£EPA
EPA/601 /R-14/001 I May 2015 I www.epa.gov/hfstudy
United States
Environmental Protection
Agency
Review of State and Industry Spill Data:
Characterization of Hydrauiic Fracturing-Related Spills
United States Environmental Protection Agency
Office of Research and Development
-------
Characterization of Hydraulic Fracturing-Related Spills May 2015
[This page intentionally left blank.]
ii
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Review of State and Industry Spill Data:
Characterization of Hydraulic Fracturing-Related Spills
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC
May 2015
EPA/601/R-14/001
iii
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency policy
and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
Preferred Citation: U.S. Environmental Protection Agency. 2015. Review of State and Industry Spill Data:
Characterization of Hydraulic Fracturing-Related Spills. Office of Research and Development,
Washington, DC. EPA/601/R-14/001.
iv
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table of Contents
Disclaimer iv
Table of Contents v
List of Tables vii
List of Figures viii
Preface ix
Authors and Contributors x
Acknowledgements x
Executive Summary 1
1. Objective 3
2. Introduction 3
3. Methods 5
3.1. Data Sources 5
3.2. Search Methods 7
3.3. Data Compilation and Analysis 8
3.4. Quality Assurance and Quality Control 9
4. Results: Spill Characterization 9
4.1. Volumes Spilled 11
4.2. Spilled Materials 13
4.3. Spill Sources 15
4.4. Spill Causes 16
4.5. Environmental Receptors 19
5. Results: Containment and Response 21
5.1. Containment 21
5.2. Response 23
6. Discussion 24
6.1. Study Limitations 26
7. Conclusions 27
References 28
Appendix A: Standardization of Spill Characteristics 32
A.l. Volumes 32
A.2. Spilled Materials 33
A.3. Spill Sources 33
A.4. Spill Causes 33
A.5. Environmental Receptors 33
Appendix B: Hydraulic Fracturing-Related Spills Table 34
v
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Glossary 35
Glossary References 36
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
List of Tables
Table 1. Data elements available in nine state data sources 6
Table 2. Number and percentage of hydraulic fracturing-related spills identified from state and
industry data sources used in this study by state 10
Table 3. Definitions and examples of spilled materials 14
Table 4. Number of hydraulic fracturing related-spills, total reported volume spilled, and reported
volume per spill by material type 14
Table 5. Definitions and examples of spill sources 16
Table 6. Number of hydraulic fracturing-related spills, total reported volume spilled, and reported
volume per spill by source type 16
Table 7. Definitions and examples of spill causes 17
Table 8. Number of hydraulic fracturing-related spills, total reported volume spilled, and reported
volume per spill by cause type 18
Table 9. Number of hydraulic fracturing-related spills, total reported volume spilled, and reported
volume per spill by environmental receptor 20
Table 10. Examples of breaches of berms and dikes, as described in records from hydraulic
fracturing-related spills 22
Table 11. Examples of immediate responses used to stop spills or contain spilled fluids, as described
in records from hydraulic fracturing-related spills 24
Table Al. Categorization method for reported volumes of spilled materials 32
Table A2. Categorization method for environmental receptors 33
vii
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
List of Figures
Figure 1. Percent of reviewed spills determined to be related or unrelated to hydraulic fracturing 10
Figure 2. Availability of reported information in records from hydraulic fracturing-related spills for
each spill characteristic assessed 11
Figure 3. Distribution of reported hydraulic fracturing-related spills by reported volume spilled 12
Figure 4. Percent distribution of fluid fate by total reported volume for hydraulic fracturing-related
spills 13
Figure 5. Summary of materials spilled by (a) number of hydraulic fracturing-related spills and
(b) total reported volume spilled 15
Figure 6. Summary of spill sources by (a) number of hydraulic fracturing-related spills and
(b) total reported volume spilled 17
Figure 7. Summary of spill causes by (a) number of hydraulic fracturing-related spills and
(b) total reported volume spilled 18
Figure 8. Number of hydraulic fracturing-related spills in which spilled fluids reached or did not
reach surface water, ground water, or soil 20
viii
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Preface
The U.S. Environmental Protection Agency (EPA) is conducting a study of the potential impacts of
hydraulic fracturing for oil and gas on drinking water resources. This study was initiated in Fiscal Year
2010 when Congress urged the EPA to examine the relationship between hydraulic fracturing and
drinking water resources in the United States. In response, the EPA developed a research plan (Plan to
Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources) that was reviewed by
the agency's Science Advisory Board (SAB) and issued in 2011. A progress report on the study (Study of
the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources: Progress Report), detailing
the EPA's research approaches and next steps, was released in late 2012 and was followed by a
consultation with individual experts convened under the auspices of the SAB.
The EPA's study includes the development of several research projects, extensive review of the
literature and technical input from state, industry, and non-governmental organizations as well as the
public and other stakeholders. A series of technical roundtables and in-depth technical workshops were
held to help address specific research questions and to inform the work of the study. The study is
designed to address research questions posed for each stage of the hydraulic fracturing water cycle:
• Water Acquisition: What are the possible impacts of large volume water withdrawals from
ground and surface waters on drinking water resources?
• Chemical Mixing: What are the possible impacts of surface spills of hydraulic fracturing fluids on
or near well pads on drinking water resources?
• Well Injection: What are the possible impacts of the injection and fracturing process on drinking
water resources?
• Flowback and Produced Water: What are the possible impacts of surface spills of flowback and
produced water on or near well pads on drinking water resources?
• Wastewater Treatment and Waste Disposal: What are the possible impacts of inadequate
treatment of hydraulic fracturing wastewaters on drinking water resources?
This report, Review of State and Industry Spill Data: Characterization of Hydraulic Fracturing-Related
Spills, is the product of one of the research projects conducted as part of the EPA's study. It has
undergone independent, external peer review in accordance with agency policy and all of the peer
review comments received were considered in the report's development.
The EPA's study will contribute to the understanding of the potential impacts of hydraulic fracturing
activities for oil and gas on drinking water resources and the factors that may influence those impacts.
The study will help facilitate and inform dialogue among interested stakeholders, including Congress,
other Federal agencies, states, tribal government, the international community, industry, non-
governmental organizations, academia, and the general public.
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Authors and Contributors
Susan Burden, PhD, US EPA
Maryam A. Cluff, MS, Student Services Contractor for the US EPA, under contract EP-13-H-000438
Leigh E. DeHaven, MS, US EPA
Cindy Roberts, MS, US EPA
Susan L. Sharkey, MA, US EPA
Alison Singer, MA, Student Services Contractor for the US EPA, under contract EP-13-H-000474
Acknowledgements
The EPA would like to acknowledge all of the organizations and individuals that provided data and
information used for this study, including Kendra Zamzow (AAAS Policy Fellow). The agency would also
like to acknowledge Claudia Meza-Cuadra (Student Services Contractor, under contract EP-13-
H-000054), Matt Richards (US EPA), and Stephen Watkins (US EPA) for their efforts in the
development of this report. An independent, external peer review of this report was conducted
through the Eastern Research Group, Inc., under contract EP-C-12-021. The contractors' role did not
include establishing agency policy.
x
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Executive Summary
Advances in hydraulic fracturing and horizontal drilling technologies have led to increased oil and
gas exploration and production activity in different regions of the United States. Hydraulic
fracturing is a technique used to enable or enhance the production of hydrocarbons from
underground rock formations. It involves the injection of hydraulic fracturing fluids (typically a
mixture of water, proppant, and chemical additives) under pressures great enough to fracture the
targeted hydrocarbon-bearing formations. The volumes and chemical compositions of hydraulic
fracturing fluids and flowback fluids (i.e., fluids that return to the surface after hydraulic fracturing)
managed on oil and gas production well pads have led to concerns about potential human health
and environmental impacts from surface spills of these fluids. The objective of this study was to
characterize hydraulic fracturing-related spills that may reach surface or ground water resources
using spill reports obtained from selected state and industry data sources.
Data gathered from selected state and industry data sources were used to characterize hydraulic
fracturing-related spills with respect to volumes and materials spilled, sources and causes of spills,
environmental receptors, and spill containment and response activities. For the purposes of the
study, hydraulic fracturing-related spills were defined as those occurring on or near the well pad
before or during the mixing and injection of hydraulic fracturing fluids or during the post-injection
recovery of fluids. Because the main focus of this study was to characterize hydraulic fracturing-
related spills on the well pad that may reach surface or ground water resources, the following
topics were not included: transportation-related spills, drilling mud spills, and spills associated
with disposal through underground injection control wells.
Data on spills that occurred between January 2006 and April 2012 were obtained from nine states
with online spill databases or other data sources, nine hydraulic fracturing service companies, and
nine oil and gas production well operators. The data sources used in this study contained over
36,000 spills. Spill records from an estimated 12,000 spills (33 percent of the total number of spills
reviewed) contained insufficient information to determine whether the spill was related to
hydraulic fracturing. Of the spills with sufficient information, the EPA identified an estimated
24,000 spills (66 percent) as not related to hydraulic fracturing and 457 spills (approximately 1
percent) as related to hydraulic fracturing. The 457 hydraulic fracturing-related spills occurred in
11 different states over the period of time studied.
For the 457 hydraulic fracturing-related spills included in the study, the most commonly reported
information obtained from state and industry data sources was the type of material spilled
(reported in 97 percent of the hydraulic fracturing-related spills), followed by the volume spilled
and then the source and cause of the spill. In approximately 90 percent of the hydraulic fracturing-
related spills, information was available on whether or not spilled fluids reached at least one
environmental receptor (surface water, ground water, and/or soil). The EPA did not determine
whether spilled fluids affected the quality of surface or ground water resources.
The hydraulic fracturing-related spills were characterized by numerous low volume spills (up to
1,000 gallons) and relatively few high volume spills (greater than 20,000 gallons). The most
common material spilled was flowback and produced water, and the most common source of spills
1
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
was storage units. More spills were caused by human error than any other cause. There were 300
spills (66 percent of the 457 spills included in this study) in which spilled fluids reached at least one
environmental receptor. Twenty-four of these spills reached multiple environmental receptors. Soil
was the most commonly reported environmental receptor, with spilled fluids reaching soil in over
half (64 percent) of all hydraulic fracturing-related spills. Spilled fluids were reported to have
reached surface water in 32 hydraulic fracturing-related spills (7 percent); the median volume per
spill for these spills was 3,500 gallons, and volumes per spill ranged from 90 gallons (5th percentile)
to 45,000 gallons (95th percentile). There was one spill in which spilled fluids were reported to have
reached ground water (0.2 percent). Spilled fluids were reported as not reaching surface or ground
water in 186 spills (41 percent).
The spills characterized in this study were likely a subset of the total number of hydraulic
fracturing-related spills that occurred in the United States between January 2006 and April 2012.
Although spill data were obtained from nine states that are among the top oil and gas producing
states in the country, similar data from other oil and gas producing states were not included. The
state data sources used in this study may not have included all spills related to hydraulic fracturing
because some spills may not have met the spill reporting requirements that were in place at the
time of the spill. Additionally, some reported spills may not have been identified as related to
hydraulic fracturing due to insufficient information in the data sources. The quantitative
characterization of hydraulic fracturing-related spills presented in this report (e.g., the percentages
in the paragraph above) may have been different if more hydraulic fracturing-related spills could
have been identified from the data sources used in this study.
