United States
Environmental Protection
Agency
Office of
Underground Storage Tanks
Washington, D.C. 20460
EPA/600/M-86/020
August 1986
c/EPA
Underground Storage Tanks
Summary of State Reports
on Releases from
Underground Storage Tanks
-------
ANALYSIS OF THE NATIONAL DATA BASE OF
UNDERGROUND STORAGE TANK RELEASE INCIDENTS
Submitted by:
Versar Inc.
P.O. Box 1549
Springfield, Virginia 22151
With Subcontractor Support From
SCS Engineers and Franklin Associates, Ltd,
Submitted to:
Angela NiIkes
Office of Solid Waste
Waste Treatment Branch
U.S. Environmental Protection Agency
401 M St., S.W.
Washington, D.C. 20460
In Response to:
EPA Contract No. 68-01-7053
Work Assignment No. 48
July 21, 1986
-------
DISCLAIMER
This report was prepared under contract to an agency of the United States
Government. Neither the United States Government nor any of its
employees, contractors, subcontractors, or their employees makes any
warranty, expressed or implied, or assumes any legal liability or
responsibility for any third party's use of or the results of such use of
any information, apparatus, product, or process disclosed in this report,
or represents that its use by such third party would not infringe on
privately owned rights.
-------
TABLE OF CONTENTS
Page No.
1. INTRODUCTION AND KEY FINDINGS 1-1
1.1 Introduction 1-1
1.2 Summary of Key Findings 1-2
1.2.1 Distribution by Year and by Region 1-3
1.2.2 Facility and Material Stored Information 1-3
1.2.3 Tank System Information 1-4
1.2.4 Release Information 1-5
1.2.5 Leak Detection Information 1-6
1.2.6 Impact and Corrective/Remedial Action
Information 1-7
1.3 Data Variability for Selected States 1-8
1.3.1 Summary of Individual State Results 1-8
2. METHODOLOGY 2-1
2.1 Data Collection Procedures 2-1
2.2 Data Analysis Procedures 2-3
2.3 Strengths and Limitations 2-4
3. DATA BASE DESCRIPTION 3-1
3.1 Subtitle I Incidents 3-1
3.2 Distribution by State 3-4
3.3 Distribution by Year 3-6
4. FACILITY AND MATERIAL STORED INFORMATION 4-1
4.1 Facility Type 4-1
4.2 Facility Operating Status 4-1
4.3 Material Stored 4-4
4.4 Facility Type and Material Stored for Releases
Greater Than 50,000 Gal Ions 4-7
5. TANK SYSTEM INFORMATION 5-1
5.1 Tank System Age at Time of Release 5-1
5.2 System Age for Tank Leaks Compared to Pipe Leaks 5-4
5.3 Material of Tank Construction 5-6
5.4 Tank Volume Distribution 5-9
-------
TABLE OF CONTENTS (Continued)
Page No.
6. RELEASE INFORMATION ......................................... 6-1
6. 1 Location of Release .................................... 6-1
6.2 Cause of Release ....................................... 6-3
6.3 Comparison of Structural Failure and Corrosion-Related
Incidents .............................................. 6-3
6.4 Causes of Structural Failure ........................... 6-5
6.5 Cause of Release by Location of Release ................ 6-6
6.6 Cause of Release by Age at Time of Report .............. 6-6
6.7 Cause of Release by Material of Tank Construction ...... 6-6
6.8 Quantity Released ...................................... 6-10
6.9 Quantity Released by Location of Release ............... 6-12
7 . LEAK DETECTION INFORMATION .................................. "7-1
7.1 Method of Initial Leak Detection ....................... 7-1
7.2 Quantity Released by Detection Method .................. 7-3
7.3 Use of Integrity Tests to Confirm Leaks ................ 7-3
8 . IMPACTS AND REMEDIAL ACTIONS
8.1 Documented Media Impacted .................. ........... 8-1
8.2 Documented Impacts ..................................... 8-1
8.3 Corrective/Remedial Action ............................. 8-8
9 . CHEMICAL RELEASES
9.1 Chemical Releases by State ............................. 9-1
9.2 Number of Chemical Releases by Year .................... 9-3
9.3 Facility and Material Information ...................... 9-3
9.3.1 Facility Type ................................... 9-3
9.3.2 Facility Operating Status ....................... 9-3
9.3.3 Materials Stored ................................ 9-7
9.4 Tank System Age Information ............................ 9-7
9.5 Release Information .................................... 9-7
9.5.1 Location of Release ............................. 9-7
9.5.2 Cause of Subsurface Release ..................... 9-11
9.6 Leak Detection Information ............................. 9-11
9.7 Corrective/Remedial Action/Impacts ..................... 9-11
9.8 Summary of Comparison between Chemical and Total
Results ................................................ 9-15
9.9 Additional Chemical Release Data ....................... 9-17
IV
-------
TABLE OF CONTENTS (Continued)
Page No.
APPENDIX A Number and Percent of Release Incidents
with Documentation in Each Field for Which
Information Was Sought A-l
APPENDIX B State Data Sources B-l
-------
LIST OF TABLES
Page No.
Table 4-1 Facility Type and Material Stored for Releases
Exceeding 50,000 Gallons 4-8
Table 5-1 Summary of Subtitle I Tank Age Data by Region 5-3
Table 8-1 Types of Impacts for Documented Subtitle I
Incidents by Media 8-3
Table 9-1 Types of Impacts for Documented Chemical Incidents
by Med 1 a 9-56
-------
LIST OF FIGURES
Page No.
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Figure 4-1
Figure 4-2
Figure 4-3
Figure 4-4
Figure 5-1
Figure 5-2
Figure 5-3
Figure 5-4
Figure 6-1
Figure 6-2
Figure 6-3
Figure 6-4
Figure 6-5
Documented Underground Tank Releases That Would Be
Regulated by RCRA Subtitle I
Comparison of Release Incidents for Subtitle I Tanks
in the Data Base and Tanks in the Universe
Reported Release Incidents by State
Number of Release Incidents by Year
Faci 1 ity Type
Operating Status of Facility at Time of Release
Report
Quantity Released by Facility Operating Status
Material Stored in Tanks
Age of Tank System at Time of Subsurface Release
Incident
Age of Tank and Piping at Time of Release
Incident
Material of Construction for Tanks
Tank Volume Distribution
Location of Release
Cause of Release
Cause of Release Where Location of Release Is
Specified as Tank or Piping
Cause of Release by Age for Subsurface Releases
Cause of Release by Material of Construction
3-2
3-3
3-5
3-7
4-2
4-3
4-5
4-6
5-2
5-5
5-8
5-10
6-2
6-4
6-7
6-8
6-9
vii
-------
LIST OF FIGURES (Continued)
Page No.
Figure 6-6 Quantity Released 6-11
Figure 6-7 Quantity Released by Location of Release 6-13
Figure 7-1 Method of Initial Leak Detection for Subsurface
Release Incidents 7_2
Figure 7-2 Quantity Released Where Initial Detection Was
Accomplished by Product Inventory or Detection
Equipment 7.4
Figure 8-1 Documented Health/Environmental Impacts Reported
for Each Medium 3_2
Figure 8-2 Corrective/Remedial Measures 8-9
Figure 9-1 Comparison of Reported Release Incidents by
State 9_2
Figure 9-2 Comparison of Total and Chemical Release Incidents
by Year g_4
Figure 9-3 Comparison of Facility Type for Total and
Chemical Release Incidents g_5
Figure 9-4 Comparison of Facility Operating Status for Total
and Chemical Release Incidents 9-6
Figure 9-5 Material Stored in Leaking Tanks for Total and Chemical
Release Incidents g_8
Figure 9-6 Comparison of Age of Tank System at Time of Report for
Total and Chemical Incidents 9.9
Figure 9-7 Comparison of Location of Release for Total and
Chemical Incidents 9_10
Figure 9-8 Comparison of Cause of Release for Total and Chemical
Release Incidents g_12
Figure 9-9 Comparison of Methods of Initial Leak Detection for
Total and Chemical Release Incidents 9-13
Figure 9-10 Comparison of Corrective/Remedial Measures for
Total and Chemical Release Incidents 9-14
VI 1 1
-------
1. INTRODUCTION AND KEY FINDINGS
1.1 Introduction
After the passage of the RCRA amendments In October 1984, the EPA
Office of Solid Waste concluded that a compilation of documented
Information concerning underground storage tank release incidents would
be most useful in determining causes, impacts, ages, and other relevant
information associated with such incidents. Because of time and funding
constraints, the Agency decided that the initial data collection would be
undertaken at the State level, with similar data to be collected at the
county and local level in 1986.
This report is based on information compiled for 12,444 underground
storage tank release incidents in all 50 States. Most of the data were
obtained by visiting State offices. The goals of the study were:
To identify the number of documented underground storage tank
release incidents and the degree of documentation in State files;
To analyze these data with respect to cause, impact, age, tank
type, and other relevant factors; and
To determine whether any trends based on geographic distribution,
tank age, material stored, or other factors were apparent from the
data.
The analyses presented in this report successfully achieve each of
the stated goals, and significantly improve our understanding of the
circumstances surrounding underground tank leaks. In fact, the data base
compiled in this study contains by far the most comprehensive information
concerning leaking underground tanks that is currently available. The
1-1
-------
12,444 documented release incidents identified in State files are not
necessarily a statistically representative subset of all underground
storage tank releases. There are some inherent biases regarding the
incidents likely to be reported (i.e., differing State enforcement
procedures). In addition, since many releases are not reported, the
incidents analyzed herein represent only a fraction of the total number
of underground storage tank releases. Therefore, while the results of
this study substantially enhance our knowledge and understanding of
circumstances surrounding underground storage tank leaks, the reader
cannot statistically relate the results to the universe of all
underground storage tank releases. Rather, the results are intended to
be used in conjunction with those of other studies currently being
performed by EPA.
The key findings are summarized in Section 1.2. Section 2 explains
the methodology used in the data collection and manipulation process.
Section 3 presents an overview of the release incidents in the data
base. Sections 4 through 8 contain detailed results for selected
analyses of the national data base of underground storage tank release
incidents. Section 9 presents analyses for the chemical subset of the
data base. Appendix A is a list of all the data fields for which
information was sought in the questionnaires, showing the number of
responses and the percent of total responses for each field.
1-2
-------
1 .2 Summary of Key Findings
Analysis of the national data base of underground storage tank
release incidents reveals that approximately 83 percent of the 12,444
release incidents documented at the State level involved tanks covered by
RCRA Subtitle I. The majority of release incidents exempt from
Subtitle I involved tanks used to store heating oil. The following
findings are based on the entire data base of 12,444 release incidents.
1.2.1 Distribution by Year and by Region
The data reflect a continuous increase in the number of release
incidents reported to State agencies annually from 1970 to 1984 (the most
recent year for which complete data are available). While this is
believed to be primarily due to improved State reporting procedures
during this period, an increase in the actual number of release incidents
during the period cannot be ruled out at this time.
There is a higher frequency of reported release incidents in the
Northeast, Mid-Atlantic, and Great Lakes States than in most other areas
of the country. Although these areas tend to exhibit relatively
corrosive soils (i.e., high moisture and low pH), other factors including
a high tank population and the procedures used to report and document
leaks also must be taken into consideration. The geographic distribution
of chemical releases is similar to that of the chemical industry.
1-3
-------
1.2.2 Facility and Material Stored Information
Approximately 65 percent of the documented incidents involved
releases from retail gasoline stations, including convenience stores
selling gasoline. This development could stem from a number of factors,
such as proximity of gasoline stations to population centers and/or the
low threshold of odor and taste detection for gasoline compared to other
stored materials.
Only about 3 percent of the documented release incidents involved
chemicals other than petroleum fuels. This compares to about 4 percent
of the total population of regulated tanks currently believed to contain
chemicals other than petroleum fuels. The majority of the chemical
releases occurred at manufacturing facilities. "Unspecified solvent" was
the most common chemical released.
Ninety-five percent of the reported release incidents involved
operating facilities as opposed to abandoned facilities. Also, the
quantity released from operating facilities was usually greater than that
for abandoned facilities.
1.2.3 Tank System Information
The mean and median tank system ages at the time the subsurface
release was documented were 17 years. The range varied from less than 1
to more than 50 years, however.
