
automotive repair shops and other service establishments. Because
of their size, few (if any) are likely to have the fill-discharge
system of piping, valves, pumps, etc., that is included in the
"typical" facility in the fault-tree analysis. Therefore, it is
assumed that only the probability of a loss from the tank and its
ancillary eguipment should be included in estimating the number
of spills from small above-ground tank storage. Since the ancil-
lary eguipment would usually consist of a bottom drain with a
valve, the probability of a loss from small tank storage is
2-23
-------
assumed to be the combined probabilities of a tank wall rupture,
a bottom drain leak, and tank overflow. These probabilities,
which are individually cited in the fault-tree analysis, were
simply added together to obtain an estimated combined failure
probability. This annual combined failure probability from
small tank storage is 2.1 percent. The corresponding annual
failure probability in a large tank with more extensive ancillary
equipment is estimated to be 5.0 percent.
Based on a conservative assumption that 75 percent of the
smaller above-ground tanks are at facilities that have no sec-
ondary containment, the estimated annual number of noncontained
spills from these facilities would be approximately 550. This
figure seems high in comparison with the just over 4000 oil/
hazardous waste nontransportation storage spills reported to the
EPA and the Coast Guard from 1975 to 1980, of which only about 60
per year were waste oil spills. Part of the discrepancy may be
due to the size of the spills. Although many small spills prob-
ably go unreported, the fault-tree analysis appears to account
for small leaks as well as larger losses. Again, secondary
containment reduces the number of noncontained spills dramatical-
ly. The installation of secondary containment at all the listed
above-ground waste oil storage facilities would reduce the esti-
mated noncontained spills to less than one spill per year.
Table 2-3 presents a summary of nontransportation storage
spill incidents (distributed by size and failure causes) reported
2-24
-------
TABLE 2-3. STORAGE SPILL PERCENTAGES FROM EPA AND U.S. COAST GUARD DATA BASES3
(percent of total in each size category)
Containment devices:
Tank rupture/leak
Tank corrosion
Subtotal
Operations:
Tank overflow
Other
Subtotal
Ancillary:
Pipes
Pumps
Valves
Secondary containment
Other
Subtotal
Other:
Fire/ explosions
Weather-related/vandalism/
other
Subtotal
Totals
Amount spilled, gallons
0-49
4.3
1.4
5.7
16.2
24.2
40.4
25.5
3.8
3.4
0.7
9.8
43.2
3.3
7.4
10.7
100%
50-99
4.3
1.2
5.5
18.5
22.0
40.5
27.5
4.0
9.0
5.3
45.8
1.0
7.2
8.2
100%
100-249
4.7
3.1
7.8
21.2
16.2
37.4
32.3
2.9
5.4
0.3
4.3
45.2
2.8
6.8
9.6
100%
250-499
3.8
1.4
5.2
19.0
16.8
35.8
36.8
2.9
7.6
4.3
51.6
0.2
7.2
7.4
100%
500-999
5.1
3.3
8.4
18.2
15.1
33.3
32.6
3.3
10.2
4.7
50.8
0.4
7.1
7.5
100%
1,000-10,000
6.2
3.0
9.2
15.2
17.9
33.1
28.6
2.5
8.9
0.4
5.7
46.1
1.0
10.6
11.6
100%
>10,000
10.7
1.6
12.5
8.9
17.8
26.7
23.0
5.8
10.2
2.2
0.9
42.1
4.9
13.8
18.7
100%
All
spill
sizes
5.1
2.2
7.3
17.1
19.3
36.4
29.3
3.3
6.9
0.5
6.1
46.1
2.0
8.2
10.2
100%
ro
I
10
Derived from PIRS and SPCC nontransportation storage spills reported in Reference 21; represents 1975 to
1980.
-------
to the EPA and the Coast Guard. The distribution of failure
causes is by percentage for each spill size category and for all
spill sizes. Although ancillary equipment failures and opera-
tional errors are clearly the dominating causes of failure in
each spill size category, operational errors become progressively
less dominant with increasing spill sizes. Containment device
failures and failures from other causes are both more significant
causes of failure for the larger (greater than 1000 gallons)
reported spills.
As noted before, most of the above-ground waste oil tanks
hold 500 gallons or less and do not include much ancillary equip-
ment. For this reason, the high percentage of spill failures
from ancillary equipment (in Table 2-3) is believed to be con-
siderably overstated for most above-ground tank storage of waste
oil. Based on the fault-tree analysis, ancillary equipment
failures are estimated to represent less than 15 percent of the
above-ground waste oil losses in small tank storage. On the
other hand, operational errors appear to account for well over 50
percent of such losses. A rupture or leak in the tank, from
whatever cause, accounts for the remaining small tank storage
spills.
Causes for failure in the larger above-ground waste oil
tanks (5000 gallons and larger) are expected to be related more
closely, by percentage distribution, to those shown in Table 2-3;
however, most of these larger tanks are believed to be surrounded
by secondary containment.
2-26
-------
A review of the "Damage Incidents Resulting From Used Oil
Mismanagement" (from EPA) provided little additional insight into
the causes of spills from above-ground storage. Spills of both
oil and other materials are reported, but the causes and quanti-
ties are seldom noted. A faulty valve is noted in one tank spill
incident, a tank rupture in another, and corrosion from tanks
resting in oily water in another. These were the only causes of
failure given for failures in above-ground tanks.
2.2.1.3 Drums—
Essentially no data were found on the frequency of waste oil
losses from drums, which are used primarily by establishments
that generate smaller quantities of waste oil. Industrial genera-
tors represent the largest sector using drum storage for waste
oil. The general practice is to use drums if waste oil volumes
are less than 500 gallons per month and to use below-ground tanks
for greater volumes. Establishments using drums to store waste
oil are estimated to have an average of three 55-gallon drums
each, which equals an average storage capacity of 165 gallons.
Some drums are stored inside a building, but secondary contain-
ment systems are believed to be rare at drum storage sites.
Because they are portable and relatively small, drums could be
easily overfilled and overturned. They do not require the
piping, pumps, valves, etc., associated with the filling and
discharging of large tanks, however, and thus are not subject to
the errors attendant with a mechanical fill-discharge system. It
2-27
-------
also seems reasonable to assume that, on the average, only about
50 percent of the available storage drums actually contain waste
oil.
An estimate of drum spills can be made by applying the small
tank loss probability (2.1 percent) from the fault-tree analysis
to the estimated average number of drums containing waste oil.
This produces an estimate of nearly 4500 spills annually, or just
over 1 percent of all the drums used for waste oil storage.
Whereas the number of spills seems large, a 1 percent probability
of a spill of any size does not seem implausible.
It is also important to note that the maximum spill from a
single drum is 55 gallons; therefore, although the total number
of noncontained waste oil spills from drums may be greater than
those from above-ground tanks, the point-source pollution dangers
are generally not as great.
2.2.2 Magnitude of Losses
2.2.2.1 Below-Ground Tanks—
The investigation of underground waste oil leaks revealed
that much less attention is being given to these leaks than to
underground gasoline leaks. Consequently, much of the informa-
tion on underground waste oil losses has been a byproduct of the
attention given to gasoline losses. Some of the more significant
reported gasoline leaks and other petroleum leaks from below-
ground tanks in recent years are discussed.
2-28
-------
In 1979, a loss of 30,000 gallons of gasoline from a service
station in East Meadow, New York, was discovered. ' The loss,
which occurred over a period just exceeding one year, was from
two tanks that held approximately 4000 gallons each. In 1980, a
loss in excess of 30,000 gallons of gasoline from a service
station in a suburb of Denver, Colorado, was discovered. This
loss occurred over a period of 3 to 4 years. A leak discovered
in 1978 is speculated to have involved several million gallons of
gasoline, fuel oil, and naphtha under Brooklyn, New York. Some
years ago a leak of 20,000 gallons of No. 2 oil was discovered
under the terminal of a large oil company in the Boston, Massa-
chusetts, area.
The usual volume of gasoline lost in below-ground tanks in
service stations is reported to range between 200 and 600 gal-
lons. One source says such losses generally are confined to
1000 gallons or less and that they occur over a period of no more
than 3 or 4 months. Nevertheless, it is evident that much
larger losses could occur. Leaks often go undetected until the
taste or smell of gasoline is noted in the water supply inside
homes.
Despite the fact that considerable data exist on gasoline
and other petroleum product losses from below-ground tanks,
little data are available on waste oil losses. It is known,
however, that several of the major oil companies now have tank
testing and replacement programs for service stations under their
2-29
-------
ownership or control and that they test and replace waste oil
27 28 29
tanks as well as gasoline tanks. ' ' The gasoline tanks are
of much greater concern because a typical station will have four
or five such tanks with a total capacity of 20,000 gallons or
more, compared with a single waste oil tank with a capacity of
500 gallons. Also, gasoline is much less viscous than oil and
spreads through the soil much faster.
One oil company representative estimated that their service
stations experience only about six incidents per year in which
waste oil leaks migrate beyond the station's property bounda-
27
ries. He further indicated that most leaks involve only a few
gallons per month before being detected. It is a common practice
for this company (and others) to replace leaking waste oil tanks
(now mostly steel) with new fiberglass units that are not troubled
with corrosion problems.
Although the major oil companies have extensive tank replace-
ment programs underway, they apparently control no more than 50
27 28 29
percent of the service stations ' ' and less than 20 percent
of all the below-ground waste oil tanks. Also, waste oil tanks
used by industrial generators are normally many times larger than
those in service stations. It is believed that the large quantity
of underground waste oil storage at locations other than service
stations receives little attention with respect to leakage until
an obvious problem is detected.
Because most below-ground losses result from tank corrosion,
leaks are probably very slow in the beginning and grow larger as
2-30
-------
the perforations caused by corrosion increase in size. The
surrounding soil also can slow down the rate of flow from the
tank. Some below-ground losses may go undetected for several
years, as opposed to losses from above-ground tanks, which are
usually more rapid and readily apparent.
More than 90 percent of the below-ground tanks have capaci-
ties of a few hundred gallons (typically 500 to 600 gallons).
Discussions with waste oil collector-processors indicate that
these tanks are usually emptied every 4 to 8 weeks and are never
more than three-fourths full. This suggests that no more than
375 gallons of waste oil is placed in a 500-gallon tank each
month and that this amount would represent the maximum potential
loss. It is also probable that a leak of this magnitude would be
readily noticed and reported by the waste oil collector.
Application of this logic to the typical 5000-gallon waste
oil tank used by an industrial generator places the upper limit
of potential loss at 3750 gallons per month. The probability of
an underground waste oil loss of this magnitude seems to be very
low compared with the larger gasoline losses listed earlier. The
30,000-gallon gasoline loss in East Meadow, New York, for example,
is believed to have averaged no more than 2500 gallons per month,
and this loss reportedly occurred from two 4000-gallon tanks.
Although it is not currently possible to estimate the total
quantity of waste oil losses from below-ground storage tanks, it
is conceivable to assume that individual losses can range from
2-31
-------
the lower limits of detection (35 to 40 gallons per month) to 375
gallons per month from 500-gallon tanks and 3750 gallons per
month from 5000-gallon tanks. Sufficient data are not available
for obtaining an average loss, but the enormous number of leaking
below-ground waste oil tanks makes underground waste oil losses a
matter of utmost concern.
2.2.2.2 Above-Ground Tanks—
Very little documentation of waste oil losses is readily
available, and such documentation rarely includes specifics on
the magnitude of the losses. Several state agencies with ap-
parently active waste oil regulatory and/or control programs were
contacted. ' ' In general, they reported the occurrence of a
few oil spills, but indicated that details of these spills were
not readily accessible from their files. Two recent spills were
reported in Illinois; one (estimated to involve 500 gallons of
waste oil) was caused by sabotage at a processor's operation, and
34
the other (quantity unknown) was fire-related. A groundwater
contamination inventory in Michigan documents a few cases of
apparent waste oil losses, but provides no quantity details.
Some of these losses may be from below-ground waste oil storage.