This report presents the results of a broad review of state and industry spill data from 457
hydraulic fracturing-related spills. Data from these spills were used to characterize volumes and
materials spilled, spill sources and causes, and environmental receptors. There were several key
findings. Spills related to hydraulic fracturing were most often characterized by numerous, low
volume events (up to 1,000 gallons) and relatively few high volume events (greater than 20,000
gallons). The most common material spilled was flowback and produced water, and the most
common source of spills was storage units. More spills were caused by human error than any other
cause. Over half of the spills associated with hydraulic fracturing reached an environmental
receptor, with 33 instances of spilled fluids reaching surface or ground water resources. These
results, as well as other information on spill characteristics and containment and response
activities, provide important insights into the nature of hydraulic fracturing-related spills in several
key states with hydraulic fracturing.
2
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
1. Objective
The objective of this study was to characterize hydraulic fracturing-related spills that may reach
surface or ground water resources using spill reports obtained from selected state and industry
data sources. For the purposes of the study, hydraulic fracturing-related spills were defined as
those occurring on or near the well pad before or during the mixing and injection of hydraulic
fracturing fluids or during the post-injection recovery of fluids. This study did not determine if or
how spilled fluids may have affected surface or ground water quality, nor did it evaluate spill
reporting requirements.
The analysis described in this report was conducted in support of the U.S. Environmental Protection
Agency's (EPA) Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources (US EPA, 2011a). The Study Plan identified the importance of understanding the possible
impacts on drinking water resources from surface spills of hydraulic fracturing fluids or flowback
and produced water on or near well pads. The analysis presented in this report provides
information on volumes and materials spilled, sources and causes of spills, environmental
receptors, and spill containment and response activities for hydraulic fracturing-related spills.
2. Introduction
Hydraulic fracturing is a technique used to enable or enhance the production of hydrocarbons from
underground rock formations. It involves the injection of hydraulic fracturing fluids (typically a
mixture of water, proppant, and chemical additives) under pressures great enough to fracture the
targeted hydrocarbon-bearing formations (Gregory etal., 2011; Vidic etal., 2013). After the
injection pressure is released, fluids flow through the fractures back out of the well, leaving behind
proppants (often fine-grained sand) that hold open the newly-created fractures. The fractures allow
oil and gas to flow from pores within the formation to the production well [Ground Water
Protection Council (GWPC) and ALL Consulting, 2009; Soeder, 2010],
On-site fluid management is a typical practice associated with hydraulic fracturing. Hydraulic
fracturing base fluids, most commonly water, are typically stored in large volume tanks on the well
pad. Chemicals additives can be stored on a flatbed truck or van enclosure that holds a number of
chemical totes. The most common chemical totes are 200 to 400 gallon polyethylene containers
(New York State Department of Environmental Conservation, 2011). Pumps and hoses are used to
move the base fluid and chemical additives to a blender that mixes the fluids. The fluid is then
transferred to a manifold for delivery to the wellhead for injection (Malone and Ely, 2007). As fluids
are transferred and moved around the well pad and through various pieces of equipment, faulty
equipment or human error may create opportunities for spills of the various components of
fracturing fluid (Colorado Department of Natural Resources, 2014).
The type and amount of fluids stored on-site is largely determined by the characteristics of the
formation being fractured, as well as by economics, production goals, and availability of chemical
additives. Estimates of water needs per well have been reported to range from 50,000 gallons for
coalbed methane production to 13 million gallons for shale gas production (US EPA, 2004; Vengosh
et al., 2014). Approximately 1 to 2 percent or less of the volume of water-based hydraulic fracturing
3
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
fluid is composed of chemical additives (GWPC and ALL Consulting, 2009; Lee etal., 2011; US EPA,
2015), which suggests that approximately 500 to 260,000 gallons or less of chemical additives may
be brought on-site for hydraulic fracturing. Chemical additives can be composed of one or more
chemicals and can be used in hydraulic fracturing fluids as acids, friction reducers, surfactants,
scale inhibitors, iron control agents, corrosion inhibitors, and biocides (Arthur et al., 2009; Gregory
etal., 2011; GWPC and ALL Consulting, 2009; US EPA, 2015).
When the pressure applied during hydraulic fracturing is released, fluid flows back from the well.
The initial fluid that returns to the surface is often called "flowback." Fluid that flows from the well
along with oil and gas during the production phase is often referred to as "produced water."
Flowback and produced water are stored at the well pad before disposal or reuse. Typical storage
facilities include closed containers and open air impoundments (GWPC, 2009). Leaks and spills of
flowback and produced water may occur on the well pad due to human error, equipment failure, or
well blowouts (Colorado Department of Natural Resources, 2014).
The volume and chemical composition of fluids that return to the surface after the rock is fractured
can vary widely. Between 10 and 70 percent of injected fluid comes back up the well as flowback
(GWPC and ALL Consulting, 2009; US EPA, 2011b). General compositions and on-site volumes of
flowback fluids vary among targeted formation types (e.g., shale versus sandstone) (Alley et al.,
2011) and within formations ofthe same type (e.g., southwest versus northwest Marcellus Shale)
(Barbot et al., 2013). Flowback fluids are typically, but not always, characterized as highly saline
(Blauch etal., 2009; Neff etal., 2011) and often contain major anions and cations, metals, and
naturally occurring radionuclides (Chapman etal., 2012; Rowan etal., 2011). Flowback fluids may
also contain organic chemicals from injected fluids, formation waters, and formation solids (Orem
etal., 2007; Sirivedhin and Dallbauman, 2004; Strong etal., 2014).
Concerns have been raised about potential human health and environmental impacts associated
with surface spills of fluids managed on oil and gas production well pads (Rozell and Reaven, 2012;
Stringfellow et al., 2014; Vengosh et al., 2014). In particular, spilled fluids associated with hydraulic
fracturing may flow into nearby surface waters or infiltrate into ground water and alter water
quality (Olmstead etal., 2013; Stringfellow et al., 2014; Vengosh etal., 2014). For example,
Papoulias and Velasco (2013) demonstrated that hydraulic fracturing fluid spilled into surface
water likely contributed to the distress and death of Blackside Dace fish in Kentucky by lowering
the pH and increasing the conductivity of the stream. Using data from post-spill sampling reports in
Colorado, Gross etal. (2013) identified concentrations of benzene, toluene, ethylbenzene, and
xylene in ground water samples collected in response to operator-reported spills. Gross and
colleagues attributed the presence of these compounds in ground water to numerous hydraulic
fracturing-related spills.
To better understand spills associated with hydraulic fracturing, this study used data from selected
state and industry data sources to characterize hydraulic fracturing-relakted spills. Information on
hydraulic fracturing-related spills was analyzed with respect to volumes and materials spilled,
sources and causes of spills, environmental receptors, and containment and response activities.
4
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
3. Methods
3.1. Data Sources
Data used in this study were obtained from both state and industry data sources.1 States often
maintain spill databases or spill records that are designed to track inspection results, complaints,
and/or violations. These spills can be associated with a wide variety of activities, including, but not
limited to, hydraulic fracturing. Incidents found in state data sources are typically self-reported or
can be identified through citizen complaints and routine inspections by state officials.
States were selected based on the number of oil and gas wells that were reported by nine oil and
gas service companies to have been hydraulically fractured between approximately September
2009 and September 2010 (US EPA, 2012).2 The EPA used the service company information to
identify the ten states with the most hydraulic fracturing activity reported during that time period:
Arkansas, Colorado, Louisiana, New Mexico, North Dakota, Oklahoma, Pennsylvania, Texas, Utah,
and Wyoming. State spill data sources were identified for all states except North Dakota.3 Online,
publicly accessible spill databases were identified for Arkansas, Colorado, New Mexico, and
Pennsylvania. Offline, publicly accessible spill data were obtained from Louisiana, Oklahoma, Texas,
Utah, and Wyoming. The state spill data sources varied in accessibility, searchability, and the types
of information available. Table 1 summarizes the data elements (e.g., volume spilled, spill cause)
available for each state data source.
Industry data were obtained from responses to information requests that were sent to nine
hydraulic fracturing service companies4 and nine oil and gas well operators5 in September 2010
and August 2011, respectively (US EPA, 2010, 2011c).6 Although the service companies were not
specifically asked for information on spills, some companies provided information relevant for this
report The EPA asked the well operators for spill reports associated with 350 well identifiers
corresponding to wells that were reported to have been hydraulically fractured between
approximately September 2009 and September 2010. Some of the industry data were submitted as
confidential business information under the Toxic Substances Control Act. The EPA worked with
1 Information on spills can also be obtained from the U.S. Coast Guard's National Response Center. The National Response
Center records information on oil spills, chemical releases, and maritime security incidents. Data sources maintained by
states were considered more appropriate for this study.
2 The nine hydraulic fracturing service companies included: BJ Services Company; Complete Production Services, Inc.;
Halliburton Energy Services, Inc.; Key Energy Services; Patterson-UTI Energy; RPC, Inc.; Schlumberger Technology
Corporation; Superior Well Services; and Weatherford International. The companies reflected a range of company sizes
and geographic diversity (US EPA, 2011a, 2012].
3 No central database from North Dakota was available at the time of this research. The North Dakota Department of
Health now records and makes reports available at www.ndhealth.gov/EHS/Spills/.
4 See footnote 2.
5 The nine oil and gas operators included: Clayton Williams Energy, Inc.; ConocoPhillips; EQT Corporation; Hogback
Exploration, Inc.; Laramie Energy II, LLC; MDS Energy, Ltd.; Noble Energy, Inc.; SandRidge Exploration and Production,
LLC; and Williams Production Company, LLC. The operators had wells in diverse geographic areas and were chosen to
reflect a range of company sizes (US EPA, 2012].
6 The Paperwork Reduction Act of 1995 limits the number of information requests to nine entities per set of queries,
unless pre-approved by the Office of Management and Budget.
5
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table 1. Data elements available in nine state data sources. Data elements varied by state due to differences in state reporting mechanisms.
State (Reference)
Data Element
Arkansas
Colorado
Louisiana
New Mexico
Oklahoma
Pennsylvania
Texas
Utah
Wyoming
(ADEQ,
(COGCC,
(LDEQ,
(NMEMNRD,
(OCC,
(PADEP,
(TRC,
(UDEQ,
(WOGCC,
2012)
2012)
2013)
2012)
2013)
2012)
2013)
2013)
2012)
Incident/report
number
X
X
X
X
X
X
X
X
Incident date
X
X
X
X
X
X
Report received/
inspection date
X
X
X
X
X
X
X
API number*
X
X
County
X
X
X
X
X
X
X
X
X
State
X
X
X
X
X
X
X
X
X
Incident description
X
X
X
X
X
X
X
X
X
Volume spilled
X
X
X
X
X
X
X
X
X
Volume recovered
X
X
X
Material spilled
X
X
X
X
X
X
X
X
X
Spill cause
X
X
X
X
X
X
X
X
Media impacted
X
Surface water nexus
X
X
X
X
X
X
X
Ground water impact
X
X
Containment
X
Action/remedy taken
X
X
X
X
X
* American Petroleum Institute (API) numbers are unique 10-digit numbers that are generally assigned to wells by state oil and gas agencies.
