oEPA
United States
Environmental Protection
Agency
EPA/601/R-14/Q02 I May 2015 I www.epa.gov/hfstudy
Review of Well Operator Files
for Hydraulically Fractured
Oil and Gas Production Wells:
Well Design and Construction
United States Environmental Protection Agency
Office of Research and Development
-------
Well Design and Construction
May 2015
[This page intentionally left blank.]
ii
-------
Well Design and Construction
May 2015
Review of Well Operator Files for
Hydraulically Fractured Oil and Gas Production Wells:
Well Design and Construction
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC
May 2015
EPA/601/R-14/002
iii
-------
Well Design and Construction
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 Well Operator Files for
Hydraulically Fractured Oil and Gas Production Wells: Well Design and Construction. Office of Research
and Development, Washington, DC. EPA/601/R-14/002.
iv
-------
Well Design and Construction
May 2015
Table of Contents
Disclaimer 4
Table of Contents 5
List of Tables 7
List of Figures 8
Preface 10
Authors and Contributors 11
Acknowledgements 12
Executive Summary 1
1. Introduction 6
2. Well Construction Overview and Definitions 6
3. Research Methods 9
3.1. Service Company Well List 10
3.2. Survey Design 12
3.3. Well Operator Information Request 16
3.4. Data Extraction and Analysis 17
3.4.1. Well Characteristics 20
3.4.2. Construction Characteristics 21
3.4.3. Drinking Water Resources 22
3.5. Estimates of Well Design and Construction Characteristics 23
3.6. Quality Assurance and Quality Control 25
4. Analytical Results 26
4.1. Well Characteristics 26
4.2. Construction and Completion Characteristics 30
4.2.1. Casing Installation 30
4.2.2. Cement Evaluation 35
4.3. Well Construction Characteristics and Drinking Water Resources 37
4.3.1. Surface Locations of Drinking Water Resources 37
4.3.2. Protected Ground Water Resources 38
5. Potential Subsurface Fluid Movement Pathways 42
5.1. Potential Pathway A: Inside to outside 44
v
-------
Well Design and Construction
May 2015
5.2. Potential Pathway B: Along the outside of the well 47
6. Representativeness Analysis 51
7. Study Limitations 54
8. Conclusions 56
References 60
Appendix A: Survey Design and Confidence Intervals 66
A.l. Optimization Algorithm 66
A.2. Variance Estimation and Confidence Intervals 67
References 69
Appendix B: Well Operator Information Request 70
Glossary 73
Glossary References 77
vi
-------
Well Design and Construction May 2015
List of Tables
Table 1. Number of well identifiers included or not included in this study 16
Table 2. Comparison between well identifiers from the service company well list and the 323
study wells 18
Table 3. Summary of data elements generally obtained from well files provided by oil and gas well
operators 19
Table 4. Quality assurance summary for calculated cement bond indices 26
Table 5. Information submitted by operators to report the depths of protected ground water
resources at each study well location 39
vii
-------
Well Design and Construction May 2015
List of Figures
Figure 1. Generic well diagram illustrating conductor, surface, intermediate, and production
casing 9
Figure 2. Casing configurations reported in the well operators' files 10
Figure 3. Well completion types reported in the well operators' files 11
Figure 4. Counties with oil and gas production wells that were reported by nine service
companies to have been hydraulically fractured between approximately September
2009 and September 2010 13
Figure 5. Locations of the 323 study wells 17
Figure 6. (a) Age, (b) production type, and (c) orientation of oil and gas production wells
hydraulically fractured by nine service companies between approximately September
2009 and September 2010 27
Figure 7. True vertical depths of oil and gas production wells hydraulically fractured by nine
service companies between approximately September 2009 and September 2010 28
Figure 8. Measured depths of the point of shallowest hydraulic fracturing in oil and gas
production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010 29
Figure 9. Lithologies hydraulically fractured for oil and gas by nine service companies between
approximately September 2009 and September 2010 30
Figure 10. Measured depths of the bottoms of surface, intermediate, and production casing in oil
and gas production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010 31
Figure 11. Type of cementing operation conducted for each casing type found in oil and gas
production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010 32
Figure 12. Degree of cementing determined for each casing type found in oil and gas production
wells hydraulically fractured by nine service companies between approximately
September 2009 and September 2010 33
Figure 13. Percentage of the evaluated casing length that was cemented, calculated only for
partially cemented casings shown in Figure 12 34
Figure 14. The 5th, 25th, 50th, 75th, and 95th percentiles of cement bond indices calculated from
standard acoustic cement bond logs 36
Figure 15. Drinking water resources within 0.5 miles of oil and gas production wells hydraulically
fractured by nine service companies between approximately September 2009 and
September 2010 38
viii
-------
Well Design and Construction
May 2015
Figure 16. Depths of operator-reported protected ground water resources for oil and gas
production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010 40
Figure 17. Comparison of casing and cement to operator-reported protected ground water
resources among oil and gas production wells hydraulically fractured by nine service
companies between approximately September 2009 and September 2010 41
Figure 18. Estimated separation distance between the point of shallowest hydraulic fracturing
and the depth of operator-reported protected ground water resources for oil and gas
production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010 43
Figure 19. Potential well construction pathways for subsurface fluid movement 44
Figure 20. (a) Concentric casings between the inside of the well and the wellbore at different
measured depths along the well, (b) Illustrative examples of wells with different
numbers of concentric casings at example Measured Depths 1, 2, and 3 along the well 46
Figure 21. (a) Cement sheaths between the inside of the well and the wellbore at different
measured depths along the well. An estimated 2,200 wells (1,000-3,400) contain
cement whose depth interval is uncertain and are not included in panel a. (b)
Illustrative examples of wells with different numbers of cement sheaths at Measured
Depths 1, 2, and 3 along the well 47
Figure 22. (a) Cemented casing strings along the outside of the well between the point of
shallowest hydraulic fracturing and the operator-reported protected ground water
resources, (b) Illustrative examples of wells with different cemented casing strings
between the point of shallowest hydraulic fracturing and operator-reported
protected ground water resources 49
Figure 23. Comparison of well estimates developed from Drillinglnfo to well estimates developed
from well files used in this study 52
-------
Well Design and Construction
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 Operator Files for Hydraulically Fractured Oil and Gas Production Wells: Well
Design and Construction, 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.
x
-------
Well Design and Construction
May 2015
Authors and Contributors
Glen Boyd, PhD, The Cadmus Group, Inc., under contract EP-C-08-015
Susan Burden, PhD, US EPA
Guy W. Cole, MS, MPP, Student Services Contractor for US EPA, under contract EP-12-H-000389
Charles J. Hillenbrand, PhD, US EPA
Andrea T. Joffe, BS, Randstad Technologies, under contract EP-C-10-023
David Marker, PhD, Westat, under contract EP-C-10-023
Gregory Oberley, MS, US EPA
Susan L. Sharkey, MA, US EPA
Chuck Tinsley, BS, US EPA
Jose E. Torres, MS, US EPA
Nathan M. Wiser, MA, US EPA
xi
-------
Well Design and Construction
May 2015
Acknowledgements
The EPA would like to acknowledge the companies that provided data and information for this report.
The agency would also like to acknowledge Stephen Watkins (US EPA) for his efforts in the development
of this report. Assistance was provided by The Cadmus Group, Inc., under contract EP-C-08-015, Westat
under contract EP-C-10-023, and the following student services contractors: Claudia Meza-Cuadra
(contract EP-13-H-000054), Alison Singer (EP-13-H-000474), and Liabeth Yohannes (EP-14-H-000455). 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.
The EPA would also like to acknowledge the participants of the US EPA-States Technical Meeting on the
Well File Review for their insights on well design and construction. The meeting was held in Dallas,
Texas, on July 15, 2014, and included representatives from the following state agencies:
Arkansas Department of Environmental Quality
Arkansas Natural Resource Commission
Colorado Department of Public Health and Environment
Colorado Oil and Gas Conservation Commission
Louisiana Department of Natural Resources
Louisiana Office of Conservation
New Mexico Energy, Minerals, and Natural Resources Department
North Dakota Industrial Commission
Oklahoma Corporation Commission
Oklahoma Department of Environmental Quality
Oklahoma Water Resources Board
Pennsylvania Department of Environmental Protection
Railroad Commission of Texas
Texas Commission on Environmental Quality
Wyoming Department of Environmental Quality
xii
-------
Well Design and Construction
May 2015
Executive Summary
Hydraulically fractured oil and gas production wells are designed, constructed, and completed to
access and extract hydrocarbons from targeted geologic formations. Well components, such as the
casing and cement used to construct production wells, can block pathways for unintended
subsurface gas and liquid movement to ground water resources. To help understand the role of well
design and construction practices in preventing pathways for subsurface fluid movement, the EPA
conducted a survey of oil and gas production wells hydraulically fractured by nine oil and gas
service companies in the United States during 2009 and 2010. The objective of the study was to
describe, for these wells: (1) well design and construction characteristics of hydraulically fractured
oil and gas production wells, (2) the relationship of well design and construction characteristics to
drinking water resources,1 and (3) the number and relative location of well construction barriers
(i.e., casing and cement) that can block pathways for potential subsurface fluid movement.
A statistically representative sample of 323 study wells was selected from a list of well identifiers
corresponding to onshore oil and gas production wells that were reported by the nine service
companies. Drilling, construction, completion, and operation information for the selected wells was
collected from nine well operators and summarized. Results of the survey are presented as rounded
estimates of the frequency of occurrence of hydraulically fractured production well design or
construction characteristics with 95 percent confidence intervals. The results are statistically
representative of an estimated 23,200 (95 percent confidence interval: 21,400-25,000) onshore oil
and gas production wells hydraulically fractured in 2009 and 2010 by the nine service companies.
Well Design and Construction Characteristics of Hydraulically Fractured Oil and Gas Production Wells.
Oil and gas production wells hydraulically fractured by the nine service companies in 2009 and
2010 were predominately vertical wells drilled between 2000 and 2010, butalso included other
well orientations (i.e., horizontal and deviated) and wells drilled prior to 2000. True vertical depths
of the wells ranged from less than 2,000 feet below ground surface to more than 11,000 feet below
ground surface. Hydraulically fractured rock formations included sandstone, shale, carbonate, coal,
and chert
Despite variations in well design and construction practices, data collected for this study indicate
that some well design and construction characteristics are common. All wells represented in this
study had surface casing, 39 (18-65) percent2 had intermediate casing, and 94 (67-99) percent3 had
production casing. The most common casing configuration, used in 55 (33-75) percent4 of wells,
1 The EPA's Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources defined "drinking
water resources" as any body of water, ground or surface, that currently serves or in the future could serve as a source of
drinking water for public or private water supplies.
2 9,100 (2,900-15,400] wells
3 21,900 (19,200-24,600] wells
4 12,800 (7,600-18,000] wells
1
-------
Well Design and Construction
May 2015
was reported to be a surface and a production casing. Eighty-seven (68-96) percent5 of wells had
casing in the production zone that was cemented and perforated for hydraulic fracturing.
The presence or absence of cement around casing was evaluated for surface, intermediate, and
production casings. Casings were found to be either fully cemented, partially cemented, or
uncemented. Fully cemented casings were defined, in this study, as casings that had a continuous
cement sheath from the bottom of the casing to at least the next larger and overlying casing (or the
ground surface, if surface casing). Conversely, casings with no cement anywhere along the casing,
from the bottom of the casing to at least the next larger and overlying casing (or ground surface),
were defined as uncemented. Partially cemented casings were defined as casings that had some
portion of the casing that was cemented from the bottom of the casing to at least the next larger and
overlying casing (or ground surface), but were not fully cemented. Ninety-three (87-96) percent6 of
surface casings were fully cemented, and 80 (57-92) percent7 of intermediate casings were fully
cemented. Fifty-two (33-70) percent8 of production casings were partially cemented, while 36 (I8-
60) percent9 were fully cemented. When partially cemented, generally more than 50 percent of the
measured length of the production casing (between the bottom of the casing and the next larger
and overlying casing) was cemented.
Relationship to Drinking Water Resources. Drinking water resources (i.e., surface water bodies,
public water supply intakes, and ground water wells) were commonly found within a 0.5 mile
radius of study wellhead locations. Eighty-two (63-92) percent10 of the wells represented in this
study were located within 0.5 miles of a surface water resource (i.e., lake, pond, or river); this
report does not identify whether surface water resources currently serve as drinking water
resources. Fewer production wells were located within 0.5 miles of either a private ground water
well [13 (7-23) percent11] or a public water supply well [2 (1-10) percent12]. An estimated 74 (51-
88) percent13 of wells had more than one water feature within the 0.5 mile radius.
Since there is variation among states with respect to the definition of protected ground water, this
study relied solely on depths to protected ground water resources that were provided by well
operators. In the majority of cases, operator-reported depths to protected ground water resources
were based on state or federal authorization documents (e.g., permits to drill or permit
applications). Operator-reported depths to protected ground water resources ranged from just
5 20,200 (17,500-23,000] wells
s 21,500 (19,500-23,600] wells
7 7,300 (600-13,900] wells
8 10,900 (6,900-14,900] wells
9 8,300 (3,800-12,800] wells
i° 18,900 (15,200-22,600] wells
11 3,000 (900-5,100] wells
12 600 (10-1,500] wells
13 17,100 (12,400-21,900] wells
2
-------
Well Design and Construction
May 2015
below ground surface to 8,000 feet deep. Ninety-two (81-97) percent14 of wells passed through
operator-reported protected ground water resources no greater than 2,000 feet deep.
Cement was placed in the annular space between the outside of the casing and the protected
ground water resources reported by well operators in most wells. Surface casing, which was found
to be fully cemented in 93 (87-96) percent15 of wells, extended below the base of the operator-
reported protected ground water in 55 (35-73) percent16 of wells. In an additional 28 (10-55)
percent17 of wells, the operator-reported protected ground water resources were fully covered by
the next cemented casing string. A portion of the annular space between casing and the operator-
reported protected ground water resources was uncemented in 3 (0.5-13) percent18 of wells.
Casing and Cement as Barriers to Potential Subsurface Fluid Movement. Subsurface fluid movement
depends on a number of factors, including the existence of a pathway, the presence of a fluid, and a
driving force (e.g., pressure differential). This report focuses only on the potential for wellbore
pathways to exist and does not address other possible pathways or the role of other factors. The
presence of pathways alone does not indicate that subsurface gas and liquid movement will occur
or is occurring, because fluid movement also depends on the other factors identified above. Two
potential pathways related to well construction were considered: from the inside of the well to the
outside and along the outside of the well. Subsurface fluid movement to ground water resources
may occur though either potential pathway or a combination of both potential pathways. The
presence or absence of casing and cement along the well was evaluated to determine the number of
well construction barriers to subsurface fluid movement via these potential pathways.
The first potential pathway—from the inside of the well to the outside—can occur at any point
along the length of the well and depends, in part, on the presence or absence of casing and cement.
At depths corresponding to the targeted geologic formation, this pathway is created when casing
and cement are perforated for hydraulic fracturing or when a well is completed with an open hole.
At other depths, multiple casing and cement barriers can prevent this pathway from forming,
because all barriers would need to fail in order for a pathway from the inside of the well to the
outside to occur. Wells represented in this study had a range in the number of casing and cement
barriers, from zero in open hole completions to six when surface, intermediate, and production
casing were each cemented to the surface. The most common number of barriers, at any point along
a well, was either two (one casing string and one cement sheath) or three (two casing strings and
one cement sheath). Nearly all wells represented in this study had perforations used for hydraulic
fracturing that were placed deeper than the base of the operator-reported protected ground water
resources; 0.4 (0.1-3) percent19 of wells had perforations placed shallower than the base of the
protected ground water resource reported by the operator.
" 21,400 (18,900-24,000] wells
is 21,500 (19,500-23,600] wells
is 12,600 (8,000-17,300] wells
17 6,400 (500-12,300] wells
is 600 (10-1,800] wells
w 90 (10-300] wells
3
-------
Well Design and Construction
May 2015
The second potential pathway considered—along the outside of the well—can also occur at any
point along the well's length and depends upon the presence or absence of cement and the quality
of the cement present. More than half of the wells represented in this study had two or more
cemented casings between the point of shallowest hydraulic fracturing and the operator-reported
protected ground water resources. This indicates that there were often two or more barriers to
subsurface fluid movement along the outside of a well, from the targeted geologic formation to
protected ground water resources reported by well operators. Multiple cemented casings between
the shallowest point of hydraulic fracturing and operator-reported protected ground water
resources did not preclude the presence of uncemented intervals along the outside of the well,
which can be pathways for fluid movement along the outside of the well. Sixty-six (44-83) percent20
of wells had one or more uncemented intervals along the outside of the well, from the bottom of the
well to the ground surface, while 29 (13-53) percent21 were fully cemented over the same interval.
Study Limitations. The survey design and data collection process may have implications for the
interpretation of the results presented in this report The results are statistically representative of
oil and gas production wells hydraulically fractured by nine oil and gas service companies in the
continental United States during 2009 and 2010. The extentto which these results may be
statistically representative of all production wells hydraulically fractured in the United States
during the same time period could not be determined. Nevertheless, comparisons between wells in
this study and other data on oil and gas production in the United States during 2009 and 2010
suggest that observations made in this report are likely indicative of oil and gas production wells
hydraulically fractured during this time period.
Estimates of the frequency of occurrence of well design and construction characteristics are
presented at the national scale. Estimates may be different for different regions of the country,
because of differences in local geologic characteristics, state regulations, and company preferences.
It is also possible that the estimates presented in this report may not apply to wells constructed and
hydraulically fractured after 2010, if well design and construction practices have changed (e.g., a
greater proportion of horizontal well completions). Additionally, the results presented in this
report are generated from data provided by oil and gas well operators. The EPA did not attempt to
independently and systematically verify data supplied by operators. Consequently, the study
results, which include comparisons of operator-reported protected ground water resources to well
construction characteristics, are of the same quality as the supplied data.
Key Findings. This report presents the results of a survey of onshore oil and gas production wells
hydraulically fractured in the continental United States during 2009 and 2010, using data provided
by well operators. Two potential pathways for subsurface fluid movement were examined—from
the inside of the well to the outside and along the outside of the well. The following key findings
contribute to an understanding of the role of well design and construction practices with respect to
these pathways:
20 15,300 (10,500-20,100] wells
21 6,800 (1,600-11,900] wells
4
-------
Well Design and Construction
May 2015
The wells generally had multiple layers of casing and cement that can act as barriers to
subsurface fluid movement by interrupting pathways for potential subsurface fluid movement.
Multiple casing and cement barriers can prevent pathways from forming, because all
barriers would need to fail in order for a pathway for potential subsurface fluid movement
to occur. The most common number of barriers to potential subsurface fluid movement
from inside of the well to the outside, at any point along the well, was either two (one casing
string and one cement sheath) or three (two casing strings and one cement sheath).
Additionally, there were often two or more barriers (i.e., cemented casings) to potential
subsurface fluid movement along the outside of a well, from the targeted geologic formation
to operator-reported protected ground water resources.
While multiple barriers were often present in hydraulically fractured oil and gas production
wells; pathways for potential subsurface fluid movement were identified in some wells.
Uncemented intervals have been shown to be pathways for subsurface fluid movement
along the outside of the well. An estimated 66 (44-83) percent22 of wells had one or more
uncemented intervals, and 3 (0.5-13) percent23 of wells had uncemented intervals within
the operator-reported protected ground water resources. Casing perforations placed at
depths shallower than the base of operator-reported protected ground water resources can
create a pathway for fluids to flow from the inside of the well to a ground water resource, if
a ground water resource is present at that depth. An estimated 0.4 (0.1-3) percent24 of wells
had perforations used for hydraulic fracturing that were placed shallower than the base of
the operator-reported protected ground water resource.
These results, as well as other information on well characteristics, the relationship between
drinking water resources and hydraulically fractured wells, and casing and cement barriers that can
prevent subsurface fluid movement, highlight important factors that should be considered when
assessing the potential impacts of hydraulically fractured oil and gas production wells on drinking
water resources.
22 15,300 (10,500-20,100] wells
23 600 (10-1,800] wells
24 90 (10-300] wells
5
-------
Well Design and Construction
May 2015
1. Introduction
The importance of oil and gas production well design and construction in isolating and protecting
ground water resources is well-known [Ground Water Protection Council (GWPC), 2014; GWPC and
ALL Consulting, 2009; King and King, 2013], Several studies, however, suggest that the construction
of oil and gas production wells may introduce pathways along which fluids may move, potentially
leading to impacts to drinking water resources (Harrison, 1983,1985; Jackson etal., 2013a; Jackson
etal., 2013b; Ohio Department of Natural Resources, 2008; Osborn etal., 2011; Van Stempvoortet
al., 2005; Watson and Bachu, 2009). For example, the Ohio Department of Natural Resources (2008)
determined that inadequately cemented casing contributed to natural gas migration to a ground
water resource by creating a pathway that connected a high pressure gas zone to the ground water
resource. As demonstrated by this case, subsurface fluid movement depends on many factors,
including the existence of a pathway, the presence of a fluid, and a driving force (e.g., pressure
differential).
The U.S. Environmental Protection Agency's (EPA) Plan to Study the Potential Impacts of Hydraulic
Fracturing on Drinking Water Resources identified the potential importance of well construction
practices in preventing pathways for subsurface fluid movement before, during, and after hydraulic
fracturing [US Environmental Protection Agency (US EPA), 2011], To help understand the role of
well design and construction practices in preventing pathways for potential subsurface fluid
movement, the EPA conducted a statistical survey of oil and gas production wells hydraulically
fractured by nine service companies in the United States during 2009 and 2010. The objective of
the study was to describe, for these wells: (1) well design and construction characteristics of
hydraulically fractured oil and gas production wells, (2) the relationship of well design and
construction characteristics to drinking water resources,25 and (3) the number and relative location
of well construction barriers (i.e., casing and cement) that can block pathways for potential
subsurface fluid movement This report focuses on the potential for wellbore pathways to exist and
does not address other possible pathways or the role of other factors (i.e., the presence of a fluid or
a driving force) in subsurface fluid movement Results from this study provide a summary of well
design and construction characteristics associated with oil and gas production wells hydraulically
fractured in the United States during 2009 and 2010, using data provided by well operators.