1-4
-------
An analysis of the mean tank system age in each of eight regions of
the country showed that seven of the eight regions exhibited mean tank
ages of between 14 and 18 years. Only the Southwest, with a mean tank
age of 28 years (based on very limited data), substantially differed from
other regions.
Tank, leaks exhibited a mean age of 17 years, while piping leaks had a
mean age of 11 years. Poor installation practices can contribute to
early pipe failures. This may also be due in part to the fact that pipe
walls are normally thinner than tank walls, and are therefore more
subject to corrosion-related leaks.
Of the incidents specifying material of tank construction, 81 percent
involved steel tanks and 19 percent involved fiberglass tanks.
1.2.4 Release Information
The most frequent documented location of release is through the tank
into the surrounding soil (43 to 58 percent). Piping accounts for
between 20 and 35 percent of the reported release locations;
overfills/spills account for 15 percent of those reported; and the
remainder are due to pump and miscellaneous causes. These data are not
sufficient to conclude that tank leaks are more common than pipe leaks;
however, it is sufficient to conclude that an effective leak control
program must address both tank and pipe leaks.
Although most incidents did not specify a cause of release, corrosion
and structural failure, followed by loose fittings and improper
installation, are the most commonly documented causes of tank system
1-5
-------
failure. Structural failure becomes a more significant cause of release
among incidents involving releases of more than 10,000 gallons. There
was essentially no overlap among those incidents specifying structural
failure compared to those specifying corrosion.
Loose fittings and improper installation were far more common in tank
systems 10 years old or less than in tank systems more than 10 years
old. Structural failure was more common among tank systems 10 years old
or less and tank systems more than 20 years old than for those from 11 to
20 years old. Conversely, corrosion was more common among those tank
systems greater than 20 years old.
The documented quantity released from overfills is less than
100 gallons in most cases. On the other hand, the documented quantity
released from most subsurface tank leaks and from most subsurface pipe
leaks is greater than 500 gallons.
1.2.5 Leak Detection Information
Visual appearance and odor of product or impacts are the two most
commonly cited means of leak detection. Conversely, only about 10
percent of the incidents with a documented detection method cited
integrity test or detection equipment as the means of initial detection.
Integrity tests were commonly used to confirm a leak after it had been
detected.
1-6
-------
Product inventory control was cited as the method of initial leak
detection in about 14 percent of the cases in which a leak detection
method was specified. An analysis of quantity released, however, showed
that incidents detected by product inventory control tended to be those
that exhibited relatively large releases.
1.2.6 Impact and Corrective/Remedial Action Information
Of the 12,444 documented releases, 68 percent cited releases to soil,
45 percent to ground water, and 22 percent to surface water. Releases to
air and other media (i.e., basements) were reported in 15 and 12 percent
of the incidents, respectively.
Most of these documented incidents did not cite a specific impact.
Those that did, however, involved contamination of more than 700 private
wells; contamination of 40 municipal wells; 100 incidents of human
illness; 155 cases of fire/explosion resulting in two human deaths; 202
cases of aquatic, plant, or wildlife damage; and 908 cases of combustible
fumes in confined areas.
The most commonly specified corrective/remedial actions were tank
system replacement and/or repair and the installation of monitoring
wells. Conversely, only 10 percent of the documented incidents reported
soil excavation, and only 6 percent involved recovery wells. The
completeness of the corrective/remedial action data in State files,
however, is uncertain.
1-7
-------
The chemical release subset differed from the entire data base in
that 25 percent of the chemical incidents involved soil excavation. This
finding implies that soil excavation is more common among chemical
releases than among petroleum fuel releases.
Documentation for these and related findings is provided in the
following sections of this report; however, much remains to be done. Ne
believe that the data base can continue to provide valuable information
to EPA through (1) the integration of the analyses herein (or of
additional analyses) with other EPA studies; (2) detailed analyses of
selected incidents; and (3) continuation of the data collection and
analysis process at the county/local level where more specific
information may be available.
1.3 Data Variability for Selected States
A parallel set of selected analyses was performed on data from four
individual States to see how the results for each State compared to the
national results. The four States evaluated were New Jersey, Maryland,
Minnesota, and Colorado. These States were selected based on the number
of cases reported, the level of detail reported, and geographic
considerations.
1.3.1 Summary of Individual State Results
The results for each State showed trends similar to the national
results for the following analyses: facility type, material stored, and
tank age.
1-8
-------
The number of release incidents reported by year showed a trend of
increase in each State; however, the length of the record-keeping period
varied from State to State. This is probably a reflection of differing
startup times for each State's reporting and record-keeping procedures.
Three of the four States exhibited location of release patterns
similar to the nation as a whole, with tanks being the most common
location of release. In Colorado, however, piping was the most commonly
reported location of release. This could conceivably reflect less
corrosive soils and therefore a greater dominance of structural piping
failures in Colorado. This hypothesis is supported by the fact that
Colorado also exhibited a higher percentage of structural failure
releases compared to the nation.
The corrective/remedial actions documented in each State's files
varied significantly from each other and from the nation. For example,
monitoring wells were commonly reported in Maryland and New Jersey, but
infrequently reported in Minnesota and Colorado. Conversely, soil
excavation was reported more often in Minnesota than in any of the other
three States, or the nation as a whole. These results imply that the
relative frequency with which specific remedial actions are employed
differs from State to State.
Based on the above, results from the four selected States appear to
be similar to the national results for facility type, tank age, and
material stored. Conversely, there appear to be some differences between
1-9
-------
the individual State data and the national data regarding length of
record, location/cause of release, and corrective/remedial action. These
differences must be viewed with caution when one attempts to draw
State/regional conclusions because of differences in State reporting
procedures.
1-10
-------
2. METHODOLOGY
2.1 Data Collection Procedures
Following passage of the RCRA amendments in October 1984, the EPA
Office of Solid Waste concluded that a compilation of documented
underground storage tank release incidents would be most instructive in
determining the causes, impacts, and other relevant factors associated
with underground tank releases. Because of time and funding constraints,
the Agency decided that initial data collection efforts would be focused
at the State level. The Agency is currently collecting similar data at
the county and local level.
Initial data were collected via phone calls to officials in all 50
States to obtain preliminary estimates of the number of underground tank
release incidents on file and of the level of detail available. This
preliminary information indicated that up to 15,000 cases were on
record. Because a few States overestimated the number of releases on
file and some release records could not be positively documented as
involving underground tanks, this study is based on information compiled
from 12,444 underground tank release cases. The data were obtained by
visits to agencies in most States and review of data from their files.
(Data from several States were obtained by mail.) It should be
recognized that many release incidents are believed to go unreported and
that the total number of incidents is unknown. The State agencies from
which data were obtained for this study are summarized in Appendix B.
A detailed coding sheet was developed to transfer the information in
State files to a common reporting format. The coding sheet consisted of
2-1
-------
126 fields with coded responses for each field. The fields were grouped
into the following categories:
Facility identification information;
Tank/piping system information;
Tank-specific information;
Piping-specific information;
Release detection information;
Health/environmental impact information; and
Legal/remedial action information.
The frequency with which information for the specific data fields was
reported was quite varied. For example, more than 75 percent of the
incident files contained information for the following data fields:
Operating status;
Material stored;
Percent volume below grade;
Date of release;
Location of release;
Cause of release; and
Corrective/remedial measures.
In comparison, between 25 and 75 percent of the files had information
on:
Tank volume;
Detection method;
Health/environmental impacts;
Legal action;
Status of cleanup; and
Quantity released.
Finally, less than 25 percent of the incident reports contained data
on the following fields:
Tank age;
Tank construction material;
Soil type;
Duration of leak;
Litigation (third party); and
Corrective action cost.
2-2
-------
The specific topics addressed by each of the 126 data fields and the
frequency with which information on each is available in the State
reports are presented in Appendix A.
For each field, a standard set of expected answers was developed. In
the unusual cases where none of the standard answers applied, the
appropriate response was handwritten on the coding form. In the majority
of cases, a copy of the original State files was made and kept for
reference.
2.2 Data Analysis Procedures
The completed coding sheets were computerized using an IBM/AT
personal computer with 60 megabytes of storage capacity. The IBM/AT was
selected because of the wide variety of software available, its
sufficient storage capacity, and its speed.
The data were input using the Infostar software package. For ease in
data entry, Versar manipulated the software to reflect the exact format
of the coding form. As a quality control measure, each field was
designed to accept only the alpha-numeric characters that were proper
responses to each data entry field. For example, if a field had possible
coded responses of 1, 2, 3, and 4, only those digits would be acceptable;
any other input would result in an error message. Also, each release
incident was given a unique case number. Case numbers entered were
carefully checked against those received to be sure that the computerized
data base accurately reflected the paper files.
2-3
-------
Once input, the data were manipulated using the R-base software
package. R-base was selected to allow a high degree of versatility in
data manipulation with a minimum amount of programming. This allowed
efficient evaluation of numerous relationships among the various fields.
2.3 Strengths and Limitations
We believe this study provides the most comprehensive and
representative data base available at this time concerning leaking
underground storage tanks. As such, it can provide meaningful insights
into the nature of the underground storage tank problem in the United
States. These insights involve important characteristics, including the
distribution of facility types where leaks tend to occur; the stored
materials involved in those leaks; the age of the tank systems; the
causes of the release; the resulting health and environmental impacts;
and the corrective actions taken. By better defining the nature of the
problem, these insights will provide important input into the regulatory
decision-making process for the development of technical tank standards
and corrective action standards.
While this data base represents a major step forward in
characterizing the underground tank problem, it also has limitations that
must be recognized to ensure its proper use. Among these limitations are:
The release incidents in the data base are those reported in State
files. As such, they are not a statistically selected sample of
the universe of tank releases, nor of the tank universe.
2-4
-------
The 12,444 release incidents included in the data base are
believed to represent a fraction of the universe of tank releases;
the size of the universe of tank releases cannot be accurately
estimated from this data base. A concurrent study being performed
by EPA/OTS and involving actual testing of a randomly selected
group of tanks will allow an estimate of the number of leaking
tanks.
The data base requires qualitative interpretation of the results,
since certain types of releases and certain types of information
may be more likely to be reported in State files than others. For
example, releases involving third party complaints may be more
likely to be reported in State files than other releases. This
may tend to weight the data base towards facilities located near
population centers and/or toward materials with a low odor
detection level. While such rational interpretation of the
results can be made, the quantitative impact of these factors on
the results cannot be determined.
The reporting procedures and the quality of information provided
vary from State to State. This will affect both the number of
incidents on file in a given State and the level of detail
provided on each incident. Only those incidents that could be
positively identified as underground tank system releases were
used in the data base.
Therefore, while the large size and the comprehensiveness of the data
base provide far more insight into the nature and character of the
underground storage tank problem than has been available previously, the
above limitations must be kept in mind during result interpretation.
2-5
-------
3. DATA BASE DESCRIPTION
3.1 Subtitle I Incidents
Although the data base consists of 12,444 underground storage tank
release incidents, not all of these incidents are covered by the
provisions in RCRA Subtitle I, since certain tanks are exempted from
those provisions. A total of 10,300 release incidents in the data base
are covered by RCRA Subtitle I. This is illustrated in Figure 3-1.
In collecting the information from State files, the only criterion
used to determine which incidents to include was that the release come
from an underground tank. Therefore, some of the incidents in the data
base are currently exempt from RCRA Subtitle I. The categories of exempt
tanks include:
Farm tanks storing motor fuels where the tank volume is less than
1,100 gallons;
Residential tanks storing motor fuels where the tank volume is
less than 1,100 gallons;
Tanks used for storing heating oil for use on the premises; and
Tanks containing wastes (which are covered by RCRA Subtitle C).
EPA is required to recommend to Congress whether or not to regulate
heating oil tanks under Subtitle I. Note that heating oil tanks are the
largest single class of exempt tanks in the data base.