6
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
the service companies and well operators to summarize and present the data used in this report in
a way that protects their claims of confidentiality.
3.2. Search Methods
Data sources were initially searched to identify spills that occurred between January 1, 2006, and
April 30, 2012. This timeframe encompassed a period of rapid increase in hydraulic fracturing
across the United States (US Government Accountability Office, 2014).
Each data source was searched separately because of differences among data sources. Data sources
were searched using a combination of available filters (e.g., material spilled), keywords, and line-
by-line reviews, depending on the searchability of each data source. Keywords included, but were
not limited to: "hydraulic fracturing," "frac," "flowback," "guar gum," "glycol," "quartz,"
"hydrochloric acid," and some names of companies known to conduct hydraulic fracturing
activities. The list of keywords was revised as each data source was searched. Keyword searches
also depended on the available search methods because some data sources used fixed terms for
certain fields. For example, the New Mexico Oil and Conservation Division Spills Database had fixed
terms for the "spill material" field that included "gelled brine (frac fluid)" and "produced water," but
not "flowback."
Incident descriptions of all spills within the study timeframe and identified through filters, keyword
searches, and line-by-line reviews were reviewed to determine whether the spill was related to
hydraulic fracturing and occurred on or near the well pad.7 Spills were identified as related to
hydraulic fracturing if the incident description indicated that the spill occurred immediately before
or during mixing and injection of hydraulic fracturing fluids, or during flowback. Hydraulic
fracturing-related spills were generally identified through the use of "hydraulic fracturing,"
'Tracking," or "flowback" in incident descriptions. Spills were also identified as related to hydraulic
fracturing if chemical additives identified in the reports were specific to hydraulic fracturing
activities (e.g., crosslinkers or gelling agents).
After identifying hydraulic fracturing-related spills within the study's scope, each state, service
company, and well operator was given a list of spills compiled from its own data. Each data owner
reviewed the data and provided further information where possible, including identifying spills
that had been incorrectly designated as being related to hydraulic fracturing. These spills were
removed and not included in any analyses.
Search Limitations. Searching individual state data sources by keywords has limitations. While
common spelling variations were applied for various keywords (e.g., "frac," "frack," and "frak"),
incidents could have been missed if a word was incorrectly spelled within the data source (e.g.,
"glycal" instead of "glycol"). Additionally, a more expansive list of keywords may have identified
additional spills that could have been related to hydraulic fracturing.
7 Because the main focus of this study was to characterize hydraulic fracturing-related spills on the well pad that may
reach surface or ground water resources, the following topics were not included: transportation-related spills, drilling
mud spills, and spills associated with disposal through underground injection control wells.
7
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
3.3. Data Compilation and Analysis
Hydraulic fracturing-related spill data obtained from the state and industry data sources were
compiled into a single table that contained information on both the spill (e.g., spill cause, spilled
material, and volume) and spill containment and response. Cells in the table were populated using
the original data from state and industry sources. Data were generally pulled from the relevant field
in the data source records. However, differences in state reporting mechanisms left data gaps for
some information used in this report, including volumes of spilled and recovered fluids, spill causes,
and environmental receptors. These data gaps were filled with information from the incident
descriptions where possible. In some cases, the same spill was entered multiple times in a single
data source or was found in both industry and state data sources. In these instances, information
from each entry was combined into a single entry, and the duplicate entries were removed.8 When a
spill was reported in both a state and an industry data source, the EPA did not find discrepancies
between the industry-provided data and the state data source; rather, the industry-provided data
supplemented information from the state data source.
Each data source characterized spills in a different way and provided slightly different information
(Table 1). Therefore, data related to hydraulic fracturing-related spills were standardized prior to
analysis. Section 4 and Appendix A provide information about the standardization of volumes,
materials spilled, sources, causes, and environmental receptors (i.e., the environmental media
reached by a spill, such as surface water, ground water, and/or soil). Appendix B contains the
standardized hydraulic fracturing-related spill table.
Analyses were performed on the data included in Appendix B to characterize hydraulic fracturing-
related spills according to volumes and materials spilled, sources and causes of spills, and
environmental receptors (Section 4). The number of spills associated with a given characteristic, as
well as total and per spill reported volumes, were determined. Median, 5th percentile, and 95th
percentile per spill volumes were calculated from information provided in 370 spill records (81
percent of the hydraulic fracturing-related spills); the remaining spill records did not contain
volume information. Calculated volumes were rounded to one significant figure if less than 100
gallons, or to two significant figures if greater than or equal to 100 gallons. Volumes reported in the
spill records included discrete values (e.g., 100 gallons), ranges (e.g., 50 to 100 gallons), and upper
or lower bounds (e.g., less 100 gallons).9 The reported or calculated volumes in this report should
be considered to be estimates. Some spill volumes, for example, were extrapolated from a storage
container with a known volume, but the precise amount of material that spilled out of the storage
container was not known. While spills from storage containers were more likely to be quantified,
spills from other sources (e.g., lines or hoses) were more uncertain because the volume associated
with those sources was likely to be unknown.
Information related to spill containment and response was reviewed and summarized to provide
additional context on fluid and spill management (Section 5).
8 Spills identified in both state and industry data sources are listed by the state data source in Appendix B.
9 Appendix A describes how each of these entries were treated during the development of the table included in
Appendix B.
8
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
3.4. Quality Assurance and Quality Control
The EPA does not make any claims as to the quality or accuracy of the data gathered from the state
and industry data sources used in this study. Quality assurance and quality control measures were
used to ensure that the analyses performed were properly conducted and that the data used in this
report accurately represent the original data obtained from state and industry data sources.
Data from each state source were evaluated by a single reviewer to identify spills potentially
related to hydraulic fracturing. Those spills categorized as "related to hydraulic fracturing" were
then reviewed by a small group of additional reviewers. For each spill, the group of reviewers
reached a consensus about whether the spill could be identified as related to hydraulic fracturing
based on the information provided in the spill reports. As noted in Section 3.2, all data owners were
offered an opportunity to review their data and provide updated or additional information,
including identifying spills incorrectly designated as related to hydraulic fracturing. Approximately
six percent of the spills were identified by the data owners as unrelated to hydraulic fracturing and
were subsequently removed from Appendix B.
Each field of Appendix B was standardized by a single reviewer. A second individual conducted a
complete review of the standardization. Any differences were resolved by both reviewers looking at
the original data. Lastly, the final, standardized data were compared with the original data and any
discrepancies were addressed. States and companies were also provided a draft of the
categorization methods for comment; their comments were considered during the development of
the table used for analysis (Appendix B).
Additional quality assurance information on this project can be found in the Quality Assurance
Project Plan for Hydraulic Fracturing Surface Spills Data Analysis; which was approved on August 2,
2012, and was later updated on September 9, 2013 (US EPA, 2013). The project underwent a
technical systems audit by the designated EPA Quality Assurance Manager on August 27, 2012, and
no corrective actions were identified.
4. Results: Spill Characterization
The EPA reviewed an estimated 36,000 spills that were reported in the state and industry data
sources during the study's timeframe. Roughly 66 percent of the spills were determined to be not
related to hydraulic fracturing (Figure 1), and approximately 1 percent (457) of the spills were
identified as being related to hydraulic fracturing. Information available for an estimated 33
percent of the approximately 36,000 spills reviewed was insufficient to determine whether or not
the spill was associated with hydraulic fracturing.
The 457 hydraulic fracturing-related spills in this analysis occurred in 11 states (Table 2) over six
years.10 State data sources identified a total of 394 hydraulic fracturing-related spills, and industry
sources identified 63 additional spills related to hydraulic fracturing. Data provided by industry
10 The reported number of hydraulic fracturing-related spills per year increased from approximately 27 spills in 2006 to
110 spills in 2011. The increase over this timeframe could have been related to an increase in hydraulic fracturing as well
as changes in state reporting requirements and industry reporting trends.
9
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Unrelated
¦ Unknown
¦ Related
66%
Figure 1. Percent of reviewed spills determined to be related and unrelated to hydraulic fracturing. State and
industry data sources reported an estimated 36,000 spills between January 2006 and April 2012. Spills identified as
related to hydraulic fracturing occurred on or near the well pad and were related to either the mixing or injection
of hydraulic fracturing fluids or the management of flowback (i.e., fluids returned to the surface after hydraulic
fracturing). Information available for roughly 33 percent of the approximately 36,000 spills was insufficient to
determine whether or not the spill was associated with hydraulic fracturing; these spills are categorized as
"unknown."
Table 2. Number and percentage of hydraulic fracturing-related spills identified from state and industry data
sources used in this study by state. Spills in West Virginia and North Dakota were identified only through industry-
provided data.
State
Number (Percent) of
Hydraulic Fracturing-Related Spills
Colorado
174 (38%)
Pennsylvania
87 (19%)
Oklahoma
55 (12%)
Arkansas
39 (9%)
Texas
36 (8%)
Louisiana
34 (7%)
New Mexico
14 (3%)
Wyoming
9 (2%)
Utah
4 (1%)
West Virginia
3 (1%)
North Dakota
2 (0.4%)
Total
457(100%)
included hydraulic fracturing-related spills in two states: North Dakota (two spills) and West
Virginia (three spills). Spills occurring in Colorado made up the largest proportion (38 percent) of
spills included in Appendix B, with the most spills per county reported in Garfield County and Weld
County. The data for Colorado do not necessarily indicate that this state had more hydraulic
fracturing-related spills than other states. Rather, spill reports from the Colorado Oil and Gas
10
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Information System were the most detailed spill reports from among the state data sources used in
this analysis and generally provided more of the information needed to determine whether a spill
was related to hydraulic fracturing (Table 1). Because of the preponderance of Colorado data, the
results presented below are more representative of hydraulic fracturing-related spills that were
identified from the Colorado Oil and Gas Information System than of spills identified from other
state and industry data sources.
Information from each hydraulic fracturing-related spill was analyzed to describe the following
spill characteristics: volumes and materials spilled, sources and causes of spills, and environmental
receptors. The availability of reported information on each of these characteristics is shown in
Figure 2. Data were not always available for each spill characteristic for each spill. For example, the
material spilled was reported in almost all instances, but fewer than half of the reports for each spill
included the amount of fluid recovered. The inconsistent data mean that results from analyses of
certain spill characteristics are more robust than others.
97%
™ l ™ 75*
I 67% 64%
11111111
Spilled Recovered Surface Ground Soil
water water
Volumes Spilled Spill Spill Environmental receptors
materials sources causes
Categories of Characteristics
Figure 2. Availability of reported information in records from hydraulic fracturing (HF)-related spills for each spill
characteristic assessed. The percent of the number of hydraulic fracturing-related spills (out of 457) is noted above
each column.