2. Well Construction Overview and Definitions
Most oil and gas production well locations have ground water and/or hydrocarbons in the pore
spaces of rock formations in the subsurface. Ground water is often variable in quality, generally
containing greater amounts of dissolved solids with greater depth, although localized geologic
structure and hydrogeologic history can also influence its quality (Chebotarev, 1955; Freeze and
Cherry, 1979). In this report, a ground water resource is considered to be any geologic formation
containing ground water. Certain ground water resources may be considered for protection as
drinking water resources, depending on ground water quality and local regulatory requirements.
25 The EPA's Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources (2011] defined
"drinking water resources" as any body of water, ground or surface, that currently serves or in the future could serve as a
source of drinking water for public or private water supplies (US EPA, 2011].
6
-------
Well Design and Construction
May 2015
The type and amount of hydrocarbons also vary in the subsurface, with certain geologic formations
containing more hydrocarbons than others. Depending on the price of hydrocarbon resources, it
maybe considered economically viable to extract hydrocarbons atone or more depths. These
resources become the targeted geologic formations for oil and gas companies. An oil and gas
production well that penetrates through a ground water resource can act as a pathway for potential
subsurface fluid movement by connecting the resource to other formations that may contain
hydrocarbons, saline water, or both. Well components, such as the casing and cement used to
construct production wells, can block pathways for unintended subsurface gas and liquid
movement to ground water resources. When these pathways are blocked, the well is said to have
"zonal isolation." (Baker, 1979; Bellabarba etal., 2008; Smith, 1976)
Oil and gas production wells are designed, constructed, and completed to access and extract
hydrocarbons from targeted geologic formations. Well construction includes drilling, casing
installation, and cementing the casing to the wellbore wall. Activities associated with these steps
are generally determined based on the subsurface environment, including the subsurface pressures
encountered between the surface and the bottom of the well (Ross and King, 2007; Spellman,
2013). Industry best management practices and state regulations that address regional and local
characteristics can also influence well design and construction (GWPC, 2014; Spellman, 2013). Well
completion activities prepare a well for production and include casing perforation, stimulation
(including, but not limited to, hydraulic fracturing), and installation of production tubing if desired.
Production well completion is generally influenced by target zone characteristics, such as porosity,
permeability, and lithology (Martin and Valko, 2007; Ross and King, 2007). Existing oil and gas
production wells can be recompleted in hydrocarbon-bearing zones that were not the original
targeted geologic resource (International Association of Drilling Contractors, 2015). Original well
design and construction characteristics may change when a well is recompleted. For example, well
files reviewed for this study showed changes in the location of casing perforations, the addition of
cement, and the removal of uncemented casing.
In general, oil and gas production wells are constructed by repeating several basic steps. In the first
step, a hole is drilled to a pre-determined depth or until a geologic target is reached. In the second
step, a steel pipe is lowered into the hole. When the steel pipe extends from the bottom of the hole
to the ground surface, it is referred to as a "casing." When the steel pipe extends from one
subsurface depth to another subsurface depth within the well, it is often referred to as a "liner." In
the last step, cement is placed between the outside of the casing and the inside of the drilled hole
(i.e., the wellbore) to seal off this annular space (Baker, 1979; Smith, 1976). The cement also
provides support behind the casing and to protect it from corrosive fluids in the subsurface (Baker,
1979; Smith, 1976). The process repeats, with the next hole drilled using a smaller diameter drill bit
that fits inside the existing casing.
Casings installed at different points in time and at different depths have different names. For this
report, casings are defined as follows:
Conductor casing is generally less than 100 feet deep and is often uncemented. Its purpose
is to prevent the in-fill of unconsolidated dirt and rock in the uppermost few feet of the
drilled hole (Baker, 1979; Devereux, 1998).
7
-------
Well Design and Construction
May 2015
Surface casing is the shallowest cemented casing, with the widest diameter. Cemented
surface casing generally serves as an anchor for blowout protection equipment and to seal
off certain ground water resources (Baker, 1979).
Intermediate casing, when cemented, generally seals off intermediate depths and geologic
formations that may have considerably different reservoir pressures than deeper zones to
be drilled (Baker, 1979; Devereux, 1998). This can include coal mine zones, gas storage
zones, intermediate-depth hydrocarbon production zones, or ground water resources
located at depths below the surface casing.
Production casing is the deepest casing set and serves primarily as the conduit for
producing fluids. When cemented to the wellbore, this casing can also serve to seal off other
subsurface zones, including ground water resources (Baker, 1979; Devereux, 1998).
Production casing is secured to the wellbore using either cement or formation packers,
which are devices that inflate or mechanically seal the annular space between the casing
and the wellbore. For this report, any casing or liner with production perforations was
considered to be a "production casing." In some cases, this can include casing identified as
"intermediate" by the operator, but later perforated for production purposes.
Figure 1 provides a simplified, generic well diagram illustrating the casings defined above. Note
that not all hydrocarbon-bearing zones in the subsurface are necessarily produced or stimulated
using hydraulic fracturing and that not all ground water resources may be considered protected.
This figure is meant to show relationships between well casings and subsurface resources, but
should not be interpreted to imply that all wells exhibit similar conditions in the subsurface.
Oil and gas production wells are constructed using different combinations of the casings and liners
described above. Figure 2 illustrates the different variations of casings and liners observed in this
study. Given the array of possible casing configurations, Figure 2 may not capture all configurations
used in hydraulically fractured oil and gas production wells.
Well completion in the targeted geologic formation can also vary. Figure 3 depicts the three types of
well completions observed in this study: cemented casing, formation packer, and open hole
completions. Cemented casing completions have both casing and cement in the targeted geologic
formation. In these completions, the casing and cement are typically perforated, and hydraulic
fracturing is performed within the perforated interval. Formation packer completions also have
casing, but use formation packers that swell to seal the annulus between the production casing and
the wellbore. In these completions, the bottom portion of the production casing, located between
formation packers, is equipped with ports that can be opened using pressure applied through the
casing. Hydraulic fracturing fluid is pumped through these ports to fracture the surrounding rock
formation. Open hole completions contain no casing and, therefore, no cement Hydraulic fracturing
in an open hole completion occurs within the open hole interval and may be conducted using a
temporary casing (i.e., a temporary frac string). A temporary frac string is lowered into the well and
secured at its base using a packer or by latching into equipment existing deeper in the well.
Hydraulic fracturing takes place through the temporary frac string, which is removed when
8
-------
Well Design and Construction
May 2015
Diagram not to scale
Ground Surface
Conductor
Protected ground
water resource
Surface
Hydrocarbon-bearing zone
Water-yielding zone
Intermediate
Legend
¦ Cement
^ Casing
— Wellbore
Induced fracture
Targeted geologic zone
Gas or oil
Production
Figure 1. Generic well diagram illustrating conductor, surface, intermediate, and production casing. Relative
positions of geologic zones and well construction components vary from well to well and are included only for
illustrative purposes. Although shown in this diagram as occurring in singular zones, water is frequently found
throughout the entire geologic column in different quantities and qualities, including in the zones containing
hydrocarbons. Hydrocarbon-bearing zones may be considered uneconomic and may be left undeveloped.
Hydraulic fracturing takes place for the purpose of producing hydrocarbons in targeted geologic zones.
hydraulic fracturing is complete. Temporary frac strings may also be used in wells with casing that
may not withstand the pressures applied during hydraulic fracturing.
After hydraulic fracturing a smaller diameter steel pipe, called "tubing," may be set within the
production casing. In open hole settings, the tubing may extend further into the well than the
deepest cemented casing. Depending on the installation of the tubing, production fluids may flow
up only the inside of the tubing or may flow up both the inside and outside of the tubing to the
wellhead.
3. Research Methods
A survey of onshore oil and gas production wells in the continental United States was conducted to
better understand well design and construction characteristics associated with hydraulically
fractured wells and the relationship between those characteristics and drinking water resources. A
statistically representative sample of production wells was selected from a list of oil and gas
production wells that were reported by nine service companies to have been hydraulically
9
-------
Well Design and Construction
May 2015
Ground Surface
Surface
Legend
J Casing or liner
0 Formation packer
Surface
Production
Surface
Intermediate
Production
Surface
Intermediate
Liner
Note: Not to scale. Conductor casing, wellbore holes, and cement not shown.
i |
Surface
Intermediate
Production
Liner
Surface
Intermediate
Liner
Production
Figure 2. Casing configurations reported in the well operators' files. Casings are steel pipes that extend from
ground surface to a predetermined depth and include surface, intermediate, and production casing. Liners are
steel pipes that are attached to the base of a casing and extend to a predetermined depth. Casing configurations
are depicted for vertical well orientations, but may occur in wells with different orientations (e.g., horizontal).
fractured between approximately September 2009 and September 2010. Drilling, construction,
completion, and operation information from the selected wells was collected from well operators
and summarized.26 Results of the survey are presented as estimates of the frequency of occurrence
of hydraulically fractured production well design or construction characteristics with 95 percent
confidence intervals. This section describes the survey design, the information requested, and the
analyses conducted.
3.1. Service Company Well List
Information needed to compile a comprehensive list of hydraulically fractured oil and gas
production wells in the United States is not generally available. Therefore, a list of oil and gas
production wells that were reported by nine service companies to have been hydraulically
fractured between approximately September 2009 and September 2010 was compiled
(subsequently referred to as the "service company well list"). The information was provided to the
EPA by the nine service companies in response to an information request sent in September 2010
(US EPA, 2010). As part of the information request, each company was asked to identify all sites
where they had provided hydraulic fracturing services in the year prior to September 2010.
26 While a portion of the data needed for this project are reported to state oil and gas agencies, the complete dataset is
available only in the records kept by oil and gas well operators.
10
-------
Well Design and Construction
May 2015
Ground Surface
Legend
¦ Cement
^ Casing
— Wellbore
Inducedfracture
B Formation packer
Cemented Casing
Completion
Perforations
11
Formation Packer
Completion
Open Hole
Completion
Lili
J
Note: Not to scale. Conductor casing not shown.
Figure 3. Well completion types reported in the well operators' files. Well completions are depicted for horizontal
well orientations, but may occur in wells with different orientations (e.g., vertical).
The nine service companies were selected from a list of 14 hydraulic fracturing service companies
identified by the U.S. House of Representatives' Committee on Energy and Commerce (Waxman et
al., 2011).27 The companies selected by the committee included, at that time, the three major
hydraulic fracturing service companies, five smaller companies that comprised a growing share of
the market, and six additional companies that were selected to assess a broader range of industry
practices (Waxman and Markey, 2010; Waxman et al., 2011). Of the 14 hydraulic fracturing service
companies identified by the committee, the EPA selected the three major service companies, which
included BJ Services Company; Halliburton Energy Services, Inc.; and Schlumberger Technology
Corporation. The six remaining service companies were selected to reflect a range of company sizes
and geographic diversity (US EPA, 2011; US EPA, 2012). The nine companies selected to receive the
EPA's September 2010 information request 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.28
27 The EPA was limited to nine service companies (and nine well operators] because of the Paperwork Reduction Act of
1995, which restricts the number of information requests to nine entities per set of queries, unless pre-approved by the
Office of Management and Budget.
28 Four of the nine service companies that reported information to the EPA have been acquired by other companies since
2010. Baker Hughes, Inc., purchased BJ Services Company, Inc., and Patterson-UTI Energy purchased the pressure
pumping business from Key Energy Services in 2010. Superior Energy Services acquired Complete Production Services,
Inc., in 2012. As of March 2015, Superior Well Services is part of C&J Energy Services.
11
-------
Well Design and Construction
May 2015
The nine hydraulic fracturing service companies identified 24,925 well identifiers in response to
the EPA's information request29 Service companies provided, for each well identifier, the well
operator's name, the well's state and county location, and the date(s) of hydraulic fracturing.30 Well
identifiers were reported in 30 of the 33 states (Figure 4) that produced crude oil and/or natural
gas in 2009 and 2010 [US Energy Information Administration (US EIA), 2014a, b]. The service
company well list did not contain well identifiers in Florida, Missouri, and Oregon, which produced
crude oil and/or natural gas in 2009.31
Four of the nine service companies, including the three major service companies, individually
reported well identifiers in 17 or more states; well identifiers from these four companies made up
82 percent of the well identifiers on the service company well list.32 The same four companies
reported hydraulic fracturing operations in 29 of the 33 oil and gas producing states in the United
States in 2009. As such, the service company well list likely resembled the geographic distribution
of oil and gas production in the United States at that time.33 It was not possible to determine
whether the service company well list was representative of all hydraulic fracturing operations that
occurred between approximately September 2009 and September 2010, because the information
needed to compile a comprehensive list of hydraulically fractured oil and gas production wells in
the United States does not generally exist.
3.2. Survey Design
A two-stage stratified sampling approach was used to select a diverse set of wells following the
principles described in Lohr (2010). As described below, well operators of various sizes were
selected in the first stage, while their well identifiers from different geographic areas were selected
in the second stage.34 To the extent that well design, construction, or hydraulic fracturing
characteristics vary by company size or geography, this stratification improves the accuracy of the
29 Well identifiers were either API numbers, well names, job numbers, or other codes assigned by the service company.
API numbers are unique 10-digit numbers generally assigned to oil and gas production wells by state oil and gas agencies.
30 Approximately 75 percent (18,798] of the 24,925 well identifiers in the service company well list had reported dates
for hydraulic fracturing between September 1,2009, and September 30,2010. There were 626 well identifiers that had
reported hydraulic fracturing dates before September 1,2009, and there were 1,213 well identifiers that had reported
hydraulic fracturing dates after September 30,2010. There were 4,288 well identifiers with no reported hydraulic
fracturing date. These wells were assumed to have been hydraulically fractured between September 2009 and September
2010, because the EPA requested that the service companies provide a list of wells for which hydraulic fracturing was
conducted within one year prior to the date of the information request letter (September 2010].
31 Collectively, production from these three states was less than 0.1 percent of all crude oil or natural gas production in
the United States in 2009 and 2010 (US EIA, 2014a, b].
32 The three major service companies contributed 17,065 well identifiers to the service company well list (68 percent of
24,925],
33 The particular well densities shown in Figure 4 would be expected to change if the six smaller service companies
chosen by the EPA were different. However, since the three major service companies provided the majority of the well
identifiers, which were spread across the United States, the overall geographic distribution of wells would not be
expected to change very much.
34 The sample of well identifiers from the nine operators was selected using a combination of a two-stage and two-phase
sampling approach [Chapters 5 and 12 in Lohr (2010]]. The variance estimator used is appropriate for the combined type
of sample (Appendix A].
12
-------
Well Design and Construction
May 2015
Mr
Number of Hydraulically Fractured Oil and Gas Wells per County
1-20
21-100
M101 - 250
251-500
501 - 2300
Shale Basins
0 250 500 Miles
1 I I I I I I I I
I I I I I I
400
I
800 Kilometers
Projection: Albers Equal Area Conic
Figure 4. Counties with oil and gas production wells that were reported by nine service companies to have been
hydraulically fractured between approximately September 2009 and September 2010. Fewer than 20 well
identifiers were reported in Alaska.
estimates presented in this report If the characteristics are unrelated to company size or
geography, the estimates are as accurate as those from an unstratified sample design.
To setup the two-stage stratified sampling, all well identifiers on the service company well list
were assigned to one of 24 geographic areas using an April 2011 map of current and prospective
shale oil and gas plays within the lower 48 states (subsequently referred to as the "shale play map")
(US EIA, 2011). This created an operator-shale play combination for each well identifier based on
the well identifier's reported county and state location. Counties containing at least one well
identifier from the service company well list were assigned to a geographic area defined as either a
13
-------
Well Design and Construction
May 2015
single shale play, a cluster of shale plays,35 or no shale play. If any portion of a county was within
one of the shale play boundaries on the map, the entire county was assigned to that shale play or
shale play cluster. If no portion of the county was within a shale play, the county was not assigned
to any shale play. Because hydraulic fracturing is used to stimulate oil and gas production from
many kinds of geologic formations (Gupta and Valko, 2007), well identifiers located in a county
overlying a shale play were not assumed to correspond to wells producing oil or gas from the
designated shale play. The shale play map was used solely to identify geographic areas that were
later used to select well identifiers for the survey.
Well identifiers in counties assigned to a shale play or a shale play cluster were subsequently
grouped into larger geographic regions—either East, South, or West. Well identifiers in counties
outside of the mapped shale play boundaries were grouped into a separate region (i.e., the "Other"
region). Forty-six well identifiers had unknown counties; these well identifiers were not assigned to
any shale play or cluster of shale plays and were assigned to the "Other" region. Geographic regions
were used to identify well operators, while the shale play designations influenced the selection of
their well identifiers, as described below.
Operator Selection. Nine well operators of various sizes were selected from 1,146 well operators
identified in the service company well list Operator size was defined by the number of well
identifiers each company had on the service company well list Well operators with 10 or more well
identifiers were categorized as either "large," "medium," or "small." Large operators were defined
as those that accounted for the top 50 percent of the well identifiers on the service company well
list, medium operators for the next 25 percent, and small operators for the last 25 percent As a
result, there were 17 large operators, 86 medium operators, and 163 small operators, for a total of
266 well operators that contributed 22,573 different well identifiers to the service company well
list (91 percent of 24,925). Each large operator was assigned to the geographic region (i.e., East,
South, West, and Other) that contained the largest proportion of its well identifiers to ensure that
the final selection of well identifiers would have geographic diversity among large well operators.36
There were 880 well operators with fewer than 10 well identifiers on the service company well list.
These operators, and their corresponding well identifiers, were excluded from the operator
selection process. This ensured that selected operators would each have at least 10 well identifiers
eligible for selection in the second stage, which improved the accuracy of the estimates presented in
this report The 880 operators retained a statistical presence in the analysis through the
assumption that their wells had characteristics similar to wells from the 163 small operators, as
described in Section 3.5.
35 Shale play clusters were defined as follows: Marcellus, Utica, and Devonian as Marcellus; Floyd-Chattanooga, Floyd-
Neal, Chattanooga, and Conasauga as Floyd-Chattanooga; Avalon-Bone Spring and Barnett-Woodford as Avalon-
Woodford; Woodford and Woodford-Caney as Woodford-Caney; Monterey and Monterey-Temblor as Monterey; Hermosa
and Mancos as Hermosa-Mancos.
36 Two of the 17 large operators were assigned to the region with their second-largest proportion of wells so that each
geographic region had four large operators.
14
-------
Well Design and Construction
May 2015
Large, medium, and small operators were selected from the list of 266 well operators. One large
well operator was randomly chosen from each of the geographic regions (i.e., one large operator
from each of the East, South, West, and Other regions), for a total of four large operators. Two
medium operators and three small operators were also chosen at random, with no preference for
geographic region. The nine selected 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.37'38
Well Identifier Selection. The nine operators had a total of 2,455 well identifiers in 15 different shale
plays or shale play clusters (including the "non-shale play").39 Four hundred well identifiers from
the nine selected operators were initially chosen for this study. The selection of 400 well identifiers
required balancing two goals: maximizing the geographic diversity of wells and maximizing the
precision of any forthcoming statistical estimates. The number of well identifiers to be selected
from each operator was determined using the optimization algorithm described in Appendix A.
Briefly, the algorithm evaluated the statistical precision of different sample sizes selected from 31
unique operator-shale play combinations. Once the sample size for each operator-shale play
combination was identified, a random sample of well identifiers was selected from all well
identifiers assigned to that shale play or shale play cluster from that operator.
Due to resource and time constraints, 50 of the 400 selected well identifiers were randomly
removed,40 and information was collected from the nine well operators for the remaining 350 well
identifiers. Data for 323 well identifiers that corresponded to hydraulically fractured oil and gas
production wells were ultimately reviewed. Table 1 summarizes the number of well identifiers
excluded from this study and provides explanations for their exclusion. The 323 study wells reflect
the geographic distribution of well identifiers reported by the nine hydraulic fracturing service
companies (Figure 5).
Because only nine operators were selected, not all shale plays or shale play clusters were able to be
sampled. Table 2 provides a comparison between the service company well list and the 323 study
wells.41 When compared with selecting more operators and their wells, the nine operator
37 ConocoPhillips; EQT Corporation; Noble Energy, Inc.; and Williams Production Company, LLC, were considered "large"
operators. Clayton Williams Energy, Inc., and SandRidge Exploration and Production, LLC, were considered "medium"
operators. Hogback Exploration, Inc.; Laramie Energy II, LLC; and MDS Energy, Ltd., were considered "small" operators.
38 As of March 2015, Williams Production Company, LLC, is WPX Energy, Inc.
39 The 2,455 well identifiers were in 31 unique operator-shale play or operator-shale play cluster combinations (including
an operator-non-shale play combination]. The final sample of 323 oil and gas production wells included well identifiers
from 27 ofthe 31 combinations.
40 Well identifiers were first sorted by operator and shale play, and then every eighth well identifier was dropped. This
process ensured that well identifiers were removed independent of well operator and geographic location.
41 The nine operators were reported to have well identifiers in 15 shale plays or shale play clusters, including the non-
shale play. The 350 well identifiers initially selected included all 15 shale plays or shale play clusters. The 323
hydraulically fractured oil and gas production wells are located in 13 ofthe 15 shale plays or shale play clusters. Two
shale plays or shale play clusters were not included, because the selected well identifiers corresponded to wells not
hydraulically fractured or not operated by one ofthe nine companies (Table 1].
15
-------
Well Design and Construction
May 2015
Table 1. Number of well identifiers included or not included in this study.