EPA has prepared preliminary estimates of the numbers and types of
tanks to be regulated under RCRA Subtitle I. Figure 3-2 compares
information for Subtitle I tanks in the release incident data base to
3-1
-------
15%
Exempt Tanks Storing Heating
Oil for Use on the Premises
Other Exempt Tanks
;;- ' / [Tanks Regulated by RCRA Subtitle 11
'' : '' ^1
83ฐ/c
* Farm or residential tanks used for
storing motor fuels where tank volume
is less than 1,100 gallons and tanks
that contained recycled oil or wastes
Figure 3-1 Documented Underground Tank Releases That
Would Be Regulated by RCRA Subtitle I
3-2
-------
Estimated Number of Subtitle I Tanks in Universe = 1,350,000
(Preliminary EPA Estimate of the Universe of Regulated Tanks)
Subtitle I Tanks in Data Base = 10,300
| Retail Gas Station Tanks ง
75%
Retail Gas Station Tanks
Tanks in Data Base
Tanks in Universe
Figure 3-2 Comparison of Release Incidents for Subtitle
I Tanks in the Data Base and Tanks in the Universe
-------
these preliminary estimates of the universe of tanks regulated by
Subtitle I. The figure shows that chemical tanks represent less than
5 percent of the Subtitle I tanks in the data base as well as in the
universe of Subtitle I tanks. Retail gasoline stations, however, are
more largely represented in the data base than in the regulated universe.
3.2 Distribution by State
Figure 3-3 illustrates the geographic distribution of the 12,444
documented release incidents by State. The results show that while all
50 States are represented, States in the Northeast, Mid-Atlantic, and
Great Lakes regions exhibit a higher number of reported releases than
most other States. Texas, Colorado, and California also have relatively
high numbers of reported release incidents.
The total number of release incidents by State is not known; however,
three factors that may contribute to this geographic distribution of
reported incidents are as follows: (1) the number of cases on file is
dependent upon the State's procedures for receiving, responding to, and
maintaining records of underground storage tank release incidents, as
well as the length of time those procedures have been in effect;
(2) States with larger populations tend to have more tanks and thus more
potential tank releases than States with smaller populations; and
(3) soils in the Northeast tend to exhibit a lower pH and higher moisture
content than those in many other areas of the country, and therefore may
be more conducive to corrosion.
3-4
-------
Total Incidents
Reported = 12,442
FIGURE 3-3 Reported Release
Incidents by State
NUMBER OF INCIDENTS
BY STATE
9-50
51-250
251-1,000
>1,000
-------
The specific relationships between these and other possible factors
in determining the number of reported release incidents cannot be
precisely established at this time. Nevertheless, the aggregate analyses
that follow are based on the above geographic distribution of release
incidents.
3.3 Distribution by Year
Figure 3-4 summarizes the release incident data by year. The data
show a continuous annual increase in reported release incidents from 1970
to 1984. Since the data for 1985 were collected between April and
October, they reflect only that portion of the year. If data were
available for the entire year, however, the total most likely would
depict a continued increase. For example, the data for California were
collected on two visits separated by a period of several months. Between
these two visits, more than 100 additional releases were reported.
The increase in reported incidents by year is believed to stem
primarily from a growing awareness of the underground tank problem and
improved enforcement/reporting procedures in the various States. There
may also be an increase in the number of leaking tanks. We have no
reason to believe that the trend has begun to level off, but only
continued data compilation and analysis will confirm this conclusion.
3-6
-------
o,uuu
2,500-
2,000-
i
0)
S
^ 1 500
iซ
e
3
1,000-
500-
0-
_f"-"1l._
Total Release Incidents Reporting Age =
n R
R r M
FS?3 Is! 1 1 iv*J
1 1 1 1 1
0
1
1
1
*~
}
\
r1
!
1 2,075 1
|
WHB
I
1
J
-
'<>
r
4>
ป:,
&,
i"'
x-
:-s
^
^
si
fe'
>
\
'>,
Vซ**i
-
s%
C
V1
ss.
*l
1
^
*.%
%.
;
;
X
^>
\
\
1
|
\
<
_
- :
..
T I I i i i i
o,uuu
p ROD
-2,000
1 unn
-1,000
-
-500
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
Year
Note: Partial results
shown for 1985
Figure 3-4 Number of Release Incidents by Year
3-7
-------
4. FACILITY AND MATERIAL STORED INFORMATION
4.1 Facility Type
Figure 4-1 shows the distribution of release incidents by facility
type. About 65 percent of the reported release incidents involved retail
gasoline stations (including convenience stores and other multipurpose
retailers). Other commercial establishments accounted for 11 percent of
the releases; manufacturing facilities, 5 percent; and municipal
facilities, 4 percent. All remaining facility types accounted for about
15 percent of the total.
The predominance of retail gasoline stations is due in part to their
relatively large number in comparison with other types of underground
tank facilities (see Figure 3.2). Another possible factor, however, is
that gasoline stations tend to be located near population centers, which
may, in turn, increase third party complaints and State involvement when
a release occurs. The relative importance of these or other factors has
not been determined.
4.2 Facility Operating Status
Figure 4-2 shows the operating status distribution for facilities
reporting release incidents. About 95 percent of the reported incidents
involve operating facilities; the remaining 5 percent involve abandoned
facilities. Thus, the results do not show that abandoned facilities
contribute significantly to the underground tank problem.
4-1
-------
c
0>
"0
_c
T3
CD
c
8.
o
a:
"o
"c
03
a
70%-r
60%-
50% -
40%-
30% -
20% -
10%-
0%
1-70%
Total Release Incidents Reporting Facility Type = 11,861
1
* Other includes: Federal (Non-Military), State,
Apartment/Condominium Complex, Private Residence,
Farm/Agricultural, Commercial Transportation, Dry
Cleaners, Office Building, Bulk Storage, Airport, Military I
I
Retail Gas Station
Other Commercial Establishment Municipal
Other Manufacturing Facility
Facility Type
-60%
-50%
-40%
-30%
-20%
-10%
0%
Figure 4-1 Facility Type
4-2
-------
1
a
CD
oc
CD
o
4*
CD
0-
100%-r
90%-
80%-
70%-
60%-
50%-
40%-
30%-
20%-
10%-
0%
V '* '
Total Incidents Specifying Status = 10,960
,-100%
-90%
-80%
-70%
-60%
-50%
-40%
-30%
-20%
-10%
-0%
Operating Abandbned
Operating Status
Figure 4-2 Operating Status of Facility at Time of Release Report
4-3
-------
It is hypothesized that the reason abandoned tanks appear to be a
small part of the problem is that the total amount released from an
abandoned tank is limited to the volume remaining at the time of
abandonment; thus, the amount likely to be spilled is substantially less
than would be the case for an operating system. As shown in Figure 4-3,
the release incident data base supports this hypothesis; however, the
data for "quantity released" are somewhat subjective and must be viewed
with caution. On the other hand, the location of abandoned tanks is less
likely to be known (and reported) than the location of operating tanks,
and therefore many release incidents might be unknown.
4.3 Material Stored
Figure 4-4 presents the distribution of materials stored in reported
underground tank releases. Gasoline accounts for more than 70 percent of
the reported release incidents. The next most common materials stored in
underground tanks are heating oil, diesel fuel, recycled oil, and
solvents. This distribution again reflects the predominance in the data
base of gasoline stations storing gasoline and diesel fuel in underground
tanks. Additional factors that may affect this distribution are that
gasoline and other petroleum products tend to be easy to detect by smell,
taste, and appearance, and gasoline stations are often located near
populated areas. These materials would therefore be more likely to be
reported than, for example, materials requiring chemical analyses to
identify their presence in the environment. Also, many States require
certain petroleum products to be stored underground, whereas other
materials may be stored above ground at lower cost.
4-4
-------
45% -i
r-45%
-a
s
I
1
V
a
40%-
35%-
30% -
25%-
20%--
15%--
10%--
5%
0-100
Incidents Reporting Facility Operating at
Time of Release = 3,545
\
-40%
-35%
Incidents Reporting Facility Abandoned
at Time of Release = 78
I
-30%
-25%
-20%
-15%
-10%
-5%
I I I
101-500 501-2,500 2,501-10,000 >10,000
Quantity Released (gallons)
0%
Figure 4-3 Quantity Released by Facility Operating Status
4-5
-------
OUVo
70% -
60%-
-o 50%-
I
03
CC
ฃ2
1 40% -,
o
"o
1
j?>
Q. ^n0/.
20%-
10%-
....
y * s
I*
(
, &\
^eeff |
W,'*'*'*
Jk* *
&
1
< s %
Total Incidents Reporting Material Stored = 12,214 |
Jl!
\
\
jff
I
1
J \
ff :
*o
fue
mis
*-?*<
,-&>'
'?*, & >
ther includes kerosene, jet fuel, motor b
, aviation gasoline, propane and 1
cellaneous petroleum products. 1
1 1 1 "Imnmni 1 "mฅm
ou /o
-70%
-60%
-50%
-40%
-30%
-20%
-10%
Gasoline Diesel Fuel Recycled Oil Specified as Unknown
Heating/Fuel Oil Other* Chemicals Wastes
Material Stored
Figure 4-4 Material Stored in Tanks
4-6
-------
4.4 Facility Type and Material Stored for Releases Greater Than
50.000 Gallons
We also separately investigated the incidents of extremely large
releases, which were defined as those involving 50,000 gallons or more.
The investigation compared facility type and material stored for larger
releases with facility type and material stored for all incidents
reported. The results are shown on Table 4-1. Of the 26 large release
incidents, only 42 percent were from gasoline stations and 50 percent
involved gasoline. (The remainder included manufacturing, commercial,
military, and residential facilities.) Therefore, gasoline and gasoline
stations are less dominant among reported releases in excess of 50,000
gallons than among all reported releases.
4-7
-------
1312S
Table 4-1 Facility Type and Material Stored for
Releases Exceeding 50,000 Gallons
State
NC
NV
CA
MN
CO
CO
NY
CT
NY
NY
PA
KS
WI
IA
TX
WA
DE
DE
HI
HI
HI
GA
PA
ID
VT
Facility type
Gasoline station
Commercial transportation
Commercial transportation
Manufacturing facility
Manufacturing facility
Bulk storage
Gasoline station
Unspecified
Gasoline station
Gasoline station
Other commercial establishment
Gasoline station
Municipal
Other conmercial establishment
Other conmercial establishment
Gasoline station
Gasoline station
Gasoline station
Military
Military
Military
Gasoline station
Apartment/condo
Gasoline station
Gasoline station
Material stored
Gasol ine
Diesel fuel
Diesel fuel
Heating oil
Heating oil
Gasol ine
Gasol ine
Motor fuel
Gasol ine
Gasol ine
Heating oil
Gasoline
Heating oil
Gasoline
Diesel fuel
Gasol ine
Gasoline
Gasoline
Gasoline
Aviation gas
Other
Gasoline
Heating oil
Diesel fuel
Gasoline
4-8
-------
5. TANK SYSTEM INFORMATION
5.1 Tank System Age at Time of Release
Figure 5-1 presents the tank age for Subtitle I incidents at the time
the subsurface release was reported. Note that only about 10 percent of
the release incidents reported age for either tank or piping. Of these,
both the mean and the median reported ages at the time of the release
incident report are 17 years. One factor to consider in interpreting
these results is that the age of newer tanks may have been known (and
therefore reported) more frequently than the age of older tanks, which
would tend to bias the results toward the lower age brackets. The
existence or magnitude of such a bias, however, cannot be verified at
this time.
It has been suggested that soil conditions in different regions of
the country significantly affect corrosion rates and tank lifetimes. For
example, the Northeast exhibits relatively moist, acid soils, which would
tend to have a high corrosivity. In order to see whether any geographic
patterns emerged from the tank system age data, the mean tank system age
was calculated for eight regions of the country. These calculations are
depicted in Table 5-1.
The results show that the mean tank age for the nation as a whole is
17 years, while the mean age for seven of the eight regions ranges from
14 to 18 years. Moreover, the regions in the Northeast are found at the
upper portion of the range, with means of 17 and 18 years. These data do
not support the contention that the high soil moisture and low pH in the
Northeast result in shorter tank system life.
5-1
-------
TJ
O
c
O
Q.