4.1. Volumes Spilled
Spilled volumes were reported and categorized for 81 percent of the hydraulic fracturing-related
spills (370 spills; Figure 2). Reported volumes per spill ranged from fewer than 5 gallons to over 1.3
million gallons, with a median volume per spill of 730 gallons. As shown in Figure 3, hydraulic
fracturing-related spills were characterized by numerous low-volume spills and comparatively
fewer high-volume incidents. Fifty-six percent of the hydraulic fracturing-related spills with
reported spill volumes resulted in a release of 1,000 gallons or less. These smaller spills, the
majority of which did not exceed 500 gallons (Figure 3 inset), released a total reported volume of
500
450
400
% 350
o
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
0-500 501-1,000
Reported Volume per Spill (gallons)
Reported Volume per Spill (gallons)
Figure 3. Distribution of hydraulic fracturing (HF)-related spills by reported volume spilled. There were 87 spills
with no volumes spilled reported. The inset shows a further breakdown of hydraulic fracturing-related spills for
low volume spills (up to 1,000 gallons spilled). The percent of the number of hydraulic fracturing-related spills (out
of 370 spills with reported volumes) is noted above each column. Percentages do not sum to 100 percent due to
rounding.
approximately 73,000 gallons. Small spills, however, accounted for only 3 percent of the total
reported volume spilled between January 2006 and April 2012 (approximately 2.3 million gallons).
The majority (57 percent) of the total reported volume spilled was from a single spill in which 1.3
million gallons of flowback and produced water spilled from a pit with a split liner.11
The ability to determine the fate of spilled fluids was limited, as only 211 hydraulic fracturing-
related spills (57 percent of spills with reported volumes spilled) contained information regarding
volumes recovered or not recovered in the course of the spill response.12 Of the approximately 2.3
million gallons reported to have been spilled, at least 1.6 million gallons were calculated to have
been unrecovered, and about 480,000 gallons were reported to have been recovered (Figure 4).
There were 32 hydraulic fracturing-related spills for which no spilled fluids were reported to have
been recovered, including the 1.3 million gallon spill of flowback and produced water. No
information regarding volumes recovered was provided for 159 hydraulic fracturing-related spills
11 Line number 320 in Appendix B.
12 Factors that may inhibit the total recovery of spilled materials include evaporation, precipitation, and the mixing of fluid
with the impacted media (i.e., soil or water).
12
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
11%
70%
Fluid Fate
Total Reported
Volume (gallons)
Unrecovered
1,600,000
Recovered
480,000
Unknown
250,000
Total Spilled
2,300,000
Unrecove red
Recovered
Unknown
Figure 4. Percent distribution of fluid fate by total reported volume for hydraulic fracturing-related spills. Volumes
spilled were reported for 370 hydraulic fracturing-related spills, and 211 spill records contained information on
volumes recovered. "Unknown" volumes were determined from spill records in which the volume spilled was
reported, but no information was included about volumes recovered. "Unrecovered" volumes were calculated
from data provided in the spill records for volumes spilled and volumes recovered. In some cases, reported
recovered volumes were larger than reported spilled volumes. Therefore, the sum of unrecovered, recovered, and
unknown fluid fate volumes is greater than the total volume spilled.
in which a spilled volume was reported. The fate of approximately 250,000 gallons from these spills
without recovery information was therefore unknown.
4.2. Spilled Materials
Materials spilled were identified and categorized for 97 percent of hydraulic fracturing-related
spills (443 spills; Figure 2). The types of spilled materials reported in the state and industry data
sources, along with definitions and examples, are shown in Table 3. Table 4 summarizes spilled
materials by number of spills, total reported volume spilled, median reported volume per spill, and
the 5th and 95th percentile reported volume per spill for each material type. The percent distribution
of the number of hydraulic fracturing-related spills and the total reported volume spilled by
material type are presented in Figure 5.
Table 4 and Figure 5 show that flowback and produced water was the most common type of fluid
reported to have been spilled (48 percent of 464 spills of different materials). Flowback and
produced water also accounted for the largest total volume of spilled material (85 percent), with
approximately 2 million gallons reported to have been spilled. Much of the estimated total volume
spilled of flowback and produced water was from 13 spills, each over 10,000 gallons and totaling
approximately 1.7 million gallons. There were 88 reported spills of fracturing fluid, with a median
reported volume per spill of 820 gallons and a total of approximately 140,000 gallons of fluid
reported to have been spilled. The 34 spills of frac water involved some of the largest hydraulic
fracturing-related spills, as evidenced by high median (1,800 gallons) and 95th percentile (11,000
gallons) reported volumes per spill.
13
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table 3. Definitions and examples of spilled materials.
Material Type
Definition
Examples
Flowback and
produced water
Fluids that return after the pressure
applied during hydraulic fracturing is
released
Flowback, flowback containing oil,
produced water, produced water
containing condensate, saltwater
Fracturing fluid
Fluid injected downhole
Frac sand, frac fluid (containing gel), frac
fluid (containing WFR-55LA, WBK-143L,
BI05000), frac fluid with diesel*
(containing HCI, clay, stabilizer, diesel,
friction reducer), KCI water
Chemicals and
products
On-site materials used in hydraulic
fracturing fluids
Acid, KCI,f biocide (diluted), friction
reducer, scale inhibitor, cross-linker (BC-
200UC), WGA15, gel
Frac water§
Water used in hydraulic fracturing
operations; may be recycled, treated, or
untreated
Treated frac water, untreated frac water
Hydrocarbons
Petroleum-related fluids released
through hydraulic fracturing operations
Diesel, oil, petroleum, condensate, gas
well liquid
Equipment fluids
Fluids from on-site equipment involved
in hydraulic fracturing activities
Antifreeze, hydraulic fluid, diesel
Unknown
Unknown which fluid type was spilled;
not reported
Unknown
* "Diesel" is included in both "fracturing fluid" and "equipment fluids" categories. "Frac fluid with diesel" was considered
a fracturing fluid, whereas "diesel" was placed under equipment fluids if it was related to on-site equipment.
f "KCI" is included in both "chemicals and products" and "fracturing fluid" categories. "KCI" was considered a chemical,
whereas "KCI water" was considered a fracturing fluid.
§ Unlike fracturing fluid, frac water may not include individual chemicals and/or chemical products, whereas fracturing
fluid is expected to contain individual chemicals and/or chemical products.
Table 4. Number of hydraulic fracturing related-spills, total reported volume spilled, and reported volume per spill
by material type. The number of spills sums to 464, as six incidents reported multiple fluids spilled; each material
type was counted as a separate spill.
Reported Volume per Spill (gallons)
Material Type
Number of
Spills
Total Reported
Volume Spilled
(gallons)
5th
Percentile
Median
95th
Percentile
Flowback and produced water
225
2,000,000
40
990
14,000
Fracturing fluid
88
140,000
80
820
8,400
Chemicals and products
63
44,000
20
230
4,200
Frac water
34
85,000
350
1,800
11,000
Hydrocarbons
24
40,000
10
710
6,300
Equipment fluids
16
1,400
20
60
280
Unknown
14
48,000
210
1,500
17,000
The least commonly reported spilled materials were fluids from on-site equipment and
hydrocarbons, such as diesel and petroleum. Equipment fluid spills were typically low-volume (all
less than 1,000 gallons). Spills of hydrocarbons were larger and similar in magnitude to spills of
chemicals and products, as shown by the median and percentile reported volumes per spill.
As shown in Figure 3, most hydraulic fracturing-related spills (56 percent) resulted in a release of
1,000 gallons or less. For spills of 1,000 gallons or less, the most commonly reported spilled
14
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
(a) HF-Related Spills (b) Total Reported Volume
by Material Type by Material Type
¦ Flowbackand produced water ¦ Fracturing fluid ¦ Chemicals and products
¦ Frac water ¦ Flydrocarbons ¦ Equipmentfluids
¦ Unknown
Figure 5. Summary of materials spilled by (a) number of hydraulic fracturing (HF)-related spills and (b) total
reported volume spilled. There were six hydraulic fracturing-related spills in which more than one fluid was spilled.
These spills were counted multiple times (once for each fluid spilled). Similarly, total volumes spilled were counted
multiple times (once for each fluid spilled). Percentages do not sum to 100 percent due to rounding.
material was flowback and produced water (88 spills), with a median reported volume per spill of
420 gallons. The next most commonly reported spilled material (48 spills), when 1,000 gallons or
less, were spilled were chemicals and products, with a median reported volume per spill of 200
gallons.
4.3. Spill Sources
The spill source was identified and categorized for 77 percent of all hydraulic fracturing-related
spills (351 spills; Figure 2). Source types, definitions, and examples are shown in Table 5. Table 6
summarizes spill sources by number of spills, total reported volume spilled, median reported
volume per spill, and the 5th and 95th percentile reported volumes per spill for each source type.
Figure 6 presents percent distributions of the number of spills and the estimated total volume
spilled by each source type.
Storage units, such as tanks or pits, were the most commonly reported source of spills (46 percent
of 458 spills from different sources). The total volume reported to have been spilled from storage
units was approximately 1.7 million gallons (75 percent of the total reported volume spilled for all
hydraulic fracturing-related spills), with a median reported volume per spill of 840 gallons. Fewer
spills were associated with wells or wellheads, but these spills had the greatest median and
percentile reported spill volumes compared to all other sources, including those from storage
containers.
15
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table 5. Definitions and examples of spill sources.
Source Type
Definition
Examples
Storage
Containers or structures that physically
hold fluids
Pit, tank, tote, trailer
Equipment
On-site machinery used in hydraulic
fracturing operations
Blender, manifold, pump
Hose or line
Connections that join on-site equipment
and storage
Chemical transfer, flowback line, water
transfer line, water transfer between
pads
Well or wellhead
Structural component of the well at the
surface
Well, wellhead
Unknown
Unknown sources; not reported
Unknown, blank field
Table 6. Number of hydraulic fracturing-related spills, total reported volume spilled, and reported volume per spill
by source type. The number of spills sums to 458, as one incident reported multiple spill sources; each source type
was counted as a separate spill.
Reported Volume per Spill (gallons)
Source Type
Number
of Spills
Total Reported
Volume Spilled
(gallons)
5th
Percentile
Median
95th
Percentile
Storage
210
1,700,000
80
840
8,400
Equipment
61
56,000
20
300
4,400
Hose or line
59
210,000
80
1,300
15,000
Well or wellhead
22
210,000
220
6,300
47,000
Unknown
106
100,000
7
420
7,100
Relatively small volume hydraulic fracturing-related spills (less than or equal to 1,000 gallons
spilled) occurred most often (98 spills) from storage units, with a median reported volume per spill
of 420 gallons. Note that containers (e.g., totes and tanks) often hold small volumes compared to
pits, which limits the maximum volume of fluid that can be spilled. Pieces of equipment were the
next most commonly reported source (47 spills) for small volume spills. The median reported
volume per spill for relatively small volume equipment spills was 175 gallons.
4.4. Spill Causes
The causes of hydraulic fracturing-related spills could be determined and categorized for 75
percent of the spills (343 spills; Figure 2). Definitions and examples of each cause type are provided
in Table 7. Table 8 summarizes spill causes by number of spills, total reported volume spilled,
median reported volume per spill, and 5th and 95th percentile reported volumes per spill for each
cause type. Figure 7 presents the percent distribution of the number of spills and the total reported
volume spilled by cause type for the 457 hydraulic fracturing-related spills.
Among the spills for which the cause was reported, the most common causes were human error (33
percent of 457 spills) and equipment failure (27 percent). Spills caused by a failure of container
integrity (e.g., holes or seal failures in storage units), which was the cause identified for 11 percent
of spills, were generally associated with larger spill volumes. The total reported volume spilled for
these spills was approximately 1.5 million gallons, compared to the combined total volume of
approximately 660,000 gallons reported for spills caused by human error and equipment failure.