Category
Number of Well Identifiers
Included in this study
Oil and gas production wells
Hydraulically fractured
323
Not included in this study
Oil and gas production wells
Duplicate well
13
Not hydraulically fractured
6
Not drilled
1
Not located in the United States
1
Operated by a different company
1
Hydraulically fractured injection well*
4
Not an oil and gas production well
1
Total number of well identifiers
350
* Class II underground injection control wells
restriction and its subsequent limitation on potential well locations contributed to a lower
precision in the results of the well characteristics estimated in this report Any well characteristics
not identified by the study wells are not represented in this report.
3.3. Well Operator Information Request
An information request letter was sent in August 2011 to the nine operators identified above. For
each of the 350 well identifiers, the EPA requested 24 distinct items organized into five topic areas:
(1) geologic maps and cross sections; (2) drilling and completion information; (3) water quality,
volume, and disposition; (4) hydraulic fracturing procedures and reports; and (5) environmental
releases. The items requested by the EPA are listed in Appendix B. Operators were asked to certify
that, to the best of their knowledge, the information submitted in response to the requested items
was true, accurate, and complete.
Approximately 9,670 electronic files and four paper files were received in response to the August
2011 information request The information in these files was compiled into a single "well file" for
each well. In September 2013, follow up letters were sent to each of the operators asking for
information not found in the original submissions. Additional information provided by the well
operators was added to the corresponding well files. In a few limited cases, the EPA successfully
searched state oil and gas websites for information that was requested, but not provided by the
operator.42
42 The EPA searched websites for additional information on basic well characteristics, such as latitude/longitude
coordinates, production type, or well orientation.
16
-------
Well Design and Construction
May 2015
•W
• Study Wells
Number of Hydraulically Fractured Oil and Gas Wells per County
1-20
21-100
¦H 101 - 250
251-500
= 501 - 2300
Shale Basins
0 250 500 Miles
1 I I I I I I I I
I I I I I ¦ I I I
0 400 800 Kilometers
Projection: Albers Equal Area Conic
Figure 5. Locations of the 323 study wells. The study wells were selected from a list of wells that were reported by
nine service companies to have been hydraulically fractured between approximately September 2009 and
September 2010. County-level well densities from the service company well list are shown for comparison. Fewer
than 20 well identifiers were reported in Alaska; no study wells were located in Alaska.
Some of the data received were claimed as confidential business information under the Toxic
Substances Control Act Through two separate letters (sent in August 2012 and September 2013),
the EPA worked with the operators to summarize and present the data in this report in a way that
protects their claims of confidentiality.43
3.4. Data Extraction and Analysis
This report summarizes data from two of the five topic areas included in the information request
(geologic maps and cross sections, and drilling and completion information) and includes some
43 Non-confidential business information provided by the well operators is available at http://www.regulations.gov/
#!searchResults;rpp=10;po=0;s=epa-hq-ord-2010-0674.
17
-------
Well Design and Construction
May 2015
Table 2. Comparison between well identifiers from the service company well list and the 323 study wells. Well
identifiers were assigned to geographic areas defined by shale plays, shale play clusters, or no shale play. Wells
assigned to these geographic areas are not assumed to produce oil and gas from the assigned shale play(s). "NA"
indicates not applicable.
Geographic
Region
Geographic Area (Shale Plays
or Shale Play Clusters)
Number of
Well Identifiers
Number of
Study Wells
East
Antrim
262
Marcellus
3,484
50
New Albany
23
South
Avalon-Woodford
1,634
34
Barnett
2,939
33
Bend
<10
Eagle Ford
860
<10
Excello-Mulky
40
Fayetteville
873
14
Floyd-Chattanooga
273
Haynesville-Bossier
1,383
<10
Tuscaloosa
24
Woodford-Caney
606
<10
West
Bakken
778
<10
Cody
<10
Gammon
40
Hermosa-Mancos
864
<10
HilliardBaxterMancos-Niobrara
1,096
<10
Lewis
320
17
Monterey
726
Mowry
76
Niobrara Fm
2,288
56
Niobrara-Mowry
121
Pierre-Niobrara
50
Other
NA
6,161
91
Total
NA
24,925
323
information from a third topic area (hydraulic fracturing procedures and reports).44 Data were
extracted from various documents contained in the well files, including driller's reports,45 wellbore
diagrams, maps, and completion reports (Table 3). Data obtained from the well files were also used
to calculate or determine other values, including cement bond indices and some stimulated
lithologies. The extracted and calculated data were recorded and organized in a database that was
then used to develop the figures and tables presented throughout this report
44 Information collected about environmental releases was used in the EPA's Review of State and Industry Spill Data:
Characterization of Hydraulic Fracturing-Related Spills (US EPA, 2015].
45 Driller's reports are daily logs of the activities at a well and include details on the well's drilling, casing, and cementing
history from surface to total depth.
18
-------
Well Design and Construction
May 2015
Table 3. Summary of data elements generally obtained from well files provided by oil and gas well operators. Data
used in this study were either extracted directly from an information source or calculated using data provided in
the information sources.
Data Element
General Information Source(s)
Extracted or
Calculated
Well Characteristics
Well location
Maps, plats, state oil and gas agency
websites
Extracted
Production type
Completion reports, state oil and gas
agency websites
Calculated /
Extracted
Wellbore orientation
Driller's reports, wellbore diagrams,
deviation surveys, state oil and gas
agency websites
Calculated /
Extracted
Well depth
Driller's reports, deviation surveys
Calculated /
Extracted
Perforation locations and purpose
Driller's reports, completion reports
Extracted
Stimulated lithology
Open hole logs, mud logs
Calculated /
Extracted
Construction Characteristics
Casings used and depth
Driller's reports, wellbore diagrams,
casing tallies
Extracted
Primary cement used behind casing
Driller's reports, cementing tickets
Extracted
Secondary cement used behind casing
Driller's reports, cementing tickets,
completion reports
Extracted
Location of cemented intervals behind
casing
Driller's reports, wellbore diagrams,
cement evaluation logs
Extracted
Cement bond index
Cement bond logs
Calculated
Well completion type (cemented,
formation packer, open hole)
Driller's reports, wellbore diagrams
Extracted
Ground Water Resources
Depth of protected ground water
resources
State/Federal well authorization
documents, aquifer maps*
Extracted
Location of private drinking water wells
Complaint reports, spill reports
Extracted
* Table 5 provides additional details on the types of information provided by well operators with respect to the depths of
protected ground water resources.
19
-------
Well Design and Construction
May 2015
3.4.1. Well Characteristics
To better understand the kinds of oil and gas production wells that were hydraulically fractured,
the following information was identified for each well: well age, production type, well orientation,46
true vertical depth, measured depth at the point of shallowest hydraulic fracturing, and stimulated
lithologies. Together, these characteristics provide insight into the variability among oil and gas
production wells hydraulically fractured by nine service companies in 2009 and 2010.
Well Age. Well age was calculated by comparing the spud date (i.e., the date the first hole was
drilled) to the last date in the study's timeframe, September 30, 2010. This provided an estimate of
the age of the well at the time of the hydraulic fracturing event that occurred within the timeframe
of the EPA's service company well list (hydraulically fractured between approximately September
2009 and September 2010).
Production Type. Production type was determined based on the ratio of initial production volumes
of oil and gas. In most cases, initial production data were provided in the well files; wells with gas-
to-oil ratios greater than 20,000 cubic feet of gas per barrel of oil were considered "gas" wells [US
Geological Survey (USGS), 2012], Approximately 14 percent of the well files did not contain initial
production data. In these cases, production types were determined from other sources of
information in the well file or classifications provided by state websites.
Well Orientation. Wells were classified as either vertical, horizontal, or deviated based on
information from driller's reports, wellbore diagrams, and deviation surveys. For this study,
"vertical" wells were defined as wells with a bottom-hole location within 500 lateral feet of the
surface wellhead location. "Horizontal" wells were defined as wells intentionally completed with
one or more boreholes drilled laterally to follow the targeted geologic formation. "Deviated" wells
were defined as non-horizontal wells with a bottom-hole location more than 500 lateral feet from
the surface wellhead location. For example, deviated wells may have an "S" shape, starting and
ending relatively vertical, but with an intermediate portion drilled at a significant angle, or they
may have a relatively constant angle along the entire drilled length.
True Vertical Depth. The true vertical depth of a well is defined as the vertical depth at the bottom-
most point of the well. True vertical depth was determined from driller's reports, wellbore
diagrams, deviation surveys, and state websites. Often this value had been calculated during drilling
and was extracted from driller's reports. In some instances, this value was calculated by the EPA
using trigonometry, given the wellbore length as well as the angle of deviation from vertical, and by
assuming a fairly straight wellbore.
Measured Depth at the Point of Shallowest Hydraulic Fracturing. Measured depth represents length
along the wellbore, which may be a straight vertical distance below ground or may follow a more
complicated path, if the wellbore is not straight and vertical. The measured depth of the shallowest
46 In this study, well orientation refers to the degree a well diverges from vertical, not the ordinal direction (e.g., north or
south] of the bottom-hole location relative to the wellhead.
20
-------
Well Design and Construction
May 2015
point of hydraulic fracturing was identified to determine the shortest distance along the wellbore
between hydraulic fracturing and operator-reported depths to protected ground water resources.
In cemented cased completions, the point of shallowest hydraulic fracturing was equal to the
shallowest measured depth of a production perforation used for hydraulic fracturing. In
uncemented cased completions, the point of shallowest hydraulic fracturing was equal to the
measured depth of the shallowest formation packer used to isolate a hydraulic fracturing interval.
In open hole completions, the point of shallowest hydraulic fracturing was equal to the uppermost
measured depth interval where hydraulic fracturing occurred.
Hydraulically Fractured Lithologies. For this report, hydraulically fractured lithologies are geologic
formations that were hydraulically fractured at any point in the well's history. Open hole logs, in
combination with mud logs, were reviewed to determine zone lithology using accepted principles
and methods, such as those described in Dewan (1983), Krygowski (2004), and Schlumberger
(1991). When these log types were not available, hydraulically fractured lithologies were identified
from other data sources in the well files that included a lithologic description of the zone
hydraulically fractured.
3.4.2. Construction Characteristics
Data on casing, cementing, and completion activities for each hydraulically fractured well were
compiled from information included in the well files (Table 3).
Casing. Casing strings were identified from reported casing tallies, driller's reports, completion
reports, wellbore diagrams, and occasionally from forms submitted to oil and gas regulatory
agencies. The measured depths of the top and bottom of the casing string were recorded for
surface, intermediate, and production casing. Conductor casing or other shallow, often uncemented
casings, were excluded.
Cement. Cement placement behind surface, intermediate, and production casing was determined
from various sources, including cement evaluation logs,47 driller's reports, cement job tickets,
wellbore diagrams, and forms submitted to oil and gas regulatory agencies. Cement tops were
determined from cement evaluation logs, when available. If cement evaluation logs were
unavailable, other reported data, including driller's reports, cement job tickets, wellbore diagrams,
and forms submitted to oil and gas regulatory agencies were assessed collectively to determine
cement tops. Cement bottoms were identified using the point of placement of cement through the
casing.
Cement Bond Indices. Cement bond indices, h, were calculated from standard acoustic cement bond
logs using a straight-line formula derived from the semi-log plot representing the
Tenneco/Fitzgerald technique, which relates amplitude and bond index, as outlined in Fitzgerald et
al. (1985) and Smolen (1996):
47 Cement evaluation logs included standard acoustic cement bond logs, ultrasonic image tool logs, and temperature logs.
21
-------
Well Design and Construction
May 2015
^ Iogi4(100)-logi4(0)
Iog4(d)-log4(0)
(1)
In equation 1, A[d) is equal to the measured amplitude at depth d, .4(0) is equal to the maximum
amplitude recorded on the log and represents a point of zero bonding (i.e., free pipe), andv4(100) is
equal to the minimum amplitude recorded on the log and represents the point of best bonding.
Thus, it can take on values from 0 to 1. A value of 0 indicates the weakest bonding, while a value of
1 indicates the strongest point of bonding measured in the well (Smolen, 1996).
The bond index was calculated at 10 foot intervals for the first 100 feet immediately above the
point of shallowest hydraulic fracturing (if that part of the wellbore was logged) and at 5 0 foot
intervals for the remaining up-hole portion of the log.
3.4.3. Drinking Water Resources
A geospatial analysis of surface features related to surface and ground water resources was
conducted to summarize the spatial relationship of drinking water resources to oil and gas
production wells hydraulically fractured by the nine service companies.48 Additional information
on ground water resources in the immediate area of the study wells was obtained from data
supplied by well operators.
Geospatial Analysis of Water Resources. The locations (latitude/longitude coordinates) of the 323
wells were used to identify water resources near hydraulically fractured oil and gas production
wells. Coordinates were obtained in either the North American Datum of 1927 (NAD27) or North
American Datum of 1983 (NAD83). All NAD27 coordinates were transformed to NAD83 coordinates
using an ArcGIS built-in utility that follows the NADCON model (2011) and plotted using ESRI's
ArcGISv.10 software (2010).
A 0.5 mile radius circle was delineated around each surface wellhead location (referred to as a "well
buffer") using an Albers Equal Area projection appropriate for the continental United States.49 The
0.5 mile radius around each wellhead location provided a consistent approach to search for nearby
water resources, although local topographical conditions might support use of a different geometry
at any specific well site. Individual well buffers were compared to available spatial data identifying
the presence of surface and ground water resources as well as current drinking water resources.
Surface water resources (e.g., streams and lakes) were obtained from the United States Geological
Survey's National Hydrography Dataset and used to calculate the total stream length and area of
open water in each buffer (USGS, 2014). This report does not identify whether surface water
resources currently serve as drinking water resources. The presence or absence of current drinking
water resources was determined using locations of public water system wells and surface water
intakes (US EPA, 2014a), as well as private drinking water wells identified in the well files.
48 The EPA's Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources defined "drinking
water resources" as any body of water, ground or surface, that currently serves or in the future could serve as a source of
drinking water for public or private water supplies (US EPA, 2011].
49 Central Meridian: -96, Standard Parallel 1: 29.5, Standard Parallel 2: 45.5, Latitude of Origin: 37.5.
22
-------
Well Design and Construction
May 2015
Protected Ground Water Resources. Since depths to protected ground water resources were not
included in most well files, well operators were asked to provide this information for each well
location in September 2013. As noted by the Ground Water Protection Council:
"There is a great deal of variation between states with respect to defining protected ground
water. The reasons for these variations relate to factors such as the quality of water, the
depth of Underground Sources of Drinking Water, the availability of ground water, and the
actual use of ground water." (GWPC, 2009)
As a result, the identification of protected ground water resources relied solely on information
provided to the EPA by well operators. To better understand the data provided by the well
operators, the EPA also asked them to identify the source of the depths they provided.
Some operators reported protected ground water resources, while other operators provided the
depth to ground water resources without identifying whether the resource was protected.
Operator-provided depths were assumed to be depths to protected ground water resources. Data
sources used by operators to identify protected ground water resources generally included state or
federal regulations or permit requirements, state or federal ground water maps or databases, and
well records from nearby water wells or oil and gas production wells. Well records included either:
(1) depths of nearby water supply wells, or (2) depths of freshwater zones identified by operators
during drilling or from induction logs. Reported depths to ground water resources included both
specific values (e.g., surface to 200 feet) and estimates (e.g., no deeper than 200 feet). Based on the
information provided, the EPA did not identify any cases in which water samples were collected
before or after drilling to determine whether water from a specific zone or found at a specific depth
met designated protected water quality criteria.
Operator-provided depths to protected ground water resources are compared to different well
construction characteristics in Section 4.3.2. Using depths based on different definitions of
protected ground water resources would likely affect the results presented in Section 4.3.2.
3.5. Estimates of Well Design and Construction Characteristics
Each analysis presented in this report used data collected from 323 study wells to estimate the
frequency of occurrence of well design and construction characteristics, which can be used to better
understand onshore well design and construction practices in the United States. Estimates are
presented at the national level50 and are statistically representative of the oil and gas production
wells hydraulically fractured by the nine service companies.51 As described in Section 3.1., the
service company well list likely resembled the overall geographic distribution of oil and gas
50 As described in Section 3.2, study wells were selected from the service company well list, which contained county and
state locations of well identifiers and well operator names. It was not possible to choose a representative sample of wells
based on hydraulically fractured geological formations, because these data were not included in the service company well
list. Consequently, results specific to individual targeted geologic formations were not calculated. Results calculated for
individual targeted geologic formations would have larger 95 percent confidence intervals than the results from analyses
conducted for this report.
51 The survey design ensured that the 323 study wells were statistically representative of all wells on the service company
well list, regardless of state.
23
-------
Well Design and Construction
May 2015
production in the United States in 2009 and 2010. The estimates presented in the report, therefore,
are likely indicative of well design and construction characteristics of onshore oil and gas
production wells hydraulically fractured across the continental United States during the timeframe
examined in this study.
In general, each of the 323 wells was assigned to a category (or characteristic) defined by a given
analysis. Statistical weights were then used to estimate the number of wells out of the sampled
population that were within each category (i.e., point estimates). Ninety-five percent confidence
intervals were calculated for each point estimate, as described in Appendix A.52
Statistical Weights. Each study well carried a statistical weight equal to the number of wells it
represented on the service company well list. Statistical weights ranged from 4 to 190. Therefore,
each study well represented between 4 and 190 hydraulically fractured oil and gas production
wells on the service company well list The sum of the statistical weights for the 323 study wells
was 23,195.
Statistical weights were calculated from the probability of a well identifier being chosen in the two-
stage stratified selection process and being retained when 50 well identifiers were randomly
removed from the 400 selected well identifiers (Section 3.2). Additional adjustments to the weights
were made to account for: (1) operators with fewer than 10 well identifiers on the service company
well list,53 (2) a change in the number of unique wells in the original service company well list,54
and (3) one well identifier, noted in Table 1, that was misidentified and is not a well from one of the
nine sampled operators.
To the extent that wells operated by the same company in the same geographic area had similar
well histories, the survey design ensured that point estimates were more precise. Also, to the extent
that wells from companies of similar sizes and, for large companies, geographic regions, had similar
well design, construction, and hydraulic fracturing characteristics, there is further improvement in
the precision of the results. Point estimates are more precise and smaller confidence intervals
occur, therefore, when more wells share similar characteristics.
52 The algorithm used to design the sample was optimized to provide the smallest 95 percent confidence interval for the
sample of 350 well identifiers selected from the service company well list. Confidence intervals reflect observed
variability in well design and construction characteristics. This variability may be associated with different company
policies, geology, state requirements, or other factors. The point estimate is the center of the confidence interval and
represents the best estimate of the true number of wells in a category given this sample of hydraulically fractured oil and
gas production wells.
53 The statistical weights for wells from the small operators were adjusted so that they represented themselves and wells
from operators that had fewer than 10 well identifiers. This assumed that wells from the three sampled small operators
had similar characteristics to wells operated by companies with fewer than 10 well identifiers on the service company
well list. To the extent that differences in characteristics existed, this is not reflected in survey estimates. Since these very
small companies operated less than 10 percent of all well identifiers (2,352 out of 24,925], the potential for biased
estimates was limited.
54 The number of wells in the original population decreased from 24,925 to 24,019 after further review of the data
provided by the nine hydraulic fracturing service companies identified duplicate entries.
24
-------
Well Design and Construction
May 2015
3.6. Quality Assurance and Quality Control
The EPA does not make any claims on the quality or accuracy of the data or information received
directly from the operators in response to the information request. In a few, limited cases,
discrepancies between data publicly available through state websites and operator-provided data
were observed. In these cases, data provided by the well operators were used. A comprehensive
comparison of operator-provided data and data available through state websites was not
conducted.
Extracted data for all wells and calculated data for up to two example wells were shared with the
well operators in September 2013. In total, 20,203 data elements were provided to the operators
for their review. Operators identified changes to 774 data elements (less than 4 percent). Some of
these changes were corrections to extracted or calculated data; in many cases, the operators
provided additional data that were not found in the original submissions. All updates provided by
the operators were incorporated.
Quality assurance and quality control measures were used to ensure that data in the database were
of sufficient quality for analysis and that the analyses performed were conducted properly. The
quality of the data extraction process was tested in two ways: (1) re-reviewing portions of the well
files and (2) designing queries to detect inconsistencies in the database. Inconsistencies found
while querying the database were resolved by referring to the original data provided by the
operators.
At least 10 percent of records from each operator relating to well location, cementing records, and
cement bond logs were re-reviewed, and 100 percent of driller's reports were re-reviewed. No
differences in well location and cementing records were identified. Any differences found in the re-
review of the driller's reports were resolved through discussions with the original reviewer.
Cement bond indices were calculated for 203 wells for which standard acoustic cement bond logs
were provided. Cement bond logs from 30 wells were randomly selected for re-review. Among
these wells, a total of 2,721 cement bond indices at different measured depths were re-calculated
(Section 3.4.2). In 64 instances (2.4 percent of the re-reviewed cement bond indices) from 7 wells,
the re-reviewed cement bond index was different from the original value by 0.05 or more bond
index points. The maximum error, which occurred one time, resulted in a difference of 0.64 in the
calculated bond index. The causes of these differences are summarized in Table 4.
Cement bond indices from all wells were used to calculate the distribution of cement bond indices
at discrete heights above the shallowest point of hydraulic fracturing (Section 4.2.2). The
distribution of cement bond indices at each height is based on data from at least 22 distinct wells,
with the most populous height (350 feet) having cement bond indices from 185 distinct wells. The
number of incorrect cement bond indices at any given height was determined to be between one
(for heights with 22 wells) and four (for heights with 185 wells), assuming that the 2.4 percent
overall error rate calculated in Table 4 was evenly distributed across the number of cement bond
25
-------
Well Design and Construction
May 2015
Table 4. Quality assurance summary for calculated cement bond indices. There were four instances in which two of
these types of errors were made in calculating the same cement bond index, explaining how 68 occurrences of
error resulted in 64 erroneous calculated bond indices. "NA" indicates not applicable.