Discussions with collectors and collector-processors of
waste oil also yielded relatively little information on waste oil
3—19 38
spills. ' One reported a spill of 500 gallons at his facili-
ty a few years ago. Another indicated a spill at his facility
(prior to his ownership) of 10,000 gallons. Sabotage was the
2-32
-------
reported cause of the latter. Other spills at collector-processor
facilities include one (reported by the owner) of 1000 to 1500
gallons and another reported to be between 3000 and 5000 gallons.
Because these latter two spills were both contained on site,
however, they reflect only potential tank losses.
Though minimal quantitative data were obtained on waste oil
losses from above-ground tanks, a review of known waste oil
storage characteristics and spills reported to the EPA and the
U.S. Coast Guard proved to be useful.
The vast majority of above-ground waste oil storage tanks
have capacities of 500 gallons or less, and collectors indicate
that the contents of these tanks are usually emptied before the
tanks are more than three-fourths full. Thus, maximum losses
from 500-gallon tanks generally would be no greater than 375
gallons per incident if the losses were to occur just before
scheduled collection of the tank's contents.
An estimated 10 to 15 percent of the above-ground waste oil
tanks are much larger (5000 gallons and up). Many collector-
processors have tanks that hold 20,000 to 50,000 gallons, and
12-19
some have tanks that hold 250,000 gallons or more. Discus-
sions with selected collector-processors suggest that nearly all
of these facilities have secondary containment, which lessens the
12-19
probability of many offsite waste oil losses. On the other
hand, some establishments using primarily 5,000- to 10,000-gallon
tanks may not have secondary containment. Again, based on the
assumption that these tanks are not often filled to capacity, the
2-33
-------
maximum loss from a 10,000-gallon tank probably would be about
7500 gallons. Such a high loss would be expected only if failure
occurred just before a scheduled waste oil collection.
Table 2-4 presents the nontransportation storage spill
incidents reported to EPA (3000) and the U.S. Coast Guard (1300)
by size and failure and the percentage of total spill incidents
within each indicated spill size range. Table 2-5 presents these
percentages both separately and in combination for the EPA and
U.S. Coast Guard data.
Over 50 percent of the nontransportation storage spills
reported to the EPA and the U.S. Coast Guard were under 250
gallons. Of the spills reported to the EPA, exactly 50 percent
were less than 250 gallons; of the spills reported to the U.S.
Coast Guard, 61 percent were less than this amount. Approximately
25 percent of the combined spills represented 1000 gallons or
more. This would undoubtedly be high for waste oil spills because
of the predominance of smaller waste oil tanks and the prevalence
of secondary containment around the larger tanks. Even if each
of the estimated 150 annual spills (losses) previously indicated
for larger waste oil tanks amounted to 1000 gallons or more, this
would represent only 21 percent of the total estimated spills
from above-ground waste oil tanks.
The indicated distribution of spills is probably skewed
toward the larger spill sizes because of a tendency for facili-
ties not to report smaller spills of waste oil. Also, the spills
reported in Tables 2-4 and 2-5 are judged to be from typically
2-34
-------
TABLE 2-4. STORAGE SPILL INCIDENTS FROM EPA AND U.S. COAST GUARD DATA BASES'
(percent of total in each size category)
Containment devices:
Tank rupture/leak
Tank corrosion
Subtotal
Operations:
Tank overflow
Other
Subtotal
Ancillary:
Pipes
Pumps
Valves
Secondary containment
Other
Subtotal
Other:
Fire/explosions
Weather- related/ vandal ism/
other
Subtotal
Totals
Percent of Total Spill
Incidents
Amount spilled, gallons
0-49
51
17
68
194
291
485
306
45
41
9
118
519
39
89
128
1,200
28
50-99
17
5
22
74
88
162
110
16
36
-
21
183
4
29
33
400
9
100-249
34
22
56
153
117
270
233
21
39
2
31
326
20
49
69
721
17
250-499
16
6
22
79
70
149
153
12
31
-
18
214
1
30
31
416
10
500-999
23
15
38
82
68
150
147
15
46
-
21
229
2
32
34
451
10
1,000-10,000
57
28
85
140
165
305
265
23
82
4
53
427
9
98
107
924
21
>10,000
24
4
28
20
40
60
52
13
23
5
2
95
11
31
42
225
5
Total
222
97
319
742
839
1,581
1,266
145
298
20
264
1,993
86
358
444
4,337
100
U)
in
Based on PIRS and SPCC nontransportation storage spills reported in Reference 31.
-------
TABLE 2-5. SIZE DISTRIBUTION OF SPILLS REPORTED TO EPA
(1975-1980) AND U.S. COAST GUARD (1974-1980)
Spill size,
gallons
0-49
50-99
100-249
250-499
500-999
1,000-9,999
>10,000
Total
Spills
reported
to EPA,
percent
24
9
17
10
12
23
5
100
Spills
reported to
U.S. Coast Guard,
percent
37
9
15
8
8
18
5
100
Weighted
average,
percent
28
9
17
10
10
21
5
100
A weighted average is represented to reflect the much greater
number of spills reported to EPA.
2-36
-------
larger storage facilities than those predominantly used for waste
oil storage because of the large number of reported spills result-
ing from ancillary equipment failures. Thus, it is quite likely
that the median waste oil spill from above-ground tanks is con-
siderably below 250 gallons. This is further substantiated by
the fact that the typical 500 gallon above-ground waste oil tank
is expected to be only about 30 percent full on the average. The
loss of the entire contents from a tank that is at 30 percent of
capacity would constitute only 150 gallons.
A reasonable estimate of the total quantity of waste oil
losses from above-ground tanks is not obtainable from currently
available data. Such a determination might be possible if tank
size were included in the EPA and Coast Guard spill data. Deter-
mination of typical time frames over which losses occur is also
impossible without additional data.
2.2.2.3 Drums—
It was previously estimated that 4500 spills (or losses) of
waste oil from drums occur annually. The maximum loss from a
single typical drum is 55 gallons, but a given incident could, of
course, involve more than one drum. On the average, establish-
ments that use drums for waste oil storage are estimated to have
three drums each, but some have a much larger number.
The following 11 reported drum spill incidents were documented
by EPA31:
2-37
-------
Amount spilled Number of spills
0-49 gallons 5
50-99 gallons 2
100-249 gallons 2
500-999 gallons 1
1,000-10,000 gallons 1
Although these reported drum spills are too few in number to
serve as a valid indication of the magnitude of most drum spill
incidents, they do indicate that such incidents can involve 500
to 1000 gallons or more. If it is assumed that the full 55-gallon
capacity of each of the 4500 drums (estimated as spilled annually)
is lost, the total waste oil loss from drums would be 247,500
*
gallons, which is as much as (or possibly more than) the expected
annual loss from above-ground oil tanks. It is unrealistic,
however, to assume that each drum spill results in the loss of
the full 55 gallons it can hold.
Generally, the magnitude of individual drum spill incidents
is judged to be less than that of spill incidents involving
above-ground tanks. Therefore, there is less potential for
severe point-source pollution from drum storage spills.
2.3 SUMMARY
This section has characterized waste oil storage in the
United States. The following subsections summarize the informa-
tion presented in this chapter.
2-38
-------
2.3.1 Sources and Composition of Waste Oil
Waste oil can generally be divided into two categories:
automotive/diesel and industrial. Automotive/diesel waste oils
are primarily crankcase oils generated by cars, trucks, and other
vehicles. Contaminants in these oils include barium, chromium,
and lead, the contaminant of greatest concern. Polynuclear
aromatic compounds (PNA's) are also contained.
Industrial waste oils are generated from industrial sources
and include metal working, hydraulic process, electrical, refrig-
eration, and turbine oils. These waste oils may contain a wide
range of contaminants, including halogenated solvents, aromatic
solvents, PCB's, and heavy metals (cadmium, chromium, and zinc).
2.3.2 Waste Oil Storage Types and Quantities
Waste oil is stored by the generators and those who collect
and process the oil. The total amount of waste oil in storage is
estimated to be 173 million gallons. Automotive/diesel oil
generators store an estimated 64 million gallons; industrial
generators store an estimated 41.7 million gallons; and collec-
tors and collector-processors store an estimated 67.8 million
gallons of both waste oil categories.
Waste oil is stored in below-ground tanks, above-ground
tanks, and 55-gallon drums. Most storage tanks are made of
steel, but major oil companies have recently been replacing steel
below-ground tanks with fiberglass tanks. Although tank sizes
vary widely, the majority of tanks (both below-ground and above-
ground) hold 500 gallons or less. Many industrial waste oil
2-39
-------
generators, however, use tanks that hold several thousand gallons,
and some collector-processors use tanks that hold a few hundred
thousand gallons.
Below-ground tanks account for nearly one-half of the stored
waste oil, and above-ground tanks for about 44 percent. The
remainder is stored in drums.
2.3.3 Waste Oil Losses
Losses of stored waste oil occur far more frequently from
below-ground tanks than from above-ground tanks or drums. It is
conservatively estimated that 12 to 14 percent of the below-
ground tanks containing waste oil (or 43,500 tanks) are leaking,
whereas, the estimated annual probability of loss from above-
ground waste oil storage facilities is under 2 percent (or 700
loss events annually). The estimated annual probability of loss
from drums is just over 1 percent (or 4,500 loss events).
Losses from below-ground waste oil storage tanks are caused
primarily by external tank corrosion. Losses from other causes
(including internal tank corrosion) are few by comparison. The
causes of losses from above-ground waste oil storage tanks are
numerous. In general, operational errors account for the majority
of failures in above-ground storage facilities with smaller tanks
(under 500 gallons each) and drums, whereas failures of ancillary
equipment (pumps, valves, pipes, etc.) are the predominant causes
in facilities with larger above-ground tanks.
Little information is available on the magnitude of waste
oil losses, and realistic estimates cannot be made of typical or
2-40
-------
average losses. Those who collect waste oil indicate that tanks
are seldom more than 75 percent full at the time of collection,
which limits the maximum spill from an above-ground tank. The
smaller tanks are estimated to be only 30 percent full, on the
average, which limits the average spill size. Many spills are
probably stopped before a tank is emptied, and others may be
stopped before the loss reaches a significant proportion.
Losses from below-ground tanks are different from those from
above-ground tanks. Most below-ground losses begin as slow
leaks. The size of these leaks increases with time, but the
leaks may go undetected for years. Few of these leaks are expected
to receive any attention until an obvious problem occurs or the
collector observes a noticeable change in the amount collected.
2-41
-------
REFERENCES FOR SECTION 2
1. Development Planning and Research Associates, Inc. Risk/
Cost Analysis of Regulatory Options for the Waste Oil Man-
agement System. Volumes I and II. Office of Solid Waste,
U.S. Environmental Protection Agency. January 1982.
2. Personal communication. M. L. Marino, Development Planning
and Research Associates, Inc., Manhattan, Kansas, October
and November 1982.
3. Personal communication. M. Kerran, Double Eagle Refining
Co., Oklahoma, City, Oklahoma, November 1982.
4. Personal communication. H. K. Robertson, Jackson Oil Pro-
ducts Co., Jackson, Mississippi, November 1982.
5. Personal communication. G. Davis, Davis Refining Co.,
Tallahassee, Florida, November 1982.
6. Personal communication. J. O'Connell, Motor Oils Refining
Co., McCook, Illinois, November 1982.
7. Personal communication. M. Pierce, Central Refining Co.,
Springfield, Illinois, November 1982.
8. Personal communication. T. Manning, R.T.I., California,
November 1982.
9. Personal communication. K. Morris, Cam-Or, Inc., Indiana-
polis Indiana, November 1982.
10. Personal communication. A. L. Warden, Warden Oil Co.,
Minneapolis, Minnesota, November 1982.
11. Personal communication. G. Booth, III, Booth Oil Co.,
Buffalo, New York, November 1982.
12. Personal communication. M. Adams, Waste Oil Collector/
Processor, Denver, Colorado, October 1982.
13. Personal communication. R. Lipscomb, Eco Oil Co., St.
Louis, Missouri, October 1982.
2-42
-------
14. Personal communication. E. Baumgardner, Waste Oil Services,
Fayetteville, Pennsylvania, November 1982.