O
rr
w
-f-ป
c
45
1-5 11-15 21-25 31-35 41-45
Age (years)
Figure 5-1 Age of Tank System at Time of Subsurface Release Incident
* Incidents reporting both tank age and piping age are only counted once
5-2
-------
1348s
Table 5-1 Summary of Subtitle I Tank Age Data by Region
Region
New England
Mid-Atlantic
Southeast
Great Lakes
South Central
Northwest
Southwest
Pacific
National
Number of incidents
States* reporting system age
ME, NH. VT, MA, RI, CT
NY, NJ, PA, DE, MD, VA, WV
NC, SC, MS, TN, KY
OH, MI, IN, WI, MN, IA, IL
LA. TX, OK
KS. NE, SD, MT, CO, ID
NM, AZ
WA, OR, CA
All of the above
132
473
38
222
9
63
4
107
1,053
Mean age at time
of report
17
17
16
18
14
14
28
18
17
The following states did not document system age for any release incidents: GA, FL, AL,
MO, HI, PR, AR. ND, WY, UT, NV, AK.
5-3
-------
The only region of the country that significantly differs from the
national average is the Southwest. The average documented tank age in
New Mexico and Arizona was 28 years; however, this figure was derived
from only four documented incidents with age data out of a total of 86
documented incidents in those States. The low moisture content in those
States may tend to retard corrosion and extend tank life. The fact that
only four release incidents reported tank age may indicate that tank
systems last longer in the Southwest, and that the age is not known in
most cases.
In summary, the mean tank age at the time the release incident was
reported was 17 years. There was no correlation between region and tank
system age. It is conceivable that soil conditions in the Southwest
result in longer tank life; however, additional data are needed to
quantify this observation.
5.2 System Age for Tank Leaks Compared to Pipe Leaks
Figure 5-2 compares the age of tank leaks with the age of piping
leaks for all documented releases, where both age and location of leak
were reported. The data reveal a distinctly different trend for tank
leaks and piping leaks; i.e., piping leaks generally occur sooner than
tank leaks.
Specifically, the median reported age for piping leaks is 8 years,
while that for tank leaks is 17 years. Similarly, the mean reported age
for subsurface tank releases is 17 years, while that for subsurface pipe
5-4
-------
DC
&
8.
35% -r
30% -
25%-
20%--.
15%-
10%-
5%
0%
Total Incidents Reported = 1,200 (Tanks)
= 113 (Piping)
-30%
-25%
-20%
-15%
-10%
-5%
I I I I I I I I I I
<1 6-10 16-20 26-30 36-40 >45
1-5 11-15 21-25 31-35 41-45
Age (years)
Figure 5-2 Age of Tank and Piping at Time of Release Incident
5-5
-------
releases is 11 years. Further, the number of piping-related leaks tends
to decrease with age, whereas the number of tank leaks tends to increase
until the age reaches the 16- to 20-year bracket, at which point tank
leaks begin to decrease.
Factors that influence the observed trend of pipe failures' occurring
sooner than tank failures are as follows:
Pipe failures resulting from settling or loose fittings caused by
improper installation would tend to occur early in the life of the
tank system.
Piping walls are usually thinner than tank walls, which may cause
corrosion-related piping leaks to occur sooner than
corrosion-related tank leaks. (Specifically, schedule 40, 2-inch
diameter pipe has a wall thickness of 0.154 inch, whereas the wall
thickness of most steel tanks is 0.25 inch.)
For this study, reports for 78 subsurface piping release incidents
included pipe age, while reports for 671 subsurface tank release
incidents included tank age. (Note that these are the incidents that
identified both leak location and age.) Therefore, continued collection
and evaluation of tank system age data (particularly for piping-related
releases) are needed to lend additional insight into our understanding of
this issue.
5.3 Material of Tank Construction
Based on a survey of randomly selected tanks, EPA/OTS has estimated
that 89 percent of the tanks in the regulated tank universe are made of
steel and 11 percent are made of fiberglass. These estimates may be
compared to tank material information from the release incident data base
5-6
-------
shown in Figure 5-3. The data show that steel tanks represent 81 percent
of the release incidents and fiberglass tanks represent 19 percent, based
on the 10 percent of the incidents reporting material of construction.
This may reflect a greater tendency among State inspectors to report
material of construction for fiberglass tanks, since they are less common
and generally newer (i.e., age may be more likely to be known). The mean
age for the 462 documented steel tank releases reporting age was 17
years, while that for the 87 documented fiberglass tank release incidents
reporting age was 5 years.
The longer life for steel tanks may be due in part to the fact that
corrosion is a primary cause of release, and corrosion-related tank
failure tends to occur between 16 and 20 years after installation. The
shorter life for fiberglass tanks may be partially explained by
structural failure caused by improper installation. Another factor
influencing the trend, however, is that fiberglass tanks have only been
used for the past 25 years, and their use has increased during this
period. Therefore, since the age of the universe of fiberglass tanks is
less than that for steel tanks, the above results are not directly
comparable.
A similar analysis was attempted for piping releases; however, the
number of piping-related release incidents that include both pipe age and
pipe material (27) was not sufficient to draw even tentative conclusions
at this time.
5-7
-------
Statistically Selected Group of Underground
Storage Tanks (EPA-OTS) *
Release Incidents in Data Base
Incidents Reporting Material of Construction = 1,220
in
CO
19%
81%
Preliminary estimate for underground motor
fuel tank universe (not necessarily leaking)
provided by EPA-OTS excluding farm tanks
Figure 5-3 Material of Construction for Tanks
-------
5.4 Tank Volume Distribution
Figure 5-4 compares the tank volume distribution for release
incidents reporting volume to the tank volume distribution for a
statistically selected group of underground tanks as reported by
EPA/OTS. The figure shows that for both cases, more than half of the
tanks fall in the 4,000- to 11,990-gallon range, and both cases exhibit a
similar proportion of tanks in the other ranges shown. Based on the
above, large tanks are as likely to leak as medium or small tanks.
About 7 percent of the reported release incident tanks were of less
than 1,000-gallon capacity. Similarly, less than 1 percent of the tanks
reporting volume were bulk storage tanks with capacities in excess of
50,000 gallons. Because of economies of scale and design requirements,
these very small and very large tanks may require leak controls that are
different from the majority of tanks.
A parallel analysis was done relating the quantity released to tank
volume for release incidents in the data base. A very weak relationship
was found, indicating that large leaks are almost as likely to result
from small tanks as from large tanks.
5-9
-------
Total Incidents Reporting Tank Volume in Data Base = 4,001
10%
tn
iป
o
11.000-3.990 gallons |
[4,000-11,990 gallons |
49%
[4,000-11,990 gallonsj
58%
Documented Leaking Tanks (OUST Data Base)
All Motor Fuel Tanks (EPA-OTS)
Figure 5-4 Tank Volume Distribution
-------
6. RELEASE INFORMATION
6.1 Location of Release
More than 90 percent of the reported release incidents identified the
source of the release. These data are summarized in Figure 6-1.
Approximately 43 percent identified the "tank" as the location of the
release, while 18 percent identified the "piping." Another 13 percent
said the location of the release was either the "tank or piping," or
"both the tank and piping." The large number of releases from pipes
suggests that any leak control system must address both piping and tanks.
Approximately 17 percent of the reports specified tank "overfills" as
the source of the release. The record-keeping procedures in some States
did not distinguish between overfills of vehicle fuel tanks and overfills
of underground storage tanks. These records were not used. The number
of overfills would probably be increased if this distinction could have
been made.
Overfills appear to represent a significant portion of the releases,
yet pose very different problems from those associated with releases from
underground storage systems. When an overfill occurs, for example, the
volume released is limited and cleanup can usually take place quickly.
The volume released from tank systems is difficult to identify and clean
up.
6-1
-------
1
s.
0)
cc
o
45%-,-
40% -...
35%-
30%-
25%--
20%-..-
15%-.-
10%-
5%--
fc'}
0%
-45%
Total Incidents Reporting Location = 11,778
-40%
-35%
-30%
-25%
-20%
-15%
-10%
-5%
I "" I I T
Tank Overfilling Pump Unknown
Piping Tank or Piping Tank and Piping Other
Location of Release
-0%
Figure 6-1 Location of Release
6-2
-------
6.2 Cause of Release
Figure 6-2 presents the specified causes for subsurface tank
releases. Structural failure is the most commonly specified cause,
followed by corrosion, loose fittings, improper installation, and natural
phenomena. The figure also indicates that a wide variety of causes were
reported, many of which (e.g., "leak") are not specific enough to clearly
characterize the leak..
The majority of reported subsurface release incidents did not
positively identify the cause. It is conceivable that when the cause was
corrosion-related, it was less likely to be reported than when it was
known to be structural failure, loose fittings, etc. While we can say
that corrosion and structural failure are both significant causes, the
data are not strong enough to support a specific estimate of the
frequency of these causes or of other less frequently occurring causes.
6.3 Comparison of Structural Failure and Corrosion-Related Incidents
The coding form used in this study was designed to allow up to five
different causes to be specified for a given release incident. Of those
incidents reporting structural failure and at least one other cause, the
other cause was seven times more likely to be improper installation or
natural phenomena than corrosion. Based on the above, there appears to
be little overlap between structural failure releases and corrosion-
related releases.
6-3
-------
0%
Structural
Failure
Corrosion
Loose Fittings
Improper
Installation
Natural
Phenomenon
10%
i
0%
10%
20%
30%
I
40%
I
50%
? ,-; '-s .
- r - - -- ' - ซ - <". ' v
*A" < '"> ' * '
Number of Times One of the Specified
Causes Was Reported = 2,496
Note: Many reports indicated a
nonspecific cause of
release, for example "leak.1
20%
30%
40%
50%
Percent of Causes Reported
Figure 6-2 Cause of Release
6-4
-------
6.4 Causes of Structural Failure
The coding form was also designed to allow the reviewer to make hand-
written comments either to expand upon a standard response or to enter a
nonstandard response.
There were a total of 158 failure incidents with explanatory comments
pertaining to structural failure. Analysis of these comments revealed
that causes cited for structural failure incidents included: vehicle hit
pump; high water table lifted tank; rupture caused by excavation; rupture
due to excessive air pressure during tank tightness test; external load
caused by vehicular traffic; frost damage; and pipe hit by snow plow.
Based on these comments, structural failure includes a variety of
specific causes. The comments also support the above finding that
relatively few of the structural failure incidents involve
corrosion-related problems.
More than half of the comments indicated that the structural failure
occurred in the piping, about 30 percent involved the tank, and 15
percent involved the pump. Of the structural failures in the piping, 51
percent involved the supply line, 22 percent involved pipe fittings, and
9 percent involved the vent pipe.
The above comments illustrate that a program to minimize tank leaks
must address a variety of causes of structural failure in addition to
corrosion-related releases. The comments also illustrate that a
comprehensive leak control program must include the piping and the pumps
as well as the tank itself.
6-5
-------
6.5 Cause of Release by Location of Release
Figure 6-3 presents the cause of release by location of release for
subsurface releases. The figure shows that although structural failure
and corrosion are the two dominant causes for subsurface tank releases,
subsurface pipe releases are more likely to result from a variety of
specified causes, including structural failure, loose fittings,
corrosion, and improper installation.
6.6 Cause of Release by Age at Time of Report
Figure 6-4 presents the cause of release for various tank system
ages. Corrosion is the dominant cause of failure for tank systems more
than 10 years old. For example, 71 percent of the failures of tank
systems between 11 and 20 years old, and 60 percent of those more than 20
years old, were caused by corrosion. Conversely, tank systems that
failed in the first 10 years after installation are almost as likely to
be caused by structural failure, loose fittings, or improper installation
as by corrosion.
We emphasize that these findings are based only on those incidents
that reported both cause of release and age information.
6.7 Cause of Release by Material of Tank Construction
Figure 6-5 compares the cause of release for release incidents
involving fiberglass tanks to that for steel tanks. The results show
that structural failure is the most common cause in incidents involving
fiberglass tanks, contributing to approximately 50 percent of the
6-6
-------
0% 10% 20% 30% 40% 50% 60%
i ! . ! . I . i.i.i
Structural
Failure
Corrosion
Natural
Phenomenon
Improper
Installation
Loose Fittings
Major Specified Causes Reported:
1137 for Tanks
864 for Piping
Note: Many incidents reports indicated a
nonspecific cause of release, for
example "leak".
i i :
' i i i i i j
0% 10% 20% 30% 40% 50%
Percent of Causes Reported
60%
Figure 6-3 Cause of Release Where Location of Release Is Specified as
Tank or Piping
6-7
-------
<ง
0% 10% 20% 30% 40% 50% 60% 70%
, i i I . i
Structural
Failure
Corrosion
Loose Fittings
Improper
Installation
Natural
Phenomenon
' ." V4#$fr '"'' ' ฃ
Vs"'' f.