Description
Occurrences
Error
Rate
Range of Errors in Calculated
Bond Index
Different values chosen for/Aioo orAo
43
1.6%
0.0504-0.3673
Different interpolations of the amplitude value
12
0.4%
0.0520-0.3413
Misunderstanding of the scale on the cement
bond log
9
0.3%
0.0538-0.6402
Incorrect inclusion of amplitude values that
were affected by a casing collar
4
0.1%
0.0697-0.0970
Overall error rate
68
2.4%
NA
indices calculated at each particular height55 Conservatively applying the maximum error of 0.64 in
the calculated bond index to between one and four calculated bond indices would move the
erroneous values toward (or away from) one of the extreme ends of the range of calculated bond
indices, depending on whether the error is added or subtracted. While, for a single point, the 0.64
error is large, its occurrence rate is small (less than 0.04 percent of 2,721 cases). Therefore,
factoring in this error would have a minor impact on the distributions presented in Section 4.2.2.
This work was conducted following the methods and procedures contained in the project-related
quality assurance projectplans (The Cadmus Group, 2013; US EPA, 2013b, 2014c; Westat, 2013).
The project underwent a series of technical systems audits by the designated EPA Quality
Assurance Manager between April and August of 2012. No corrective actions were identified.
4. Analytical Results
Well design and construction characteristics for an estimated 23,200 (95 percent confidence
interval: 21,400-25,000) onshore oil and gas production wells hydraulically fractured by nine
service companies in 2009 and 2010 are summarized below. Throughout the remainder of this
report, figures show point estimates with 95 percent confidence intervals. Data tables and in-text
calculations report rounded point estimates and 95 percent confidence intervals.
4.1. Well Characteristics
Age, Production Type, and Orientation. The nine service companies hydraulically fractured both
recently drilled wells and older wells, oil and gas wells, and wells of different orientations (Figure
6).
Most of the wells were drilled within 10 years of the end of the study's timeframe (September 30,
2010). An estimated 1,400 wells (600-2,200) were drilled more than 10 years before the end of the
55 For example, the most populous height has calculated cement bond indices from 185 wells. Multiplying 185 by the
error rate, 0.024, results in 4.44 (which rounds to 4.0].
26
-------
Well Design and Construction
May 2015
w
"a3
01
-Q
£
20,000
15,000
10,000
5,000
0
I
<1 Year
1-10 years
Age of Well
>10 years
(b)
01
01
¦Q
E
20,000
15,000
10,000
5,000
0
Gas Oil
Production Type
(c) 20,000
15,000
10,000
5,000
0
w
ai
-------
Well Design and Construction
May 2015
study's timeframe, with the oldest well among the study wells spudded in 1959.56 The presence of
older wells suggests that either targeted geologic formations accessed through existing oil and gas
production wells were re-fractured or that existing oil and gas production wells were recompleted
using hydraulic fracturing to recover oil and gas from hydrocarbon-containing formations
previously drilled through.
An equal proportion of oil and gas production wells were hydraulically fractured. An estimated
13,100 wells (7,400-18,700) initially produced gas, while an estimated 10,100 wells (4,500-15,800)
initially produced oil (panel b in Figure 6). More than half of the wells (15,200; 11,200-19,200) in
this study were vertical wells, with fewer deviated and horizontal wells (panel c in Figure 6).
Depths. True vertical depths ranged from less than 2,000 feet to greater than 11,000 feet The
majority of wells had a true vertical depth greater than 5,000 feet (Figure 7). Among the study
wells, the shallowest well had a true vertical depth of just over 1,600 feet, while the deepest well
had a true vertical depth slightly greater than 13,000 feet.
Number of Wells
0 2,000 4,000 6,000 8,000 10,000 12,000
<1,999
2,000-2,999
3,000-3,999
4->
S 4,000-4,999
£ 5,000-5,999
01
- 6,000-6,999
ro
u
¦B 7,000-7,999
(y ' '
>
« 8,000-8,999
9,000-9,999
10,000-10,999
>11,000
Figure 7. True vertical depths of oil and gas production wells hydraulically fractured by nine service companies
between approximately September 2009 and September 2010. True vertical depth is defined as the vertical depth
of the bottom-most point of the well. Error bars display 95 percent confidence intervals.
56 Instances in this report that refer to oldest, deepest, or other extreme values from among the 323 study wells should
not be interpreted to mean that these are the most extreme values among the 23,195 wells in the sampled population,
since it is unlikely that the sample of 323 study wells happened to include the most extreme values from the 23,195 wells.
For this example, 1959 is the year when the oldest well from among the 323 study wells was drilled, but there are likely
older wells in the population of 23,195 oil and gas production wells that were hydraulically fractured by nine service
companies between approximately September 2009 and September 2010.
28
-------
Well Design and Construction
May 2015
The depth at which hydraulic fracturing is conducted is different than the true vertical depth of a
well, because hydraulic fracturing typically occurs within a specific portion of the well. Hydraulic
fracturing depths were identified using the measured depth of the shallowest point at which
hydraulic fracturing occurred. Measured depths of hydraulic fracturing ranged from less than 1,000
feet along the length of the well to more than 11,000 feet along the length of the well (Figure 8).
More than half of the wells had points of shallowest hydraulic fracturing deeper than 5,000 feet
below the wellhead, as measured along the length of the well. The shallowest point of hydraulic
fracturing in the study wells was recorded at a measured depth of roughly 600 feet, and the deepest
point of shallowest hydraulic fracturing was found to be at a measured depth of approximately
13,500 feet.
0
a.
01
13D
T3
>
a. ^
oi To
Q O)
T3 |
11,000
Number of Wells
2,000 4,000 6,000 8,000 10,000 12,000
Figure 8. Measured depths of the point of shallowest hydraulic fracturing in oil and gas production wells
hydraulically fractured by nine service companies between approximately September 2009 and September
2010. The measured depth is the distance along the length of the well and is not always equal to the true vertical
depth. The point of shallowest hydraulic fracturing is the uppermost depth interval where hydraulic fracturing
occurred. Error bars display 95 percent confidence intervals.
Hydraulically Fractured Lithologies. Hydraulic fracturing is used to enhance oil and gas production
from many different types of lithologies, or rock formations. Lithologies hydraulically fractured by
the nine service companies included sandstone, shale, carbonate, coal,57 and chert (Figure 9).
57 This included coalbed methane production wells. Unlike gas production from tight sands and shales, coalbed methane
production typically occurs after sufficient connate water is removed to release methane from coal seams (Christen,
2003; US EPA, 2004]. Because hydraulic fracturing was used to stimulate methane production from these wells, they are
included in this study.
29
-------
Well Design and Construction
May 2015
16,000
14,000
« 12,000
I" 10,000
M-
Z 8,000
Q)
£ 6,000
3
2 4,000
2,000
0
-r
1 1
1
h"
1
¦
11
11
L
Sandstone Shale Carbonate
Coal Chert Unable to
determine
Hydraulically Fractured Lithologies
Figure 9. Lithologies hydraulically fractured for oil and gas by nine service companies between approximately
September 2009 and September 2010. An estimated 8,100 wells (5,400-10,800) had one hydraulically fractured
lithology, and 9,000 wells (6,100-11,900) had more than one lithology hydraulically fractured. Wells with
insufficient information to determine hydraulically fractured lithologies are represented by the "unable to
determine" bar. Error bars display 95 percent confidence intervals.
An estimated 8,100 wells (5,400-10,800) had one lithology hydraulically fractured, while 9,000
wells (6,100-11,900) had more than one hydraulically fractured lithology. Hydraulically fractured
lithologies could not be identified for 7,700 wells (5,900-9,500) due to insufficient data in the well
files. Because hydraulic fracturing occurred in different lithologies within the same well (depending
on the stratigraphic layering in the area), the number of wells displayed in Figure 9 is greater than
the total number of wells identified in this study (i.e., wells could be counted more than once in this
analysis).
4.2. Construction and Completion Characteristics
Oil and gas production wells are constructed in sequential stages using casing and cement, as
described in Section 2. Casing and cement allow a desired flow to take place in a controlled fashion
and create zonal isolation. Wells represented in this study were completed in the targeted geologic
formation with an open hole or with casing secured to the formation using either cement or
formation packers (Figure 3). An estimated 20,200 wells (17,500-23,000) had cemented casing
completions, while 1,500 wells (1,400-1,600) had formation packer completions; another 1,500
wells (10-4,800) had open hole completions.
4.2.1. Casing Installation
Oil and gas production wells in this study had between one and three casing strings (excluding
conductor casing), as illustrated in Figure 2. An estimated 23,200 wells (21,400-25,000) had
surface casing while 9,100 wells (2,900-15,400) had intermediate casing and 21,900 wells (19,200-
24,600) had production casing. The most common casing configuration, used in 12,800 wells
(7,600-18,000), was a surface and production casing. Another 8,100 wells (2,300-13,900) had a
30
-------
Well Design and Construction
May 2015
surface, intermediate, and production casing.58 Together, these two casing configurations
accounted for nearly all of the production wells in this study.
Measured depths of the bottom of surface, intermediate, and production casings are shown in
Figure 10. The range of measured depths for each type of casing is larger for deeper casings. For
example, the measured depths of surface casing ranged from less than 1,000 feet to between 6,000
and 6,999 feet, while the measured depths of production casing ranged from less than 1,999 feet to
greater than 15,000 feet. The median measured depths of the bottom of surface, intermediate, and
production casing were calculated to be 800, 5,000, and 8,000 feet, respectively.
<999
1,000-1,999
2,000-2,999
3,000-3,999
4,000-4,999
5,000-5,999
6,000-6,999
7,000-7,999
8,000-8,999
w 9,000-9,999
re
Ol
§ 10,000-10,999
11,000-11,999
12,000-12,999
13,000-13,999
14,000-14,999
>15,000
Ol
01
Q.
01
Q
T3
01
5,000
Number of Wells
10,000 15,000
20,000
25,000
I Surface
Intermediate
Production
Figure 10. Measured depths of the bottoms of surface, intermediate, and production casing in oil and gas
production wells hydraulically fractured by nine service companies between approximately September 2009 and
September 2010. Not all wells had all types of casing: 23,200 wells (21,400-25,000) had surface casing, 9,100 wells
(2,900-15,400) had intermediate casing, and 21,900 wells (19,200-24,600) had production casing. Error bars display
95 percent confidence intervals.
58 This excludes the casing configurations shown in Figure 2 that included liners.
31
-------
Well Design and Construction
May 2015
An estimated 52,800 casing strings (45,800-59,700) were cemented.59 These casing strings were
cemented to the wellbore through either primary or secondary cementing operations (Figure 11).
Primary cementing operations occur at the time the well is drilled and constructed, with the cement
being delivered from the surface down through the casing passing around the casing shoe (i.e., the
bottom of the casing), and moving upward into the annular space behind the casing. An estimated
48,300 (38,500-58,100) of all cement placement methods for cemented casings were primary
cementing operations. This included instances where cement staging tools were used at the time
the casing was originally cemented (i.e., multi-stage cementing operations).
25,000
20,000
01
5 15,000
01
¦Q
E
10,000
5,000
0
I Primary
Secondary
No cement (formation packer)
Surface
Intermediate
Casing Type
Production
Figure 11. Type of cementing operation conducted for each casing type found in oil and gas production wells
hydraulically fractured by nine service companies between approximately September 2009 and September
2010. Primary cementing operations occurred at the time the well was drilled and constructed. Secondary
cementing operations were considered to be any non-primary cementing operation. Casings with "no cement"
were not designed to have cement placed, because the wells were completed using formation packers. Error bars
display 95 percent confidence intervals.
Secondary cementing operations,60 which can occur at various times following primary cementing
accounted for an estimated 4,500 (1,500-7,500) of all cement placement methods for cemented
casings. For this analysis, secondary cementing operations were considered to be any non-primary
cementing operation and included cement squeezes, use of cement baskets, and pumping cement
down the annulus. Reasons for secondary cementing operations were not always clearly stated in
the well files; however, contextual information from the well files was used to identify situations
addressed through secondary cementing operations. Types of situations included: (1) adding
cement to fracture a different zone, (2) plugging open perforations that were no longer wanted, (3)
placing cement behind casing when primary cementing operations had failed to place it to a desired
59 An estimated 54,200 casing strings (47,300-61,200] were installed using cement or packers.
60 Secondary cementing operations are typically referred to as "remedial" cementing operations (Nelson and Guillot,
2006; Schlumberger, 2014].
32
-------
Well Design and Construction
May 2015
height, and (4) sealing off a formation thought to contribute to pressure detected in an uncemented
annulus between the wellbore and the casing.
Figure 12 displays the number of wells that had fully cemented, partially cemented, and
uncemented casings, by casing type (surface, intermediate, or production). Fully cemented casings
were defined, in this study, as casings that had a continuous cement sheath from the bottom of the
casing to at least the next larger and overlying casing (or the ground surface, if surface casing).61
Conversely, casings with no cement anywhere along the casing from the bottom of the casing to at
least the next larger and overlying casing (or ground surface), were defined as uncemented.
Partially cemented casings were defined as casings that had some portion of the casing that was
cemented from the bottom of the casing to at least the next larger and overlying casing (or ground
surface), but were not fully cemented.
25,000
20,000
01
5 15,000
01
¦Q
E
10,000
5,000
0
Surface
I Fully cemented
Partially cemented
Uncemented (formation packer)
I Uncertain
J
L
Intermediate
Casing Type
Production
Figure 12. Degree of cementing determined for each casing type found in oil and gas production wells hydraulically
fractured by nine service companies between approximately September 2009 and September 2010. The presence
or absence of cement was evaluated from the bottom of the casing to either the ground surface or the next larger
and overlying casing. Fully cemented casings were defined, in this study, as casings that had a continuous cement
sheath over the distance evaluated. Partially cemented casings were defined as casings that had some portion of
the casing that was cemented over the distance evaluated, but were not fully cemented. Uncemented casings
were not designed to have cement placed, because the wells were completed using formation packers. Due to a
lack of sufficiently detailed cement information in the well files, it could not be determined if some casings were
fully cemented or partially cemented; this is reflected by the "uncertain" category. Error bars display 95 percent
confidence intervals.
Figure 12 shows that surface and intermediate casings were fully cemented in most wells, while
production casings were generally either fully or partially cemented. An estimated 21,500 wells
(19,500-23,600) had fully cemented surface casing, and 7,300 wells (600-13,900) wells had fully
cemented intermediate casings. Production casing was fully cemented in 8,300 wells (3,800-
61 Continuous cement sheaths were defined by the presence of cement, which was based on the determination of cement
tops and bottoms (Section 3.4.2]. This analysis does not incorporate cement quality derived from cement evaluation logs.
33
-------
Well Design and Construction
May 2015
12,800) and partially cemented in 10,900 wells (6,900-14,900). Also shown in Figure 12 are
formation packer completions, which, by design, had uncemented production casings. Open hole
completions are not shown. Due to a lack of sufficiently detailed cement information in the well
files, it could not be determined if some casings were fully or partially cemented; this is reflected by
the "uncertain" category in Figure 12.
An estimated 6,800 wells (1,600-11,900) were fully cemented along the outside of the well, from
the bottom of the well to ground surface. In these wells, all casing strings had a continuous sheath
of cement from the bottom of each casing to either the surface or into the next overlying casing.
Approximately 15,300 wells (10,500-20,100) had one or more uncemented intervals along the
outside of the well, from the bottom of the well to the ground surface. Partially cemented casing
strings (identified in Figure 12) were further examined to determine the proportion of the casing
length cemented. Figure 13 displays, for each casing type, the percentage of the evaluated casing
length that was cemented and shows that the majority of partially cemented casings had more than
50 percent of the evaluated casing length cemented. Of the partially cemented casings in Figure 13,
lengths of uncemented intervals ranged from 5 to 7,800 feet, with a median value of 1,300 feet.
Eighty percent of all uncemented intervals behind a single casing were between 100 and 5,300 feet
in length (10th to 90th percentiles).
7,000
6,000
in 5,000
"5
^ 4,000
M- '
O
jj 3,000
E
Z 2,000
1,000
0
I Surface
Intermediate
Production
<25 25-49 50-74
Percent of Casing Cemented
>75
Figure 13. Percentage of the evaluated casing length that was cemented, calculated only for partially cemented
casings shown in Figure 12. Error bars display 95 percent confidence intervals.
Uncemented intervals in partially cemented casings can occur when it is technically difficult to
return cementto the surface or when casing is expected to be replaced or reused (GWPC, 2014).62 It
can be technically difficult to place cement in zones that: (1) could fracture under the pressures
62 Many sources recommend that cement be placed across zones containing corrosive fluids, water in need of protection,
or hydrocarbons (Nelson and Guillot, 2006; Ross and King, 2007; Smith, 1976], which suggests that there maybe
uncemented intervals along zones that do not meet these criteria. Criteria for zones not needing to be cemented could not
be found.
34
-------
Well Design and Construction
May 2015
applied during cementing operations, (2) are highly permeable and relatively under-pressured (e.g.,
mine voids), (3) cause water loss from the cement slurry, (4) contribute significant fluid into the
wellbore, or (5) have poor removal of drilling muds (Holt and Lahoti, 2012; Smith, 1976). It may
also be difficult to place cement in zones that are pressurized (Nelson and Guillot, 2006;
Schlumberger, 1984; Smith, 1976); this may include gas storage zones and wastewater disposal
zones. The occurrence of some of these situations was identified from driller's reports. An
estimated 800 (10-1,900) were drilled through a mine void, and 90 wells (50-100) were drilled
through either a gas storage zone and/or a wastewater disposal zone. Driller's reports also
identified instances of losses of circulation in 1,800 wells (10-3,600).63 Losses of circulation occur
when drilling fluids are lost to surrounding zones and may identify zones in which it could be
difficult to place cement (Devereux, 1998).
4.2.2. Cement Evaluation
Once cement has been placed, operators often run one or more logs to evaluate whether the cement
formed an effective seal (i.e., cement evaluation logs). An effective seal depends on the full
circumferential coverage of the cement around the casing and good bonding between the cement,
casing, and wellbore (Brufatto etal., 2003; Devereux, 1998; Holt and Lahoti, 2012). An estimated
18,000 wells (14,900-21,200) had atleastone kind of cement evaluation log run, and 5,100 wells
(1,300-9,000) had no cement evaluation log run. The most common type of cement evaluation log
run was a standard acoustic cement bond log. Other types of cement evaluation logs run included
segmented bond, ultrasonic imaging, and temperature logs. Segmented bond logs and ultrasonic
imaging logs can provide circumferential cement coverage information, while temperature logs
detect the presence or absence of cement in newly-cemented wells.
Standard acoustic cement bond logs, the most common type of cement evaluation log run in the
wells represented by this study, are used to evaluate both the extent of the cement placed along the
casing and the cement bond between the cement, casing, and wellbore. The logs are run in the
cemented casing on an electric wireline and record the attenuation of an emitted acoustic signal as
it travels between a sound source and one or more receivers located at a fixed distance from the
source. A relatively large recorded amplitude indicates a weaker bond, while a relatively small
recorded amplitude indicates a stronger bond. The recorded amplitudes can be used to calculate a
cement bond index using the Tenneco/Fitzgerald technique described in Section 3.4.2. (Fitzgerald
etal., 1985; Smolen, 1996)
Cement bond indices were calculated for individual study wells at different heights above the point
of shallowest hydraulic fracturing and then extrapolated to the sampled population. Figure 14
shows the percentiles of the weighted distributions of all cement bond indices calculated at specific
heights above the point of shallowest hydraulic fracturing. In total, the data in Figure 14 represent
15,500 wells (12,300-18,800) in which standard acoustic cement bond logs were run. Each height
above the point of shallowest hydraulic fracturing represents between 1,300 wells (200-2,400;
63 Losses of circulation occurred in an estimated 200 holes (180-210] drilled for surface casing, 140 holes (90-190] drilled
for intermediate casing, and 1,300 holes (10-2,900] drilled for production casing. While most volumes of lost drilling fluid
were not reported, the maximum reported volume of lost drilling fluids for a single well was greater than 252,000 gallons
(6,000 barrels].
35
-------
Well Design and Construction
May 2015
5,000 feet above the point of shallowest hydraulic fracturing) and 14,800 wells (11,400-18,200;
350 feet).
« .E
Q) U
> ro
£ -
to «
01 RJ
S!
5
Q 5
"a re
Q) £
= 1/1
iS o
QJ 4-»
M
a.
5,000
4,500
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
*
*
€
%
*.
I
&
vt'-X
*-X
:x
vx
-K
se-'*~
k
i
?
V.
5.
-ir
i,+;
'+¦»¦
-H-
V
r*
~+V,.
¦A-
-K-".
J.
+C
:-+
-hi
?+
-ft;'
,7+
+j-
-KT
-Kt-
%
\
2f
2L_
0.2 0.4 0.6 0.8
Calculated Cement Bond Index
5th
25th
50th
75th
95th
Figure 14. The 5th, 25th, 50th, 75th, and 95th percentiles of cement bond indices calculated from standard acoustic
cement bond logs.
Figure 14 shows that the 50th percentile bond index was approximately 0.7 at the point of
shallowest hydraulic fracturing and decreased over the next 5,000 feet of overlying measured depth
to approximately 0.4. Similar trends were observed for the 5th, 25th, 75th, and 95th percentiles. The
calculated cement bond indices shown in Figure 14 depend on the recorded amplitudes (equation
1). There are many conditions that can increase or decrease the amplitude signal. An increased
amplitude signal results in a lower cement bond index than actually exists. This has been observed
when microannuli are present, when thin cement sheaths (less than 0.75 inches) are present, when
the first signals to arrive at the tool pass through acoustically fast formations, and when insufficient
time has passed for the cement to cure (Albert et al., 1988; Fitzgerald et al., 1985; Flournoy and
Feaster, 1963; Jutton et al., 1991; Pilkington, 1992). A decreased amplitude signal results in a higher
cement bond index than actually exists. Conditions that result in decreased amplitude signals
include out-of-center positioning of the tool within the casing and casing that is in contact with the
side of the wellbore (Albert et al., 1988; Fitzgerald et al., 1985; Flournoy and Feaster, 1963; Nutt,
36
-------
Well Design and Construction
May 2015
2012). Other conditions, including those associated with improper tool setting, have been shown to
either increase or decrease the recorded amplitude (Albert et al., 1988; Fitzgerald et al., 1985;
Jutton etal., 1991).