15. Personal communication. Mrs. W. John, Bill John's Waste
Oil, Jacksonville, Florida, November 1982.
16. Personal communication. C. Crutchfield, Petroleum Recycling
Corp., Knoxville, Tennessee, November 1982.
17. Personal communication. C. Hayward, California Oil Recyclers,
Fremont, California, November 1982.
18. Personal communication. S. Puchtel, Steve's Oil Service,
Minneapolis, Minnesota, November 1982 and April 1983.
19. Personal communication. M. Stewart, Edgington Oil Co., Long
Beach, California, November 1982.
20. Personal communication. B. Mayhew, Port Mobil Oil, New
York, New York, April 1983.
21. Personal communication. S. Calill, Caddell Shipyard & Tank
Cleaning, New York, New York, April 1983.
22. Personal communication. Captain Lawrence, Port Allen Marine
Services, Baton Rouge, Louisiana, April 1983.
23. Personal communication. M. Keller, Chemical Processors,
Inc., Seattle, Washington, April 1983.
24. Warren Rogers Associates, Inc. Report on the Statistical
Analysis of Corrosion Failures of Unprotected Underground
Steel Tanks. American Petroleum Institute. January 15,
1982.
25. Personal communication. W. Rogers, Warren Rogers Associates,
Newport, Rhode Island, November and December 1982.
26. Correspondence from Shell Oil Company, Houston, Texas,
December 8, 1982.
27. Personal communication. A representative of Exxon Co.,
U.S.A., Houston, Texas, October and November 1982.
28. Personal communication. A representative of Shell Oil Co.,
Houston Texas, November 1982.
29. Personal communication. A representative of Texaco, U.S.A.,
Houston, Texas, November and December 1982.
2-43
-------
30. Shaner, J. R. The Tank Leak Mess. National Petroleum News.
July 1982.
31. JRB Associates. Failure Incident Analysis: Evaluation of
Storage Failure Points. (Draft.) Office of Solid Waste,
U.S. Environmental Protection Agency. March 1982.
32. U.S. Department of Transportation, United States Coast
Guard. Polluting Incidents in and Around U.S. Waters,
Calendar Years 1975 through 1980.
33. Osgood, J. S. Hydrocarbon Dispersion in Groundwater:
Significance and Characteristics. Groundwater. November
and December 1974.
34. Personal communication. J. Perry and J. Langley, Illinois
Environmental Protection Agency Emergency Response Team,
December 1982.
35. Personal communication. T. Manor, Wisconsin Department of
Natural Resources, December 1982.
36. Development Planning and Research Associates, Inc. State
Regulation of Waste Oil. Office of Solid Waste, U.S. Envi-
ronmental Protection Agency, July 1982.
37. Michigan Department of Natural Resources, Groundwater Qual-
ity. Assessment of Groundwater Contamination: Inventory of
Sites, July 1982.
38. Case study No. 5, Pierce Waste Oil, Springfield, Illinois,
prepared by E. Hillenbrand and B. Burgher, Illinois Envi-
ronmental Protection Agency, July 21, 1982.
2-44
-------
SECTION 3
ENVIRONMENTAL FATE OF WASTE OIL LOST FROM OIL STORAGE SITES
This section is concerned with the fate of the waste oil
that is released into the environment by leaks or spills from
storage sites. The rate and degree of contamination from typical
and worst-case spills are examined. Losses may be from slow
leaks, which go undetected for long periods of time, or from
catastrophic spill incidents. The models used to estimate envi-
ronmental pollution consider the spill conditions, including the
oil and soil types, distance to groundwater, and time from spill
to detection.
3.1 MECHANISMS OF WASTE OIL MOVEMENT
Waste oil may leave an oil storage site by one or more of
three mechanisms: evaporation, surface runoff, and seepage into
the soil. The degree of concern for the environment as waste oil
and its contaminants leave an oil storage site by one or more of
these mechanisms depends on the type and composition of waste
oil, type of spill or leak, type of oil storage system, and the
climatic conditions.
The processes of evaporation, runoff, and seepage occur
simultaneously, and all may be of environmental concern in the
event of surface spills. Evaporation and runoff are rarely of
3-1
-------
concern in subsurface leaks, which occur primarily from below-
ground tanks, but sometimes from the bottom of above-ground
tanks. Evaporation and seepage are continuous processes fol-
lowing a surface spill, whereas runoff is more intermittent in
nature, e.g., rainfall runoff.
Each of the three movement mechanisms is discussed in the
following subsections. Because seepage is the primary mechanism
for the movement of spill oil, it is examined in more detail than
the other two.
3.1.1 Evaporation
Evaporation is the process by which waste oil and contami-
nants are vaporized, which enables them to leave the oil storage
site in gaseous form. The organic vapors from waste oil spills
can cause deterioration of air quality.
Ambient air concentrations of organic vapors from oil-covered
road surfaces have been modeled in a related study. Both the
rate of evaporation of the waste oil components and the distribu-
tion and resultant ambient air concentrations were modeled.
Environmental contamination levels were calculated for the follow-
ing components: arsenic, barium, cadmium, chromium, lead, zinc,
dichlorodifluoromethane, trichlorotrifluoroethane, trichloroeth-
ane, tetrachloroethylene, benzene, toluene, xylene, benzo(a)an-
thracene, benzo(a)pyrene, naphthalene, and PCB's. Model outputs
indicate that ambient levels are not likely to pose a threat to
human health.
The mechanism of evaporation from surface spills of waste
oil will differ from that of evaporation of waste oil applied to
3-2
-------
a road surface because of the formation of pools of oil within
secondary containment systems. Nevertheless, the overall magni-
tude of oil evaporation should be similar to or less than evapora-
tion from road surfaces. Because evaporation of spilled waste
oil is not likely to pose a threat to human health, it is not
considered further within this report.
3.1.2 Rainfall Runoff
Oil and waste oil components that are spilled from above-
ground tanks and drums may contaminate surface waters. Rainfall
and subsequent surface runoff may carry colloidal oil, dissolved
oil components, and oil adsorbed onto soil particles. Oil may be
washed from the spill area or it may be carried with water as a
surface film or as a colloid. Oil that has seeped into the soil
cannot be easily or rapidly washed off by rainfall, but as the
rain seeps into the soil, it can displace the oil and cause it to
float to the surface, where it can be washed away. Rainfall run-
off is generally limited to spills that occur where there is no
secondary containment system. Most commonly it will involve
small spills from 55-gallon drums. Most large spills are already
regulated under the Spill Prevention, Control, and Countermeasures
(SPCC) plan. Because this study is concerned with contamination
from unregulated oil spills, environmental contamination due to
rainfall runoff is not considered further in this report.
3.1.3 Seepage
Seepage of oil (i.e., slow movement of oil through the pore
spaces of the soil) from waste oil storage sites is the primary
3-3
-------
mechanism of oil movement from a waste oil storage site. Seepage
is especially important because it is often a gradual process
that may go undetected for years and can result in significant
environmental contamination.
3.1.3.1 Oil Movement—
Movement of oil through the soil is limited either by the
rate of oil loss from the storage system or by the nature of the
soil environment. When the rate at which oil enters the soil
exceeds the rate at which it can pass through the soil, soil
characteristics control the rate of movement. In some situations,
however, the rate of oil release may be the variable controlling
oil migration rates, e.g., when the leak rates are very low or
the soil is highly permeable.
As oil enters the soil environment, it gradually coats
available soil surfaces, and if enough oil is present, it fills
the pore spaces within the soil. The oil then moves gradually
downward or laterally, following the path of least resistance
until a boundary is encountered. One such boundary is the ground-
water table. When the oil encounters the groundwater, it spreads
laterally, floating on the groundwater surface. Although some
dissolution of the oil components in water will occur, this is a
slow process because of the low solubilities of oil components.
The solubility of oil and its contaminants in water is not ad-
dressed in this study.
3-4
-------
3.1.3.2 Fate of Oil Components—
During oil migration, some interaction can be expected to
occur between oil components and soil particles. This process,
known as attenuation, usually involves the retention of selected
oil components on soil particles while the main body of oil
continues to migrate. This results in the reduction of the
concentrations of some oil components during soil migration. The
extent of such reductions cannot be predicted readily because of
the variety of factors that influence the interactions between
oil and soil particles. Two major generalizations can be made,
however: 1) positively charged ions and some polar organics
(e.g., trichloroethylene and most organo-metallic compounds) tend
to be adsorbed onto soil particles that have a negatively charged
surface (e.g., clay); and 2) nonpolar organic compounds (e.g.,
benzene or xylene) are more readily attenuated by soils that are
high in organic matter. Because organic soils are generally
located within a few feet of the surface, attenuation of nonpolar
organics should be fairly limited. In the case of clay-type
soils (which extend much deeper), metals attenuation and attenua-
tion of some polar organics should continue throughout the oil
migration processes.
3.1.3.3 Effect of Organics on Soil Permeability—
Recent research has provided evidence that several organic
liquids may have a significant effect on the permeability of clay
234
soil. Both laboratory and field studies have been undertaken
to examine the permeability of compacted clay liners exposed to
3-5
-------
organic liquids. Initial laboratory tests examined permeability
changes when the liners were exposed to a xylene paint solvent
waste and a contaminated acetone waste. The three types of clay
liners tested consisted of selected clay minerals mixed with
sandy loam soil to achieve permeabilities to water normally
considered acceptable for waste impoundments (^10 cm/s). The
results of the tests indicated that permeabilities of the clay
liners were 2 to 3 orders of magnitude greater for these waste
materials than for water; thus the liquids continued to move
through the soils at a relatively rapid rate.
Subsequent tests were conducted for several other organic
liquids, including diesel fuel and parafin oil. ' All of the
liquids have reportedly caused changes in the clay soil. The
soil changes that resulted in liner failures may have been caused
by reactions that dissolve portions of the soil or by reactions
that remove water from the soil and produce changes in soil
volume changes. Reactions caused by the organic liquid may
increase soil pore openings and/or cracks in the soil and thereby
increase the permeability.
The measured permeabilities with parafin oil and diesel fuel
were also approximately 2 to 3 orders of magnitude greater than
for water. This suggests that oil could flow through a clay
liner between 100 and 1000 times faster than water.
The results of this recent research are believed to be of
considerable importance to this study with respect to the esti-
mated effectiveness of clay liners in containing waste oil spills.
3-6
-------
Whereas the use of equations may indicate that soil liners of
only a few inches thickness should contain an oil spill for years
(or perhaps even hundreds of years), evidence would seem to
indicate the contrary. Actually, it may be optimistic to count
on a soil liner to contain a waste oil spill.
Although these experiments ' were not performed on waste
oil, they certainly raise important questions about the influence
of contaminated waste oil on soil characteristics. The remainder
of this analysis is based on the assumption that the use of
traditional soil parameters is valid; however, future research
may show that this is not the case.
3.1.3.4 Residual Saturation—
As soon as oil is spilled onto a soil surface, it begins to
seep into the soil. Once the pool of oil at the surface has been
exhausted, saturated flow (i.e., flow in which all the voids are
filled with oil) ceases and unsaturated flow begins. Even though
all the pores are not filled in the saturated flow mode, oil will
continue to migrate downward, coating soil particles as it travels.
Some residual oil will remain coated on the particles and some
will fill the dead-end spaces. Eventually, oil migration stops
completely because all of the oil is coated on soil particles or
trapped in pore spaces. The oil that remains after migration has
ceased is referred to as residual saturation. In theory, stability
occurs when the residual saturation level is reached. Whereas
the potential for some soluble components to reach the water
3-7
-------
table remains, the threat is far less than it would be had the
oil reached the groundwater surface. Therefore, reduction to the
residual -saturation level before reaching the water table removes
much of the pollution risk. The amount of residual oil in soil
pores varies according to type of oil, soil type, and moisture
content, but up to 20 percent of the void space may be occupied
by residual oil.