Major Specified Causes Reported:
0-10 Years-191
11-20 Years-134
>20 Years-101
Note: Many incident reports indicated a
nonspecific cause of release, for
example "leak."
i
0% 10% 20% 30% 40% 50% 60%
Percent of Incidents Reported
70%
Figure 6-4 Cause of Release by Age for Subsurface Releases
6-8
-------
0% 10% 20% 30% 40% 50% 60% 70%
Structural
Failure
Corrosion
Loose Fittings
Improper
Installation
Natural
Phenomenon
OP/o
Major Specified Causes Reported:
Steel Tanks - 334
FRP Tanks-111
Percentages total more than 100% since
some incidents reported multiple causes.
1 - - 1
30% 40%
r
10% 20% 30% 40% 50% 60%
Percent of Incidents Reporting above Causes
70%
Figure 6-5 Cause of Release by Material of Construction
6-9
-------
incidents specifying cause. For steel tank releases, however, corrosion
is the most commonly specified cause, contributing to more than 60
percent of the incidents specifying cause. Other causes that are more
commonly reported for fiberglass tanks than for steel tanks include
improper installation and natural phenomena.
Based on the preceding information, it appears that the greater
resistance to corrosion and the lesser structural strength of fiberglass
tanks increase the importance of manufacture- and installation-related
causes and decrease the importance of corrosion-related causes in
comparison to steel tanks.
6.8 Quantity Released
Figure 6-6 presents the quantity released distribution for documented
release incidents. The figure shows that about 33 percent of the
incidents involve releases of 100 gallons or less, about 65 percent
involve releases of 500 gallons or less, and about 97 percent involve
releases of less than 10,000 gallons.
The quantity released is one of the more subjective variables
reported in the data base. This is because it involves an estimate of
the rate of release and duration of release based on inventory data and
observation of the tank system and impacts. The value was commonly
reported, however, and is believed to be meaningful on an
order-of-magnitude scale.
6-10
-------
1
a
o>
CC
c
I
s
I
10,000
Quantity Released (Gallons)
Figure 6-6 Quantity Released
1-35%
-30%
25%
-20%
-15%
-10%
- -5%
-0%
6-11
-------
6.9 Quantity Released by Location of Release
Figure 6-7 compares the reported quantity released for surface
spills, subsurface tank leaks, and subsurface pipe leaks. The data
indicate that subsurface piping leaks and tank leaks exhibit a similar
pattern of quantity released, increasing to a peak at the 500- to
2,500-gallon range, and then decreasing. Surface spills exhibit a
significantly different pattern, with the majority of leaks in the less
than 100-gallon range, sharply decreasing as the spill amount increases.
Quantity released is believed to be one of the more subjective
parameters in the data base. Nevertheless, these results support the
logical finding that surface spills usually involve smaller quantities
than do subsurface spills.
The data also indicate that the quantity released from subsurface
tank leaks is similar to that for subsurface pipe leaks, even though
piping may have a smaller volume than the tank and may not always hold
the product. When it does, however, the line is often under pressure and
the product is forced out, resulting in release quantities much larger
than the pipe volume.
6-12
-------
70%-,
r-70%
1
cc
c
03
32
o
c
10,000
Quantity Released (gallons)
Figure 6-7 Quantity Released by Location of Release
6-13
-------
7. LEAK DETECTION INFORMATION
7.1 Method of Initial Leak Detection
Figure 7-1 summarizes the methods by which subsurface releases were
detected based on information from more than half of the subsurface
release incidents. Sight is the most commonly reported detection method,
followed by smell. As examples, sight might include evidence of gasoline
at the ground surface, in surface water, or in excavation holes, while
smell would include odors in basements or in drinking water. Together,
sensory detection methods account for more than 70 percent of the
reported methods. Therefore, the majority of reported tank releases
appear to be detected unintentionally. On the other hand, inventory
control accounts for only 11 percent of the reported methods. This
finding implies that inventory control alone may not be effective in
identifying leaks. (We are not certain of the percent of incidents where
inventory control was practiced.) Similarly, tank integrity testing and
detection equipment accounted for less than 10 percent of the reported
methods. This could be due to the fact that relatively few tanks are
tested prior to some other indication that a leak has occurred. Water in
the tank accounted for 7 percent of the releases and would be expected
only in areas of high water table where water would flow into the tank.
Based on this information, it appears that a significant number of
documented tank releases were first detected unintentionally and that a
planned, effective leak detection program would therefore reduce the
associated impacts and costs.
7-1
-------
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
c
o
;=
O
0)
4-*
0)
Q
^:
ctS
T3
O
Q)
Sight
Smell
Product
Inventory
Water in Tank
Integrity Test
Other
Detection
Equipment
Taste
Chemical
Analysis
Total Incidents Reporting
Detection Method = 5 845
0% EP/o 10% 15% 20% 25% 30% 35% 40% 45%
Percent of Incidents Reported
Figure 7-1 Method of Initial Leak Detection for Subsurface Release Incidents
7-2
-------
It should be noted that releases that were initially detected by
inventory control or detection methods may have been corrected without
notifying State officials, and therefore would not be included in this
data base. Detection equipment and chemical analyses were more often
reported as methods of initial detection in 1984 and 1985 than they were
in 1970 to 1983. This trend is expected to grow in future years, due in
part to regulatory requirements.
7.2 Quantity Released by Detection Method
Figure 7-2 compares the quantity released in incidents detected by
product inventory with that discovered by detection equipment {including
integrity tests). The figure shows a distinctly different pattern
between the two methods. Product inventory control accounts for a
smaller percentage of all releases of 500 gallons or less and a larger
share of all releases of more than 500 gallons. Although the quantity
released is probably one of the more subjective elements in the data
base, the trend indicates that inventory control is less sensitive than
leak detection equipment. This supports our earlier finding that
relatively few small leaks were detected by inventory control.
Additional information from tank systems using both inventory control and
periodic tank testing would be most useful in quantifying the relative
effectiveness of detection methods.
7.3 Use of Integrity Tests to Confirm Leaks
The use of integrity tests to confirm leaks appears to be much more
common than their use as an initial detection method. For example, while
7-3
-------
45% -]
40% -
1-45%
35% -
30%--
25%--
20%--
15%--
10%-
0%
0-100
Incidents Reporting Product Inventory and Quantity Released = 671
Incidents Reporting Detection Equipment and Quantity Release =40
] Product Inventory
I
I
101-500 501-2,500
Quantity Released (Gallons)
I 1
2,501-10,000
>10,000
Figure 7-2 Quantity Released Where Initial Detection Was
Accomplished by Product Inventory or Detection Equipment
-40%
-35%
-30%
-25%
-20%
-15%
-10%
-5%
-0%
7-4
-------
only about 8 percent of the incidents reporting detection method showed
that an integrity test was used for initial detection, approximately half
of the incidents reporting detection method used an integrity test to
confirm the leak after it had been detected by another means.
7-5
-------
8. IMPACTS AND REMEDIAL ACTIONS
8.1 Documented Media Impacted
Figure 8-1 summarizes the number of incidents with documented impacts
to a given medium. (A documented impact is one in which the record
specifically states that a release affected a given medium.)
The data indicate that soil impacts were found in 67 percent of the
documented incidents and that ground water was contaminated in 45 percent
of these incidents. Releases to surface water, air, and other media each
occurred in 22, 15, and 12 percent of the documented incidents,
respectively. The other media specified included sewer lines, pavement,
basements, and electrical conduit/vault. Impacts to media other than
soil are a significant subset of all impacts. Therefore, the potential
economic costs associated with remediating these impacts (and the savings
associated with minimizing releases) can be substantial.
8.2 Documented Impacts
Table 8-1 summarizes the specific impacts to each medium. Note that
83 percent of the documented releases did not identify the impact, other
than to say that a release to a given medium had occurred. Of those
incidents where the impact was specified, some of the impacts reported
were:
Combustible fumes in confined areas (908 incidents);
Contamination of municipal potable ground-water supply
(40 incidents);
8-1
-------
ฃ3
I
c
o
3
O
"c
70% -
60% -
50% -
40%-
30% -
20% -
10%-
0%
^ ฃ -
i
: J"> ,
';l *
t -** '
"
fotaJ Documented Incidents = 12,444 k
* Note: Other med a include pavement,
sewer lines, electrical conduits, and
basements
V
^
%*^y
',
""^
>
-------
1257s
Table 8-1 Types of Impacts for Documented Subtitle I Incidents by Media
Number of incidents (percent)
Health/environmental impact
Media impact "unspecified"
Contamination of private water supply
Contamination of municipal water supply
Contamination of surface waters
Contamination of other ground waters
Human illness
Human death
Damage to aquatic life
Damage to wildlife
Damage to plant life/crop loss
Damage from corrosion
f Damage from fire/explosion
*** Contamination of private potable surface water
Contamination of municipal potable surface water
Contamination of industrial surface water
Contamination of non-contact recreational surface water supply
Contamination of contact recreational surface water supply
Combustible fumes in confined area
Other
Total impacts"
Ground
water Soil
4,423 (79) 8,070 (98)
749 (13)
40 (1)
126 (2)
177 (4)
12 ซ1) 25 ซ1)
1 ซD
6 (<1) 10 <<1)
2 ซ1) 9 (4)
8 ซ1) 99 (1)
5 ( <1 ) 3 ( <1 )
13 ซ1) 27 (<1)
32 ซ1) 10 ซD
5,593 8,254
Surface
water
2,242
12
103
26
22
3
8
11
27
17
114
28
129
2,742
(82)
ซD
(4)
(1)
ซD
ซ1>
ซD
ซD
(1)
ซD
(4)
(1)
(5)
Air Other
1,770 (97) 322 (22)
33 (2) 28 (2)
1 ซD 1 ซD
3 ซ1) 14 (1)
3 (<1)
2 <<1) 10 ซ1)
23 (2)
19 (1) 101 (7)
908 (63)
6 ซ1) 39 (3)
1,834 1,449
Total
impacts*
16,827 (85)
749 (4)
40 (<1)
126 ซ)
177 (1)
110 ซ1)
3 (<1)
136 (1)
40 (<1)
141 (1)
34 (<1)
168 (1)
11 (
-------
Contamination of private wells (749 incidents);
Fire/explosion (155 incidents);
Human illness (100 incidents);
Aquatic life impacts (103 incidents);
Plant life/crop loss (99 incidents);
Corrosion of structures (34 incidents); and
Human death (2 incidents).
These data begin to put into perspective the impacts associated with
tank releases. Caution is required in projecting these data to the
universe of tanks, however, because there is still a high degree of
uncertainty in several key areas
The data are based on 12,444 documented release incidents;
however, the total number of releases is not known.
The impact information on file may be incomplete.
Release incidents that are documented in State files are likely to
include impacts resulting in third party complaints.
Therefore, the above information may be used to document the types
and the minimum number of impacts associated with releases. Additional
data would be required to estimate current and potential impact
associated with the universe of tanks.
A total of 1,120 incident coding sheets contained some type of
handwritten comment pertaining to a health/environment impact.
Approximately 20 percent of these cases were reviewed, providing insight
into the content of the comments that were not accommodated by the
standard wording on the coding sheet.
8-4
-------
The media receiving comments included ground water, soil, surface
water, air, and "other." Soil and surface water accounted for the
greatest number of comments, 65 percent and 60 percent, respectively,
while ground water, "other," and air received 38 percent, 29 percent, and
11 percent, respectively.
The specific content of the comments varied from descriptive and
informative accounts to vague and insignificant comments. Some of the
more significant findings include the following:
More than 60 percent of the comments were repetitious of coded
information. The major exception was comments on surface water
where approximately 90 percent included a brief location
description (i.e., name and/or type of water).
Six cases of human illness were documented, related to
fire/explosion or combustible fumes/vapors in a confined area. No
further information was provided for these cases.
Two human deaths were attributed to fire/explosion. The details
surrounding these deaths were provided.
Approximately 90 percent of the soil comments were "unspecified."