16
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
(a) HF-Related Spills (b) Total Reported Volume
by Source Type by Source Type
Storage "Equipment i Hose or line ¦ Well or wellhead i Unknown
Figure 6. Summary of spill sources by (a) number of hydraulic fracturing (HF)-related spills and (b) total reported
volume spilled. There was one spill in which more than one source was identified. This instance was counted
multiple times (once for each source). Similarly, the total volume spilled for this spill was counted multiple times
(once for each source). Percentages in graph (b) do not sum to 100 percent due to rounding.
Table 7. Definitions and examples of spill causes.
Cause Type
Definition
Examples
Human error
Human error as listed by the state or
determined to be the root spill cause
Valve left open, miscommunication,
failure to monitor or equalize tanks
Equipment failure
Equipment failure as listed by the state
or determined to be the root spill cause
Blowout preventer failure, corrosion or
washout, valve failed
Failure of container
integrity
Holes, leaks, and seal failures in storage
units
Hole, leak, seal failure
Other
Assorted causes of spills
Blowout, weather, vandalism, well
communication
Unknown
Unknown or unspecified cause; not
reported
Unknown, not specified, unanticipated
flowback
Appendix B includes hydraulic fracturing-related spills that were caused by two well blowouts13
and ten well communication events,14 which were included in the "Other" or "Equipment failure"
categories. Well blowouts and well communication events can both lead to high volume spills. The
13 Line numbers 326 and 339 in Appendix B.
14 Line numbers 163,236,265,271,286,287, 375,376,377, and 380 in Appendix B.
17
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table 8. Number of hydraulic fracturing-related spills, total volume spilled, and reported volume per spill by cause
type. No spills identified more than one cause.
Reported Volume per Spill (gallons)
Cause Type
Number
of Spills
Total Reported
Volume Spilled
(gallons)
5th
Percentile
Median
95th
Percentile
Human error
150
270,000
80
740
6,300
Equipment failure
124
390,000
20
700
14,000
Failure of container integrity
50
1,500,000
100
760
29,000
Other
19
86,000
200
2,900
17,000
Unknown
114
100,000
20
420
7,200
(a) HF-Related Spills (b) Total Reported Volume
by Cause Type by Cause Type
4%
Humanerror ¦ Equipmentfailure ¦ Failure of container integrity "Other "Unknown
Figure 7. Summary of spill causes by (a) number of hydraulic fracturing (HF)-related spills and (b) total reported
volume spilled. No spills identified more than one cause. Percentages in graph (b) do not sum to 100 percent due
to rounding.
reported volume per spill for well communication events ranged from approximately 250 gallons to
greater than 19,000 gallons. No volumes were reported for either of the well blowout spills.
Additionally, two cases of vandalism15 and six cases of inclement weather16 led to spills included in
this study. The reported volumes spilled due to vandalism were 840 and 4,200 gallons. Reported
15 Line numbers 18 and 194 in Appendix B.
16 Line numbers 50,82,183,223,248, and 293 in Appendix B.
18
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
volumes spilled due to weather spills ranged from 60 to 16,000 gallons per spill; these spills were
usually caused by freezing conditions that led to frozen valves and ruptured lines.
Small volume hydraulic fracturing-related spills (1,000 gallons or less of fluid released) were most
commonly caused by human error (78 spills). The median reported volume per spill from human
error was 420 gallons. Sixty-eight spills were caused by equipment failure, which was the second
most frequent cause of small volume spills. The median reported volume per spill due to equipment
failure was 270 gallons.
4.5. Environmental Receptors
Environmental receptors are the environmental media reached by spilled fluids. Three types of
environmental receptors were considered for this analysis: surface water, ground water, and soil.
These environmental receptors are of particular interest when considering potential impacts to
drinking water resources from hydraulic fracturing-related spills. Soil was included because spilled
fluids may infiltrate soil and percolate into ground water (Bodvarsson et al., 2000; Schwarzenbach
et al., 2002; US EPA, 1996). Surface and ground water resources could be currently used as drinking
water resources or may provide drinking water in the future. The EPA did not determine whether
any surface or ground water environmental receptors currently serve as drinking water resources.
Appendix B includes 411 spills (approximately 90 percent of the total number of hydraulic
fracturing-related spills) for which information regarding whether or not the spilled fluid reached
an environmental receptor was available. Information about whether or not spilled fluids reached
any of these environmental receptors was not available for 46 spills. Figure 8 shows the number of
hydraulic fracturing-related spills in which spilled fluids reached surface water, ground water, or
soil. Also shown in Figure 8 is the number of times spilled fluids were reported as not reaching an
environmental receptor or when it could not be determined if any environmental receptor was
reached from the available data.
There were no spills in which spilled fluids were reported as not reaching any of the three
environmental receptors. In 186 spills, spilled fluids were reported as not reaching surface or
ground water. Of these 186 spills, spilled fluids were reported to have reached soil in 107 spills. It
was unknown whether spilled fluids reached soil or were contained in the other 79 spills.
There were 300 hydraulic fracturing-related spills (approximately 65 percent of the total number
of hydraulic fracturing-related spills) in which spilled fluids reached at least one environmental
receptor; 24 of these reached multiple environmental receptors. Soil was the most commonly
reported environmental receptor, with spilled fluids reaching soil in over half of all spills in
Appendix B. The median reported volume per spill for these spills was 630 gallons (Table 9). Spilled
fluids reached surface water in 32 hydraulic fracturing-related spills (approximately 7 percent of all
hydraulic fracturing-related spills); the median reported volume per spill was 3,500 gallons. There
was one spill in which spilled fluids reached ground water (0.2 percent of spills). The spill that
19
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
350
« 300
£ 250
Ol
"3 200
f„
o
| 100
£
Z 50
n
¦ Yes
1 Unknown
il IV
Surface water Ground w
Environmental
ater Soil
Receptors
Figure 8. Number of hydraulic fracturing (HF)-related spills in which spilled fluids reached (yes) or did not reach
(no) surface water, ground water, or soil. "Unknown" refers to hydraulic fracturing related spills for which
environmental receptors were specified as unknown or not were identified.
Table 9. Number of hydraulic fracturing-related spills, total reported volume spilled, and reported volume per spill
by environmental receptor. There were 300 hydraulic fracturing-related spills that reached environmental
receptors. Twenty-four of these 300 spills reached both soil and surface water receptors and were counted as
having reached two separate receptors. Therefore, the number of receptors reached sums to 324. "NA" indicates
"not applicable."
Reported Volume per Spill (gallons)
Environmental Receptor
Number
of Spills
Total Reported
Volume Spilled
(gallons)
5th
Percentile
Median
95th
Percentile
Soil
291
540,000
30
630
8,400
Surface water
32
200,000
90
3,500
45,000
Ground water
1
130
NA
NA
NA
reached ground water occurred when 130 gallons of flowback and produced water were released
under a well pad due to an unknown cause.17
Data sources used for this study did not generally contain water quality monitoring results.
Therefore, the EPA did not determine whether or how spilled fluids affected ground or surface
water resources. Additionally, it may take a long time for spilled fluids to reach ground water
(Bodvarsson et al., 2000; Glass et al., 2005). Some of the data sources in this study did not contain
long-term monitoring information.
17 Line number 363 in Appendix B.
20
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
5. Results: Containment and Response
Fluid containment and spill response affect the path of spilled fluids and are important to
understand when assessing the potential for impacts to drinking water resources. Approximately
25 percent of the spill records associated with hydraulic fracturing-related spills included
information on containment systems, while approximately 60 percent described response
activities. This section summarizes containment and response activities intended to prevent
impacts to drinking water resources once a spill has occurred.
5.1. Containment
Containment systems are used to hold fluids or to stop the flow of spilled fluids. They can include
primary, secondary, and emergency containment systems. Primary containment systems are the
storage units, such as tanks or pits, in which fluids are intentionally kept Pre-planned secondary
containment systems, such as liners and berms, are installed before a spill occurs and are intended
to contain spilled fluids until they can be cleaned up. Emergency containment systems are often
temporary and are not in place before a spill occurs, but rather are implemented in response to a
spill. The most common types of containment systems mentioned in the hydraulic fracturing-
related spill records included pre-planned secondary or emergency systems such as berms, booms,
dikes, liners, and pits.
Secondary containment systems surround primary containment systems and are generally
intended to provide temporary containment of any spilled fluids until appropriate actions are taken
to stop the spill and remove the fluid. Due to the limited information on pre-planned secondary
containment in the hydraulic fracturing-related spill records, it was not possible to evaluate the
extent to which the measures taken to contain spills were effective. In some cases, "effectiveness"
appeared to be determined by whether the spilled fluid was contained on the site or within the pre-
planned secondary containment unit. For example, a spill record from Pennsylvania noted that an
unspecified volume of "frac fluid spilled before going downhole" and "no evidence of any fluid
leaving containment was observed."18 Another record from Wyoming was more specific: 1,260
gallons of produced water spilled around a tank and was contained in a berm. The spill record also
noted that a "small portion soaked into [the] ground."19 This indicated that the containment did not
completely prevent spilled fluids from reaching an environmental receptor.
In at least two instances in Colorado, the pre-planned secondary containment systems appeared to
be successful at both preventing fluid migration off-site and preventing spilled fluid from entering
soil or water through the use of a liner. The first incident report, in 2010, described a tank overflow
resulting in the release of 420 gallons of flowback fluid. The report stated:
18 Line number 360 in Appendix B.
19 Line number 388 in Appendix B.
21
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
The entire release was contained within the lined containment cell the upright tanks
are located in. None of the fluids were recovered because the lined containment, as
it is constructed, allows any fluids to migrate back into the frac pit.20
A similar incident occurred in 2011. In this case, a tank overflow resulted in a release of 3,990
gallons of flowback fluid. The report for this case noted that:
The entire release was contained within the lined containment around the upright
tanks. Due to the design of the secondary containment, the flowback water flowed
back into the frac pit. No fluids migrated off the location.21
There were instances in which the hydraulic fracturing-related spill reports noted that the pre-
planned secondary containment systems were breached. Breaches of berms and dikes were most
commonly reported. Examples of these types of incidents are provided in Table 10. Leaks from pre-
planned secondary containment systems were also reported. Causes of pre-planned secondary
containment failures were generally not specified in the data sources used in this study.
Table 10. Examples of breaches of berms and dikes, as described in records from hydraulic fracturing-related spills.
Incident Description
Reported Volume
Spilled (gallons)
Line Number
(Appendix B)
"...flowed out onto the frac pad and eventually
breached the perimeter berm on the south side..."
11,130
81
"...tank overflowed during flowback operations,
filled the bermed containment area, ran over the
top of the berm and onto the pad..."
4,200
192
"...because of a weakness in the berm, some of
the water flowed under the berm and onto a
farmer's field..."
1,470
165
"...the frac pit breached the berm and
overflowed..."
1,050
170
"...the existing dirt retention dike behind the
tanks was not consolidated dirt, large rocks are
mixed in with the dirt. The water from the
overflow channeled through the rocks..."