3.2 ABOVE-GROUND TANKS
Oil loss from above-ground tanks may result from surface
spills or leaks in the tank bottoms. Surface spills may be
uncontrolled or they may be contained within a secondary contain-
ment system. Uncontrolled spills are not considered in this
report because most are already regulated by SPCC Plans. Seepage
is the mechanism of oil movement that is of concern when tank
bottom leaks or contained surface spills occur. Evaporation of
surface spills will occur, but it is of less concern for reasons
described previously.
3.2.1 Surface Spills Within Secondary Containment Systems
Surface spills that occur within a secondary containment
system are prevented from leaving the site by a berm (earthen
ridge), lagoon, or some other system. The oil generally forms a
pool, the deepest part of which is adjacent to the berm or dike
wall. In many cases, the bottom of the diked area consists of
natural soils, so some seepage can be expected to occur.
3-8
-------
Seepage is affected by soil characteristics, oil properties,
the nature of the spill, and time until oil cleanup. The Green-
6
Ampt equation can be used to describe the seepage process (Appen-
dix B). The Green-Ampt model, developed in 1911 and normally
used for water, approximates the dynamics of the liner infiltra-
tion event. In this analysis, the Green-Ampt model has been used
to calculate the effectiveness of the secondary containment
system in preventing the spilled oil from reaching the ground-
water.
Three major soil types are considered: clay, silt, and
sand. For each soil type, values of soil parameters (porosity,
soil moisture, and relative conductivity) were chosen to simulate
low-permeability, high-permeability, and average-permeability
soil. Literature agreement is poor regarding the value of one of
the parameters of the Green-Ampt equation—capillary force.
Capillary force is a surface tension phenomenon that tends to
attract liquids to soil and therefore increases the speed of oil
propagation. Capillary forces are largest for dry soils. Because
of the scatter in the literature data, a range of capillary force
values (negative numbers listed in Appendix B) was used in this
analysis.
Table 3-1 presents the Green-Ampt calculations in terms of
the time it would take for a spill with a liquid depth above the
soil of 10, 30.5 , or 50 centimeters to penetrate the soil to a
depth of 30.5 centimeters (12 inches). The large variation of
times for a specific soil and spill depth is a consequence of the
uncertainty in capillary force values. High-permeability soils
3-9
-------
TABLE 3-1. TIME REQUIRED FOR PENETRATION OF SPILLED OIL.
TO A DEPTH OF 30.5 cm (12 inches) FOR VARIOUS SOIL TYPESfi
Soil
type
Clay
Silt
Sand
Spill
depth,
cm
10.0
30.5
50.0
10.0
30.5
50.0
10.0
30.5
50.0
Low soil
permeability
Low High
Time, years
5,700 176,000
5,620 115,000
5,530 87,100
Time, years
23.5 202
21.9 122
20.6 89.5
Time, days
2.23 3.98
1.61 2.34
1.28 1.69
High soil
permeability
Low High
Time, years
8.56 264
8.42 173
8.29 131
Time, days
1.37 11.8
7.28 7.17
1.20 5.21
Time, min
7.55 13.5
5.46 7.92
4.32 5.73
Average soil
permeability
Low High
Time, years
38.2 1,180
37.6 772
37.0 583
Time, days
3.24 27.9
3.02 16.9
2.84 12.3
Time, min
12.5 22.3
9.03 13.1
7.16 9.48
Calculations based on soil seepage factors listed in Table B-4 of
Appendix B.
3-10
-------
represent the worst-case scenarios, and the highly negative
values for capillary force result in the fastest rates of oil
penetration into the soil. Worst-case times for clays indicate a
minimum of 8 years to reach a 12-inch (30.5-cm) soil depth,
whereas times for sand are as low as 6 minutes, and for silt, are
generally measured in days.
3.2.2 Tank Bottom Leaks
Above-ground tanks may develop leaks anywhere on their
surfaces, but visible leaks should be detected and repaired quite
rapidly. Leaks that occur on the tank bottoms, however, may go
undetected for long periods of time. Seepage of oil as it leaves
the tank bottom can be described by use of the Green-Ampt equa-
tion. The ranges of soil and oil characteristics are the same as
described for above-ground spills (Section 3.2.1), but the head
within the tank greatly exceeds that which could occur in a
secondary containment system. Ranges of typical tank sizes and
levels of contained oil are given in Table 3-2. These values
have been used to determine probable times of migration to the
groundwater table for depths of 100 and 1000 centimeters (Tables
3-3 and 3-4). The range in times listed for a given type of soil
results from using a range of capillary force values, as described
in the previous section.
Leak rates from the tanks have not been estimated for this
part of the calculations because it has been assumed that soil
characteristics will control the rate of oil flow under worst-
case conditions.
3-11
-------
TABLE 3-2. ABOVE-GROUND TANK SIZES AND TYPICAL OIL LEVELS
Vertical tank
Horizontal tank
Tank
size,
gallons
10,000
25,000
50,000
200,000
10,000
25,000
Typical
configuration,
height:diameter
3:1
2:1
1.5:1
1:1
1:3
1:2
Typical
height,
meters
9
8
8
6
3
4
Typi ca 1
oil level,
meters
6
5
5
4
2
2.5
3-12
-------
TABLE 3-3. OIL MIGRATION TIME FROM AN ABOVE-GROUND TANK
TO A WATER TABLE 100 CENTIMETERS DEEP3
Soil
type
Clay
Silt
Sand
Tank
head,
cm
200
500
750
200
500
750
200
500
750
Low soil
permeability
Low
High
Time, years
51,500
42,700
37,400
274,000
131,000
91,300
Time, years
137
86.7
66.4
264
124
86.3
Time, days
4.49
2.20
. 1.54
4.88
2.29
1.59
High soil
permeability
Low
High
Time, years
77.2
64.1
56.2
411
196
137
Time, days
8.02
5.07
3.88
15.7
7.25
5.04
Time, min
15.2
7.46
5.24
16.5
7.76
5.38
Average soil
permeability
Low
High
Time, years
344
286
251
1,830
877
611
Time, days
18.9
12.0
9.16
36.4
17.1
11.9
Time, min
25.2
12.3
8.67
27.4
12.8
8.91
Calculations based on soil seepage factors listed in Table B-4 of
Appendix B.
3-13
-------
TABLE 3-4. OIL MIGRATION TIME FROM AN ABOVE-GROUND TANK
TO A WATER TABLE 1,000 CENTIMETERS DEEP9
Soil
type
Clay
Silt
Sand
Tank
head,
cm
200
500
750
200
500
750
200
500
750
Low soil
permeability
Low
High
Time, years
3,710,000
3,220,000
2,910,000
9,530,000
6,730,000
5,460,000
Time, years
6,700
5,160
4,340
9,030
6,360
5,140
Time, days
161
114
166
117
93.0 94.6
i
High soil
permeability
Low
High
Time, years
5,560
4,840
4,366
14,300
10,100
8,190
Time, days
391
301
254
527
371
300
Time, min
9.08
6.46
5.25
9.39
6.60
5.34
Average soil
permeability
Low
High
Time, years
24,800
21,600
19,500
63 ,800
45,100
36,600
Time, days
2.53
1.95
1.64
3.41
2.40
1.94
Time, min
15.0
10.7
8.70
15.5
10.9
8.85
Calculations based on soil seepage factors listed in Table B-4 of Appendix B.
3-14
-------
Migration times are strongly influenced by both soil type
and permeability. The longest migration calculated times are for
clay soils, followed by silt. Migration of oil through sand is
the most rapid. In general, low soil permeabilities result in
much longer migration time. Factors that reduce the permeability
of soils to oil are high moisture content, low porosity, and low
relative conductivity.
There is a nonlinear relationship between the water table
depth and the time to reach the water table. Tables 3-3 and 3-4
show that it takes much more than 10 times longer for oil to
migrate 1000 centimeters than to migrate 100 centimeters. Actu-
ally, it usually takes almost 100 times as long. This nonlin-
earity results from the fact that the hydraulic gradient (rate of
change of head with depth below the surface) varies in a nonlinear
fashion.
3.3 BELOW-GROUND TANKS
Storage tanks that are located underground are subject to
develop leaks. The rate at which oil leaves the tank is deter-
mined by the head of oil within the tank, the size of the leak,
and soil characteristics. For the purposes of this analysis, it
has been assumed that soil characteristics rather than leak size
are the primary factors controlling the rate of oil migration.
Once again, the Green-Ampt equation (Appendix B) is used to
describe oil movement through the soil environment. The range of
soil and oil characteristics is the same as used in previous
analyses (Table 3-5).
3-15
-------
TABLE 3-5. OIL MIGRATION FROM A BELOW-GROUND TANK
TO A WATER TABLE 100 CENTIMETERS DEEP*
Soil
type
Clay
Silt
Sand
Tank
head,
cm
60
120
180
60
120
180
160
120
180
Low soil
permeability
Low
High
Time, years
56,900
54,500
52,200
568,000
388,000
296,000
Time, years
189
163
143
338
227
171
Time, days
8.81
6.22
4.82
10.7
7.09
5.32
High soil
permeability
Low
High
Time, years
85.4
81.7
78.3
851
582
443
Time, days
11.0
9.49
8.34
32.9
22.0
16.6
Time, min
29.9
21.1
16.3
35.6
23.8
17.9
Average soil
permeability
Low
Time, j
381
365
350
High
fears
3,800
2,598
1,980
Time, days
26.0
22.4
19.7
77.7
52.1
39.3
Time, min
49.4
34.9
27.1
59.0
39.3
29.6
Calculations based on soil seepage factors listed in Table B-4 of
Appendix B.
3-16
-------
Results are very similar to those for above-ground tanks, as
would be expected. Once again, soil permeability is the major
factor influencing migration rates. The differences between
predicted migration times for above- and below-ground tanks are
the result of differences in the head or oil depth typically
found in the two tank types. Above-ground tanks are taller, and
the increased head results in slightly faster migration times
than those predicted for below-ground tanks.
3.4 SPILLS FROM CONTAINERS AND DRUMS
This analysis evaluates the extent of soil contamination
resulting from spills or leaks from establishments that store
waste oil in drums and have no secondary containment systems.
Approximately 12 million gallons of waste oil is stored in drums
(primarily 55-gallon drums) at more than 141,000 sites. The
contaminated soil volume and depth of oil penetration resulting
from spills are calculated.
Two spill scenarios are considered: 1) a catastrophic spill
resulting in a sudden loss of waste oil stored on the site, and
2) the cumulative effect of a number of sequential spills such as
might occur during a drum-filling operation.
3.4.1 Catastrophic Spills
The average number of waste oil drums per site is three,
although some have as many as 10. Thus the maximum amount of
oil that could be spilled in a single incident (e.g., vandalism
or fire) is 550 gallons. A typical drum storage guantity is
3-17
-------
considered to be four drums (220 gallons), and that is the amount
used in the models described in this section.
When a spill occurs at a site that has no secondary contain-
ment, oil will flow in the direction of the slope of the land.
It will spread and form a thin layer over the surface; the area
covered will depend on the type of land surface and the viscosity
of the oil. The model used for this analysis is a simple one
that was developed for Arctic regions based on data from spills
in Canada.
Ae = 53.5 V °'89 (Eq. 1)
5 5
where
A is the spill area in square meters.
5
V is the spill volume in cubic meters
5
In the more temperate regions of the continental United
States, the spill area may be considerably larger because of the
warm weather and the resulting lower oil viscosity. When the
spill area from Equation 1 is used, a worst case for oil penetra-
tion into the soil may be approximated, because as the soil area
decreases with a constant volume, the depth of penetration in-
creases. The spill area (as calculated from Equation 1) for
spills up to 550 gallons is shown in Figure 3-1 as a function of
spill volume. Typical spill areas range from 13 square meters
(140 square feet) for a 55-gallon spill to 45 square meters (490
square feet) for a 220-gallon spill. These areas result in an
average spill depth of 1.6 cm (0.63 in.) for a 55-gallon spill
and 1.8 cm (0.71 in.) for a 220-gallon spill. The model assumes
3-18
-------
100-
80.