About 50 percent of the incidents reviewed contained comments
describing the given impact. Most of these comments were short and many
were uninformative. In several instances, however, in which the
responses did provide some insight into the extent of the impact. The
following case summaries are representative of those presenting useful
and interpretive information.
A facility in Northern Virginia reported a chemical leak.
Ground-water impact was described as "contamination of 23 private
wells and city (of Manassas) well."
8-5
-------
A service station in Barrington, Illinois, was responsible for a
gasoline leak. Health impact included human illness. Descriptive
information indicated: soil - "75 yds^ of soil excavated"; other
- "fan installed to dissipate gas vapors must have pulled carbon
monoxide into bank, facility."
A company in St. Paul, Minnesota, reported a mineral spirits
leak. Ground-water medium impact was described as "contamination
of other ground waters, i.e., a parched aquifer only - no adverse
environmental effects on local ground water."
A facility in Charleston, West Virginia, was responsible for a
diesel fuel leak. Other (miscellaneous) impact description
included "a family evacuated from their house and the children
became sick."
The majority of the coding sheets reviewed contained handwritten
comments that were often repetitious of the coded responses. Those
responses that provided additional data usually contained a description
and/or a personal viewpoint of the extent of impact.
The above discussion was based only on handwritten comments from the
coding sheets. The following examples illustrate the level of detail
that can be achieved by analyzing individual State files. The three
files discussed were chosen based on a high reported corrective action
cost.
Facility Type: Retail Gas Station
Pennsylvania
Date of Release: February 2, 1980
Capital Cost: $500,000
This incident reported that 5,600 gallons of gasoline had been
released from a 3,000-gallon tank, which resulted in the contamination of
a private well. The report noted that the legal and/or regulatory
actions since the release included remedial measures and monitoring
8-6
-------
requirements. The file recorded corrosion to be the cause of release,
but did not note the age of the tank. The release resulted in the
contamination of the ground water, soil, and a basement, and was detected
offsite by odor/vapor.. The remedial measures included
recovery/interceptor wells, artificial recharge, biological techniques,
monitoring wells, and tank or piping removal/replacement.
Facility Type: Private Establishment
Pennsylvania
Date of Release: August 8, 1975
Capital Cost: $464,000
This incident involved the release of 94,000 gallons of heating fuel
oil that contaminated the soil and surface water. The duration of the
release was noted as 450 days. There were no legal or regulatory
requirements prior to or after the leak was detected. The cause of the
incident was noted as "leak." The detection was offsite by physical
inspection/sheen. The remedial actions included soil excavation, tank or
piping repair, and recovery (unspecified).
Facility Type: Municipal
Massachusetts
Date of Release: May 7, 1984
Capital Cost: $400,000
This multiple tank incident involved the release of an unspecified
amount of heating fuel oil from two 10,000-gallon tanks. The release
impacted the ground water and soil on the site of a public high school.
The legal and regulatory actions required once the leak was detected
included remedial measures, release detection, and containment.
8-7
-------
These examples illustrate in graphic detail the impacts that can
result from underground storage tanks without leak controls.
8.3 Corrective/Remedial Action
Figure 8-2 shows the distribution of corrective/remedial action
measures for those incidents that reported taking some action. The
figure shows that relatively simple remedial actions such as tank
replacement/repair and monitoring wells were more common than actions
involving recovery and treatment.
Replacement of tank and/or piping is the most common type of action
reported, followed by monitoring wells. The "other" category represents
actions that were not among the standard answers. The most commonly
reported "other" actions included tank pumped out, cleanup, hired
contractor, pumped from hole, and sewer flushed.
Soil excavation and recovery wells were both cited as corrective
actions, but neither accounted for more than 15 percent of the reported
incidents. Ground-water treatment techniques such as steam stripping
(17 cases) were extremely rare. This information is based only on data
specified in the State files, and in many cases the remedial action may
not have been reported. Conversely, the release incidents reported in
State files would probably include those involving major cleanup
activity. Therefore, the data imply that the majority of underground
tank release incidents to date have involved no cleanup action. This
8-8
-------
0% 5ฐ
Tank/Piping
Remove/Repl.
Monitoring
Wells
Soil Excavation
Absorbant
Material
Tank/Piping
Repair
Recovery
-i- :
) ""' ',1 j
|
I
, _. ~i. :>,4.^^S-> - |
. ,' f ' "
4f
" &$$$>*
-!/," l^j
- * ^ ]
;-r i
!
[Total Documented Incidents = 12,
|
i
444 |
n i i i i | i
0% 5% 10% 15% 20% 25% 3C
I
%
Percent of Documented Incidents
Figure 8-2 Corrective/Remedial Measures
8-9
-------
conclusion is tentative, however, because of uncertainty regarding the
comprehensiveness of State records dealing with cleanup.
8-10
-------
9. CHEMICAL RELEASES BY STATE AND BY YEAR
EPA has the authority under RCRA Subtitle I to establish regulations
for underground storage tanks that may vary depending on the material
stored. The mandated dates for implementation of these regulations may
also vary depending on whether the storage tank is used for petroleum
products or for chemicals. This portion of the report compares the
results of the subset of regulated chemical release incidents (excluding
wastes) to the entire data base. The majority of incidents in the entire
data base are petroleum releases.
9.1 Chemical Releases by State
Figure 9-1 presents a comparison of release incidents for all
materials per State in the entire data base with the subset of chemical
releases reported per State in the entire data base. The States with
large numbers of reported chemical releases were among those with the
most chemical manufacturing facilities (i.e., California, New Jersey, and
Texas). One exception is Louisiana, which has a high number of chemical
facilities but reports no chemical release incidents.
Minnesota reports a large number of petroleum and chemical release
incidents due, in part, to the rigorous reporting and enforcement
procedures that have been in place since the early 1970s. Colorado has
mandated that petroleum tank releases be reported since 1973, but only
recently implemented similar requirements for chemicals. State reporting
requirements have apparently had an influence on the number of documented
release incidents, as illustrated by the results of these States.
9-1
-------
3 :: MO
Number of Incidents
by State
D 9-50
51-250
251-1,000
.000
TOTAL RELEASE INCIDENTS
BY STATE
Number of Incidents
by State
CHEMICAL RELEASE INCIDENTS
BY STATE
Figure 9-1 Comparison of Reported Release Incidents by State
9-2
-------
9.2 Number of Chemical Releases by Year
Figure 9-2 compares the number of chemical releases by year to the
total reported releases by year. In general, the number of incidents
reported is Increasing each year. The variations observed for the
chemicals released are probably due to the relatively small size of this
subset for this part of the data base. Since the number of chemical
incidents is small (248 incidents) in comparison to the total number of
reported incidents (12,075 incidents), a difference of a few chemical
incidents in a given year can influence the results for that year
significantly. One difference that may be significant is that about 19
percent of the chemical releases were reported in 1985, while only about
8 percent of all the total documented releases (of all materials) were
reported in 1985.
9.3 Facility and Material Information
9.3.1 Facility Type
Figure 9-3 compares the facility types for total and chemical release
incidents. The pattern is distinctly different on each graph. The
majority of chemical releases occurred at manufacturing facilities
(~657.) whereas the majority of the total release incidents occurred at
retail gasoline stations (~65%).
9.3.2 Facility Operating Status
Figure 9-4 compares the facility operating status at the time of
release for total and chemical release incidents. The results are
virtually identical with more than 97 percent of the reported releases
coming from operating facilities.
9-3
-------
Number of Incidents
_- _-* _l\> fO G
en o 01 o in c
o o o o o c
3 O O 0 O O C
. 1.1.1,1,1.1
I Total Release Incidents Reported =
^ m n n n fl I
1 2,075 1
rm
WT
:
...
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
-3,000
-2,500
-2,000
-1,500
-1,000
-500
Year
Note: Partial results
shown for 1985
I
ou
45-
40-
35-
30-
25-
20-
15-
10-
5-
u\ I
70
Chemical Release Incidents Repotted = 248
n
;
r^*
(
ff
1
I
\
Wfซ
--
&
;;
*w
^
5
'i;
s
!
*
...
...
...
"p
.i:
f
1 1 I I I 1 I I 1 I I I I I
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
-t)U
-45
-40
-35
-30
-25
-20
-15
-10
-5
Year
Note: Partial results
shown for 1985
Figure 9-2 Comparison of Total and Chemical Release Incidents by Year
9-4
-------
I
Q)
8.
a>
CC
0)
D.
/U/o
60% -
50% -
40% -
OU /o "
20% -
10%-
* ป ^
! * 4 %
t tt ss -JUfc.
V* .
f*$,
\& j s\"'
f :
'4'
?-\ .
I
Total Release Inc
Facility Type = 1
^ S#
v5 -.;.<
,'
/> ^ v ' t
I
'"${ *
idents Reporting ta
1,861 1
HP1
1 1
inm
lU'/o
-60%
-50%
-40%
-20%
-10%
Retail Gas Station Commercial Establishment Municipal
Other * Manufacturing Facility
Facility Type
ง
OUYo
70%
60% -
50%-
40% -
30% -
20%-
10%-
Chemical Release Incidents Reporting I
Facility Type = 239 |
r !
i
<,;a
- if
s ซ !
^il<
f \
1 1 1
-BUYo
-70%
-60%
-50%
-40%
-30%
-10%
Retail Gas Stations Commercial Establishment Municipal
Other * Manufacturing Facility
Facility
* Other Includes: private residence, bulk storage/petroleum jobber, commercial transportation, military, state,
apartment/condominium complex, federal (non-military), airport, farm agricultural, dry cleaner, and office buildings.
Figure 9-3 Comparison of Facility Type for Total and Chemical Release Incidents
9-5
-------
0,
a
c
0)
8
**r
1
8.
o
cc
^2
o
'o
*;
"E
0.
1 UU /o
Qno/ _
au /o
70% -
ft no/
4 no/
1 no/ _
1 nno/ _,
I UU /o ""
7 n*/
cno/
OU /o ""
JE^no/
jinp/ _
ono/ _
ซ
i no/
1U /o
1 i H
- ff t- ^
o si<
1 . i -li
1 \
* ss
f *
, ; t
^ *"
I fff
"
S
Operating
\ || ^|x>;^ '; [,
yllpl? f^
? ^'J $ฃ f ff
}f|:' ;,
IP ^
r ^ ^
& /; ' 5' ', '
'*!*'''ซ ^
f. ^s -.-.A;! -.^^
"'if '! -"'
t" !C'' %
Total Release Incidents b
Reporting Status = 10,873 1
Fx^Sx^x^yx^r-xiVS:^:^:1:^-:??^!
i
Abandoned
Chemical Release Incidents j|
Reporting Status = 231 |
i
i uu 10
-80%
-70%
_cnoA
A no/
ono/
f\O/
mno/
B
vnซ/
_ cno/
DU /o
_ono/
_ < no/
lU/o
Operating Abandoned
Operating Status
Figure 9-4 Comparison of Facility Operating Status for Total
and Chemical Release Incidents
9-6
-------
9.2 Number of Chemical Releases by Year
Figure 9-2 compares the number of chemical releases by year to the
total reported releases by year. In general, the number of incidents
reported is increasing annually. The variations observed for the
chemicals released are probably due to the relatively small size of this
subset for this part of the data base. Since the number of chemical
incidents is small (248 incidents) in comparison to the total number of
reported incidents (12,075 incidents), a difference of a few chemical
incidents in a given year can influence the results for that year
significantly. One difference that may be significant is that about
19 percent of the chemical releases were reported in 1985, while only
about 8 percent of the total documented releases (of all materials) were
reported that year.
9.3 Facility and Material Information
9.3.1 Facility Type
Figure 9-3 compares the facility types for total and chemical release
incidents. The pattern is distinctly different on each graph. The
majority of chemical releases occurred at manufacturing facilities
(~65%), whereas the majority of the total release incidents occurred at
retail gasoline stations ("657=).
9.3.2 Facility Operating StatJS
Figure 9-4 compares the facility operating status at the time of
release for total and chemical release incidents. The results are
virtually identical, with more than 97 percent of the reported releases
coming from operating facilities.
9-7
-------
9.3.3 Materials Stored
Figure 9-5 compares the material stored for total and chemical
release incidents. Gasoline was cited in over 70 percent of the release
incidents for the total data base, followed by 17 percent for heating and
fuel oils. Solvents specified as unknown in the file, along with
toluene, account for 70 percent of the chemical release incidents. Acid
and alkaline material releases accounted for only about 10 percent of all
chemical incidents reported.