The presence of the conditions described above were noted, where possible, during reviews of
cement bond logs, using the travel time and variable density curves. Two to 27 percent of the
cement bond indices in Figure 14 may have been calculated using amplitudes affected by these
conditions.64 65 The true calculated cement bond indices could be either higher or lower, but could
not be determined from data in the well files. This type of error likely resulted in a broader
distribution of values about the median (Carroll etal., 2006). The trend observed for the 50th
percentile in Figure 14 (i.e., a decrease in the calculated bond index further away from the point of
shallowest hydraulic fracturing) is unlikely to change due to these errors. These trends suggest
that, in general, the quality of the cement bond, as indicated from calculated cement bond indices,
was likely to be better near the point of shallowest hydraulic fracturing than at shallower depths. A
similar trend was observed by Flournoy and Feaster (1963) and Watson and Bachu (2009).
Flournoy and Feaster (1963) state that the best bonding is usually found near the bottom of the
casing, with both the percentage and degree of bonding decreasing as the distance from the bottom
of the casing increases; they attributed this affect to channeling in the cement due to contamination
of the cement by drilling mud in the annulus.
4.3. Well Construction Characteristics and Drinking Water Resources
Characterizing the spatial relationship between surface and ground water resources and oil and gas
production wells helps to understand whether hydraulic fracturing could affect the quality of
drinking water resources. While proximity alone does not determine impacts, it is a factor that
should be considered when assessing the potential for hydraulic fracturing to affect drinking water
resources. This section briefly summarizes the spatial relationship of drinking water resources to
the wellhead, then focuses on well design and construction characteristics related to protected
ground water resources reported by well operators.
4.3.1. Surface Locations of Drinking Water Resources
Figure 15 shows the estimated number of oil and gas production wells hydraulically fractured by
the nine service companies that had one or more of the specified drinking water resources within
0.5 miles of the wellhead.
64 Figure 14 used 10,738 bond indices calculated from standard acoustic cement bond logs run in 203 study wells. Up to
14 percent of the 10,738 calculated bond indices may have been affected. The statistical weights among the study wells
contributing data to Figure 14 ranged from 10 to 148. If all of the affected bond indices were evenly distributed across all
of the calculated bond indices and had a statistical weight of 10, approximately 2 percent of the 822,299 points used to
calculate the percentiles in Figure 14 would be affected. If all of the affected bond indices were evenly distributed across
all of the calculated bond indices and had a statistical weight of 148, approximately 27 percent of the 822,299 points used
to calculate the percentiles in Figure 14 would be affected.
65 The use of different cements in the same well may affect the amplitude signal. The use of different cements could not be
accounted for, because the depth at which the cement type changed was unclear in many wells.
37
-------
Well Design and Construction
May 2015
25,000
20,000
01
> 15,000
*4—
01
10,000
5,000
Lakes and Rivers Groundwater PWS wells PWS surface
ponds wells water intakes
Dringing Water Resource
Figure 15. Drinking water resources within 0.5 miles of oil and gas production wells hydraulically fractured by nine
service companies between approximately September 2009 and September 2010. Drinking water resources were
defined as any body of water, ground or surface, that currently serves or in the future could serve as a source of
drinking water for public or private water supplies. This report does not identify whether surface water resources
currently serve as drinking water resources. The presence or absence of current drinking water resources was
determined using locations of public water system wells and surface water intakes. No production wells were
estimated to have a surface water intake for a public water supply (PWS) within 0.5 miles of the wellhead. Private
ground water wells were identified through information provided by well operators. The number of hydraulically
fractured production wells within 0.5 miles of a private ground water well is expected to represent a lower-bound
estimate. Error bars display 95 percent confidence intervals.
An estimated 18,900 wells (15,200-22,600) were within 0.5 miles of a surface water resource (i.e.,
lake, pond, or river). Fewer production wells were estimated to be within 0.5 miles of either a
private ground water well (3,000; 900-5,100) or a public water supply well (600; 10-1,500). The
number of hydraulically fractured production wells within 0.5 miles of a private ground water well
is likely a lower-bound estimate.66 An estimated 17,100 production wells (12,400-21,900) had
more than one water feature within the 0.5 mile radius, while 4,100 production wells (600-7,600)
had no water feature identified within 0.5 miles of the wellhead. No study wells were found to be
within 0.5 miles of a surface water intake for a public water supply system.
4.3.2. Protected Ground Water Resources
Depths of protected ground water resources were provided by well operators for nearly all of the
study wells. As described in Section 3.4.3, operators used different data sources to report depths of
protected ground water resources. Because the analyses presented below compare well
construction characteristics to the depths provided by well operators, the use of different data
sources may affect the results of the analyses. Therefore, it is important to understand how the data
66 Nationwide data on the locations of private drinking water wells are not available. Well operators were asked for
results from water quality analyses of nearby surface or ground water (Appendix B]. Private ground water wells were
identified only if sample results were provided. It is likely that additional private ground water wells were present, but
not sampled.
38
-------
Well Design and Construction
May 2015
sources used by the operators to report depths to protected ground water resources are
represented throughout the sampled population of oil and gas production wells hydraulically
fractured by the nine service companies. Table 5 shows the number of production wells
represented by each type of data source and shows that, for the majority of wells, protected ground
water resources were identified by the well operators from state or federal authorization
documents (e.g., permits to drill or permit applications).
Table 5. Information submitted by operators to report the depths of protected ground water resources at each
study well location.
Data Sources Reported by
Agency-
Base of Protected
Number of
95 Percent
Operators
based*
Ground Water
Resource Depth
Wells
Confidence
Interval
Well authorization
Yes
Yes
11,700
10,800-12,700
documentation
No1^
1,500
1,300-1,700
Aquifer maps
Yes
Yes
4,000
3,400-4,500
Data from offset
No
Yes
2,300
10-6,200
production wells
Operator experience§
No
Yes
1,900
10-6,100
Open hole log
No
Yes
1,100
10-3,700
interpretation
General agency
Yes
Yes
500
400-500
requirement
Well authorization without
Yes
Indeterminate
100
60-200
documentation*
Online database
Yes
Yes
90
50-100
* "Agency" refers to the state oil and gas agency or the U.S. Bureau of Land Management.
^ Ground water depths based on the depth of a nearby water well were not considered to represent the base of the
aquifer, since deeper water wells may be drilled nearby at a later date.
§ No exact ground water depth value was provided. Operators stated that protected ground water is located at some
depth less than the value provided (e.g., less than 200 feet).
* No exact ground water depth value was provided. Operators inferred that protected ground water depth is located at a
depth shallower than surface casing because of state agency approval.
Operators reported either the depth to the base of the protected ground water resource (i.e., "base
identified") or the depth to some portion of the protected ground water resource (i.e., "base not
identified"). Table 5 shows that the base of the protected ground water resources was identified by
the well operator for most wells (20,500 wells; 16,300-24,600). The distinction between "base
identified" and "base not identified" affects analyses that compared well construction
characteristics to operator-reported depths to protected ground water resources. The depth to the
base of the protected ground water resource is unknown for "base not identified" wells. In these
cases, it is likely that the base of the protected ground water resource is deeper than the depth
reported by the well operators.67 Results presented for these wells may be different if the depth of
the base of the protected water resources was used instead of the depth provided by the well
operator. Therefore, results are shown for wells in which the base of the protected ground water
67 Ground water depths based on the depth of a nearby water well were not considered to represent the base of the
aquifer, since deeper water wells may be drilled nearby at a later date.
39
-------
Well Design and Construction
May 2015
resource for the study well was identified (dark blue bars in Figures 16,17, and 18) and also for
wells in which the base of the protected ground water resource for the study well was not identified
(light blue bars). Well files for a small number of study wells had insufficient data to identify the
base or depth of protected ground water resources. These wells represent an estimated 100 wells
(60-200) in the sampled population of hydraulically fractured oil and gas production wells and
were excluded from analyses that used depths of operator-reported protected ground water
resources.
Operator-reported depths to protected ground water resources ranged from just below ground
surface to 8,000 feet deep. The distribution of reported depths to ground water resources is shown
in Figure 16. An estimated 21,400 wells (18,900-24,000) passed through operator-reported
protected ground water resources that were less than 2,000 feet deep.
n
U 2
f £;
0
Q.
01
0£
O (11
**
ns (0
O ?
O 3
** O
cu <5
0) "O
2,000
Number of Wells
5,000 10,000
15,000
I Base of protected ground water identified
Base of protected ground water not identified
Figure 16. Depths of operator-reported protected ground water resources for oil and gas production wells
hydraulically fractured by nine service companies between approximately September 2009 and September
2010. Well counts are grouped according to whether or not the base of the operator-reported protected ground
water resources was identified. Error bars display 95 percent confidence intervals.
Operator-Reported Protected Ground Water Resources Relative to Casing and Cement. As described in
Section 2, surface casing generally serves as an anchor for blowout protection equipment and to
seal off ground water resources. Measured surface casing depths were subtracted from operator-
reported protected ground water resource depths to determine whether surface casing was
installed deeper than protected ground water resources. The subtraction was done for individual
wells, and the results were then extrapolated to the sampled population of oil and gas production
wells; Figure 17 displays the results of this analysis.
Figure 17 shows that surface casing extended below the base of the operator-reported protected
ground water for 12,600 wells (8,000-17,300). Most surface casings were fully cemented (Figure
12); in these cases, the operator-reported protected ground water resources located at depths
shallower than the bottom of the surface casing are covered by cement in the annular space. If the
40
-------
Well Design and Construction
May 2015
¦ Base of protected ground water resource identified
¦ Base of protected ground water resource not identified
I I _ .
Surface casing set Cemented behind Uncemented Unable to
below next casing below surface determine
casing
Protected Ground Water Resources
Relative to Casing and Cement
Figure 17. Comparison of casing and cement to operator-reported protected ground water resources among oil
and gas production wells hydraulically fractured by nine service companies between approximately September
2009 and September 2010. Well counts are grouped according to whether or not the base of the operator-
reported protected ground water resource was identified. For some wells, there was either ambiguity in the depth
of the base of the protected ground water resource or in the top of the protected ground water resource. These
wells are categorized as "unable to determine." Error bars display 95 percent confidence intervals.
surface casing was partially cemented, the operator-reported protected ground water resource may
or may not be covered by cement in the annular space. An additional 6,400 wells (500-12,300) had
operator-reported protected ground water resources located deeper than the bottom of the surface
casing that were covered by the next cemented casing string (either fully cemented intermediate or
fully cemented production casing).
The remaining wells may have had some part of the operator-reported protected ground water
resources not covered by cement in the annular space (Figure 17). In these cases, the bottom of the
surface casing was shallower than the operator-reported depth to the base of the protected ground
water resource and the next casing string was either partially cemented or uncemented. These
wells were further examined to determine whether any portion of the protected ground water
resource was uncemented based on the operator-reported depth to the base of the protected
ground water resource and the location of the cement top behind the next deeper casing. There
were sufficient data in the well files to determine that a portion of the operator-reported protected
ground water resource was uncemented below the surface casing in an estimated 600 wells (10-
1,800). The well files representing the remaining 1,900 wells (1,500-2,400) did not have sufficient
data to determine whether the operator-reported protected ground water resource was
uncemented or cemented. In these cases, there was ambiguity either in the depth of the base of the
01
¦Q
E
20,000
15,000
10,000
5,000
0
41
-------
Well Design and Construction
May 2015
operator-reported protected ground water resource68 or in the top of the protected ground water
resource.
Figure 17 also shows that all wells in which the operator-provided protected ground water
resource is "base not identified" had surface casing that extended below the depth used to identify
the location of the protected ground water resource (1,500 wells; 1,300-1,700). Since the base of
the protected ground water resource is not known for these wells, it could not be determined if the
surface casing or another cemented string of casing covers the entire protected ground water
resource in the annular space.
Estimated Separation Distance. The separation distance was estimated by subtracting the depth of
the operator-reported protected ground water resources from the measured depth of the point of
shallowest hydraulic fracturing (Figure 18). In nearly all cases, measured depths of the point of
shallowest hydraulic fracturing were deeper than operator-reported depths to protected ground
water resources, as shown in Figure 18. In an estimated 90 wells (10-300), the measured depth of
the point of shallowest hydraulic fracturing was shallower than the base of the operator-reported
protected ground water resource.69
Estimated separation distances shown in Figure 18 are not always equal to the true vertical
separation distances, which is defined as the vertical distance from the point of shallowest
hydraulic fracturing to the depth of the operator-reported protected ground water resources. In
perfectly vertical wells, the estimated separation distance is the true vertical separation. True
vertical separation distances for deviated and horizontal wells are expected to be smaller than the
estimated separation distances shown in Figure 18. For wells in which the operator-reported
protected ground water resource is "base not identified," the estimated separation distance
between the point of shallowest hydraulic fracturing and the actual base of the protected ground
water resource is likely to be smaller than what is shown in Figure 18.
5. Potential Subsurface Fluid Movement Pathways
The results presented in Section 4 indicate a range of well design and construction characteristics
for oil and gas production wells hydraulically fractured by nine service companies between
approximately September 2009 and September 2010. Some well design and construction
characteristics can act as barriers (e.g., casing and cement) that block pathways for subsurface fluid
movement between ground water resources and other formations. Subsurface fluid movement
depends on many factors, including the existence of a pathway, the presence of a fluid, and a driving
force. Pathways, in this report, are defined as any opening along which fluids can flow and can exist
regardless of whether fluids are flowing. Fluids can include gases or liquids naturally present in the
subsurface (e.g., brine, methane, or other hydrocarbons) or gases or liquids that are introduced into
the subsurface during well construction and completion (e.g., hydraulic fracturing fluid). Naturally
68 In these cases, operators indicated that the base of the protected ground water resource was less than or equal to a
given depth. The maximum value provided was used in this analysis.
69 In these cases, the top of the protected ground water resource was also identified and was shallower than the measured
depth of the point of shallowest hydraulic fracturing.
42
-------
Well Design and Construction
May 2015
0 2,000
Number of Wells
4,000 6,000 8,000 10,000 12,000 14,000
01
01
Negative
0-999
Q. 1,000-1,999
0)
Q
"§ 2,000-2,999
<5 3,000-3,999
§
™ 4,000-4,999
<->
c
2 5,000-5,999
O 6,000-6,999
g_ 7,000-7,999
01
to
T3
§ 8,000-8,999
9,000-9,999
>10,000
H
I Base of protected ground water identified
Base of protected ground water not identified
Figure 18. Estimated separation distance between the point of shallowest hydraulic fracturing and the depth of
operator-reported protected ground water resources for oil and gas production wells hydraulically fractured by
nine service companies between approximately September 2009 and September 2010. Well counts are grouped
according to whether or not the base of the operator-reported protected ground water resources was identified.
The "negative" category indicates the number of wells in which the point of shallowest hydraulic fracturing is at a
measured depth that is less than the operator-reported depth of protected water resources. Error bars display 95
percent confidence intervals.
present fluids can originate from the targeted geologic formation or from intermediate water-
and/or hydrocarbon-bearing zones (Council of Canadian Academies, 2014). Subsurface fluid
movement can occur due to gravity, a pressure differential, or gas buoyancy (Council of Canadian
Academies, 2014; Flewelling and Sharma, 2014; Kissinger etal., 2013; Myers, 2012).
Impacts to drinking water resources may occur when fluids flow through a continuous pathway
that extends from a source zone to a ground water resource. A continuous pathway could involve
pathways along the well that are introduced during well construction, pathways through the
overlying geology, or some combination of the two (Council of Canadian Academies, 2014;
Kissinger etal., 2013; Myers, 2012; US EPA, 2012; Vengosh etal., 2014). The discussion below
focuses only on the potential for well construction pathways to exist and does not address other
potential pathways or the presence of a fluid or driving force.
43
-------
Well Design and Construction
May 2015
Figure 19 illustrates two potential pathways related to well construction that connect a well to a
ground water resource: (1) from the inside of the well to the outside (Pathway A in Figure 19) and
(2) along the outside of the well (Pathway B in Figure 19). While Figure 19 illustrates potential
pathways for fluid movement directly to ground water resources, subsurface fluid movement to
ground water resources may occur via a combination of Pathways A and B. Although combinations
of these pathways are possible, the discussions presented below focus on the presence or absence
of Pathway A and Pathway B at any point along the well.
Diagram not to scale
Ground Surface
Conductor
Protected ground
water resource
Surface
Hydrocarbon-bearing zone
Water-yielding zone
Intermediate
Legend
H Cement
^ Casing
— Wellbore
Inducedfracture
Targeted geologic zone
Gas or oil
Production
Figure 19. Potential well construction pathways for subsurface fluid movement. Pathway A illustrates fluid
movement from the inside of the well to the outside. Pathway B illustrates fluid movement along the outside of
the well. Fluid movement along these potential pathways is dependent on the existence of the pathway, the
presence of a fluid, and a pressure differential.
5.1. Potential Pathway A: Inside to outside
Pathway A can occur at any point along the well. When Pathway A directly extends to a drinking
water resource, there can be impacts to that resource. When Pathway A does not extend to a
drinking water resource, and instead extends to a point below or above the drinking water
resource, an additional pathway (e.g., Pathway B) is needed to connect Pathway A to a drinking
water resource.
The existence of Pathway A depends, in part, upon the presence or absence of casing and cement.
Casings and cement sheaths act as barriers to fluid movement by interrupting Pathway A (Davies et
44
-------
Well Design and Construction
May 2015
al., 2014; Gray et al., 2009; GWPC, 2009; King and King, 2013). Wells are often constructed with
multiple barriers to minimize or preventwell integrity failures (Davies etal., 2014; King and King,
20 1 3).70 This section summarizes the number of casings (Figure 20) and cement sheaths (Figure
21) between the inside of a well and the wellbore at different measured depths. Conductor casing
or other shallow, often uncemented casings were excluded from these analyses.
Figure 20 shows the estimated number of wells having zero, one, two, or three casing strings
separating the inside of the well from the wellbore at different measured depths.71 In general, the
number of casing strings decreased as the measured depth increased, which was expected given
the purpose of each casing string, as described in Section 2. At measured depths reflective of
operator-reported ground water resources (generally <2,000 feet, as shown in Figure 16), most
wells had either one or two casing strings. Atvery shallow depths (<500 feet), there was a
relatively large proportion of wells that had three casing strings. A smaller number of wells
(<1,500) had no casing (open hole; Figure 20b) between 700 and 4,200 feet along the length of the
well. By design, these wells had no casing or cement barriers between the inside of the well and the
surrounding environment along the open hole completion section (Figure 20b).
Figure 21 shows the estimated number of wells having zero, one, two, or three cement sheaths
separating the inside of a well from the wellbore at different measured depths. Figures 20 and 21
show that, at many measured depths below the surface, there was typically one more casing string
present than the number of cement sheaths, which suggests that the casing strings counted in
Figure 21 were often not cemented to the surface. Most wells, at any given measured depth, had
one cement sheath between the inside of the well and the wellbore. At measured depths reflective
of operator-reported ground water resources (generally <2,000 feet), most wells had one cement
sheath. Between 4,900 and 9,100 wells had zero cement sheaths between 1,000 and 4,000 feet of
measured depth. At these depths, there may be an uncemented interval behind casing or an open
hole completion (panel b in Figure 21). A small number of wells (<100) had three cement sheaths
at shallow measured depths (<400 feet).
Figures 20 and 21 depict a range in the number of casing and cement barriers, from zero in open
hole completions to six when surface, intermediate, and production casing were each cemented to
the surface. The most common number of casing and cement barriers, at any point along a well, was
either two (one casing string and one cement sheath) or three (two casing strings and one cement
sheath). Under this scenario, Pathway A would have to extend through at least one casing and one
cement sheath to allow fluid to move from the inside of the well to the wellbore. A pathway through
70 A barrier failure occurs when a single element within a well fails, but there is no evidence of fluid migration to the
surrounding environment or surface. A well integrity failure occurs when no barriers are left to prevent fluid leaking into
the environment (King and King, 2013].
71 The analyses presented here do not consider well components that can serve as barriers, but are either temporary or
easily removed. Temporary well components that can serve as additional barriers may be installed during hydraulic
fracturing, such as a frac string, or after hydraulic fracturing, such as tubing set on a retrievable packer.
45
-------
Well Design and Construction
May 2015
Q)
-------
Well Design and Construction
May 2015
-------
Well Design and Construction
May 2015
cemented casings in preventing subsurface fluid movement, this section first summarizes the
presence and absence of cement along the outside of a well, and then briefly addresses the quality
of any cement present.
Presence or Absence of Cement. An estimated 6,800 wells (1,600-11,900) were fully cemented along
the outside of the well, from the bottom of the well to the ground surface. In these wells, all casing
strings were continuously cemented either to the surface or into the next overlying casing. An
estimated 15,300 wells (10,500-20,100) had one or more uncemented intervals along the outside of
the well, from the bottom of the well to the ground surface. Figure 12 indicates that uncemented
intervals were common for production casing, and to a lesser extent, intermediate casing.
Several studies that involved monitoring of uncemented annuli indicate that uncemented intervals
can be pathways for gas and liquid movement (Brufatto etal., 2003; Ohio Department of Natural
Resources, 2008; Watson and Bachu, 2009).72 After reviewing 142 wells in Alberta, Canada, Watson
and Bachu (2009) determined that the factor most strongly correlated with annular gas flow is the
length of the uncemented annulus. Additionally, Harrison (1985) demonstrated how a well with an
uncemented annulus between its casing and formation wall can unintentionally contribute
contaminants to subsurface drinking water supplies from intermediate depths and may do so
despite wellhead monitoring meant to detect it The Ground Water Protection Council (2014) also
identified the potential movement of fluids along the casing/formation annulus as a risk to ground
water. They noted that "the most effective means of protecting ground water from upward
migration in the annulus is the proper cementation of well casing across vertically impermeable
zones and ground water zones."