;60-
40-
20-
10-
1200
1000
800
400-
200
100
55 110
220 330
SPILL VOLUME.galIons
330
440
Figure 3-1. Spill area versus spill volume for oil spills without secondary containment.
-------
that the spill is so abrupt that the liquid spreads out before
significant quantities penetrate the soil.
The movement of oil through soil has been described by Van
Q Q in
Dam, Schwille, and Dietz. The Green-Ampt equation is not
used for noncontained spills because the liquid head above the
surface is continuously changing and is equal to zero soon after
the spill. After the spilled oil enters the soil, it begins to
seep vertically downward under the influence of gravity (with
some lateral movement) until saturated flow conditions are met.
When all of the spilled oil has entered the soil, the oil will
continue to move downward under unsaturated flow conditions, and
some of the oil will be left behind as it passes through the
soil. Oil movement will eventually cease, and coated soil with
part of the pore spaces filled will be left behind. Residual
saturation levels for oil in soil have been measured for several
oil types. These levels, which are expressed as a fraction of
total spaces in the soil that are filled with oil, are 0.2 for
lube oil and heavy fuel oil; 0.15 for diesel and light fuel oil;
and 0.10 for light oil or gasoline.
Assuming the spilled oil has completely soaked into the soil
and that the groundwater level is deeper than the soil penetra-
tion depth, the volume of oil-contaminated soil depends on the
soil porosity and the residual saturation according to the fol-
lowing relationship:
V soil = Vs (Eq. 2)
3-20
-------
where
V soil = Volume of oil-contaminated soil, m
V = Spill volume, m
s
n = Soil porosity or ratio of void space to total soil
volume after correction for moisture content
S = Residual saturation
In Figure 3-2, the volume of contaminated soil is plotted as
a function of spill volume for soil porosity of 50 percent (typi-
cal for clay) and residual saturations of 0.2, 0.15, and 0.1.
The worst case for volume of contaminated soil occurs with light
oil because the residual saturation of this oil is the lowest.
Typical volumes for light oil range from 4.2 cubic meters (148
cubic feet) for a 55-gallon spill to 16.7 cubic meters (590 cubic
feet) for a 220-gallon spill.
The volume of contamination is a function of the type of
soil, in that a more porous soil holds more oil and, therefore,
less total volume of soil is contaminated as a result of a spill
of a given size.
Table 3-6 lists representative porosity ranges for dry
gravel, sand, silt, and clay. The effective porosity for the
purposes of calculating saturated soil volume is obtained by
subtracting the soil moisture content from the porosity of the
dry soil.
The effect of porosity of contaminated soil volume is shown
in Figure 3-3. A spill volume of 220 gallons was assumed for
this analysis. Over 33 cubic meters (1165 cubic feet) of low-
porosity sand or gravel may be contaminated with a 220-gallon
3-21
-------
CO
I
to
to
40-
r30-
20-
10-
1500
1250 -
1000 -
£ 750 -
0 55 110
220 330
SPILL VOLUME, gallons
440
Figure 3-2. Volume of saturated soil versus spill volume for a soil porosity of 50 percent.
-------
I
K>
10
40-
«•»_ 30-
i
§20-
10-
1500
1250
1000
+• 750
500
250
[*— GRAVEL *\
SAND
RESIDUAL SATURATION'
0.10 (light oil)
RESIDUAL SATURATION >
0.15 (residual oil)
RESIDUAL SATURATION-
0.20 (lube oil)
10
20 30 40
SOIL POROSITY, percent
50
60
70
Figure 3-3. Volume of contaminated soil versus soil porosity for a spill volume of 220 gallons.
-------
catastrophic spill of light oil or gasoline. On the other hand,
if the spill is on a highly porous clay, the contaminated soil
volume would be only 12 cubic meters (424 cubic feet). The range
for a lube oil spill of 220 gallons is from 16.7 cubic meters
(590 cubic feet) for a low-porosity sand to 6 cubic meters (212
cubic feet) for a porous clay.
TABLE 3-6. RANGE OF VALUES OF POROSITY3
Soil type
Gravel
Sand
Silt
Clay
Porosity, %
25-40
25-50
35-50
40-70
a Source: Reference 11, p. 37.
The maximum depth reached by the spilled oil is obtained by
assuming there is no lateral flow and that the groundwater level
is too deep to be reached by the oil. In this case, the contami-
nated soil volume can be approximated by a cylinder. Because the
volume of a cylinder is equal to the area of its base times its
height, Equation 2 can be rewritten as follows to determine the
depth:
d = Vsoil = Vs (Eq. 3)
As nSrAs
where
d = Spill penetration depth, m
s
V ., = Volume of oil-contaminated soil, m
soil
3-24
-------
V = Spill volume, m
s
n = Soil porosity corrected for moisture
S = Residual saturation
A = Area of the oil spill, m
S
The depth of penetration for a 220-gallon spill is shown as
a function of soil porosity in Figure 3-4 for residual satura-
tions of 0.1, 0.15, and 0.20. Typical porosity ranges for gravel,
sand, silt, and clay are indicated on the figure. The maximum
penetration depth is obtained when light oil is spilled on a
low-porosity sand or gravel, where the depth of penetration is
0.73 meter (2.4 feet). At the other extreme, a spill of lube oil
on a porous clay results in a penetration depth of 0.26 meter
(0.86 foot).
In most oil spills some horizontal spreading will occur as a
result of capillary forces and produce a coning effect; therefore,
penetrations will be shallower than shown in Figure 3-4. The
angle of the resulting cone of contaminated soil depends on soil
characteristics. Low-permeability soils (such as clay and some
silts) generally result in more horizontal spreading. Quantita-
tive experimental data are limited, but laboratory model experi-
g
ments by Schwille have demonstrated that minor differences an
permeabilities laterally or vertically can produce strong distor-
tions in the shape of the oil migration zone. Undisturbed soils
may exhibit permeabilities 10 or more times greater in one direc-
tion than the other. In general, fluvial deposits (deposits laid
down by physical processes in river channels or floodplains) have
3-25
-------
u>
I
KJ
0.8-
0.7-
vt
t-
0.6-1
0.5-
0.3-
0.2-
0.1-
0
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
GRAVEL -
RESIDUAL SATURATION
0.1 (light Oil)
RESIDUAL SATURATION
0.15 (dlesel oil)
RESIDUAL SATURATION
0.20 (lube oil)
SAND
10
20 30 40
SOIL POROSITY, percent
50
60
70
Figure 3-4. Depth of spill penetration versus soil porosity for a catastrophic spill of 220 gallons.
Spill area is 45 square meters (489 square feet).
-------
higher horizontal permeabilities and aeolian deposits (materials
that are transported and deposted by wind) have higher vertical
permeabilities. The shape of the contaminated soil volume in
an aeolian soil can be closely approximated by a cylinder. In a
homogeneous, isotropic soil, a cone angle of approximately 45
degrees, with a rounding of the leading front, may be expected.
Of course, fractures, root channels, and animal burrows can cause
a secondary permeability in the vertical direction that would
result in even much greater depths than those calculated by
assuming a cylindrically shaped zone.
The sensitivity of penetration depth to cone angle can be
calculated by the following equation:
1/3
3V
soil + r3
tan2<|> tan3
tan
(Eg. 4)
where
d = depth of spill penetration, m
5
V ., = volume of oil-contaminated soil, m3
41 = 1/2 angle of the cone (see Figure 3-5)
r = spill radius on the surface, m (see Figure 3-5)
A derivation of Eguation 4 is based on the cone as shown in
Figure 3-5. See Appendix C for the derivation.
3-27
-------
Spill Surface
Figure 3-5. Geometry of cone assumed for calculating penetration depths.
Figure 3-6 shows how the depth of spill penetration varies
with cone angle for a residual saturation of 0.1 (typical for
light oil or gasoline). For the rather shallow depths under
these conditions (less than 1 meter), a cone angle up to 45
degrees affects the maximum depth by less than 20 percent.
Based on the analysis above, a single catastrophic spill
from a waste oil drums storage location should not result in
groundwater contamination. A worst-case spill, where four 55-
gallon drums are spilled on low-porosity sand or gravel, results
in a large area of surface spread, but even without lateral
3-28
-------
NJ
VO
0.8-
.
s
0.7-
0.6-
&0.4-
0.3-
0.2-
0.1-
30 40
POROSITY, percent
50
60
70
Figure 3-6. Depth of spill penetration versus soil porosity for three cone angles for residual
saturation of 0.1. Assumes a catastrophic spill of 220 gallons and spill area of
45 meters (489 square feet).
-------
spreading after entering the oil, the penetration depth would be
less than 1 meter.
3.4.2 Sequential Spills
If a site generates 220 gallons of waste oil per month and
places the oil in drums (typical for a service station) in 5-gallon
increments, oil would be added to the drums 44 times per month.
If the fill operation is a careless one, as much as 1/2 gallon of
oil could be spilled without being cleaned up (although 1 pint or
less would be more typical. Thus, the monthly quantity of spilled
oil could vary from 5 to 22 gallons.
Each of the small spills would be unlikely to occur at
exactly the same spot; instead they probably would occur at all
four drum locations (four drums are assumed for this analysis).
Thus, according to Equation 1, the spill area covered by the 44
small spills would be four times the area calculated for a single
spill. The calculated spill areas from a series of one-pint or
half-gallon spills are as follows:
Spill area, m2
Spill volume Single spill Four locations
1 pint 0.059 0.235
1/2 gallon 0.202 0.807
The volume of contaminated soil from sequential spills is
calculated by using Equation 2 in the same manner as for the
catastrophic spills, where the spill volume is the cumulative
volume of oil spilled. For example, a series of 44 one-pint
3-30
-------
spills of light oil in a month could contaminate about 0.42 m
(14.7 ft ) of clay soil with a 50 percent porosity. If the
spills occurred in half-gallon increments, the contaminated soil
3 3
volume would be 1.66 m (58.8 ft ). Spills of heavier oils would
result in less contaminated soil.
The depth of soil penetration from a month-long accumulation
of spills can be calculated by using Equation 3 (cylindrically
shaped contaminated soil volume) and Equation 4 (cone-shaped
volume). The spill volume (V ) in Equation 3 represents the
5
month's accumulation of spills, and the spill area is an area
equal to four times the area calculated from a single small spill
(four drum locations).
Figure 3-7 shows the maximum depth of spill penetration
versus soil porosity for a total spill volume of 22 gallons,
where the spill area is based on four half-gallon spills. This
should represent a worst-case month for a very careless four-drum
service station operation.
The maximum depth of 4.1 meters (13.5 feet) is obtained if
light oil is spilled on gravel or sand of low porosity. Lube oil
spilled on the same gravel or sand would penetrate only to a
depth of about 2.1 meters (6.8 feet). A spill on a high-porosity
clay would range from a depth of 1.5 meters (4.8 feet) for light
oil to a depth of 0.75 meter (2.4 feet) for heavy lube oil.
Figure 3-8 shows the sensitivity of the penetration depth to
the spreading cone angle for light oil (deepest penetration in
Figure 3-7). This graph shows that a cone angle of 45 degrees,
3-31
-------
I
U)
4-
$
S3-
2-1
1-
14
12
10
$ 8
RESIDUAL SATURATION
0.10 (light oil)
RESIDUAL SATURATION
0.15 (diesel oil)
RESIDUAL SATURATION
0.20 (lube oil)
10
20 30 40
SOIL POROSITY, percent
50
60
70
Figure 3-7. Depth of spill penetration versus soil porosity for periodic spills of one-half gallon each.
Assumes 44 spills per month at four spill areas - total spill area is 0.81 square meters
(8.69 square feet). Total spill volume is 22 gallons.
-------
U)
I
u>
Ul
4-
«ft
L.
g
53H
i
L
55 2^
&
1-
14
12
10
3 8
10
20 30 40
POROSITY, percent
50
60
70
Figure 3-8. Depth of spill penetration of soil porosity for three cone angles for residual saturation
of 0.1 (light oil). Assumes 44 sequential spills of one-half gallon each at four locations.
Total spill area is 0.81 square meters (8.69 square feet). Total spill volume is 22 gallons.