9.4 Tank System Age Information
Figure 9-6 compares tank system age at time of leak for total and
chemical subsurface release incidents. About 10 percent of the
documented chemical release incidents specified the age of the tank
system at the time the leak was reported. Both the mean and the median
of the chemical release incident ages were 17 years. This agrees with
the mean and median reported ages for the total data base, which were
also 17 years.
9.5 Release Information
9.5.1 Location of Release
Figure 9-7 presents location of release information for chemical
incidents. The overall trend for chemical releases is similar to that
for the total data base, with the tank leaks being the most common
location. The most significant difference is that documented piping
releases appear to be less frequent than overfills in chemical releases,
while the opposite was true for the total data base.
9-8
-------
&
I
OUVo
7f)o/
firw
50%
dn%
orvo/
?o% -
10% .
> "
.'*
1
s>:
, $ *
'n
$
: gt{
# ,
J,O
3'
ป>
! '
1
HI
1
n it
i i
Total Release Incidents Reporting h
Material Stored = 12,214 '
PI
( 1 1 1
-ao%
/0%
60%
Crtoy
OUvo
-40%
20%
-10%
Gasoline Diesel Fuel RecydedOil Unknown
Heating/Fuel Oil Other * Chemicals Wastes
Material Stored
* Other includes kerosene, jet fuel, motor fuel, aviation gasoline, propane,
an miscellaneous petroleum products.
70%
"m 60%
a
(T 50%-
1
o 40% -
-0 HU /o
-2 30% -
Cv 20%
10%
nv
1
!#'
* ซwซ^
>
X '
-(
1
;ซKซ:$ซ::KK:::
:j,ซซ? .ijS;S?ซi:
Chemical Release Incidents k
Reporting Material Stored = 248 |
|/ilfel|
| : 'ป>Vt\ .--^w ^ ^i-^^^^f^|
1 I
70%
-60%
-50%
-40%
-30%
- 20%
-10%
-0%
Solvents Acids
Miscellaneous Chemicals
Material Stored
Bases
Figure 9-5 Material Stored in Leaking Tanks for Total and Chemical
Release incidents
9-9
-------
O)
O.
ฃ3/0
20%-
15%-
10%-
i
-
-,
!
I
nn 1 1
*
*
c ^
Total Release Incidents jl
Reporting Age = 1,220 '""
,
-
^u:
1: EH
1 1 1 1 J 1 1 1 T "~T
-25%
-20%
-15%
-10%
-5%
-0%
6-10 16-20 26-30 36-40 >45
1-5 11-15 21-25 31-35 41-45
Age (years)
I
'o
c
I
0>
Q.
30% -
25%-
20%
15%-
10%-
0%
I
^
1
**W
1
f\"
f
?
fs ^
1 1
M**
FWW
^
-
'
Chem cal Release Im
Reporting Age = 26
:-
'$,
lilt
~?~
' ,:
;idents
1
1 J
-35%
-30%
-25%
-20%
-15%
-10%
-5%
6-10 16-20 26-30 36-40 >45
1-5 11-15 21-25 31-35 41-45
Age (years)
Figure 9-6 Comparison of Age of Tank System at Time of Report for Total and
Chemical Incidents
9-10
-------
HO /o
40% -
"8 35% -
ง 30%-
cc
(A
1 25%-
o
'o
"6
1 15%-
^ 10%-
45%-
40% -
P 35%-
1
ง 30% -
cc
I 25% ~
= 20% -
"o
| 15%-
^ 10%-
ฐ
....
...
\>h^
sj "1'
Av
Total Incidents Reporting Location b
of Release = 1 1 ,674 |
"S:
!
VA
f.
,^
|"!;::!v':!a;|
1 1 1 1 1 1
Tank Overfilling Pump Other
Piping Tank or Piping Tank and Piping
Location of Release
' &
'/<'
^:
i
^1
Chemical Release Incidents Reporting jl
Location of Release = 242 ง
ifl
"'
' ^. *
;&
&' r~i F1 F^n
1 1 1 1 1 1
-40%
-35%
-30%
-25%
-15%
-10%
-5%
-45%
-40%
-35%
-30%
-25%
-20%
-15%
-10%
-5%
Tank Overfilling Pump Other
Piping Tank or Piping Tank and Piping
Location of Release
Figure 9-7 Comparison of Location of Release for Total and Chemical Incidents
9-11
-------
9.5.2 Cause of Subsurface Release
Figure 9-8 compares causes of subsurface release for the chemical
subset to the total data base. Corrosion replaces structural failure as
the most commonly reported cause of release for chemical incidents. This
and other differences may be due at least in part to the smaller number
of chemical releases reported.
9.6 Leak Detection Information
Figure 9-9 compares the initial leak detection method for chemical
and total release incidents. Sight accounts for 63 percent of the
documented chemical release detection methods and 41 percent of the total
incident detection methods. The relative order of the initial detection
methods is similar for chemical and total releases. In both data sets,
integrity tests and chemical releases account for less than ?0 percent of
the reported detection methods. Therefore, implementation of an
effective leak detection program should significantly reduce the duration
of release for chemical tanks and for all tanks.
9.7 Corrective/Remedial Act ion/Impacts
Figure 9-10 compares documented corrective/remedial actions for
chemical releases to actions for all releases. Tank/piping removal/
replacement is the most common action reported for both data sets.
However, soil excavation was reported in 25 percent of the chemical
incident corrective actions, which is about twice as often as for the
total release incidents. Monitoring wells were also more common for
chemical incidents than for the total data base. Tank/piping repair, on
the other hand, was more common for the total data base than for chemical
incidents.
9-12
-------
M
Structural
Failure
Corrosion
Loose Fittings
Improper
Installation
Natural
Phenomenon
'o
10%
20% 30% 40% 50%
1.1,1.
: | :
li^t;,
\ *Ar
'
;
/> v ^ >,% -.j--
* 1
.' i."1
"< S ฃ i'f
0%
0%
Structural
Failure
i
j
10%
10%
I
-
- - > ^V^,^A^W^$iฃt,
. \ ^
? \v.^ f '- -
E
1
Number of Times One of the Specified h
Causes Was Reported = 2,496 1
, . , i i i
20% 30% 40% 50%
Percent of Causes Reported
20% 30% 40% 50%
| .1,1
^JJ^^ s * V'{
Corrosion
Loose Fittings
Improper
Installation
Natural
Phenomenon
J/o 10
% 20% 30% 40% 50
1 . 1 . I
1 ^:~:IA^ ;^
: r
iS?r:A '^ ?<ฃ * * ^^ {
_ ^/l
i
v'.'H'Jvr
% 10
"7"!' !
' :
^;^-^-1;^
I
I
Number of Times One of the Specified Causes k
Was Reported for Chemical Incidents = 41 |
:
t
I
i i I '
% 20% 30% 40% 50
%
;
!
i
%
Percent of Causes Reported
Figure 9-8 Comparison of Cause of Release for Total
and Chemical Release Incidents
9-13
-------
o
CD
Q
75
-
o
1
CO
^
'c
o
o
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
_| . 1 . I . I . I . 1 . I . I . 1 . 1
Sight
Smell
Product
Inventory
Water in Tank
Integrity Test
Other
Detection
Equipment
Taste
Chemical
Analysis
' 1
r
i
i
i
Total Incide
Method = 5
1
I
1
nts Reporting Detection ta
845 |
i i i i i i i i i i . i > i - i
!
i
0% 5% 10% 15% 20% 25% 30% 35% 40%
Percent of Incidents Reported
45%
Sight
Smell
Product
Inventory
Water in Tank
Integrity Test
Other
Detection
Equipment
Taste
Chemical
Analysis
oc
-
0ฐ
/o 10
% 20% 30% 40% 50% 60% 70
I I I I I
>' i J
j
1
1
i
1
Chemical Release Incidents Reporting h
Detection Method =164 1
:
1 ' 1 1 1 1 1 1
/o 10% 20% 30% 40% 50% 60% 70
Percent of Incidents Reported
Figure 9-9 Comparison of Methods of Initial Leak Detection for
Total and Chemical Release Incidents
9-14
-------
I
1
rt
'-a
SSSS&SjS&lfcSB'iy
$8$%&$:;ฎ3ฃi8Sf
nAirfiiiiin^^
I.x>fe
--.:.\...:-.-..-:-
1
ฑH
]
1
:&*ฎ8&SS^S:;:S:;:ฃS:S^^
: r
:^:9ft::5:::::W:::::ฅ::W:W:^*^::>::::5ฅ:ฅ:ฅr ."
^SWS:;:SS?^SW:%y ix >" ^V 1
ui&^'Wtti&W1:'1*^ >
S5ซSftSM:W5xSas.
fcgassss
f
f.SSSS-SS?
a^SSi?:^?
iiซa
SSSS>S!W:K:ฅ1
P
Ms:ซ:-.^Sy
:
"งงS-?ฅl
Total 1
Chem
|
:
1
13
ncidents = 12,444 jj
cal Incidents = 248 |
1 1 . 1 1 1 1 1 1 1 1
% 5% 10% 15% 20% 25% 3C
u
I
)% 35
Percent of Incidents Reported
Figure 9-10 Comparison of Corrective/Remedial Measures for Total and
Chemical Release Incidents
9-15
-------
Table 9-1 summarizes the documented Impacts associated with chemical
releases. Most documented impacts indicate only that a release occurred
to a given medium, but do not specify what the resulting impact was.
Those incidents where the impact was specified include contamination of
seven potable ground-water supplies and three potable surface water
supplies, seven incidents of damage to aquatic life, and two incidents
involving damage to plant life. At least eight incidents of combustible
chemical fumes in confined areas were documented, along with one case
involving fire and explosion.
9.8 Summary of Comparison between Chemical and Total Results
This section of the report compares parallel results for the
regulated chemical subset of the documented release incidents to all the
release incidents in the data base. The parameters compared were:
Number of releases by State;
Number of releases by year;
Facility type;
Material stored;
Tank system age information;
Release information; and
Corrective action/impacts.
The most notable difference appeared to occur in the facility type,
where the majority of chemical releases occurred at manufacturing
facilities, while the majority of all reported releases occurred at
gasoline stations. Also, the most commonly reported cause of failure for
9-16
-------
1257s
Table 9-1 Types of Impacts for Documented Chemical Incidents by Media
Number of incidents (percent)
Health/environmental impact
Ground
water
Soil
Surface
water
Air
Other*'
Total
impacts"
ID
I
Media - impact "unspecified"
Contamination of private water supply
Contamination of municipal water supply
Contamination of surface waters
Contamination of other ground waters
Human illness
Human death
Damage to aquatic life
Damage to wildlife
Damage to plant life/crop loss
Damage from corrosion
Damage from fire/explosion
Contamination of private potable surface water
81 (74)
6 (6)
1 (1)
2 (2)
12 (11)
171 (98) 49 (67) 29 (100) 8 (50)
1 ซD 5 (7)
2 (1) 1 (1)
1 (6)
338 (84)
6 (1)
1 ซD
2 ซ1)
12 (3)
6 (1)
3 (1)
1 ซD
Contamination of
Contamination of
Contamination of
Contamination of
Combustible fumes
Other
Total impacts*
municipal potable surface water
industrial surface water
non-contact recreational surface water supply
contact recreational surface water supply
in confined area
1
2
6
3
7 (6) 6
109 174 73
(1)
(3)
(8)
(4)
(8)
29
1
2
6
3
7 (44) 20
16 401
(
-------
chemical releases was corrosion, while that for all releases was
structural failure. Another difference is that soil excavation was
reported in 25 percent of the chemical releases and in only about
12 percent of all releases.
On the other hand, the mean and median tank system ages at time of
report were 17 years for both chemical and all documented incidents that
reported age. Most other parameters showed limited variability between
the two data sets, which could be due in part to the relatively small
number of documented chemical release incidents compared to all
documented incidents.
9.9 Additional Chemical Release Data
In an attempt to obtain additional information on chemical releases
from underground storage tanks, researchers for this report briefly
reviewed Superfund information compiled to date. A total of 64 Superfund
sites were identified as involving underground storage tanks. Of these
64 sites, 5 had sufficient investigation to have a Record of Decision at
this time. These sites are:
Aidex Site, Iowa;
Sinclair Refinery, New York;
Berlin and Farro Site, Mississippi;
Douglasvi1le, Pennsylvania; and
McAdoo Associates, Pennsylvania.