630
154
Several instances of emergency containment systems were described in the hydraulic fracturing-
related spill records. Berm or dike construction and boom deployment were the most commonly
reported types of emergency containment systems. Ditches or pits were also reported to have been
dug in order to capture the spilled fluid. For example, a 2011 spill report from Colorado indicated
that a "frac head failed on the wellhead during a frac operation, releasing approximately 200
barrels [8,400 gallons] of slick frac water."22 The spilled fluid was reported to have been contained
in emergency ditches and "did not migrate off-site." In another instance in Colorado, 294 gallons of
20 Line number 90 in Appendix B.
21 Line number 57 in Appendix B.
22 Line number 75 in Appendix B.
22
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
flowback and produced water leaked from a failed weld on a flowback tank.23 Consequently, "a dike
and berm [were] constructed around [the flowback] tank" to contain the release. Similar to
secondary containment reports, it was often unclear if emergency containments systems
successfully prevented an impact to environmental media.
Absorbent materials were reported to have been used to contain spilled fluids. In general, it
appears that absorbent materials were used when the estimated spill volumes were relatively small
(10 to 200 gallons) and the spilled materials were individual chemicals or chemical products (e.g.,
scale inhibitors) rather than wastewater (i.e., flowback and produced water).
5.2. Response
Approximately 60 percent of spill records for spills identified as having been associated with
hydraulic fracturing contained information about company responses, state responses, or both. The
summary below focuses on the actions taken to clean up spilled fluids, which ranged from
immediate actions to stop the spill and contain the spilled fluid to longer term actions to remediate
the affected area.
The immediate responses to spills were varied, as shown in Table 11. In general, it was reported
that actions were first taken to stop the spill (e.g., shut down operations, adjust equipment, or drain
a leaking container) and then to contain the spilled fluid (e.g., construct emergency containment). In
some situations, spills were discovered after fluids had been released from the primary
containment unit (e.g., a tank). In these cases, it appeared that the focus of the immediate response
was to contain the spilled fluid, if possible.
After a spill is stopped and contained, the response shifts to remediation. For this analysis,
"remediation" was considered to be any action taken in response to the spill, including removal of
contaminated material. The most commonly reported remediation activity, mentioned in
approximately half of the hydraulic fracturing-related spill records, was removal of either the
spilled fluid and/or affected media (typically soil). In the spill records, removal included excavation
of contaminated soil and use of absorbent material and/or vacuum trucks to remove the spilled
fluid. Removal activities were found to occur in various combinations. For example, a spill of
approximately 4,200 gallons of acid was cleaned up by first spreading soda ash to neutralize the
acid and then removing the affected soil.24
In general, a higher percentage of fluid volume was recovered for spills of up to 1,000 gallons. At
the other extreme, the report of the spill of 1.3 million gallons indicated that no fluids were
recovered. In some cases the volume recovered exceeded the volume spilled, because the spilled
fluid mixed with water or soil due to precipitation spills (i.e., rain or snow).
23 Line number 49 in Appendix B.
24 Line number 258 in Appendix B.
23
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Table 11. Examples of immediate responses used to stop spills or contain spilled fluids, as described in records
from hydraulic fracturing-related spills. Responses were often found to be used in combination (e.g., shutdown
operation and construct emergency containment).
Immediate Company
Response
Examples
Line Number
(see Appendix B)
Adjust equipment
"...fracing personnel immediately shut the valves to any
tanks which were open to the manifold, thus preventing
any further fluid loss through the open end of the
manifold..."
186
"...the 2 source tanks were isolated by closing valves..."
247
"...flow directed into another tank..."
307
Construct emergency
containment
"...booms and earthen berms were constructed around
spill area..."
292
"...diked the spill area..."
458
"...pads and trenching were used to contain the spill..."
451
Drain leaking container
or pit
"...tank was evacuated of any remaining fluid..."
184
"...leaking tank was immediately drained of water..."
231
"...8,700 barrels [365,400 gallons] were drained from the
pit within 2 hours..."
276
Shutdown operations
"...when the overflowing tanks were discovered, flowback
personnel on the [...] pad immediately halted flowback
operations..."
208
"...the fracing operation was immediately shut down..."
204
"...shut frac down; shut in well; blended pressure off..."
245
Other remediation activities noted in the spill reports included flushing the affected area with
water25 and neutralizing the spilled material. In some cases, the spill records referred to general
"remediation" and did not provide additional details on specific actions taken. In two cases, the spill
reports specifically indicated that no remediation activities were needed or occurred.26
6. Discussion
The data sources used in this study contained over 36,000 spills. Spill records from an estimated
12,000 spills (33 percent of the total number of spills reviewed) contained insufficient information
to determine whether the spill was related to hydraulic fracturing. From the remaining spills, the
EPA identified an estimated 24,000 spills (66 percent of the total number of spills reviewed) as not
related to hydraulic fracturing and 457 spills (approximately 1 percent) as related to hydraulic
fracturing. The 457 hydraulic fracturing-related spills occurred in 11 different states over six years
(January 2006 and April 2012).
Hydraulic fracturing-related spills identified in this study (Appendix B) were characterized by
numerous low volume spills (1,000 gallons or less released) and relatively few high volume spills
(20,000 gallons or more). A similar spill volume distribution was reported by Fisher and Sublette
(2005) after examining the Oklahoma Corporation Commission Complaint Database for the years
1993 to 2005. Fisher and Sublette specifically looked at oil and saltwater environmental releases
25 The spill reports did not include information on the fate of the rinsate.
26 Line numbers 255 and 291 in Appendix B.
24
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
during oil and gas operations and found that the spilled fluid volumes were predominantly below
500 gallons for oil spills and 1,700 gallons for saltwater spills, with few releases exceeding 21,000
gallons. A recent study of spills reported to the Colorado Oil and Gas Conservation Commission
found that the most common spill volume was between 462 and 2,100 gallons (11 to 55 barrels),
followed by spills between 84 and 420 gallons (2 to 10 barrels) (Colorado Department of Natural
Resources, 2014).
This report describes hydraulic fracturing-related spills in terms of the following characteristics:
volumes and materials spilled, sources and causes of spills, and environmental receptors. As
discussed below, combinations of spill characteristics (e.g., materials spilled and source) can be
used to identify common spill scenarios.
Among the entire dataset of 457 hydraulic fracturing-related spills, the most common spill scenario
found was spills of flowback and produced water due to human error (19 percent), with a median
reported volume per spill of 1,000 gallons. Fourteen percent of all spills were of flowback and
produced water due to human error at storage units, with a median reported volume per spill of
840 gallons. An example of this spill scenario is the 8,400 gallons of flowback reported to have
spilled when a valve on a flowback tank was accidentally left open.27
Spill records from 370 hydraulic fracturing-related spills contained information on volumes of fluid
spilled. Fifty-six percent of these spills reported spilled volumes of 1,000 gallons or less, which
suggests that most reported hydraulic fracturing-related spills were relatively low volume. Spill
scenarios identified for these low-volume spills were similar to spill scenarios observed for the
entire dataset Eighteen percent of low-volume spills were spills of flowback and produced water
due to human error, with a median reported volume per spill of 460 gallons. Fourteen percent of
low-volume spills were spills of flowback and produced water due to human error at storage units,
with an estimated median reported volume per spill of 550 gallons.
Large volume spills from well blowouts and well communication events have recently been
reported by the media (Atkin, 2014; Vaidyanathan, 2013). Similar types of spills were identified as
part of this study. For example, a spill in Oklahoma occurred when "fracing story #5 communicated
with story #2," leading to over 13,000 gallons of oil being released.28 In another instance, failure of
blowout prevention equipment caused a well in Pennsylvania to discharge natural gas and
flowback for 18 hours before it was contained.29 These types of spills represented a small
proportion of the hydraulic fracturing-related spills in Appendix B: two spills were caused by well
blowouts30 and ten spills were caused by well communication.31
Large and small volumes of spilled fluids were reported to have reached at least one environmental
receptor (i.e., surface water, ground water, or soil) in 300 spills. In 24 instances, spilled fluids were
reported to have reached more than one environmental receptor. Soil was the most predominant
27 Line number 390 in Appendix B.
28 Line number 286 in Appendix B.
29 Line number 326 in Appendix B.
30 Line numbers 326 and 339 in Appendix B.
31 Line numbers 163,236,265,271,286,287, 375,376,377, and 380 in Appendix B.
25
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
environmental receptor (291 instances), with a median reported volume per spill of 630 gallons.
Thirty-two instances of spilled fluids reaching surface water (median reported volume per spill
3,500 gallons) and one instance of spilled fluids reaching ground water (130 gallons) were
identified.
The extent to which spilled fluids affected the geochemistry and ecology of surface or ground water
systems was not addressed in this analysis. Effects to surface and ground water resources depend
on the site conditions, environmental conditions, chemical properties, and the volume of spilled
fluids (Schwarzenbach etal., 2002). For example, a small volume spill of an environmentally
harmful chemical into a small stream may have a more serious impact on the water resource than a
large volume spill of an environmentally benign chemical. Additionally, spill containment and
response efforts can affect the path of spilled fluids and are important to understand when
determining the potential for spilled fluids to affect surface and ground water resources.
6.1. Study Limitations
The objective of this study was to characterize hydraulic fracturing-related spills that may reach
surface or ground water resources using spill reports obtained from selected state and industry
data sources. Although spill data were obtained from nine states that are among the top oil and gas
producing states in the country (US Energy Information Administration, 2012, 2014), similar data
from other oil and gas producing states were not included.
The 457 spills used to characterize hydraulic fracturing-related spills were likely a subset of the
total number of hydraulic fracturing-related spills that could have been identified from the state
and industry data sources. State and federal reporting requirements determine which spills, and
what information about each spill, are recorded in state data sources. A review of the 63 hydraulic
fracturing-related spills recorded in industry data sources, but not found in state data sources,
offered insights into why some hydraulic fracturing-related spills may not have been recorded in
state data sources. In some instances, it appeared that the spill volume or spilled material may not
have met the state reporting requirements in place at the time of the spill. Consequently, some
hydraulic fracturing-related spills may not have been included in the state data sources. In the case
of industry data sources, the EPA did not evaluate company practices or policies with respect to the
documentation of hydraulic fracturing-related spills.
Additionally, some reported spills may not have been identified as related to hydraulic fracturing
due to insufficient information in the data sources. Incident descriptions, which were often used in
this study to identify hydraulic fracturing-related spills, varied in detail from state to state, and
within states from incident to incident There were also inconsistencies in terminology among
incident descriptions, making it difficult to distinguish hydraulic fracturing-related spills from spills
related to other oil and gas activities. Missing or inconsistent information in some data sources
likely led to spills not being identified as related to hydraulic fracturing (Commonwealth of
Pennsylvania, 2014).
The quantitative characterization of hydraulic fracturing-related spills presented in Section 4 may
have been different if more hydraulic fracturing-related spills could have been identified from the
data sources used in this study.
26
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
7. Conclusions
This report presents the results of a broad review of state and industry spill data from 457
hydraulic fracturing-related spills. Data from these spills were used to characterize volumes and
materials spilled, spill sources and causes, and environmental receptors. There were several key
findings. Spills related to hydraulic fracturing were most often characterized by numerous, low
volume events (up to 1,000 gallons) and relatively few high volume events (greater than 20,000
gallons). The most common material spilled was flowback and produced water, and the most
common source of spills was storage units. More spills were caused by human error than any other
cause. Over half of the spills associated with hydraulic fracturing reached an environmental
receptor, with 33 instances of spilled fluids reaching surface or ground water resources. These
results, as well as other information on spill characteristics and containment and response
activities, provide important insights into the nature of hydraulic fracturing-related spills in several
key states with hydraulic fracturing.