Many sources recommend that cement be placed across zones containing corrosive fluids, water in
need of protection, or hydrocarbons (Nelson and Guillot, 2006; Ross and King, 2007; Smith, 1976).
Participants at a 2013 EPA technical workshop on well construction and operation noted that full
cementing of annular spaces can enhance barrier functioning, but that cement placed to the surface
eliminates an operator's ability to monitor annular pressure for insights into well conditions during
operations (US EPA, 2013a). If annulus monitoring or testing detects fluid movement that causes
concern, a common response is to place cement in the uncemented portion of the well, if possible
(Nelson and Guillot, 2006; Smith, 1976; US EPA, 2013a).
The presence or absence of cemented casing strings between the targeted geologic formation and
protected ground water resources affects whether a direct pathway along the outside of the well
exists for fluids to move from the targeted geologic formation to protected ground water resources.
Figure 22 shows the estimated number of wells having zero, one, two, three, or four cemented
casings, independent of the quality of the cement, between the point of shallowest hydraulic
fracturing and the operator-reported protected ground water resource. The number of cemented
casings is plotted for different estimated separation distances in measured depths, as illustrated in
panel b.
72 In this report, an uncemented interval is defined as a segment of annular space along the outside of the well (between
the casing and the wellbore wall] that has no cement, while an uncemented annulus is defined as an uncemented interval
that reaches the surface.
48
-------
Well Design and Construction
May 2015
0)
¦Q
£
3
12,000
4,000
4,000
3,000
2,000
1,000
I.
k_d
I 0 cemented casings
1 cemented casing
2 cemented casings
I 3 cemented casings
4 cemented casings
£_1 I ii
^ f f <$> <£> <£> <£>' (9' <£>' ^
cF cP cF cP cP CP cP cP
V V ?>' V A'
Estimated Separation Distance In Measured Depth (feet)
(b)
cemented 3 cemented 2 cemented 1 cemented 0 cemented
casings casings casings casing casings
Ground Surface
Reported depth
- to ground
water resource
Estimated separation
distance shown in (a)
Legend
Cement
^ Casing
Wellbore
Inducedfracture
- Point of shallowest
hydraulicfracturing
Note: Not to scale. Conductor casing not shown.
Figure 22. (a) Cemented casing strings along the outside of the well between the point of shallowest hydraulic
fracturing and the operator-reported protected ground water resources. All operator-reported depths to
protected ground water resources were used, regardless of whether the depth was reported as the base of the
ground water resource. Wells in which the point of shallowest hydraulic fracturing was shallower than the base of
the operator-reported protected ground water resource are excluded from panel a. An estimated 2,200 wells
(1,000-3,400) contain cement whose depth interval is uncertain; these wells are also excluded from panel a. Error
bars display 95 percent confidence intervals, (b) Illustrative examples of wells with different cemented casing
strings between the point of shallowest hydraulic fracturing and operator-reported protected ground water
resources. The red arrows indicate cemented casing, and the green arrows show uncemented intervals.
Figure 22 shows that the number of cemented casings between the point of shallowest hydraulic
fracturing and the operator-reported protected ground water resource ranged from zero (open
49
-------
Well Design and Construction
May 2015
hole) to four, with more than half of the wells having two cemented casings. The number of
cemented casings generally increased as the estimated separation distance increased. At estimated
separation distances of less than 1,000 feet, wells had either zero cemented casings (open hole) or
one cemented casing. At the most commonly identified estimated separation distance (5,000-5,999
feet), wells had between one and three cemented casings. At estimated separation distances greater
than 9,000 feet, wells had two or three cemented casings between the point of shallowest hydraulic
fracturing and the operator-reported protected ground water resource.
As illustrated in Figure 22b, multiple cemented casings, which were commonly reported in this
study, did not preclude the presence of uncemented intervals, which have been shown to serve as
pathways for fluid movement (see above). Figure 22 also shows that more than half of the wells in
this study had two or more cemented casings that can interrupt Pathway B between the point of
shallowest hydraulic fracturing and the operator-reported protected ground water resource. This
indicates that there were often two or more barriers to potential subsurface fluid movement along
the outside of a well from the targeted geologic formation to protected ground water resources.
Quality of Cement. Cement coverage and bonding in the annulus between the outer casing and the
wellbore also influences Pathway B in Figure 19. Complete circumferential cement coverage and
bonding isolates zones that could otherwise communicate via the outside of a well (Brufatto et al.,
2003; Holt and Lahoti, 2012; Smolen, 1996). Most well files used in this study contained a standard
acoustic cement bond log, which can be used to evaluate both the extent of the cement placed along
the casing and the cement bond between the cement, casing, and wellbore. Figure 14 shows that
calculated cement bond indices were generally greater near the point of shallowest hydraulic
fracturing (the 50th percentile bond index was 0.7) than at shallower measured depths (the 50th
percentile bond index was 0.4 at 5,000 feet away from the shallowest point of hydraulic fracturing).
While standard acoustic cement bond logs can give an average estimate of bonding, they cannot
alone indicate zonal isolation, because they may not be properly run or calibrated (Boyd etal.,
2006; Flournoy and Feaster, 1963; Smolen, 1996). In addition, they cannot always discriminate
between full circumferential cement coverage by weaker cement and lack of circumferential
coverage by stronger cement (King and King, 2013; Smolen, 1996).73 A few studies have compared
cement bond indices to zonal isolation, with varying results. For example, Brown etal. (1970)
showed that among 16 South American wells with varying casing size and cement bond indices, a
cemented 5.5 inch diameter casing with a bond index of 0.80 along as little as five feet can act as an
effective seal. The authors also suggest that an effective seal in wells having calculated bond indices
differing from 0.80 are expected to have an inverse relationship between bond index and requisite
length of cemented interval, with longer lengths needed along casing having a lower bond index.
Conversely, King and King (2013) concluded that field tests from wells studied by Flournoy and
Feaster (1963) had effective isolation when the cement bond index ranged from 0.31 to 0.75.
Because of the ambiguity associated with using cement bond logs to determine zonal isolation, the
results of the cement bond log analysis conducted for this report should not be used to determine
the extent to which wells hydraulically fractured by the nine service companies had zonal isolation.
73 "Weaker cement" refers to cement that has lower attenuation properties. "Stronger cement" refers to cement that has
higher attenuation properties.
50
-------
Well Design and Construction
May 2015
The trends observed in Figure 14 suggest that the quality of the cement bond in the wells, as
indicated from calculated cement bond indices, was likely to be better near the point of shallowest
hydraulic fracturing.
6. Representativeness Analysis
The results presented in Sections 4 and 5 are representative of onshore oil and gas production
wells that were hydraulically fractured by nine service companies in the continental United States
between approximately September 2009 and September 2010. Wells fractured by the nine service
companies were predominantly vertical wells drilled after 2000 (Figure 4). Starting in about 2005,
however, horizontal drilling technologies began to be widely used to access unconventional
geologic formations for oil and gas production (Baker Hughes, 2014; GWPC and ALL Consulting,
2009).74 The reported increase in horizontal well completions raises two questions for this study:
• Are the wells hydraulically fractured by nine service companies representative of all
onshore oil and gas production wells hydraulically fractured in the continental United
States over the timeframe examined in this study?
• Would the results presented in this report change significantly if more horizontal wells
would have been sampled?
Since the information needed to compile a comprehensive list of hydraulically fractured oil and gas
production wells in the United States is not generally available, the EPA used Drillinglnfo (2012), a
commercial database compiled from data maintained by individual state oil and gas agencies, to
estimate the national number of horizontal and vertical wells hydraulically fractured in 2009 and
20 1 0.75 Well estimates from Drillinglnfo were based on the assumptions that all horizontal oil
and/or gas wells were hydraulically fractured in the year they started production and that all oil
and/or gas wells within a shale, coalbed, or low permeability formation,76 regardless of well
orientation, were also hydraulically fractured in the year they started production.77
Because hydraulically fractured well estimates from Drillinglnfo are limited by the year of initial
production (i.e., hydraulic fracturing is assumed to have occurred in the same year as the initial
production), well estimates from Drillinglnfo were compared to well estimates developed from the
subset of study wells that were hydraulically fractured within a year of being drilled. Since all wells
in this study were hydraulically fractured between approximately September 2009 and September
2010, the subset of wells used for this comparison included 179 study wells that were reported by
the well operators to have been drilled and hydraulically fractured within one year of the last date
74 In this instance, "unconventional" geologic formations refer to formations in which it is uneconomical to produce oil
and gas using traditional drilling and completion technologies. Examples include shale and tight gas formations.
75 Drillinglnfo does not track whether a well has been hydraulically fractured.
76 Identification of shale, coalbed, and low permeability formations is based on a crosswalk of information available in
Drillinglnfo with expert knowledge at the US Energy Information Administration (US EIA, 2012].
77 This is similar methodology to that used by the EPA for its Inventory of U.S. Greenhouse Gas Emissions and Sinks (US EPA,
2014b],
51
-------
Well Design and Construction
May 2015
in the study's timeframe (September 30, 2010). To compare well estimates from the study wells to
estimates derived from Drillinglnfo, well orientations for each study well were assigned using the
designations provided in Drillinglnfo (i.e., vertical, horizontal, directional, or unknown).78 Figure 23
shows the results of this comparison.
¦ Drillinglnfo
¦ This study
I I .i ¦
Vertical Horizontal Directional Unknown
Well Orientation (from Drillinglnfo)
Figure 23. Comparison of well estimates developed from Drillinglnfo to well estimates developed from well files
used in this study. Data obtained by the EPA from Drillinglnfo were used to develop estimates of the national
number oil and gas production wells drilled and hydraulically fractured in the continental United States in 2009
and 2010. Well files from this study were used to estimate the number of oil and gas production wells drilled
between approximately September 2009 and September 2010 and hydraulically fractured by nine service
companies within the same timeframe.
Using the well orientations designated in Drillinglnfo and applied to the study population, 67
percent (48-86 percent) of wells drilled in 2009 and 2010 and hydraulically fractured by the nine
service companies were estimated to be vertical wells, and 15 percent (9-21 percent) were
estimated to be horizontal. The number of oil and gas production wells hydraulically fractured in
2009 and 2010 estimated from Drillinglnfo also shows that more vertical wells were hydraulically
fractured (40 percent) in 2009 and 2010 than horizontal wells (31 percent). Differences in the
percentages of horizontal and vertical wells may arise from one or both of the following factors: (1)
incorrect or missing information in Drillinglnfo, as described below, or (2) the newly drilled wells
reported by the nine hydraulic fracturing service companies are not representative of newly drilled
and hydraulically fractured wells reported in Drillinglnfo.
The well estimates from Drillinglnfo in Figure 23 were developed using the assumptions outlined
above and then corrected after comparing well-specific information from the well files to data
78 Definitions for different well orientations in Drillinglnfo were not specified. When matching 171 study wells to data in
Drillinglnfo, 98 percent of horizontal wells, as defined in this study, were also designated as horizontal in Drillinglnfo.
Wells classified as vertical in this study were mostly designated as vertical in Drillinglnfo (78 percent of vertical wells, as
defined in this study]; 17 percent of vertical wells, as defined in this study, were categorized as unknown in Drillinglnfo.
Wells classified as deviated in this study were designated as either directional (43 percent of the deviated wells], vertical
(43 percent], or unknown (14 percent].
re
u
01
"£>
I T3
M- 01
0 3
01 +2
oo <•>
re re
01
u
OI
a.
100
90
80
70
60
50
40
30
20
10
0
52
-------
Well Design and Construction
May 2015
obtained by the EPA from Drillinglnfo. The well-specific comparisons suggested that data obtained
by the EPA from Drillinglnfo may not have been complete and that the assumptions used to identify
hydraulically fractured wells in Drillinglnfo underestimated the national number of hydraulically
fractured oil and gas production wells. As stated above, 179 of the study wells were used to assess
the national representativeness of wells surveyed in this study. Most (171) of the 179 study wells
were matched with well-specific entries in Drillinglnfo, indicating that 4 percent of oil and gas
production wells may not be found in Drillinglnfo or may not be easily identifiable in Drillinglnfo.79
Eighty-one percent of the 171 matched wells were correctly identified as hydraulically fractured
using the assumptions described above. The other 19 percent of wells were found in Drillinglnfo,
but not identified by the EPA as hydraulically fractured using the assumptions. All horizontal wells,
as defined by Drillinglnfo, were found to be correctly identified as hydraulically fractured, but
counts of hydraulically fractured vertical and directional wells were underestimated by 23 percent
and 129 percent, respectively. Given the limitations of the data in Drillinglnfo, it was not possible to
use Drillinglnfo to determine whether the oil and gas production wells hydraulically fractured by
the nine service companies between approximately September 2009 and September 2010 were
representative of all oil and gas production wells hydraulically fractured during that timeframe.
Well estimates derived from Drillinglnfo, as well as other data, suggest that the proportion of
horizontal well completions has increased since 2005 (Baker Hughes, 2014; Drillinglnfo, 2012).
Given this apparent shift in the prevalence of drilling orientation, the data from this study were
used to determine whether the results presented in Sections 4 and 5 are sensitive to different well
orientations. To do this, the statistical weights associated with each study well (Section 3.5) were
skewed such that the summed statistical weights of horizontal wells represented 90 percent of the
23,200 (21,400-25,000) oil and gas production wells estimated to have been hydraulically fractured
by nine service companies. The remaining 10 percent of the surveyed population was represented
by vertical and deviated study wells. Adjusting the statistical weights such that 90 percent of the
surveyed population is horizontal overestimates the proportion of hydraulically fractured oil and
gas production wells that are horizontal and should not be considered to produce true survey
estimates. However, it allows the identification of well design and construction characteristics that
may be different for horizontal well completions.
When 90 percent of the wells fractured by the nine service companies were assumed to be
horizontal, the majority of hydraulically fractured lithologies were shale and most of the wells were
less than a year old. Under this scenario, measured depths to the point of shallowest hydraulic
fracturing were generally deeper (>4,000 feet) and operator-reported depths to protected ground
water resources were generally shallower (<1,000 feet), when compared to Figures 8 and 16,
respectively. Consequently, the estimated separation distance between the point of shallowest
hydraulic fracturing and operator-reported protected ground water resources was generally more
than 3,000 feet, as measured along the wellbore. Additionally, the estimated number of wells in
which the point of shallowest hydraulic fracturing was shallower than the base of the operator-
reported protected ground water resources decreased.
79 This may be because not all Drillinglnfo records contained usable API numbers. Some entries in the API number field
had extra digits, while others did not appear to be API numbers.
53
-------
Well Design and Construction
May 2015
Cemented surface casing covered operator-reported protected ground water resources in nearly all
wells, when 90 percent of the wells fractured by nine service companies were assumed to be
horizontal. The number of wells having some portion of the operator-reported protected ground
water resource uncemented decreased, when compared to Figure 17. Furthermore, under this
scenario, the number of intermediate casings that were partially cemented was estimated to
increase, and the number of production casings that were set on formation packers with no cement
was also estimated to increase. The majority of partially cemented casings under this scenario were
estimated to have more than 50 percent of the evaluated casing length (i.e., from the bottom of the
casing to next overlying casing or ground surface, if surface casing) cemented, which was similar to
the trend presented in Figure 13. However, when 90 percent of the wells fractured by the nine
service companies were assumed to be horizontal, more partially cemented production casings
were estimated to have between 50 and 74 percent of the evaluated casing length cemented than
were estimated to have more than 75 percent of the evaluated casing length cemented.
Under the scenario in which 90 percent of the wells in the sampled population were horizontal,
nearly 80 percent of the wells were drilled within one year of being hydraulically fractured and
almost 100 percent were drilled within 10 years. This shows that the newer wells in the sampled
population are predominantly horizontal, which indicates that the service company well list
captured the increase in the use of horizontal drilling technologies that began in approximately
2005 (Baker Hughes, 2014; GWPC and ALL Consulting, 2009).
As previously stated, adjusting the statistical weights such that 90 percent of the wells in the
surveyed population were horizontal overestimated the proportion of hydraulically fractured oil
and gas production wells that were horizontal. Additionally, the horizontal study wells used in this
sensitivity analysis were mostly located in two sedimentary basins. Results from the sensitivity
analysis, therefore, heavily represent well design and construction characteristics in those two
basins and may not reflect characteristics across the United States. Consequently, the results from
the sensitivity analysis should not be interpreted to represent horizontal oil and gas production
wells in general. Rather, the results from the sensitivity analysis should only be used to identify well
characteristics that may be different for horizontal well completions.
7. Study Limitations
The survey design and data collection process have implications for the interpretation of the results
presented in Sections 4 and 5 of this report As described in Section 3, the EPA used a stratified
random sampling approach to select well files from nine oil and gas well operators based on well
location and operator size. The study wells were selected from a list of well identifiers provided by
nine hydraulic fracturing service companies and represent oil and gas production wells
hydraulically fractured by those companies between approximately September 2009 and
September 2010. As a result, the estimates of the frequency of occurrence of well design and
construction characteristics presented in Sections 4 and 5 are statistically representative of the
23,200 (21,400-25,000) oil and gas production wells estimated to have been hydraulically fractured
by the nine service companies in 2009 and 2010.
54
-------
Well Design and Construction
May 2015
The estimates may not be statistically representative of all oil and gas production wells
hydraulically fractured in the same time period. Comparisons between the service company well
list and other data, however, suggest that observations made in this report are likely indicative of
national well design and construction characteristics of oil and gas production wells hydraulically
fractured in 2009 and 2010. As described in Section 3.1, the geographic distribution of well
identifiers provided by the nine service companies resembled the geographic distribution of oil and
gas production in the United States in 2009 and 2010. Additionally, an analysis of data obtained
from Drillinglnfo (Section 6) estimated that more vertical wells were hydraulically fractured in
2009 and 2010 than horizontal wells, which is consistent with data reported in this study (Figure
6c). Given the similarities in geographic distribution and well orientations, the results presented in
this report are likely indicative of well design and construction characteristics of onshore oil and
gas production wells that were hydraulically fractured in the United States during the timeframe
examined in this study.
As mentioned in Section 6, the use of horizontal drilling technologies began to be widely used
around 2005 and continued to be widely used in 2014 (Baker Hughes, 2014). The sensitivity
analysis described in Section 6 identified well design and construction characteristics that may be
different for horizontal well completions than non-horizontal well completions. Since wells
hydraulically fractured by the nine service companies in this study were predominantly vertical
well completions, it is possible that the specific well estimates presented in Sections 4 and 5 would
be different if the survey had included more horizontal wells. The results of the sensitivity analysis
should not be interpreted to represent horizontal oil and gas production wells in general. It is also
possible that the estimates presented in this report may not apply to wells constructed and
hydraulically fractured after 2010, if well design and construction practices changed after 2010.
Estimates of the frequency of occurrence of well design and construction characteristics are
summarized in Section 4 and 5 at the national level. Given the survey design, it was not possible to
display results at a smaller scale, such as targeted geologic formations. Estimates of the frequency
of occurrence of well design and construction characteristics may be different for different regions
of the country, because of differences in local geologic characteristics, state regulations, and
company preferences, as described in Section 2. Despite variation in practices, the data in Sections
4 and 5 provide an overview of well design and construction characteristics of onshore oil and gas
production wells hydraulically fractured in the United States in 2009 and 2010 and indicate that
some characteristics were common.
Finally, the results presented in this report were generated from data provided by oil and gas well
operators. No attempt was made to independently and systematically verily the data supplied by
the operators, including depths to protected ground water resources. Consequently, the study
results, which include comparisons of operator-reported protected ground water resources to well
construction characteristics (Section 4.3.2), are of the same quality as the supplied data. Other
sources, such as state oil and gas agencies, may identify different values for the data elements used
in this study, which would affect the results presented in this report.
55
-------
Well Design and Construction
May 2015
8. Conclusions
This study used data from a survey of oil and gas production wells to describe: (1) well design and
construction characteristics of hydraulically fractured oil and gas production wells, (2) the
relationship of well design and construction characteristics to drinking water resources, and (3) the
number and relative location of well construction barriers that can block pathways for potential
subsurface fluid movement Results presented in this report are statistically representative of an
estimated 23,200 (21,400-25,000) onshore oil and gas production wells in the continental United
States that were hydraulically fractured in 2009 and 2010 by nine oil and gas service companies.
Oil and gas production wells hydraulically fractured by the nine service companies were
predominately vertical wells drilled between 2000 and 2010, but also included other well
orientations (i.e., horizontal and deviated) and wells drilled prior to 2000. The most common casing
configuration, used in 55 (33-75) percent80 of wells, was reported to be a surface and a production
casing. Eighty-seven (68-96) percent81 of wells had casing in the production zone that was
cemented and perforated for hydraulic fracturing. Sixty-six (44-83) percent82 of wells had one or
more uncemented intervals along the outside of the well, from the bottom of the well to the ground
surface, while 29 (13-53) percent83 were fully cemented over the same interval.
Production wells were found to be located near surface and ground water resources that are
currently used as sources of drinking water or may be used as sources for drinking water in the
future. While proximity alone does not determine impacts, it is a factor that should be considered
when assessing the potential for hydraulic fracturing to affect drinking water resources. Eighty-two
(63-92) percent84 of wells were located within 0.5 miles of a surface water resource (i.e., lake, pond,
or river). Fewer production wells were within 0.5 miles of either a private ground water well [13
(7-23) percent85] or a public water supply well [2 (1-10) percent86]. Operator-reported depths to
protected ground water resources ranged from just below ground surface to 8,000 feet deep.