An initial review of the Records of Decision indicated that the sites
involved pollution from multiple sources in addition to underground
tanks. Future efforts may involve a more detailed analysis of these or
other reports to augment information on releases from chemical tanks.
9-18
-------
APPENDIX A
Number and Percent of Release
Incidents with Documentation in Each Field for Which
Information Was sought
-------
type e:sumout.doc
Page: 1
Date: 05/30/86
No. of Incidents: 12444
-0-
Number and percent of incidents
Field
13
14
15a
15b
15c
15d
15e
16
17 a
17b
17 c
18
19
20
21
22
23
24
25
26
27
28
29
30a
3 Ob
30c
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Description
Status
Inactive Date YY-MM-DD
Other Party Involvement - a
Other Party Involvement - b
Other Party Involvement - c
Other Party Involvement - d
Other Party Involvement - e
Material Stored
Prior Leak Monitoring - 1
Prior Leak Monitoring - 2
Prior Leak Monitoring - 3
Manifolded?
Metering?
Vapor Recovery?
Temperature Controls?
Minimum Temperature Range, F
Maximum Temperature Range, F
Pumping?
Type of Pumping System
Pressure Control?
Minimum Pressure, psi
Maximum Pressure, psi
Containment?
Containment Type (1)
Containment Type (2)
Containment Type (3)
Flow Thru Rate, gpd
Percent Volume Below Grade
Is Sytem Ever In Groundwater
How Often In Groundwater
Tank Age at Leak - years
Tank Volume - gallons
Material of Construction, Tank
Number of Walls, Tank
Exterior Tank Coating
Interior Tank Lining
Cathodic Protection, Tank
Recycled Tank?
Age When Tank Repaired, years
Age When Tank Replaced, years
Age When Tank Abandoned, years
Tank Backfill Material
Tank Backfill Thickness
Age at Pipe Leak
No. of
Incidents
Percent of
Incidents
11030
104
9775
5916
2522
983
322
12214
1349
172
18
477
71
164
7
1
1
1007
658
20
0
2
85
45
4
0
23
11346
512
165
1295
4001
1221
57
81
71
12
87
29
66
164
304
41
122
88.64
0.84
78.55
47.54
20.27
7.90
2.59
98.15
10.84
1.38
0.14
3.83
0.57
1.32
0.06
0.01
0.01
8.09
5.29
0.16
0.00
0.02
0.68
0.36
0.03
0.00
0.18
91.18
4.11
1.33
10.41
32.15
9.81
0.46
0.65
0.57
0.10
0.70
0.23
0.53
1.32
2.44
0.33
0.98
A-2
-------
Page: 2
Date: 05/30/86
No. of Incidents: 12444
-0-
Number and percent of incidents
Field
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63a
63b
63C
63d
63e
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87a
87b
Description
Material of Construction
Number of Pipe Walls
Inside Pipe Diameter - inches
Exterior Pipe Coating
Interior Pipe Lining
Cathodic Protection, pipe
Pipe Length to Tank, feet
Age When Pipe Repaired, years
Age When Pipe Replaced, years
Age When Pipe Abandoned, years
Pipe Backfill Material
Pipe Backfill Thickness
Date of Release YYMMDD
Location of Release
Cause of Release (1)
Cause of Release (2)
Cause of Release (3)
Cause of Release (4)
Cause of Release (5)
How Was Leak First Detected?
First Detected On or Off Premise
Where Was Leak First Detected
How Else Was Leak Detected (1)
Was 2nd Detection On or Off P.
Where Was 2nd Detection
How Else Was Leak Detected (2)
Was 3rd Detection On or Off P.
Where Was 3rd Detection
Party First Detecting Leak
Distance, Point of Det. to Leak
First Confirmation Detection
1st Conf. Detection On or Off P.
Where Was 1st Conf. Detection
How 2nd Confirmation Detection
2nd Conf. Det. On of Off Premise
Where Was 2nd Conf. Detection
Soil pH (rounded to whole #)
Soil Restivity - ohm-cm
Sulfides Present in Soil
Duration of Leak, days
Estimated Quantity Released, gal
Integrity Test, gph
Documented Health/Env. Impact, GW (1)
Documented Health/Env. Impact, GW (2)
No. of
Incidents
192
31
163
49
11
16
16
9
8
9
80
11
12109
11779
9833
2218
253
34
14
7520
7598
7470
1411
1397
1399
180
178
180
5400
306
5641
5619
5580
1297
1291
1281
18
27
8
1375
5989
635
5271
292
Percent of
Incidents
1.54
0.25
1.31
0.39
0.09
0.13
0.13
0.07
0.06
0.07
0.64
0.09
97.31
94.66
79.02
17.82
2.03
0.27
0.11
60.43
61.06
60.03
11.34
11.23
11.24
1.45
1.43
1.45
43.39
2.46
45.33
45.15
44.84
10.42
10.37
10.29
0.14
0.22
0.06
11.05
48.13
5.10
42.36
2.35
A-3
-------
Page: 3
Date: 05/30/86
No. of Incidents: 12444
-0-
Number and percent of incidents
Impact, Soil (2)
Impact, Soil (3)
Field Description
87c Documented Health/Env. Impact, GW (3)
87d Documented Health/Env. Impact, GW (4)
88 Aquifer Use
89 Depth to Aquifer - ft.
90 Horizontal Distance to Aquifer, feet
91 Maximum Distance of Groundwater Impacts
92 Depth of Groundwater Contamination, ft.
93 Area of Groundwater Contamination, acre
94a Documented Health/Env. Impact, Soil (1)
94b Documented Health/Env.
94c Documented Health/Env.
94d Documented Health/Env. Impact, Soil (4)
95 Soil Type
96 Depth to Bedrock, feet
97 Longitudinal Distance of Impacts, feet
98 Depth of Soil Contamination, feet
99 Area of Soil Contamination, acres
lOOa Documented Health/Env. Impact, Wat.
10Ob Documented Health/Env. Impact, Wat.
10Oc Documented Health/Env. Impact, Wat.
lOOd Documented Health/Env. Impact, Wat.
101 Longitudinal Distance of Water Imp.,
102 Depth of Water Contamination, feet
103 Area of Water Contamination, acres
104a Documented Health/Env. Impact, Air
104b Documented Health/Env.
104c Documented Health/Env.
104d Documented Health/Env.
105 Longitudinal Distance of Air Impact, ft
106 Height of Air Contamination, feet
107 Area of Air Contamintaion, acres
108 Type of Other Environmental Impacts
109a Documented Health/Env. Impact, Oth.
109b Documented Health/Env. Impact, Oth.
109c Documented Health/Env. Impact, Oth.
109d Documented Health/Env. Impact, Oth.
HOa Legal Action Since Detection (1)
llOb Legal Action Since Detection
llOc Legal Action Since Detection
llOd Legal Action Since Detection
111 Litigation?
112a Legal/Regulatory Req. Before Leak (1)
112b Legal/Regulatory Req. Before Leak (2)
112c Legal/Regulatory Req. Before Leak (3)
Impact, Air
Impact, Air
(1)
(2)
(3)
(4)
ft
(1)
(2)
(3)
Impact, Air (4)
(2)
(3)
(4)
(1)
(2)
(3)
(4)
No. of
Incidents
34
3
610
964
50
136
130
26
8198
55
7
0
1082
163
264
378
110
2537
165
35
5
201
10
23
1814
20
2
1
13
1
3
3306
1372
77
12
2
4974
1153
501
214
778
198
34
3
Percent of
Incidents
0.27
0.02
4.90
7.75
0.40
1.09
1.04
0.21
65.88
0.44
0.06
0.00
8.69
1.31
2.12
3.04
0.88
20.39
1.33
0.28
0.04
1.62
0.08
0.18
14.58
0.16
0.02
0.01
0.10
0.01
0.02
26.57
11.03
0.62
0.10
0.02
39.97
9.27
4.03
1.72
6.25
1.59
0.27
0.02
A-4
-------
Page: 4
Date: 05/30/86
No. of Incidents: 12444
-0-
Number and percent of incidents
Field
112d
113a
113b
113c
113d
114a
114b
114c
115
116
117a
117b
117C
118
119
120
121
122
123
124
125
Description
Legal/Regulatory Req. Before Leak (4)
Corrective/Remedial Measures (1)
Corrective/Remedial Measures (2)
Corrective/Remedial Measures (3)
Corrective/Remedial Measures (4)
Preliminary Studies for Corr. Act. (1)
Preliminary Studies for Corr. Act. (2)
Preliminary Studies for Corr. Act. (3)
Time Lapse - Leak to Corr. Act., weeks
Duration of Corrective Action, weeks
Party Financing Cleanup (1)
Party Financing Cleanup (2)
Party Financing Cleanup (3)
Capital Cost in Correcting Incident
Annual Operating Cost for Correction
Start of Corr/Rem Action YYMMDD
Status of Cleanup
Owner Cost Percentage
Operator Cost Precentage
Federal Cost Percentage
State Cost Percentage
No. of
Incidents
Percent of
Incidents
3
9557
5301
2430
974
1179
88
25
3318
1254
673
30
1
198
54
4582
5052
408
16
11
21
0.02
76.80
42.60
19.53
7.83
9.47
0.71
0.20
26.66
10.08
5.41
0.24
0.01
1.59
0.43
36.82
40.60
3.28
0.13
0.09
0.17
E:\ >
A-5
-------
APPENDIX B
State Data Sources
-------
Table B-l state Data Sources
State
Agency
Alabama Department of Environmental Management, Water Division
Alaska Department of Environmental Conservation, Oil Spill Division
Arizona Department of Health Services
Arkansas Department of Pollution Control and Ecology
California Water Resources Control Board and Orange County Department of Environmental Health
Colorado Department of Labor and Employment
Connecticut Department of Environmental Protection, Oil and Chemical Spill Section
Delaware Department of Natural Resources and Environmental Control
Florida Department of Environmental Resources, Spill Response Division
Georgia Department of Natural Resources, Remedial Action Units
Guam Guam Environmental Protection Agency
Hawaii Department of Health
Idaho Department of Health and Welfare, Division of Environment
Iowa Department of Water, Air, and Waste Management
Illinois State Environmental Protection Agency, Emergency Response Unit
Indiana State Fire Marshall
Kansas Department of Health and Environment
Kentucky State Fire Marshall, Hazardous Materials Section
Louisiana State Department of Environmental Quality and New Orleans Fire Department
Maine Department of Environmental Protection, Oil and Hazardous Materials Bureau
Maryland Department of Natural Resources, Oil Spill Program
B-2
-------
Table B-l (continued)
State
Agency
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
New York
New Hampshire
New Jersey
New Mexico
Nevada
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
Department of Environmental Quality Engineering
Department of Natural Resources
Pollution Control Agency
Department of Natural Resources. Bureau of Pollution Control, Ground-water Quality Branch
Department of Natural Resources, Division of Environmental Quality
Department of Health and Environmental Sciences, Bureau of Water Quality
Department of Environmental Control, Water Management Division
Department of Environmental Conservation and Department of Transportation
water supply and Pollution Control Commission
Department of Environmental Protection
Albuquerque Environmental Health and Energy department
Department of Conservation and Natural Resources
Department of Natural and Economic Resources, Spill Response Division
State Department of Health
Environmental Protection Agency. Emergency Response Branch
Water Resources Board
Department of Environmental Quality
Department of Environmental Resources
Environmental Quality Board
Department of Environmental Management
B-3
-------
Table B-l (continued)
State
Agency
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virgin Islands
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Department of Health and Environmental Control. Solid and Hazardous Waste Division
Department of Water and Natural Resources
Department of Health and Environment, Water Pollution Control Division
Department of Water Resources. Texas Water Conmission, State Fire Marshall
Department of Health, Bureau of Water Pollution Control
Agency of Environmental Conservation. Office of Air, Solid, and Hazardous Waste
Department of Conservation and Cultural Affairs
State Water Control Board
Department of Ecology
Department of Natural Resources
Department of Natural Resources
Department of Environmental Quality
B-4
------- |