27
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
References
Alley, B., Beebe, A., Rodgers, J., and Castle, J.W. 2011. Chemical and Physical Characterization of
Produced Waters from Conventional and Unconventional Fossil Fuel Resources. Chemosphere
85:74-82.
Arkansas Department of Environmental Quality (ADEQ). 2012. Complaints and Inspections
Database. Available at http://www.adeq.state.ar.us/home/pdssql/complaints_inspections.aspx.
Accessed December 7, 2012.
Arthur, J., Bohm, B., Coughlin, B.J., Layne, M., and Cornue, D. 2009. Evaluating the Environmental
Implications of Hydraulic Fracturing in Shale Gas Reservoirs. SPE Americas Environmental and
Safety Conference, San Antonio, Texas, March 23-25, 2009. Society of Petroleum Engineers.
Atkin, E. 2014. Fracking Well Blowout Causes Oil and Chemical Wastewater Spill in North Dakota.
Climate Progress, February 18, 2014. Available at http://thinkprogress.org/climate/2014/02/17/
3299221/north-dakota-blowout/. Accessed May 21, 2014.
Barbot, E., Vidic, N.S., Gregory, K.B., and Vidic, R.D. 2013. Spatial and Temporal Correlation of Water
Quality Parameters of Produced Waters from Devonian-Age Shale Following Hydraulic Fracturing.
Environmental Science & Technology 47:2562-2569.
Blauch, M.E., Myers, R.R., Moore, T.R., and Lipinski, B.A. 2009. Marcellus Shale Post-Frac Flowback
Waters: Where is All the Salt Coming From and What are the Implications? SPE Eastern Regional
Meeting, Charleston, West Virginia, September 23-25, 2009. Society of Petroleum Engineers.
Bodvarsson, G., Finsterle, S., Liu, H.H., Oldenburg, C.M., Pruess, K., Sonnenthal, E., and Wu, Y.-S. 2000.
Flow and Transport Modeling of the Vadose Zone. Chapter 5. In Looney, B.B., and Falta, R.W.,
editors, Vadose Zone: Science and Technology Solutions. Battelle Press, Columbus, Ohio.
Chapman, E.C., Capo, R.C., Stewart, B.W., Kirby, C.S., Hammack, R.W., Schroeder, K.T., and Edenborn,
H.M. 2012. Geochemical and Strontium Isotope Characterization of Produced Waters from
Marcellus Shale Natural Gas Extraction. Environmental Science & Technology 46:3545-3553.
Colorado Department of Natural Resources, Oil and Gas Conservation Commission. 2014. Risk-
Based Inspections: Strategies to Address Environmental Risk Associated with Oil and Gas
Operations. Colorado Oil and Gas Conservation Commission. Publication number: OGCC-2014-
PROJECT #7948 pages.
Colorado Oil and Gas Conservation Commission (COGCC). 2012. Colorado Oil and Gas Information
System. Available athttp://cogcc.state.co.us/data.html#/cogis. Accessed May 24, 2012.
Commonwealth of Pennsylvania, Department of the Auditor General. 2014. DEP's Performance in
Monitoring Potential Impacts to Water Quality from Shale Gas Development, 2009-2012.
Fisher, J.B., and Sublette, K.L. 2005. Environmental Releases from Exploration and Production
Operations in Oklahoma: Type, Volume, Causes, and Prevention. Environmental Geosciences 12:89-
99.
Glass, R.J., Brainard, J.R., and Jim Yeh, T.-C. 2005. Infiltration in Unsaturated Layered Fluvial
Deposits at Rio Bravo: Macroscopic Anistropy and Heterogeneous Transport Vadose Zone Journal
4:22-31.
28
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Gregory, K.B., Vidic, R.D., and Dzombak, D.A. 2011. Water Management Challenges Associated with
the Production of Shale Gas by Hydraulic Fracturing. Elements 7:181-186.
Gross, S.A., Avens, H.J., Banducci, A.M., Sahmel, J., Panko, J.M., and Tvermoes, B.E. 2013. Analysis of
BTEX Groundwater Concentrations from Surface Spills Associated with Hydraulic Fracturing
Operations. Journal of the Air and Waste Management Association 63:424-432.
Ground Water Protection Council (GWPC). 2009. State Oil and Natural Gas Regulations Designed to
Protect Water Resources. Ground Water Protection Council for US Department of Energy, National
Energy Technology Laboratory. Available at http://www.gwpc.org/sites/default/files/
state_oil_and_gas_regulations_designed_to_protect_water_resources_0.pdf. Accessed September 16,
2014.
GWPC and ALL Consulting. 2009. Modern Shale Gas Development in the United States: A Primer.
Ground Water Protection Council and ALL Consulting for US Department of Energy. Available at
http://energy.gov/fe/downloads/modern-shale-gas-development-united-states-primer. Accessed
June 18, 2014.
Lee, D., Herman, J., Elsworth, D., Kim, H., and Lee, H. 2011. A Critical Evaluation of Unconventional
Gas Recovery from the Marcellus Shale, Northeastern United States. KS CE Journal of Civil
Engineering 15:67-687.
Louisiana Department of Environmental Quality (LDEQ). 2013. Personal communication, M. Molieri.
January 3, 2013.
Malone, M., and Ely, J.W. 2007. Execution of Hydraulic Fracturing Treatments. Chapter 9. In
Economides, M.J., and Martin, T., editors, Modern Fracturing: Enhancing Natural Gas Production.
Energy Tribune Publishing, Inc., Houston, Texas.
Neff, J., Lee, K., and DeBlois, E.M. 2011. Produced Water: Overview of Composition, Fates, and
Effects. In Lee, K., and Neff, J., editors, Produced Water: Environmental Risks and Advances in
Mitigation Technologies. Springer, Duxbury, MA.
New Mexico Energy, Minerals, and Natural Resources Department (NMEMNRD). 2012. Oil
Conservation Division Permitting, Spill Search. Available at https://wwwapps.emnrd.state.nm.us/
ocd/ocdpermitting//Data/Incidents/Spills.aspx. Accessed May 17, 2012.
New York State Department of Environmental Conservation. 2011. Supplemental Generic
Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program (revised
draft): Well Permit Issuance for Horizontal Drilling and High-Volume Hydraulic Fracturing to
Develop the Marcellus Shale and Other Low-Permeability Gas Reservoirs. Available at
http://www.dec.ny.gov/energy/75370.html. Accessed July 30, 2014.
Oklahoma Corporation Commission (OCC). 2013. Personal communication, Baker, T. April 11, 2013.
Olmstead, S.M., Muehlenbachs, L.A., Shih, J.-S., Chu, Z., and Krupnick, A.J. 2013. Shale Gas
Development Impacts on Surface Water Quality in Pennsylvania. Proceedings of the National
Academy of Sciences 110:4962-4967.
Orem, W.H., Tatu, C.A., Lerch, H.E., Rice, C.A., Bartos, T.T., Bates, A.L., Tewalt, S., and Corum, M.D.
2007. Organic Compounds in Produced Waters from Coalbed Natural Gas Wells in the Powder River
Basin, Wyoming, USA. Applied Geochemistry 22:2240-2256.
29
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Papoulias, D.M., and Velasco, A.L. 2013. Histopathological Analysis of Fish from Acorn Fork Creek,
Kentucky, Exposed to Hydraulic Fracturing Fluid Releases. Southeastern Naturalist 12:92-111.
Pennsylvania Department of Environmental Protection (PADEP). 2012. Oil and Gas Compliance
Report Database. Available at http://www.depreportingservices.state.pa.us/ReportServer/Pages/
ReportViewer.aspx?/Oil_Gas/OG_Compliance. Accessed April 20, 2012.
Rowan, E.L., Engle, M.A., Kirby, C.S., Kraemer, T.F. 2011. Radium Content of Oil- and Gas-Field
Produced Waters in the Northern Appalachian Basin (USA)—Summary and Discussion of Data: US
Geological Survey Scientific Investigations Report. Available at http://pubs.usgs.gov/sir/2011/
5135/pdf/sir2011-5135.pdf. Accessed May 21, 2014.
Rozell, D.J., and Reaven, S.J. 2012. Water Pollution Risk Associated with Natural Gas Extraction from
the Marcellus Shale. Risk Analysis 32:1382-1393.
Schwarzenbach, R.P., Gschwend, P.M., and Imboden, D.M. 2002. Environmental Organic Chemistry.
2nd edition. John Wiley & Sons, Hobeken, New Jersey.
Sirivedhin, T., and Dallbauman, L. 2004. Organic Matrix in Produced Water from the Osage-Skiatook
Petroleum Environmental Research Site, Osage County, Oklahoma. Chemosphere 57:463-469.
Soeder, D.J. 2010. The Marcellus Shale: Resources and Reservations. Eos 91:277-278.
Stringfellow, W.T., Domen, J.K., Camarillo, M.K., Sandelin, W.L., and Borglin, S. 2014. Physical,
Chemical, and Biological Characteristics of Compounds used in Hydraulic Fracturing. Journal of
Hazardous Materials 275:37-54.
Strong, L.C., Gould, T., Kasinkas, L., Sadowsky, M.J., Aksan, A., and Wackett, L.P. 2014. Biodegradation
in Waters from Hydraulic Fracturing: Chemistry, Microbiology, and Engineering. Journal of
Environmental Engineering 140:B4013001.
Texas Railroad Commission (TRC). 2013. Personal communication, Pope, P. July 8, 2013.
US Energy Information Administration. 2012. Rankings: Natural Gas Marketed Production, 2012.
Available at http://www.eia.gov/state/rankings/?sid=US#/series/47. Accessed July 30, 2014.
US Energy Information Administration. 2014. Rankings: Crude Oil Production, April 2014. Available
athttp://www.eia.gov/state/rankings/?sid=US#/series/46. Accessed July 30, 2014.
US Environmental Protection Agency (US EPA). 1996. Soil Screening Guidance: Technical
Background Document. Available at http://www.epa.gov/superfund/health/conmedia/
soil/toc.htm. Accessed September 16, 2014.
US EPA. 2004. Evaluation of Impacts to Underground Sources of Drinking Water by Hydraulic
Fracturing of Coalbed Methane Reservoirs Study. EPA/816/R-04/003. Available at
http://water.epa.gov/type/groundwater/uic/class2/hydraulicfracturing/
wells_coalbedmethanestudy.cfm. Accessed July 30, 2014.
US EPA. 2010. Information Request Letter Sent to Nine Hydraulic Fracturing Service Companies.
Available at http://www2.epa.gov/sites/production/files/documents/
september_2010_request_letter.pdf. Accessed May 21, 2014.
30
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
USEPA. 2011a. Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources. EPA/600/R-11/122. Available athttp://www2.epa.gov/sites/production/files/
documents/hf_study_plan_110211_final_508.pdf. Accessed July 30, 2014.
US EPA. 2011b. Proceedings of the Technical Workshops for the Hydraulic Fracturing Study: Water
Resources Management. EPA/600/R-11/048. Available athttp://www2.epa.gov/sites/production/
files/documents/HF_Workshop_4_Proceedings_FINAL_508.pdf. Accessed September 16, 2014.