Ninety-two (81-97) percent87 of wells passed through protected ground water resources less than
2,000 feet deep, as reported by well operators. Operator-reported protected ground water
resources were estimated to be covered by cement in the annular space in most wells, as described
in Section 4.3.2. Three (0.5-13) percent88 of wells had a portion of the operator-reported protected
ground water resource that was uncemented. Perforations used for hydraulic fracturing were
80 12,800 (7,600-18,000] wells
si 20,200 (17,500-23,000] wells
82 15,300 (10,500-20,100] wells
83 6,800 (1,600-11,900] wells
84 18,900 (15,200-22,600] wells
85 3,000 (900-5,100] wells
as 600 (10-1,500] wells
87 21,400 (18,900-24,000] wells
88 600 (10-1,800] wells
56
-------
Well Design and Construction
May 2015
placed deeper than the base of the protected ground water resources identified by well operators in
all but 0.4 (0.1-3) percent89 of wells.
As described in Section 5, subsurface fluid movement depends on many factors, including the
existence of a pathway, the presence of a fluid, and a driving force. This report focused only on the
potential for well construction pathways to exist and does not address the role of other possible
pathways or the role of other factors. Two potential pathways related to well construction were
considered: from the inside of the well to the outside and along the outside of the well. Subsurface
fluid movement to ground water resources may occur though either potential pathway or a
combination of both potential pathways. The presence or absence of casing and cement along the
well were evaluated to identify the number of well construction barriers to subsurface fluid
movement via these potential pathways.
Wells represented in this study had a range in the number of casing and cement barriers between
the inside of the production casing and the wellbore at any point along the well. The number of
casing and cement barriers ranged from zero in open hole completions to six when surface,
intermediate, and production casing were each cemented to the surface. The most common number
of casing and cement barriers, at any point along a well, was either two (one casing string and one
cement sheath) or three (two casing strings and one cement sheath). Multiple casing and cement
barriers can minimize or prevent well integrity failures that could result in subsurface fluid
movement, because all barriers would need to fail in order for a pathway from the inside of the well
to the outside to occur.
More than half of the wells represented in this study had two or more cemented casings between
the point of shallowest hydraulic fracturing and the operator-reported ground water resources.
This indicates that there were often two or more barriers to subsurface fluid movement along the
outside of a well, from the targeted geologic formation to protected ground water resources
reported by well operators. Multiple cemented casings between the shallowest point of hydraulic
fracturing and operator-reported protected ground water resources did not preclude the presence
of uncemented intervals along the outside of the well. Monitoring studies of uncemented annuli
indicate that uncemented intervals can be pathways for gas and liquid movement along the outside
of a well. Some companies and state regulators, however, note that cementing only across critical
zones (i.e., zones containing corrosive fluids, water in need of protection, or hydrocarbons) and
leaving uncemented annuli allows operators to monitor uncemented annuli for subsurface fluid
movement More detailed information is presented in Section 5.2.
The survey design and data collection process may have implications for the interpretation of the
results presented in this report. The results are statistically representative of oil and gas production
wells hydraulically fractured by nine oil and gas service companies in the continental United States
during 2009 and 2010. The extent to which these results may be statistically representative of all
production wells hydraulically fractured in the United States during the same time period could not
be determined. Nevertheless, comparisons between wells in this study and other data on oil and gas
production in the United States during 2009 and 2010 suggest that observations made in this
89 90 (10-300] wells
57
-------
Well Design and Construction
May 2015
report are likely indicative of oil and gas production wells hydraulically fractured during this time
period.
Estimates of the frequency of occurrence of well design and construction characteristics are
presented at the national scale. Estimates may be different for different regions of the country,
because of differences in local geologic characteristics, state regulations, and company preferences.
It is also possible that the estimates presented in the report may not apply to wells constructed and
hydraulically fractured after 2010, if well design and construction practices have changed (e.g., a
greater proportion of horizontal well completions). Additionally, the results presented in this
report are generated from data provided by oil and gas well operators. The EPA did not attempt to
independently and systematically verify data supplied by the operators. Consequently, the study
results, which include comparisons of operator-reported protected ground water resources to well
construction characteristics, are of the same quality as the supplied data.
This report presents the results of a survey of onshore oil and gas production wells hydraulically
fractured in the continental United States during 2009 and 2010, using data provided by well
operators. Two potential pathways for subsurface fluid movement were examined—from the inside
of the well to the outside and along the outside of the well. The following key findings contribute to
an understanding of the role of well design and construction practices with respect to these
pathways:
The wells generally had multiple layers of casing and cement that can act as barriers to
subsurface fluid movement by interrupting pathways for potential subsurface fluid movement.
Multiple casing and cement barriers can prevent pathways from forming, because all
barriers would need to fail in order for a pathway for potential subsurface fluid movement
to occur. The most common number of barriers to potential subsurface fluid movement
from inside of the well to the outside, at any point along the well, was either two (one casing
string and one cement sheath) or three (two casing strings and one cement sheath).
Additionally, there were often two or more barriers (i.e., cemented casings) to potential
subsurface fluid movement along the outside of a well, from the targeted geologic formation
to operator-reported protected ground water resources.
While multiple barriers were often present in hydraulically fractured oil and gas production
wells; pathways for potential subsurface fluid movement were identified in some wells.
Uncemented intervals have been shown to be pathways for subsurface fluid movement
along the outside of the well. An estimated 66 (44-83) percent90 of wells had one or more
uncemented intervals, and 3 (0.5-13) percent91 of wells had uncemented intervals within
the operator-reported protected ground water resources. Casing perforations placed at
depths shallower than the base of operator-reported protected ground water resources can
create a pathway for fluids to flow from the inside of the well to a ground water resource, if
a ground water resource is present at that depth. An estimated 0.4 (0.1-3) percent92 ofwells
90 15,300 (10,500-20,100] wells
600 (10-1,800] wells
92 90 (10-300] wells
58
-------
Well Design and Construction
May 2015
had perforations used for hydraulic fracturing that were placed shallower than the base of
the operator-reported protected ground water resource.
These results, as well as other information on well characteristics, the relationship between
drinking water resources and hydraulically fractured wells, and casing and cement barriers that can
prevent subsurface fluid movement, highlight important factors that should be considered when
assessing the potential impacts of hydraulically fractured oil and gas production wells on drinking
water resources.
59
-------
Well Design and Construction
May 2015
References
ArcGIS version 10. 2010. Environmental Systems Research, Inc. (ESRI), Redlands, California.
Albert, L.E., Standley, T.E., Tello, L.N., and Alford, G.T. 1988. A Comparison of CBL, RBT, and PET
Logs in a Test Well with Induced Channels. Journal of Petroleum Technology 40:1211-1216.
Baker, R. 1979. A Primer of Oil-Weil Drilling. 4th edition. Petroleum Extension Service, Austin,
Texas.
Baker Hughes. 2014. North America Rotary Rig Count (Jan 2000 - Current). Available at
http://phx.corporate-ir.net/phoenix.zhtml?c=79687&p=irol-reportsother. Accessed October 20,
2014.
Bellabarba, M., Bulte-Loyer, H., Froelich, B., Le Roy-Delage, S., van Kuijk, R., Zeroug, S., Guillot, D.,
Moroni, N., Pastor, S., and Zanchi, A. 2008. Ensuring Zonal Isolation Beyond the Life of the Well.
Oilfield Review 18-31.
Boyd, D., Al-Kubti, S., Khedr, O.H., Khan, N., Al-Nayadi, K., Degouy, D., Elkadi, A., and Kindi, Z.A. 2006.
Reliability of Cement Bond Log Interpretations Compared to Physical Communication Tests
Between Formations (SPE 101420). 2006 Abu Dhabi International Petroleum Exhibition and
Conference, Abu Dhabi, United Arab Emirates, November 5-8, 2006. Society of Petroleum Engineers.
Brown, H.D., Grijalva, V.E., and Raymer, L.L. 1970. New Developments in Sonic Wave Train Displays
and Analysis in Cased Holes. SPWLA 11th Annual Logging Symposium, Los Angeles, California, May
3-6,1970. Society of Petrophysicists and Well Log Analysts.
Brufatto, C., Cochran, J., Conn, L., Power, D., El-Zeghaty, S.Z.A.A., Fraboulet, B., Griffin, T., James, S.,
Munk, T., Justus, F., Levine, J.R, Montogmery, C., Murphy, D., Pfeiffer, J., Pornpoch, T., and Rishmani,
L. 2003. From Mud to Cement: Building Gas Wells. Oilfield Review 62-76.
Carroll, R.J., Ruppert, D., Stefanski, L.A., and Crainiceanu, C.M. 2006. Measurement Error in
Nonlinear Models: A Mondern Perspective. 2nd edition. Chapman and Hall/CRC, Boca Raton,
Florida.
Chebotarev, I.I. 1955. Metamorphism of Natural Waters in the Crust of Weathering—2. Geochimica
et Cosmochimica Acta 8:137-170.
Christen, K. 2003. Environmental News: Increases Projected in Coal Bed Methane Production.
Environmental Science and Technology 37:244A.
Council of Canadian Academies. 2014. Environmental Impacts of Shale Gas Extraction in Canada.
Available at http://scienceadvice.ca/en/assessments/completed/shale-gas.aspx. Accessed March
26,2015.
60
-------
Well Design and Construction
May 2015
Davies, R.J., Almond, S., Ward, R.S., Jackson, R.B., Adams, C., Worrall, F., Herringshaw, L.G., Gluyas,
J.G., and Whitehead, M.A. 2014. Oil and Gas Wells and Their Integrity: Implications for Shale and
Unconventional Resource Exploitation. Marine and Petroleum Geology 56:239-254.
Devereux, S. 1998. Practical Well Planning and Drilling Manual. PennWell Publishing Company,
Tulsa, Oklahoma.
Dewan, J.T. 1983. Essentials of Modern Open-Hole Log Interpretation. PennWell Corp, Tulsa,
Oklahoma.
Drillinglnfo. 2012. DI Desktop® August 2012 Download. Drillinglnfo, Inc.
Fitzgerald, D.D., McGhee, B.F., and McGuire, J.A. 1985. Guidelines for 90% Accuracy in Zone-Isolation
Decisions. Journal of Petroleum Technology 37:2013-2022.
Flewelling, S.A., and Sharma, M. 2014. Constraints on Upward Migration of Hydraulic Fracturing
Fluid and Brine. Groundwater 52:9-19.
Flournoy, R.M., and Feaster, J.H. 1963. Field Observations on the use of the Cement Bond Log and its
Application to the Evaluation of Cementing Problems (SPE 632). 38th Annual Fall Meeting of the
Society of Petroleum Engineers of AIME, New Orleans, Louisiana, October 6-9,1963. Society of
Petroleum Engineers of AIME.
Freeze, R.A., and Cherry, J.A. 1979. Groundwater. Prentice Hall, Inc., Englewood Cliffs, New Jersey.
Gray, K.E., Podnos, E., and Becker, E. 2009. Finite-Element Studies of Near-Wellbore Region During
Cementing Operations: Part 1. SPE Drilling & Completion 24:127-136.
Ground Water Protection Council (GWPC). 2009. State Oil and Natural Gas Regulations Designed to
Protect Water Resources. GWPC for U.S. Department of Energy. Available at http://www.gwpc.org/
sites/default/files/state_o il_and_gas_regulations_designed_to_protect_water_resources_0.pdf.
Accessed May 9, 2014.
GWPC and ALL Consulting. 2009. Modern Shale Gas Development in the United States: A Primer.
GWPC and All Consulting for US Department of Energy. Available at http://energy.gov/sites/prod/
files/2013/03/f0/ShaleGasPrimer_0nline_4-2009.pdf. Accessed June 18, 2014.
GWPC. 2014. State Oil and Natural Gas Regulations Designed to Protect Water Resources: 2014
Edition. Available at http://www.gwpc.org/sites/default/files/files/0il%20and%20Gas
%20Regulation%20Report%20Hyperlinked%20Version%20Final-rfs.pdf. Accessed October 27,
2014.
Gupta, D.V.S., and Valko, P. 2007. Fracturing Fluids and Formation Damage. In Economides, M.J., and
Martin, T., editors, Modern Fracturing: Enhancing Natural Gas Production. Energy Tribune
Publishing, Inc., Houston, Texas.
Harrison, S.S. 1983. Evaluating System for Ground-Water Contamination Hazards Due to Gas-Well
Drilling on the Glaciated Appalachian Plateau. Groundwater 21:689-700.
61
-------
Well Design and Construction
May 2015
Harrison, S.S. 1985. Contamination of Aquifers by Overpressuring the Annulus of Oil and Gas Wells.
Groundwater 23:317-324.
Holt, C., and Lahoti, N. 2012. Dynamic Cementation Improves Wellbore Construction and Reduces
the Hazards to Groundwater Contamination in Shale Plays (SPE 162739). SPE Canadian
Unconventional Resources Conference, Calgary, Alberta, Canada, October 30 - November 1, 2012.
Society of Petroleum Engineers.
International Association of Drilling Contractors. 2015. Recomplete. Drilling Lexicon: Oil and Gas
Drilling Glossary. Available at http://www.iadclexicon.org/recomplete/. Accessed March 26, 2015.
Jackson, R.B., Vengosh, A., Darrah, T.H., Warner, N.R, Down, A., Poreda, R.J., Osborn, S.G., Zhao, K.,
and Karr, J.D. 2013a. Increased Stray Gas Abundance in a Subsetof Drinking Water Wells Near
Marcellus Shale Gas Extraction. Proceedings of the National Academy of Sciences 110:11250-11255.
Jackson, RE., Gorody, A.W., Mayer, B., Roy, J.W., Ryan, M.C., and Van Stempvoort, D.R. 2013b.
Groundwater Protection and Unconventional Gas Extraction: The Critical Need for Field-Based
Hydrogeological Research. Groundwater 51:488-510.
Jutton, J., Toma, I., Morel, Y., and Ferreol, B. 1991. Integration of Cement Job Data for Better Bond
Index Interpretation (SPE 21690). SPE Production Operations Symposium, Oklahoma City,
Oklahoma, April 7-9,1991. Society of Petroleum Engineers.
King, G.E., and King, D.E. 2013. Environmental Risk Arising from Well Construction Failure:
Differences Between Barrier and Well Failure, and Estimates of Failure Frequency Across Common
Well Types, Locations, and Well Age. SPE Production & Operations 28:323-344.
Kissinger, A., Helmig, R, Ebigbo, A., Class, H., Lange, T., Sauter, M., Heitfeld, M., Kliinker, J., and
Jahnke, W. 2013. Hydraulic Fracturing in Unconventional Gas Reservoirs: Risks in the Geological
System, Part 2. Environmental Earth Sciences 70:3855-3873.
Krygowski, D. 2004. Basic Well Log Analysis. 2nd edition. American Association of Petroleum
Geologists, Tulsa, Oklahoma.
Lohr, S.L. 2010. Sampling: Design and Analysis. 2nd edition. Brooks/Cole, Boston, Massachusetts.
Martin, T., and Valko, B. 2007. Hydraulic Fracture Design for Production Enhancement. In
Economides, M.J., and Martin, T., editors, Modern Fracturing: Enhancing Natural Gas Production.
Energy Tribune Publishing, Inc., Houston, Texas.
McDaniel, J., Watters, L., and Shadravan, A. 2014. Cement Sheath Durability: Increasing Cement
Sheath Integrity to Reduce Gas Migration in the Marcellus Shale Play (SPE 168650). SPE Hydraulic
Fracturing Technology Conference, The Woodlands, Texas, February 4-6, 2014. Society of
Petroleum Engineers.
Myers, T. 2012. Potential Contaminant Pathways from Hydraulically Fractured Shale to Aquifers.
Groundwater 50:872-882.
62
-------
Well Design and Construction
May 2015
Nelson, E.B., and Guillot, D. 2006. Well Cementing. 2nd edition. Schlumberger, Sugar Land, Texas.
North American Datum Conversion (NADCON) Utility version 4.2. 2011. National Oceanic and
Atmospheric Administration.
Nutt, R.L. 2012. Some Interesting and Unusual Cement Bond (CBL) and Production Log Examples.
SPWLA 53rd Annual Logging Symposium, Cartagena, Colombia, June 16-20, 2012. Society of
Petrophysicists and Well Log Analysts.
Ohio Department of Natural Resources, Division of Mineral Resources Management. 2008. Report
on the Investigation of the Natural Gas Invasion of Aquifers in Bainbridge Township of Geauga
County, Ohio. Available athttp://oilandgas.ohiodnr.gov/portals/oilgas/pdf/bainbridge/report.pdf.
Accessed December 22, 2014.
Osborn, S.G., Vengosh, A., Warner, N.R., and Jackson, R.B. 2011. Methane Contamination of Drinking
Water Accompanying Gas-Well Drilling and Hydraulic Fracturing. Proceedings of the National
Academy of Sciences 108:8172-8176.
Pilkington, P.E. 1992. Cement Evaluation: Past, Present, and Future. Journal of Petroleum Technology
44:132-140.
Ross, D., and King, G. 2007. Well Completions. In Economides, M.J., and Martin, T., editors, Modern
Fracturing: Enhancing Natural Gas Production. Energy Tribune Publishing, Inc., Houston, Texas.
Schlumberger. 1991. Log Interpretation Principles/Applications. Schlumberger, Sugar Land, Texas.
Schlumberger. 2014. Remedial Cementing. Oilfield Glossary. Available at
http://www.glossary.oilfield.slb.com/. Accessed November 20, 2014.
Schlumberger, D. 1984. Cementing Technology. Nova Communications, Ltd., London, England.
Smith, D.K. 1976. Cementing (Monograph - Society of Petroleum Engineers; v. 4). Henry L. Doherty
Memorial Fund of AIME, Dallas, Texas.
Smolen, J.J. 1996. Cased Hole and Production Log Evaluation. PennWell Corp, Tulsa, Oklahoma.
Spellman, F. 2013. Environmental Impacts of Hydraulic Fracturing. Taylor & Francis Group, LLC,
Boca Raton, Florida.
The Cadmus Group, Inc. 2013. Supplemental Programmatic Quality Assurance Project Plan for
Work Assignment 5-58, Including Amendment 1: Well File Review in Support of Study of Potential
Impacts of Hydraulic Fracturing on Drinking Water Resources, Part 3. Available at
http://www2.epa.gov/sites/production/files/documents/cadmusspqapp-wellfiles-508.pdf.
Accessed October 27, 2014.
US Energy Information Administration (US EIA). 2011. Personal communication, S. Limerick, Z, Inc.
April 7,2011.
63
-------
Well Design and Construction
May 2015
US EIA. 2012. Formation Crosswalk. Provided July 7, 2012.
US EIA. 2014a. Natural Gas Gross Withdrawals and Production. Available at http://www.eia.gov/
dnav/ng/ng_prod_sum_dcu_r98_a.htm. Accessed November 20, 2014.
US EIA. 2014b. Crude Oil Production. Available at http://www.eia.gov/dnav/pet/
pet_crd_crpdn_adc_mbbl_a.htm. Accessed November 20, 2014.
US Environmental Protection Agency (US EPA). 2004. Evaluation of Impacts to Underground
Sources of Drinking Water by Hydraulic Fracturing of Coalbed Methane Reservoirs. Available at
http://water.epa.gov/type/groundwater/uic/class2/hydraulicfracturing/
wells_coalbedmethanestudy.cfm. Accessed October 27, 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.
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_110211_final_508.pdf. Accessed June 18, 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 October 27, 2014.
US EPA. 2013a. Summary of the Technical Workshop on Well Construction/Operation and
Subsurface Modeling, April 16-17, 2013, and Subsurface Modeling Technical Follow-up Discussion,
June 3, 2013. Available athttp://www2.epa.gov/sites/production/files/2013-10/documents/
summary-technical-workshop-well-construction-operation-subsurface-modeling.pdf. Accessed
November 20, 2014.
US EPA. 2013b. National Hydraulic Fracturing Study Evaluation of Existing Production Well File
Contents: Quality Assurance Project Plan. Available at http://www2.epa.gov/sites/production/
files/documents/wiser_508qapp_well_file_review_research_final_2012.pdf. Accessed October 27,
2014.
US EPA. 2014a. Safe Drinking Water Information System. Available at http://www.epa.gov/enviro/
facts/sdwis/index.html. Accessed May 27, 2014.
US EPA. 2014b. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2012. EPA/430/R-
14/003. Available at http://www.epa.gov/climatechange/ghgemissions/
usinventoryreporthtml#fullreport Accessed July 30, 2014.
US EPA. 2014c. National Hydraulic Fracturing Study Evaluation of Existing Production Well File
Contents: Quality Assurance Project Plan, Amendment 1. Available at http://www2.epa.gov/sites/
64
-------
Well Design and Construction
May 2015
production/files/documents/wiser_5 0 8qapp_well_file_revie w_research_final_2 012 .pdf. Accessed
October 27, 2014.
US EPA. 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.
US Geological Survey (USGS). 2012. National Assessment of Oil and Gas: Assessment of
Undiscovered Oil and Gas Resources of the Ordovician Utica Shale of the Appalachian Basin
Province, 2012. Available at http://pubs.usgs.gov/fs/2012/3116/FS12-3116.pdf. Accessed June 19,
2014.
USGS. 2014. National Hydrography Dataset Available at http://nhd.usgs.gov/. Accessed May 27,
2014.
Van Stempvoort, D., Maathuis, H., Jaworski, E., Mayer, B., and Rich, K. 2005. Oxidation of Fugitive
Methane in Ground Water Linked to Bacterial Sulfate Reduction. Groundwater 43:187-199.
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 and Technology 48:8334-8348.
Watson, T.L., and Bachu, S. 2009. Evaluation of the Potential for Gas and CO2 Leakage Along
Wellbores. SPEDrilling & Completion 24:115-126.
Waxman, H.A., and Markey, E.J. 2010. Memorandum to the Members of the Subcommittee on
Energy and Environment on Examining the Potential Impact of Hydraulic Fracturing, dated
February 18, 2010. US House of Representatives, Committee on Energy and Commerce.
Waxman, H.A., Markey, E.J., and DeGette, D. 2011. Letter addressed to US Environmental Protection
Agency Administrator Lisa Jackson, dated January 31, 2011. US House of Representatives,
Committee on Energy and Commerce.