-------
which may be representative of a homogeneous isotropic soil,
would reduce the penetration depth by a factor of 2 to 4, with
the greatest reduction in depth occurring for the low-porosity
soils.
Figure 3-9 shows spill penetration depth versus oil porosity
for a typical service station with 44 one-pint spills in a month.
This size spill is expected to be more typical for a four-drum
service station operation than the half-gallon spills. It should
be noted that although the quantity of oil spilled at each occur-
rence and the total oil spilled per month are reduced by a factor
of 4, the penetration depth is reduced by only 14 percent. The
reduction in penetration depth is much less than the reduction in
spill volume because a smaller spill area results from the one-pint
spills.
Figure 3-10 shows the sensitivity of penetration depth to
the spreading cone angle for 44 one-pint spills of light oil.
This graph, which is similar to the graph in Figure 3-8, shows
that a spreading angle of 45 degrees would reduce the penetration
depth by a factor of 2 to 4, with the greatest reduction in depth
occurring for the low-porosity soils.
This analysis represents typical spills over a month's time.
If the spills are chronic and continue over a long period of
time, the depths can be significantly greater than those shown,
and groundwater is likely to be contaminated eventually. Accord-
ing to the model assumed, after the residual saturation level of
the soil has been reached, any new spilled oil simply flows
3-34
-------
U)
U)
Ul
4-
C
5
3-
12"
fe
Q.
Ul
o
1-
14
10
£ 8
k GRAVEL H
RESIDUAL SATURATION
0.10 (light oil)
RESIDUAL SATURATION
0.15 (dlesel fuel)
RESIDUAL SATURATION
0.20 (lube oil)
10
20 30 40 50
SOIL POROSITY, percent
60
70
Figure 3-9. Depth of spill penetration versus soil porosity for periodic spills of one pint each
in four spill areas. Assumes 44 spills per month. Total spill area is 0.24 square
meters (2.53 square feet). Total spill volume is 5.5 gallons.
-------
4-
JB
I 3-
s
I—i
i
U)
I
OJ
UJ
o.
ui
O
2-
14
10
^ 8
i!
10
20 30 40
SOIL POROSITY, percent
50
60
70
Figure 3-10.
Depth of spill penetration versus soil porosity for three cone angles for a residual
saturation of 0.1 (typical for light oil). Assumes 44 spills of one pint each in
four spill areas. Total spill volume is 5.5 gallons.
-------
through the already contaminated soil and the penetration depth
continues to increase until the groundwater table or some other
boundary is reached. Once the groundwater has been reached, each
pint of soil spilled at the surface results in a pint of oil
being introduced into (or on top of) the groundwater.
The calculations of penetration depths shown in this section
are conservative and represent worst-case conditions. At least
three factors can limit the depth of contamination: 1) evapora-
tion of oil from near the surface, 2) decomposition (by bacterial
action) of the oil over a period of time, and 3) soil compaction.
The evaporation rate is temperature-dependent, and can be signifi-
cant for the lighter oils. Significant quantities of oil can be
decomposed by bacterial action, but decomposition requires several
months and depends on ground temperature, soil moisture content,
and type of oil. Under ideal conditions, such as at land dis-
posal sites where the soil is periodically mixed, the rate of
decomposition can be as high as 60 pounds of oil per cubic foot
12
of soil per month. At waste storage sites, however, natural
biodegradation is expected to be extremely slow, primarily be-
cause of the lack of oxygen. With regard to soil compaction, the
upper layer of soil will be compacted if the spill area is heavily
traveled, and the oil will spread out further before it enters
the soil. This causes the oil to stay near the surface, where
rates of evaporation and biodegradation are the highest.
3-37
-------
Thus, in some cases, sequentially spilled oil will be miti-
gated and will not reach groundwater, but in a large number of
cases, these mitigative factors will not be significant, and
small amounts of oil will reach groundwater.
3.5 SUMMARY
In this section, the fate of waste oil that has entered the
environment through spills or leaks from oil storage containers
has been examined. Such entrance of waste oil components into
the environment has been evaluated according to storage container
type, spill or leak processes, soil characteristics, and other
hydrogeological factors. This evaluation indicates that seepage
of oil into the subsurface environment is the area of greatest
concern. The following subsections summarize the findings that
were discussed in detail in this chapter.
3.5.1 Storage Tanks
Spills from storage tanks can be classified into two main
groups: slow leaks and contained surface spills. The rate of
contamination of the subsurface environment from leaks and con-
tained surface spills depends primarily on soil and hydrogeolog-
ical characteristics. The greater the soil permeability, the
higher will be the oil migration rates for spilled oil and leaks
from above- and below-ground tanks. Predictions based on the
Green-Ampt equation (Appendix B) indicate that oil penetrates
sand and some silts very rapidly, but its movement through clay
is much slower. It is assumed that passage of oil through the
3-38
-------
soil does not alter its permeability. This assumption was neces-
sary because, even though recent research indicates that some
permeability increase may occur, quantification of this effect is
not possible based on current knowledge. Green-Ampt predictions
for clay are especially likely to be much too high.
3.5.1.1 Above-ground Tanks—
Oil can escape from an above-ground tank through a leak in
the tank bottom or as a result of an above-ground spill. A tank
bottom leak or rupture that goes undetected could result in
severe groundwater contamination. If an above-ground tank is
placed on a sand bed, groundwater could become contaminated in a
matter of minutes. If placed on silt, contamination could occur
within a few weeks. A clay bed affords a significant degree of
protection if the soil structure has not been altered by interac-
tion with the oil.
A surface spill of oil from an above-ground storage tank can
result in rapid oil loss if the soil within the containment area
is not low in permeability. Even with immediate detection and
efficient cleanup, oil is likely to remain within the berm area
for several hours, during which time significant soil contamina-
tion can be expected to occur if the surrounding soil is sand;
some contamination can be expected if the soil is silt. If the
soil is not removed following cleanup of the pooled oil, ground-
water contamination may occur from the leaching of oil components.
Clay is the only soil with a slow enough rate of oil penetration
to allow for safe oil cleanup without soil removal. Also, short
3-39
-------
exposure of the clay bed to oil is not likely to cause the perme-
ability changes some researchers have observed to occur during
prolonged exposure.
3.5.1.2 Below-ground Tanks—
Leaks in storage tanks located underground will go unnoticed
until they are large enough to be detected visually or by a
monitoring system (if one is used). A severe tank failure or
rupture can result in rapid groundwater contamination. A leaking
below-ground tank could cause groundwater contamination in less
than an hour in sandy soil and in just over a week in a silty
soil. It is predicted that it could take more than 75 years for
such contamination to occur in clay soil, but this prediction
does not consider the increases in soil permeabilities that can
result from long-term exposure to organics.
Because below-ground tanks are more likely to develop leaks
than above-ground tanks are, and because these leaks are likely
to go undetected for longer periods of time, storage of waste oil
in below-ground tanks presents a greater risk to the environment.
3.5.2 Containers and Drums
Two types of spills that may be expected to occur at waste
oil drum storage facilities were analyzed: the catastrophic
spill and sequential small spills. In the catastrophic spill,
which represents one extreme, all the oil stored at a particular
site is spilled in a single incident. This type of spill results
in the spread of oil over a large surface, but usually results in
relatively shallow penetration. A single spill of 220 gallons or
3-40
-------
less would not be likely to contaminate the groundwater. On the
other hand, a series of small sequential spills will spread over
a much smaller area than the single catastrophic spill (even if
the same total quantity of oil is spilled over a month's time),
but the depth of oil penetration under these circumstances can be
quite high. Over a long period of time, the oil from small
spills at one location (such as a service station) can be expected
to reach the groundwater; however, the volume of oil that reaches
the groundwater will probably be too small to cause significant
water quality deterioration. Cleanup also may be more difficult
and expensive for the small sequential spills than for a sudden
quickly recognized spill.
3-41
-------
REFERENCES FOR SECTION 3
1. Metzler, S. C., et al. Evaluation of Health and Environ-
mental Problems Associated With the Use of Waste Oil as a
Dust Suppressant. (Draft report) Prepared by Franklin
Associates, Ltd., and PEDCo Environmental, Inc. for U.S.
Environmental Protection Agency. May 1983.
2. Brown, K. W. Impact of Organic Liquids on Clay Liners and
the Landfill of the Future. Presented at Conference, Working
Together to Manage Wastes, Chicago, Illinois, June 28 and
29, 1983.
3. Brown, K. W., and D. C. Anderson. Effects of Organic Solvents
on the Permeability of Clay Soils. Project Summary published
by U.S. Environmental Protection Agency, Cincinnati, Ohio.
4. Brown, K. W., J. W. Green, and J. C. Thomas. The Influence
of Selected Organic Liquids on the Permeability of Clay
Liners. U.S. Environmental Protection Agency Grant No.
CR-808824020. 1983.
5. Personal communication from Dr. K. W. Brown to Nicholas S.
Artz, July 6, 1983.
6. Cogley, D. R., D. J. Goode, and C. W. Young. Review of the
Transit Time Equation for Estimating Storage Impoundment
Bottom Liner Thickness. (Draft report) GCA Corporation for
U.S. Environmental Protection Agency. July 1982.
7. Mackey, D., and M. Mohtadi. The Area Affected by Oil Spills
on Land. Canadian Journal of Chemical Engineering, 53:140,
1975.
8. Van Dam, J. Migration of Hydrocarbons in a Water-Bearing
Stratum. In: The Joint Problems of the Oil and Water
Industries, Peter Hepple, ed. Institute of Petroleum, London.
1967.
9. Schwille, F. Petroleum Contamination of the Sub-Soil-
Hydrological Problem. In: The Joint Problems of the Oil
and Water Industries, Peter Hepple, ed. Institute of Petro-
leum, London. 1967.
3-42
-------
10. Dietz, D. N. Pollution of Permeable Strata by Oil Compo-
nents. In: Water Pollution by Oil, Peter Hepple, ed.
Elsiver, Amsterdam. 1971. pp. 128-142.
11. Freeze, R. A., and J. A. Cherry. Groundwater. Prentice
Hall, Inc., New Jersey. 1979.
12. Cheremisinoff, P. N.f and R. A. Young. Pollution Engi-
neering Practice Handbook. Ann Arbor Science Publishers,
Inc., Ann Arbor, Michigan. 1975. p. 598.
3-43
-------
APPENDIX A
DERIVATION OF ESTIMATED FAILURE PROBABILITIES
IN BELOW-GROUND WASTE OIL STORAGE TANKS
The first of the two approaches used to estimate failure (or leak) proba-
bilities in below-ground waste oil tanks was based on applications of the API
study mathematical model. The model was used in conjunction with assumed
conditions relevant to its application and an estimation of failure probabili-
ties in the total system of below-ground waste oil tanks. The mathematical
model predicts the mean age to failure from external corrosion as follows:
Age - 5.75 x R0'05 x s'0'018 x e(0'13 >H ' °'41M ' °'26 Su>
where
R = resistivity of soil in ohm-centimeters
S = tank capacity in gallons
pH = soil acidity
e = 2.72
M = 1 if soil saturated with water, 0 otherwise
Su = 1 if sulfides present in soil, 0 otherwise
The data in Table A-l show estimated failure probabilities for 500-gallon
underground steel tanks subject to localized corrosion based on two sets of
soil conditions. The first set of soil conditions is judged to be typical of
those encountered around below-ground waste oil tanks. A moisture-saturated
soil condition is assumed, however, whereas such soil conditions are estimated
2
to occur in only about 50 percent of the cases. The presence of moisture-
saturated soil is highly important in predictions of failures due to corro-
sion. The second set of soil conditions, which is supposed to represent the
A-l
-------
TABLE A-l. ESTIMATED FAILURE PROBABILITIES FOR 500-GALLON BELOW-GROUND STORAGE TANKS
SUBJECT TO LOCALIZED EXTERNAL CORROSION3
i
K>
Years
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
Probability
under typical
soil conditions,
percent
0.0
0.0
0.0
0.0
0.0
0.1
0.2
0.7
2.0
4.9
10.4
19.5
32.3
47.6
63.3
77.0
87.3
93.8
97.4
99.0
Cumulative
probability
under typical
soil conditions, >c
percent
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.3
0.8
1.7
3.2
5.4
8.4
12.1
16.1
20.3
24.4
28.2
31.8
Probability
under
optimum soil.
conditions,
percent
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.5
Cumulative
probability
under optimum
soil conditions, C'd
percent
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Based on Warren Rogers Associates model applied to various soil properties.(Reference 1).
Assumes the following values for parameters used in Warren Rogers Associates model: soil
resistivity = 4,000 ohm-centimeters; pH = 7.5; moisture = saturated; sulfides = none-
calculated mean age failure = 13.66 years. *
£
Assumes equal tank age distribution.
Assumes the following values for parameters used in Warren Rogers Associates model: soil
resistivity = 30,000 ohm-centimeters; pH = 8.5; moisture = nonsaturated; sulfides = none;
«an to c^j^ur^ _ 25. m ~ca,
-------
optimum for preventing corrosion, assumes that soil is not saturated with
moisture. The mean age of a tank to the outset of a leak was calculated for
each of the two assumed soil conditions by using the mathematical model devel-
oped in the API study. Under the first set of soil conditions, the mean age
of a tank at the time of failure was calculated to be 13.66 years; under the
second set of soil conditions, the mean age was 25.93 years.
The probability of external corrosion failures (Table A-l) for unpro-
tected steel tanks of various ages was based on three mean ages to failure, as
calculated by the mathematical model and the application of tables of the
standard normal distribution (with the use of the estimated 2.5 years standard
deviation) to compute failure probabilities at each age.
The probabilities of failure for tanks up to 20 years of age under both
soil conditions are shown in Table A-l. A comparison of the corresponding
probabilities of failure under these two soil conditions reveals the enormous
impact of soil conditions on tank failures from corrosion. Under the first
set of soil conditions, an underground storage tank with localized corrosion
is expected to fail within 20 years. Under the second set of soil conditions,
there is a less than 1 percent probability of failure within 20 years.
Estimated cumulative probabilities of below-ground tank failures from
external corrosion are based on an assumed equal number of tanks in all age
categories up to 20 years. Thus, for the first set of soil conditions, it is
estimated that 31.8 percent of a given number of tanks (with localized corro-
sion), uniformly distributed between 0 and 20 years of age, are leaking,
whereas essentially none of these tanks would be leaking under the second set
of soil conditions.
A-3
-------
If it is assumed that an average of the figures derived under the two
sets of soil conditions is a reasonable expectation, the cumulative proba-
bility of failure is about 16 percent. The API study estimates that 77 per-
cent of the below-ground unprotected steel tanks experience localized corro-
sion. Application of this figure results in an expectation that at least 12
percent of the tanks uniformly distributed between the ages of 0 and 20 years
are leaking. This is believed to be a conservative estimate for several
reasons. First, the 12 percent only covers expected failures from external
corrosion; failures due to leaks from other causes are not included. Second,
soil conditions generally encountered are believed to be closer to those
assumed in the first set used. Third, the age distribution of below-ground
waste oil tanks is believed to be very conservative (on the average these
2 3
tanks are believed to be older than indicated). ' Finally, the capacity of
many of the below-ground waste oil tanks is larger than 500 gallons; these
larger tanks would be expected to fail slightly earlier from external corro-
sion.
Another approach used to estimate the probability of leaks in below-
ground waste oil tanks involved the use of recent data compiled by Warren
Rogers Associates on predicted national mean tank ages to predict the occur-
rence of leaks in underground gasoline tanks at service stations (Figure A-l).
The expected average age of tanks at the time of failure due to external
corrosion is from 8 to 24 years. A marked number of tank failures occur
between the ages of 10 and 15 years and 19 and 23 years. The earlier pre-
2
dieted failures are reported to be due to moisture-saturated soil ; failures
in dry soil occur later. Nearly all such tanks are expected to have failed by
24 years of age or shortly thereafter.
A-4
-------
NATIONAL MEAN AGE TO LEAK
Percentage
12 -
11 -
10 -
9 -
8 -
7
6 -
5 -
4 -
3 -
2 -
1 -
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Mean age, years
Figure A-l. National mean age to leak.1
A-5
-------
The data in Figure A-l were used to estimate failure probability for
below-ground unprotected steel tanks experiencing localized corrosion, as
shown in Table A-2. The figures in Column 2 represent the probability of
failure for tanks of different ages. For example, the probability of leak(s)
for the tanks between 19 and 20 years old is shown at just over 60 percent.
This assumes that all below-ground tanks remain in service for a full 20 years
even though they have begun to leak earlier.
As a point of clarification, the distinction between Figure A-l and
Column 2 in Table A-2 is as follows. The data in Figure A-l represent the
mean (average) ages at which a leak (resulting from external corrosion) is
expected to begin. The probabilities of a mean age to leak occurrence are
shown for each age category. The probability figures in Column 2 of Table A-2
are based on the information from Figure A-l. For example, the probability
that a below-ground tank 15 years old is leaking is based on the cumulative
probability of a leak beginning during each previous year. This cumulative
probability may be determined from the data in Figure A-l and the application
of a normal distribution with a 2.5-year standard deviation.
The estimated cumulative probabilities of below-ground tank failures from
external corrosion shown in Table A-2 are based on the probabilities by age
group. The estimates are based on the assumption that there is an equal
number of tanks in all age categories up to 20 years. Thus, it is estimated
that more than 18 percent of the tanks with localized corrosion (uniformly
distributed between 0 and 20 years of age) are leaking. If the tanks that do
not have localized corrosion are included, 14 percent of below-ground unpro-
tected steel tanks of these ages are estimated to be leaking.
A-6
-------
TABLE A-2. ESTIMATED FAILURE PROBABILITIES FOR
UNDERGROUND STORAGE TANKS SUBJECT TO
LOCALIZED EXTERNAL CORROSION**
Probability Cumulative probability
based on based on
nationally nationally
determined ages determined ages
Tank age,
years
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
bw ACBIV
occurrence,
percent
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
2.3
7.8
12.5
18.2
27.3
35.7
43.3
48.5
52.7
55.6
60.2
kW .LCOn. «
occurrence,
percent
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.9
1.9
3.2
4.9
6.9
9.2
11.5
13.8
16.0
18.2
Based on application of Warren Rogers Associates national mean age to leak data.
Assumes equal tank age distribution.
A-7
-------
Although these figures are based on data generated to predict failure in
below-ground gasoline tanks, they are judged to be almost equally applicable
to waste oil tanks. Two factors that could result in some differences are the
relatively smaller sizes of the waste oil tanks and their generally shallower
placement underground. Comparison of predicted mean age to failure arrived at
by the API study mathematical model shows little difference between a typical
waste oil tank size and a typical gasoline tank size. Further, soil condi-
tions around the waste oil tanks in service stations are reportedly not sub-
stantially different from those around the larger gasoline tanks.
The probability of leaks in below-ground waste oil storage tanks is
estimated at 12 to 14 percent, based on the two estimation approaches
described. For reasons already cited, these figures are believed to be con-
servative.
A-8
-------
APPENDIX A
REFERENCES
1. Warren Rogers Associates, Inc. Report on the statistical analysis of
corrosion failures of unprotected underground steel tanks. American
Petroleum Institute. January 15, 1982.
2. Personal communication. W. Rogers, Warren Rogers Associates, Newport,
Rhode Island, November and December 1982.
3. Correspondence from Shell Oil Company, Houston, Texas, December 8, 1982.
A-9
-------
APPENDIX B
OIL INFILTRATION INTO SOIL USING GREEN-AMPT MODEL
Green and Ampt derived a simple model of infiltration in 1911. This
model describes the infiltration of soil moisture as a square wave moving down
the soil column (Figure B-l). Above the wetting front, the soil is completely
saturated, but below the wetting front, soil moisture remains at its original
level. The time required for water to penetrate a given depth of the soil
column can be calculated from Equation B-l.
_ n - 6. fh ibH t h+d~^ | (E(l' B-1)
' "-T^L " " n^h-*./_
where
t = time
9. = initial solid moisture content
n = porosity
d = depth of fluid (head) above the soil surface
ifr = capillary pressure
K = hydraulic conductivity
The Green-Ampt wetting front model can be used for estimating oil infil-
tration by simply using the appropriate hydraulic conductivity (K) and capil-
lary pressure (i|>) values for oil. The oil hydraulic conductivity can be
determined for a three-phase system of oil, water, and air by modifying hy-
draulic conductivity values for a one-phase oil system by using the diagram in
2
Figure B-2.
Each point within the triangle corresponds to a different degree of
saturation for air, oil, and water, as indicated on the scales along the sides
B-l
-------
Liner Top
o
Liner Bottom
Impounded Liquid
Green-Ampt
Moisture Content
Figure B-l. Green-Ampt infiltration model.
100% M M 40 20
Figure B-2. Relative permeabilities of three-phase flow.'
B-2
-------
of the triangle. The heavy solid lines, light solid lines, and dashed lines
labeled oil, air, and water, respectively, represent relative permeabilities
for each of the three phases.
For example, consider the determination of oil hydraulic conductivity for
a highly porous sand with the pore spaces filled with oil (70 percent), water
(15 percent), and air (15 percent). The phase diagram in Figure B-2 can be
used to determine that the hydraulic conductivity (permeability) of oil is
reduced to approximately 55 percent (heavy line labeled oil) of the value with
no water or air present. As shown in Table B-l, for a highly porous sand, oil
hydraulic conductivity would be about 1.4 E-2 cm/s, with no soil moisture and
55 percent of that value, or 7.7 E-3, for a soil with a 15 percent soil mois-
ture and 15 percent air content.
TABLE B-l. SEEPAGE FACTORS FOR OIL AND WATER
IN VARIOUS SOILS3
Clay
Silt
Sand
Gravel
Hydraulic conductivity (K)
Porosity
0.40 to 0.70
0.25 to 0.35
0.25 to 0.50
0.25 to 0.40
Oil, cm/s
1.4 E-12 to 1.4 E-9
1.4 E-9 to 1.4 E-5
1.4 E-5 to 1.4 E-2
1.4 E-3 to 1.4
Water, cm/s
E-10 to E-7
E-7 to E-3
E-3 to 1
E-l to E+2
Intrinsic
permeability (K)
cm2
E-15 to E-12
E-12 to E-8
E-8 to E-5
E-6 to E-3
Reference 3.
K = 100 Kg T y, where K = hydraulic conductivity (cm/s), k = intrinsic per-
meability (cm2), g = acceleration due to gravity (9.8 m/s2), y = kinematic
viscosity (0.71 cm2/s for oil and 0.01 cm2/s for water).
The Green-Ampt model is particularly sensitive to values used for capil-
lary pressure. Since the capillary forces attract the liquid molecules to the
B-3
-------
soil particles, their effect is mathematically that of a negative pressure.
Capillary forces vary according to soil type and the amount of moisture ini-
tially present in the soil. The lower the initial soil moisture, the larger
the capillary forces.
The literature values show poor agreement for the negative capillary
pressure for water in various soils. ' ' Values to -80 atmospheres (-1,200
psi) have been reported. More recent reports, however, indicate that these
4 5
extremely negative values are probably not valid. ' Values for the various
soil types have recently been calculated by Rawls et al. to be in the range
of 5 to 50 centimeters (Table B-2).
Capillary pressures for oil in soil are not readily available in the
literature. For this analysis, estimates for oil capillary pressures have
been made based on the data for water and the ratio of oil surface tension to
water surface tension. Use of this ratio should give a reasonably good esti-
mate of capillary pressures in a soil environment.
The effect of soil moisture on oil capillary forces is uncertain. It is
known that at low moistures, water remains preferentially adsorbed to soil
particles as oil passes through the soil-water system. In this situation, oil
will interact with both soil and water surfaces and the capillary forces
exerted will be influenced by these interactions.
Because of the uncertainty in capillary pressures, a range of values was
used for this analysis. The values used for water range from the low values '
in Table B-2 to the high values in a range of textbook values (Table B-3).