US EPA. 2011c. Information Request Letter Sentto Nine Oil and Gas Operators. Available at
http: //www2. epa. gov/sites/production/files/documents/August_2 01 l_request_letter.pdf.
Accessed May 21, 2014.
US EPA. 2012. Study of the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources:
Progress Report EPA 601/R-12/011. Available at http://www2.epa.gov/hfstudy/study-potential-
impacts-hydraulic-fracturing-drinking-water-resources-progress-report-O. Accessed May 21, 2014.
US EPA. 2013. Quality Assurance Project Plan for Hydraulic Fracturing Surface Spills Data Analysis.
Available at http://www2.epa.gov/sites/production/files/documents/hf-spills-analysis-qapp.pdf.
Accessed May 21, 2014.
US EPA. 2015. Analysis of Hydraulic Fracturing Fluid Data from the FracFocus Chemical Disclosure
Registry 1.0. EPA/601/R-14/003. Available at http://www2.epa.gov/hfstudy/epa-analysis-
fracfocus-l-data. Accessed April 8, 2015.
US Government Accountability Office. 2014. Oil and Gas: Updated Guidance, Increased
Coordination, and Comprehensive Data Could Improve BLM's Management and Oversight Available
at http://www.gao.gov/products/GAO-14-238. Accessed June 18, 2014.
Utah Department of Environmental Quality (UDEQ). 2013. Personal communication, Zucker, M. July
2, 2013.
Vaidyanathan, G. 2013. Hydraulic Fracturing: When Two Wells Meet, Spills Can Often Follow.
EnergyWire, August 5, 2013. Available at http://www.eenews.net/energywire/stories/
1059985587. Accessed May 21, 2014.
Vengosh, A., Jackson, R.B., Warner, N., Darrah, T.H., and Kondash, A. 2014. A Critical Review of the
Risks to Water Resources from Unconventional Shale Gas Development and Hydraulic Fracturing in
the United States. Environmental Science & Technology 48:8334-8348.
Vidic, R.D., Brantley, S.L., Vandenbossche, J.M., Yoxtheimer, D., and Abad, J.D. 2013. Impact of Shale
Gas Development on Regional Water Quality. Science 340:1235009-1-9.
Wyoming Oil and Gas Conservation Commission (WOGCC). 2012. Personal communication,
Kropatsch, T. December 21, 2012.
31
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Appendix A: Standardization of Spill Characteristics
Data on spills in each data source were recorded in a different way and provided slightly different
information (Table 1). Therefore, these data were standardized prior to analysis. Information
regarding the identification of hydraulic fracturing-related spills and the standardization of
volumes, materials spilled, sources, causes, and environmental receptors (i.e., the environmental
media reached by a spill, such as surface water, ground water, and soil) is provided in this
appendix. Appendix B contains the standardized spill table.
As shown in Table 1, state data sources often contained information in specific data fields (e.g.,
volume spilled and material spilled). The state data sources also included incident descriptions that
provided additional information about the spill. When available, data in specific data fields were
used to develop Appendix B. If a specific data field was not available in the state data source or was
not populated, relevant information from the incident description was used to develop Appendix B.
A.l. Volumes
Depending on the data source, volumes were reported for the amount spilled, amount recovered,
and amount lost; all volumes were converted to gallons. In instances where two different volumes
were reported (e.g., if a state data source and an industry data source reported different volumes
for the same spill), the volume contained in the more detailed spill report was chosen. Table A1
summarizes the categorization of volumes.
Table Al. Categorization method for reported volumes of spilled materials.
Volume Category
Definition
Any numerical value
Discrete volume (e.g., 5 gallons)
Volume range (e.g., 5-10 gallons)
< or > discrete volume (e.g., <10 gallons)
Split volumes (e.g., 5 gallons flowback and 10 gallons
HCI)
Unknown
No volumes reported
Quantitative volume not specified
"Unknown" in original database
When provided with a volume range, the mean was calculated to use as a discrete number for the
volume spilled (e.g., a 10-20 gallon spill was treated as a 15 gallon spill). When a volume greater
than or less than a discrete number was indicated, the number itself was used as the spill volume
for analysis (e.g., reported volumes of >5 gallons or <5 gallons were both treated as 5 gallons).
As described in Section 3.3, spilled volumes should be considered to be estimates. Recovered
volumes should similarly be considered estimates. Spilled volumes and recovered volumes were
used to calculate the net loss of spilled material. Some spill reports indicated that the volume
recovered exceeded the volume spilled due to precipitation or mixing spills. In these cases, the net
loss was assumed to be zero gallons.
32
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
A.2. Spilled Materials
Spilled materials were assigned to the broad categories listed in Table 3. Some materials identified
in the spill reports were not unique to hydraulic fracturing (e.g., produced water). In these cases,
spills were only added to Appendix B if other hydraulic fracturing-specific terminology (e.g., frac,
hydraulic fracturing) was identified in the spill report.
A.3. Spill Sources
Spill sources were assigned to the broad categories listed in Table 5 based on where spilled fluids
originated. In general, spills must have occurred on a well pad during the hydraulic fracturing
treatment or during flowback to be included in Appendix B.
A.4. Spill Causes
Spill causes were assigned to the broad categories listed in Table 7.
A.5. Environmental Receptors
Environmental receptors included soil, surface water, and ground water. Table A2 summarizes the
categories developed for this analysis and their descriptions. A spill could be assigned to more than
one category if it reached multiple environmental receptors.
Table A2. Categorization method for environmental receptors.
Category
Description
Surface Water
Surface water (e.g., creek, lake, pond, river, wetland)
Ground Water
Spill report specifically indicated ground water was
reached
Soil
References to soil (e.g., crops, ditch, forest, pasture,
ground, vegetation)
Unknown
Unspecified whether fluids reached an environmental
receptor
When spilled materials were reported to have reached the well pad, an "unknown" environmental
receptor was assigned, because it was often unclear whether the well pad was lined or not. A lined
well pad may have prevented fluids from reaching soil, but an unlined well pad may not have
prevented fluids from leaching into soil and possibly ground water. Additionally, secondary
containment measures did not exclude the possibility of fluids reaching environmental receptors,
and in cases where the secondary containment was earthen, soil was identified as the
environmental receptor.
When surface water was reported as the environmental receptor, the specific type of surface water
body (e.g., stream, river, wetland, pond) was identified, if sufficient information was provided. Dry
ditches and drainage were not considered water bodies, because they were constructed for erosion
and sediment control during site development Standing water, or water collected in on-site pools,
was considered a type of surface water receptor, because spilled materials mixed with these
standing waters required site removal and/or treatment comparable to other surface water
receptors.
33
-------
Characterization of Hydraulic Fracturing-Related Spills May 2015
Appendix B: Hydraulic Fracturing-Related Spills Table
[See accompanying Microsoft Excel file.]
34
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Glossary
Aquifer: An underground geological formation, or group of formations, containing water. A source
of ground water for wells and springs, (ref 5)
Blowout: An uncontrolled flow of formation fluids from a well, (ref 4)
Blowout preventer: A large valve at the top of a well that may be closed if the drilling crew loses
control of formation fluids, (ref 4)
Condensate: A natural gas liquid with lower vapor pressure than natural gasoline or liquefied
petroleum gas. It is mainly composed of propane, butane, pentane, and heavier hydrocarbon
fractions, (ref 4)
Drinking water resource: Any body of water, ground or surface, that could currently, or in the
future, serve as a source of drinking water for public or private water supplies, (ref 6)
Flowback: After the hydraulic fracturing procedure is completed and pressure is released, the
direction of fluid flow reverses, and fluids flow up the wellbore to the surface. The fluids that return
to the surface are commonly referred to as "flowback." (ref 2)
Formation: A geological formation is a body of earth material with distinctive and characteristic
properties and a degree of homogeneity in its physical properties, (ref 5)
Ground water: Water found below the surface of the land, usually in porous rock formations.
Ground water is the source of water found in wells and springs and is frequently used for drinking,
(ref 5)
Hydraulic fracturing: A stimulation technique used to increase production of oil and gas.
Hydraulic fracturing involves the injection of fluids under pressures great enough to fracture the oil
and gas production formations, (ref 6)
Hydraulic fracturing fluid: Specially engineered fluids that generally contain water, proppants,
and chemical additives. Hydraulic fracturing fluids are pumped under high pressure into the well to
create and hold open fractures in the formation, (ref 3)
Natural gas: A naturally occurring mixture of hydrocarbon and non-hydrocarbon gases that is
highly compressible and expansible. Methane is the chief constituent of most natural gas, with
lesser amounts of ethane, propane, butane, and pentane. (ref 4)
Produced water: Formation fluids, and possibly fracturing fluids, that are produced from the
hydrocarbon-bearing formation. Sometimes the term produced water is used synonymously with
flowback.
Proppant: Sized particles (e.g., sand) mixed with fracturing fluid to hold fractures open after a
hydraulic fracturing treatment, (ref 4)
35
-------
Characterization of Hydraulic Fracturing-Related Spills
May 2015
Quality assurance / quality control: A system of procedures, checks, audits, and corrective
actions to ensure that all technical, operational, monitoring, and reporting activities are of high
quality. (ref5)
Service company: A company that assists well operators by providing specialty services, including
hydraulic fracturing.
Surface water: All water naturally open to the atmosphere (rivers, lakes, reservoirs, ponds,
streams, impoundments, seas, estuaries, etc.). (ref 5)
Well communication: Spills reported in the data sources as being caused by well communication
refer to spills on the surface that occur when pressure applied during hydraulic fracturing activities
at one well affects the production and collection of fluids at a nearby or offset well.
Well operator: A company that operates oil and gas production wells.
Glossary References
1. GWPC and ALL Consulting. 2009. Modern Shale Gas Development in the US: A Primer. Ground
Water Protection Council and ALL Consulting for US Department of Energy. Available at
http://energy.gov/fe/downloads/modern-shale-gas-development-united-states-primer.
Accessed September 16, 2014.
2. New York State Department of Environmental Conservation. 2011. Supplemental Generic
Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program
(revised draft). Well permit Issuance for Horizontal Drilling and High-Volume Hydraulic
Fracturing to Develop the Marcellus Shale and Other Low-Permeability Gas Reservoirs.
Available at http://www.dec.ny.gov/energy/75370.html. Accessed July 30, 2014.
3. Oil and Gas Mineral Services. 2010. Oil and Gas Terminology. Available at
http://www.mineralweb.com/library/oil-and-gas-terms/. Accessed May 8, 2015.
4. Schlumberger. 2015. Oilfield Glossary. Available athttp://www.glossary.oilfield.slb.com/.
Accessed May 8, 2015.
5. US EPA. 2006. Terminology Services: Terms and Acronyms. Available at http://iaspub.epa.gov/
sorjnternet/registry/termreg/searchandretrieve/termsandacronyms/search.do. Accessed
May 8, 2015.
6. US EPA. 2011. Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources. EPA/600/R-11/122. Available at http://www2.epa.gov/sites/production/files/
documents/hf_study_plan_l 102ll_final_508.pdf. Accessed July 30, 2014.
36
-------
Characterization of Hydraulic Fracturing-Related Spills May 2015
[This page intentionally left blank.]
37
-------
EPA/601/R-14/001 I May 2015 I www.epa.gov/hfstudy
oEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |