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

-------