Values for oil were derived from these data and adjusted by the ratio of
surface tension. The high values should represent a worst case for infiltra-
tion times.
B-4
-------
TABLE B-2. GREEN AND AMPT PARAMETERS ACCORDING TO
SOIL TEXTURE CLASSES AND HORIZONS*'
Soil texture
class Horizon Total porosity
Sand
Loamy sand
Sandy loam
Loan
Silt loam
Sandy clay loam
Clay loam
Sllty clay loam
Sandy clay
Silcy clay
Clay
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
0.437
0.452
0.440
0.424
0.437
0.457
0.447
0.424
0.453
0.505
0.466
0.416
0.463
0.512
0.512
0.412
0.501
0.527
0.533
0.470
0.348
- c
0.393
0.407
0.464
0.497
0.451
0.452
0.471
0.509
0.469
0.475
0.430
_
0.435
-
0.479
-
0.476
0.464
0.475
-
0.470
0.483
1(0.374-0.500)
(0.396-0.508)
(0.385-0.495)
(0.385-0.436)
(0.363-0.506)
(0.385-0.529)
(0.379-0.515)
(0.372-0.476)
(0.351-0.
(0.399-0.
(O.J52-0.
(0.352-0.
(0.375-0.
(0.427-0.
(0.406-0.
(0.350-0.
(0.420-0.
(0.444-0.
(0.430-0.
(0.409-0.
(0.332-0.
-
(0 310-0.
(0.359-0.
(0.409-0.
(0.434-0.
(0.401-0.
(0.412-0.
(0.418-0.
(0.449-0.
(0.423-0.
(0.436-0.
(0.370-0.
_
(0.371-0.
-
(0.425-0.
-
(0.465-0.
(0.430-0.
(0.427-0.
-
(0.426-0.
(0.441-0.
555)
611)
580)
484)
551)
597)
616)
474)
582)
610)
636)
531)
464)
476)
455)
519)
560)
501)
492)
524)
569)
515)
514)
490)
499)
533)
507)
498)
523)
514)
525)
Effective porosity
0.417
0.431
0.421
0.408
0.401
0.424
0.412
0.385
).412
0.469
0.428
0.389
0.434
0.476
0.498
0.382
0.486
0.514
0.515
0.460
0.330
-
0.330
0.332
0.309
0.430
0.397
0.400
0.432
0.477
0.441
0.451
0.321
_
0.335
-
0.423
-
0.424
0.416
0.385
-
0.412
0.419
(0.354-0.480)
(0.375-0.487)
(0.365-0.477)
(0.365-0.
(0.329-0.
(0.347-0.
(0.334-0.
(0.323-0.
(0.283-0.
(0.330-0.
(0.271-0.
(0.310-0.
(0.334-0.
(0.376-0.
(0.382-0.
(0.305-0.
(0.394-0.
(0.425-0.
(0.367-0.
(0.396-0.
(0.235-0.
_
(0.223-0.
(0.251-0.
(0.279-0.
(0.328-0.
(0.226-0.
(0.320-0.
(0.347-0.
(0.410-0.
(0.374-0.
(0.386-0.
(0.207-0.
_
(0.220-0.
-
(0.334-0.
.
(0.345-0.
(0.346-0.
(0.269-0.
_
(0.309-0.
(0.294-0.
451)
473)
501)
490)
447)
541)
608)
585)
466)
534)
576)
614)
459)
578)
603)
643)
524)
425)
437)
413)
501)
532)
530)
480)
517)
544)
508)
516)
435)
450)
512)
503)
486)
501)
515)
544)
Netted front Hydraulic
capillary conductivity.
pressure , cent iae ter s
centimeters per hour
4.95
5.34
6.38
2.07
6.13
6.01
4.21
5.16
11.01
15.24
8.89
6.79
8.89
10.01
6.40
9.27
16.66
10.91
7.21
12.62
21.85
_
26.10
23.90
20.88
27.00
18.52
15.21
27.30
13.97
18.56
21.54
23.90
—
36.74
-
29.22
_
30.66
45.65
31.63
_
27.72
54.65
(0.
(1.
(1.
(0.
(1.
(1.
(1.
(0.
(2.
(5.
(2.
(1.
(1.
(2.
(1.
(0.
(2.
(1-
(0.
(3.
(4.
(4.
(5.
(4.
(6.
(4.
(3.
(5.
(4.
(4.
(4.
(4.
(8.
(6.
(7.
(IB
(6.
(6.
(10
97-25.36)
24-23.06)
31-31.06)
32-13.26)
35-27.94)
56-22.87)
03-17.24)
76-34.85)
67-45.47)
56-41.76)
02-39.06)
16-39.65)
33-59.38)
14-46.81)
01-40.49)
87-99.29)
92-95.39)
89-63.05)
86-60.82)
94-40.45)
42-108.0)
_
79-142.30)
51-103.75)
79-91.10) 1
13-118.9)
36-78.73)
79-61.01)
67-131.50)
20-46.53)
08-84.44)
56-101.7)
06-140.2)
—
33-162.1)
-
13-139.4)
_
15-131.5)
.27-114.1)
39-156.5)
_
21-123.7)
.59-282.0)
11.78
2.99
1.09
0.34
0.65
0.15
D.in
0.10
0.06
0.05
0.03
Reference 4.
Numbers in (), * one standard deviation.
Insufficient sample to determine parameters.
B-5
-------
TABLE B-3. CAPILLARY PRESSURES FOR WATER IN VARIOUS SOILS'
Soil
Sand:
Coarse
Fine
Silt
Clay
Colloids
Size of particles
and of
openings, mm
2.0-0.025
0.025-0.05
0.05-0.005
0.005-0.001
0.01 and finer
Capillary
pressures, cm
1.5-12
12-61
61-610
610-3,050
3,050 and more
Reference 5.
Other parameters in the Green-Ampt equation that affect infiltration rate
include: porosity (n), initial moisture (<)>.), and liquid head (h). Porosity
is a measure of the volume of void space in a given soil.
For oil infiltration calculations, the soil moisture is subtracted from
the total porosity to obtain a measure of the void space available for the oil
to occupy.
The liquid head (h), which is the depth of standing liquid from the spill
or leaking tank, is a parameter in the Green-Ampt equation because the standing
liquid exerts some pressure on the migrating oil front, which results in an
increasing rate of movement with increasing head.
A list of values used for parameters in the Green-Ampt equation to calcu-
late the time required for penetration of spilled oil in Section 4 of this
report is shown in Table B-4.
B-6
-------
TABLE B-4. GREEN-AMPT PARAMETER VALUES FOR WASTE OIL
. ,, Low High Average
Soil type Soil variables soil permeability soil permeability soil peneability
Clay
Silt
Sand
Total porosity °
Relative conductivity
(cm/sec) btC
Soil moisture
Capillary force (cm) e
Total porosity
Relative conductivity
(co/sec) b
Soil moisture 8
Capillary force (cm) e
Total porosity a
Relative conductivity
(cm/sec) b
Soil moisture
Capillary force (cm) e
0.4
4.2 E-13
0.2
-1.200 to -11
0.35
5.6 E-10
0.1
-250 to -It
0.25
9.8 E-6
0.03
-25 to -2
0.7
4.2 E-10
0.4
-1.200 to -11
0.5
4.2 E-6 C
0.2
-250 to -4
0.5
7.0 E-3
0.13
-25 to -2
0.45
6.9 E-ll
0.23
-1,200 to -11
0.4
1.6 E-6
0.13
-250 to -4
0.35
3.2 E-3
0.07
-25 to -2
Range of values from Reference 3. Average value from Reference 6.
From Table B-l and Figure B-2; all values assume 10Z of voids filled with trapped air.
C Assumes relative permeability for oil (Figure B-2) cannot be below 0.3, which is value below
which Figure B-2 shows water mobility also.
Estimated by FAL. Value equals soil saturation multiplied by total porosity.
Derived from literature values for water multiplied by ratio of surface tension of oil to
surface tension of water. Ratio used - 0.383 (References 7 and 9). Range of water values
taken from high values of Table B-3 and low values in Table B-2.
Range of values from Reference 3. Average estimated by FAL.
o
Range estimated by FAL. Average based on residual saturation of 0.33 (Reference 7).
Low and high numbers based on residual saturations of 0.10 and 0.25 respectively (Reference 8).
Average number based on residual saturation of 0.20 (Reference 7).
B-7
-------
APPENDIX B
REFERENCES
1. GCA Corporation. Fate of Hazardous and Nonhazardous Wastes in Used Oil
Recycling. Fifth Quarterly Report. DOE Contract No. DE-AC19-81BC10375.
March 1983.
2. Van Dam, J. Migration of Hydrocarbons in Water-Bearing Stratum. In:
The Joint Problems of Oil and Water Industries, Peter Hepple, ed. Insti-
tute of Petroleum, London. 1967.
3. Freeze, R. A., and J. A. Cherry. Groundwater. Prentice-Hall, Inc., New
Jersey. 1979.
4. Rawls, W. J., D. L. Brankenseik, and N. Miller. Green-Ampt Infiltration
Parameters From Soils Data. ASCE, Proc. Journal of Hydraulic Engineering,
109:1, January 1983.
5. Tschebotarioff, G. P. Soil Mechanics, Foundations and Earth Structures.
McGraw-Hill Book Company, Inc., New York. 1957.
6. Lindsley, R. K., Jr., M. A. Kohler, and J. L. H. Paulhus. Hydrology for
Engineers. McGraw-Hill Book Company, Inc., New York. 1982.
7. Personal communication. Mr. Doug Morell, PEDCo Environmental, Inc., to Mr.
Nicholas Artz. 1982.
8. DeJong, R., and K. Loebel. Empirical Relations Between Soil Components
and Water Retention at 1/3 and 15 Atmospheres. Canadian Journal of Soil
Science, 62:343-350, May 1982.
9. Corey, A. T. Mechanics of Heterogeneous Fluids in Porous Media. Water
Resources Publications, Colorado. 1977.
B-8
-------
APPENDIX C
DERIVATION OF SPILL PENETRATION DEPTH EQUATION
The purpose of this appendix is to show the derivation of the equation
for calculating the spill penetration depth (d ), given the spill area and
s
the volume of saturated soil. The spill area is related to the spill volume
according to a simple model developed from spill data in Canada.
Ag = 53.5 Vs°'89 (C-l)
The volume of saturated soil is directly proportional to the spill volume
and inversely proportional to the soil porosity and the residual saturation
level:
V .. = Vs (C-2)
5011 HT
In these two equations:
V = volume of oil spilled (cubic meters)
S
V .n = volume of saturated soil (cubic meters)
soil
n = soil porosity or ratio of void volume to total soil volume,
S = residual saturation
A = spill area (square meters)
S
It is assumed in this analysis that the oil spill does not reach the
groundwater.
Two shapes for the contaminated soil volume are considered in this
analysis. If the shape is that of a cylinder, then the depth of penetra-
tion (dg) is equal to the contaminated soil volume divided by the surface
area:
V V
d = soil = s (C-3)
A nS A
s r s
C-l
-------
If there is lateral movement of the oil as it moves through the soil,
then the shape of the contaminated soil volume may be approximated by a
truncated cone as shown in Figure G-l.
Spill Surface
Figure C-l. Geometry of contaminated soil volume, with lateral movement.
C-2
-------
. . or R
then Equation 7 can be rewritten as:
The volume of a cone is given by:
V «= - R2H, where R is the radius of the base of the cone
t 3
and H is the cone height. From the geometry in Figure C-l,
2
-------
r 2 3 n
3* v + ir/3-I-_
|_irr^ soil tan 9 J
1/3
-—5—
tan 6
or d_=r 3
5 [iftan29 ™" tan 9j taiT9
Rearranging, we have
.3V ., 3 1/3
d
[JV -i •> ~i
-SS1+ -V
irtan 6 tan 9 J
s
tan 9
U.S. Enviroixnental Protection A«enoy
Library, Room 2404 PM-211-A
401 M Street, S.W.
Washinston, DC 20489
C-A
------- |