Westat. 2013. Quality Assurance Project Plan vl for WA 3-4: Hydraulic Fracturing. Available at
http://www2.epa.gov/sites/production/files/2014-01/documents/westat-qapp.pdf. Accessed
October 27, 2014.
65
-------
Well Design and Construction
May 2015
Appendix A: Survey Design and Confidence Intervals
A.l. Optimization Algorithm
An optimization algorithm was used to determine the number of well identifiers to be selected from
each of the 31 unique operator-shale play combinations represented by the nine well operators.
The optimization algorithm was designed to select a sample of 400 well identifiers across the 31
operator-shale play combinations that best maximized the geographic diversity of the sample and
maximized the precision of any estimates calculated from the sample using the statistical weights
described in Section 3.5.
The algorithm identified the optimal solution by minimizing a penalty function, subject to the
constraints that at least two well identifiers were selected in each operator-shale play combination
(or one if there was only one) and restricting the total sample size to 400. The penalty function is
given in equation A.l.
The first two terms in equation A. 1 reflect the goals for selecting well identifiers: maximizing the
precision of statistical estimates and maximizing geographic diversity. The first term is the
coefficient of variation (cv) of the statistical weights for the sample of 400 well identifiers.93 One
important method for maximizing precision is to have the final statistical weights close to equal
across the entire sample of 400 well identifiers (i.e., cv is small); this prevents a few results from
dominating the entire estimate (Lohr, 2010). The statistical weights are the product of the inverse
of the probabilities of selection at each stage of sampling. In the first-stage selection of well
operators, the four large operators were selected with higher probabilities than the medium and
small operators (Section 3.2). Balancing the statistical weights, therefore, required selecting well
identifiers at a higher rate from the medium and small operators at the second stage.
The second term in equation A.1 includes a standard chi-squared statistic to measure the
geographic diversity of the sample. The chi-squared statistic compares the number of well
identifiers selected for each geographic region (n;) to the expected number of well identifiers for
that region assuming proportional representation (r;). A value of zero for this term indicates that
the sample of 400 well identifiers has the same geographic distribution as the service company well
list
Equation A.1 includes a third term that constrains extreme sample sizes by introducing a penalty if
the number of well identifiers selected from each operator is not between 33 and 75; this controls
the burden placed on the nine operators by limiting the number of well files requested from any
one operator. Alternative minimum and maximums were examined as part of the optimization
procedure. The penalty associated with extreme sample sizes (D) was calculated using equation A.2.
93 The coefficient of variation is defined as the ratio of the standard deviation to the mean.
Penalty = 1.0 cv + 0.5
(A.1)
66
-------
Well Design and Construction
May 2015
D = 1 + £Uck
Var(M)
(A.2)
where
di = [m - 75)3
= 0
if n, > 75
if 75 > rij > 33
= (33 - n,)3 if n, < 33
and Var[M) is equal to the variance of the number of well identifiers selected from each operator-
shale play combination (M;).
The penalty function, and thus the number of well identifiers selected from each operator-shale
play combination, depends on the coefficients in each term in equation A.l. The coefficients are
relative in the sense that they can all be increased by a factor of two or ten without affecting the
sample size allocation. As a result, the coefficient in the first term was set to 1.0 and the other two
coefficients (0.5 and 0.1) were adjusted. The exact setting of each coefficient was based on trial and
error, attempting to minimize the first term, while keeping the sample size within an acceptable
range and minimizing the chi-square function without having one geographic region contribute
many more well identifiers in the chi-square term than the other geographic regions.
The optimal number of well identifiers from each operator-shale play combination minimized the
penalty function shown in equation A.l. In the final sample of 400 well identifiers, the number of
well identifiers selected for each company ranged from 16 to 77. For this sample, the estimated
coefficient of variation was 0.496; two companies had well counts over 75 (77.6 and 76.2 before
rounding); and the expected versus achieved sample size in each geographic region, on which the
chi-squared statistic is based, were: East (expected 60.5, achieved 61.6), West (expected 102.1,
achieved 98.9), South (expected 138.6, achieved 135.8), and Other (expected 98.9, achieved 103.7).
A.2. Variance Estimation and Confidence Intervals
The sampling approach used to select 400 (subsampled down to 350) well identifiers determined
the approach used to estimate the variance, which is used to calculate the 95 percent confidence
intervals. The sampling approach used in this study is best approximated as a stratified, two-stage
sample design in which well operators were selected in the first stage and well identifiers were
selected in the second stage.
As described in Section 3.2, nine well operators were selected from six strata by simple random
sampling in the first stage. In the second stage, well identifiers for each of the nine well operators
were stratified by shale play or shale play combination, which resulted in 31 unique combinations
of well operator and shale play or shale play cluster. For eight of these combinations, or sampling
strata, all well identifiers were selected in the second stage. This was the case for the three small
well operators (all well identifiers from each small operator were selected) and for strata with one
or two well identifiers.
The number of well identifiers selected from each of the remaining 23 strata was determined using
the optimization algorithm described in Section A.l. A simple random sample of that size was then
selected from each stratum. Since the optimization algorithm accounting for the number of well
67
-------
Well Design and Construction
May 2015
identifiers was a function of the total sample across all sampled operators, the selection of well
identifiers from these remaining strata more closely resembled a two-phase sampling approach,
rather than a two-stage sampling approach [Chapters 5 and 12 in Lohr (2010); Section 4.3 and
Chapter 9 in Sarndal et al. (2003)]. Even in these strata, since the resulting sample was selected
within first-stage operators, the sample retains some of the characteristics of two-stage sample and
its variance will be less than the general formula for two-phase variance. In the terminology of
Sarndal et al. (2003), the samples of well identifiers within strata were independent, but not all
were invariant
Variance Estimation. The stratified version of the standard approximation for the variance of a two-
stage sample design provided by the first term of equation 5.24 in Lohr (2010) was used to
estimate the variance of the sample design used in this study. While a two-phase design has a
somewhat larger variance than a two-stage design, the standard approximation is reasonable in
this situation that is a combination of both.
The sample sizes reflect the final sample of 350 wells. The strata for variance estimation are the
first-stage strata used to select operators. Variance estimates measure the consistency (or
variability) between operators within the same stratum, so each variance stratum must contain at
least two operators. Thus, for variance estimation, the four large operator strata were collapsed
into two variance strata, each with two sampled well operators. The other two variance strata are
the strata for medium-sized operators and for small-sized operators. The formula used to estimate
the variance is given in equation A.3:
Hex) = T.t=1^x(i-^)x9ll (a.3)
Where 6X is the point estimator for the total of variable x, the well characteristic of interest (e.g.,
well age or true vertical depth), Mh, and mh are the first-stage frame and sample sizes for the
number of operators in variance stratum h, and
(a. 4)
frtfi i
In equation A. 4, th = m^1 an<^ estimated total of variable x for company i using the
second-stage conditional weights (e.g., the estimated total number of wells in company i with a
cement bond log).
Confidence Intervals. To compute 95 percent confidence intervals on count or continuous variables,
the estimated mean plus or minus the standard error was multiplied by a t-statistic with five
degrees of freedom:94
l@x ~ f5,0.975 @x + f5,0.975 Jv(&x)] (A5)
94 The variance formula had five degrees of freedom because nine companies were selected across four variance strata.
Degrees of freedom are calculated as the number of first-stage units (9] minus the number of variance strata (4].
68
-------
Well Design and Construction
May 2015
where £5,0.975 is the 97.5 percentile of a t-distribution with five degrees of freedom. If the lower end
of the confidence interval was below 10, it was set to 10 to avoid disclosure about the actual
sampled responses.
Variance estimates for proportions in the Executive Summary and Section 8 were handled
differently, because many point estimates were near 0 or 100 percent Wilson confidence intervals
were used to constrain the limits to remain above 0.0 percent and below 100 percent (Heeringa et
al., 2010). The formula for the 95-percent Wilson confidence intervals is provided in equation A.6.
Where n* is the effective sample size (i.e., number of wells, adjusting for the correlation of wells
within operators); t = 15,0.975, the 97.5 percentile of a t-distribution with five degrees of freedom;
and p is the estimated weighted proportion.
References
Heeringa, S.G., West, B.T., and Berglund, P.A. 2010. Applied Survey Data Analysis. Chapman and
Hall/CRC, Boca Raton, Florida.
Lohr, S.L. 2010. Sampling: Design and Analysis. 2nd edition. Brooks/Cole, Boston, Massachusetts.
Sarndal, C.-E., Swensson, B., and Wretman, J. 2003. Model Assisted Survey Sampling. Springer, New
York, New York.
(2n*p+t2)-(tyj t2 +4p(l-p)n*) (2n*p+t2)+(tyj t2 +4p(l-p)n*)
(A.6)
2(n*+£2)
2 (n*+£2)
69
-------
Well Design and Construction
May 2015
Appendix B: Well Operator Information Request
In August 2011, the EPA issued information requests to nine well operators to collect data used in
this report The requests were sent to the following companies: 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. These companies were chosen to reflect a range in operator size and geographic
diversity of oil and gas production wells hydraulically fractured by nine service companies. The
EPA requested information for a total of 350 well identifiers. The information requested is stated
below.
INFORMATION REQUESTED
Your response to the following questions is requested within thirty (30) days of receipt of this information
request:
For each well listed in Enclosure 5 of this letter (not provided here), provide any and all of the following
information:
Geologic Maps and Cross Sections
1. Prospect geologic maps of the field or area where the well is located. The map should depict, to the
extent known, the general field area, including the existing production wells within the field,
preferably showing surface and bottom-hole locations, names of production wells, faults within the
area, locations of delineated source water protection areas, and geologic structure.
2. Geologic cross section(s) developed for the field in order to understand the geologic conditions
present at the wellbore, including the directional orientation of each cross section such as north,
south, east, and west.
Drilling and Completion Information
3. Daily drilling and completion records describing the day-by-day account and detail of drilling and
completion activities.
4. Mud logs displaying shows of gas or oil, losses of circulation, drilling breaks, gas kicks, mud weights,
and chemical additives used.
5. Caliper, density, resistivity, sonic, spontaneous potential, and gamma logs.
6. Casing tallies, including the number, grade, and weight of casing joints installed.
Continued on next page
70
-------
Well Design and Construction
May 2015
Continued from previous page
7. Cementing records for each casing string, which are expected to include the type of cement used,
cement yield, and wait-on-cement times.
8. Cement bond logs, including the surface pressure during each logging run, and cement evaluation
logs, radioactive tracer logs or temperature logs, if available.
9. Pressure testing results of installed casing.
10. Up-to-date wellbore diagram.
Water Quality, Volume, and Disposition
11. Results from any baseline water quality sampling and analyses of nearby surface or groundwater prior
to drilling.
12. Results from any post-drilling and post-completion water quality sampling and analyses of nearby
surface or groundwater.
13. Results from any formation water sampling and analyses, including data on composition, depth
sampled, and date collected.
14. Results from chemical, biological, and radiological analyses of "flowback," including date sampled and
cumulative volume of "flowback" produced since fracture stimulation.
15. Results from chemical, biological, and radiological analyses of "produced water," including date
sampled and cumulative volume of "produced water" produced since fracture stimulation.
16. Volume and final disposition of "flowback."
17. Volume and final disposition of "produced water."
18. If any of the produced water or flowback fluids were recycled, provide information, including, but not
limited to, recycling procedure, volume of fluid recycled, disposition of any recycling waste stream
generated, and what the recycled fluids were used for.
Hydraulic Fracturing
19. Information about the acquisition of the base fluid used for fracture stimulation, including, but not
limited to, its total volume, source, and quality necessary for successful stimulation. If the base fluid is
not water, provide the chemical name(s) and CAS number(s) of the base fluid.
Continued on next page
71
-------
Well Design and Construction
May 2015
Continued from previous page
20. Estimate of fracture growth and propagation prior to hydraulic fracturing. This estimate should
include modeling inputs (e.g., permeability, Young's modulus, Poisson's ratio) and outputs (e.g.,
fracture length, height, width).
21. Fracture stimulation pumping schedule or plan, which would include the number, length, and location
of stages; perforation cluster spacings; and the stimulation fluid to be used, including the type and
respective amounts of base fluid, chemical additives and proppants planned.
22. Post-fracture stimulation report containing, but not limited to, a chart showing all pressures and rates
monitored during the stimulation; depths stimulated; number of stages employed during stimulation;
calculated average width, height, and half-length of fractures; and fracture stimulation fluid actually
used, including the type and respective amounts of base fluid, chemical additives and proppants used.
23. Micro-seismic monitoring data associated with the well(s) listed in Enclosure 5, or conducted in a
nearby well and used to set parameters for hydraulic fracturing design.
Environmental Releases
24. Spill incident reports for any fluid spill associated with this well, including spills by vendors and service
companies. This information should include, but not be limited to, the volume spilled, volume
recovered, disposition of any recovered volume, and the identification of any waterways or
groundwater that was impacted from the spill and how this is known.
72
-------
Well Design and Construction
May 2015
Glossary
Annulus: The space between two concentric objects, such as between the wellbore and casing or
between casing and tubing, (ref 6)
Borehole: The wellbore itself, including the open hole or uncased portion of the well, (ref 6)
Casing: Steel pipe that is lowered into a wellbore. Casing extends from the bottom of the hole to the
surface, (ref 6)
Conductor casing: Casing that is generally less than 100 feet deep and is often uncemented.
Its purpose is to prevent the in-fill of unconsolidated dirt and rock in the uppermost few
feet of drilled hole, (refs 1, 3)
Intermediate casing: Casing that seals off intermediate depths and geologic formations
that may have considerably different reservoir pressures than deeper zones to be drilled,
(refs 1, 3)
Production casing: The deepest casing set and serves primarily as the conduit for
producing fluids, although when cemented to the wellbore, this casing can also serve to seal
off other subsurface zones including ground water resources (refs 1, 3). For this report, any
casing or liner with production perforations was considered to be a "production casing." In
some cases, this can include casing identified as "intermediate" by the operator, but later
perforated for production purposes.
Surface casing: The shallowest cemented casing, with the widest diameter. Cemented
surface casing generally serves as an anchor for blowout protection equipment and to seal
off certain ground water resources, (ref 1)
Casing string: An assembled length of steel pipe configured to suit a specific wellbore. The sections
of pipe are connected and lowered into a wellbore, then fixed in place using cement or packers, (ref
6)
Casing tally: Often provided by well operators, this is a detailed list of the sections of casing placed
into the well during its construction, including the precise length of each section.
Cementing: To prepare and pump cement into place in a wellbore. Cement is placed between the
outside of the casing and the inside of the drilled hole in order to seal off the annulus. Cement also
provides support behind the casing and protects it from corrosive fluids in the subsurface, (refs 1, 6,
7)
Primary cementing: The first cementing operation performed to place a cement sheath
around a casing or liner, (ref 6)
Secondary cementing: Defined in this report as any non-primary cementing operation.
Secondary cementing operations are also referred to as "remedial" cementing operations,
(refs 5, 6)
73
-------
Well Design and Construction
May 2015
Cementing ticket: Often provided by well operators, this is a description of the cementing job. The
description usually includes how much cement of a specific type was used, how much volume the
cement occupies when cured, the amount of cement additives used, and the pumping pressures
used to place the cement
Cement bond log: A log that uses the variations in amplitude of an acoustic signal while traveling
down the casing wall to evaluate the cement bond between the cement, casing, and wellbore. (ref 6)
Cement top: Defined in this report as the height above the bottom of the casing string reached by
cement in the wellbore, as measured from the surface.
Completion: A generic term used to describe the events and equipment necessary to bring a well
into production (ref 6). Completion activities can include hydraulic fracturing.
Completion report: Often provided by well operators, this is a daily log of the completion activities
at a well after it has been drilled. Hydraulic fracturing is included in the completion report.
Deviated well: Defined in this report as a non-horizontal well where the bottom-hole location is
more than 500 lateral feet from the surface location.
Deviation survey: A completed measurement of the inclination (deviation from vertical) and
azimuth (the compass direction) of a location in a well (typically the total depth at the time of
measurement), (ref 6)
Drill bit: The tool used to crush or cut rock. Most bits work by scraping or crushing the rock as part
of a rotational motion, while some bits work by pounding the rock vertically, (ref 6)
Drilling fluid: Any of a number of liquid and gaseous fluids and mixtures of fluids and solids used
when drilling boreholes, (ref 6)
Driller's report: Often provided by well operators, this is a daily log of the activities at a well and
includes details on the well's drilling, casing, and cementing history from surface to total depth.
Drinking water resource: Any body of water, ground or surface, that currently serves or in the
future could serve as a source of drinking water for public or private water supplies, (ref 9)
Formation: A body of earth material with distinctive and characteristic properties and degree of
homogeneity in its physical properties, (ref 8)
Ground water resource: Defined in this report as any geologic formation containing ground water.
Horizontal well: Defined in this report as a well intentionally completed with one or more
boreholes drilled laterally to follow the targeted geologic formation.
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 9)
74
-------
Well Design and Construction
May 2015
Liner: A casing string that does not extend to the top of the wellbore, but instead is anchored or
suspended from inside the bottom of the previous casing string, (ref 6)
Lithology: The macroscopic nature of the mineral content, grain size, texture, and color of rocks,
(ref 6)
Carbonate: A class of sedimentary rock whose chief mineral constituents are calcite,
aragonite, and dolomite, (ref 6)
Chert: A sedimentary rock and a variety of quartz made of extremely fine-grained silica, (ref
6)
Coal: A carbon-rich sedimentary rock that forms from the remains of plants deposited as
peat in swampy environments, (ref 6)
Sandstone: A clastic sedimentary rock whose grains are predominantly sand-sized.
Sandstone has relatively high porosity and permeability, (ref 6)
Shale: A fine-grained, fissile, detrital sedimentary rock formed by consolidation of clay- and
silt-sized particles into thin, relatively impermeable layers, (ref 6)
Losses of circulation: The reduced or total absence of fluid flow up the annulus when fluid is
pumped through the drill string, (ref 6)
Measured depth: The length of the wellbore, as if determined by a measuring stick. This
measurement differs from the true vertical depth of the well in all but vertical wells, (ref 6)
Mud log: Often provided by well operators, this is a detailed profile of the drilled hole describing
the presence or absence of oil or gas in the various rock formations penetrated by the drill bit. The
drilling fluid and rock cuttings returned to the surface are tested at the surface and correlated with
their depth to produce the log. (ref 3)
Open hole: The uncased portion of a well. All wells, when first drilled, have open sections.
Additionally, some completions are open, (ref 6)
Open hole log: Often provided by well operators, this is a generic term applied to any electric
wireline log run in a part of the wellbore that has no casing set Open hole logs are generally run to
understand the hydrocarbon-bearing potential along the logged interval.
Packer: A device that can be run into a wellbore with a smaller initial outside diameter that then
expands to seal the wellbore. (ref 6)
Perforation: The communication tunnel created from the casing or liner into the targeted geologic
formation through which injected fluids and oil or gas flow, (ref 6)
Plat: Often provided by well operators, this is a map or plan-view depiction of the well site.
75
-------
Well Design and Construction
May 2015
Point of shallowest hydraulic fracturing: Defined in this report as the measured depth of the
shallowest point at which hydraulic fracturing occurred.
Protected ground water resources: This report does not use a single definition for protected
ground water resources and relies solely on information provided to the EPA by well operators.
Spud: To start the well drilling process by removing rock, dirt, and other sedimentary material with
the drill bit. (ref 6)
Stratified sampling approach: In a stratified sampling approach, the population is divided into
subpopulations, called strata. The strata do not overlap and they constitute the whole population,
so that each sampling unit belongs to exactly one stratum. An independent, random sample can
then be selected from each strata. Information from all strata are pooled to obtain overall
population estimates, (ref 4)
Targeted geologic formation: Defined in this report as the geologic formation intended for
hydrocarbon production.
True vertical depth: The vertical distance from a point in the well (usually the current or final
depth) to a point at the surface, (ref 6)
Tubing: The narrowest steel pipe set within a completed well, either hung directly from the
wellhead or secured at its bottom using a packer. Tubing is not typically cemented in the well.
Uncemented annulus: Defined in this report as an uncemented interval that reaches the surface.
Uncemented interval: Defined in this report as a segment of annular space along the outside of the
well (between the casing and the wellbore wall) that has no cement
Vertical well: Defined in this report as a well with a bottom-hole location within 500 lateral feet of
the surface location.
Wellbore: The drilled hole or borehole, including the open hole or uncased portion of the well, (ref
6)
Wellbore diagram: A schematic diagram that identifies the main completion components installed
in a wellbore. The information included in the wellbore diagram relates to the principal dimensions
of the components and the depth at which the components are located, (ref 6)
Wellhead: The surface termination of a wellbore. (ref 6)
Zonal isolation: Wells that prevent inter-formational flow are said to have "zonal isolation." (refs 1,
2, 7)
76
-------
Well Design and Construction
May 2015
Glossary References
1. Baker, R. 1979. A Primer of Oil-Weil Drilling. 4th edition. Petroleum Extension Service, Austin,
T exas.
2. Bellabarba, M., Bulte-Loyer, H., Froelich, B., Le Roy-Delage, S., van Kuijk, R., Zeroug, S., Guillot, D.,
Moroni, N., Pastor, S., and Zanchi, A. 2008. Ensuring Zonal Isolation Beyond the Life of the Well.
Oilfield Review 18-31.
3. Devereux, S. 1998. Practical Well Planning and Drilling Manual. PennWell Publishing Company,
Tulsa, Oklahoma.
4. Lohr, S.L. 2010. Sampling: Design and Analysis. 2nd edition. Brooks/Cole, Boston,
Massachusetts.
5. Nelson, E.B., and Guillot, D. 2006. Well Cementing. 2nd edition. Schlumberger, Sugar Land,
T exas.
6. Schlumberger. 2014. Oilfield Glossary. Available at http://www.glossary.oilfield.slb.com/.
Accessed December 1, 2014.
7. Smith, D.K. 1976. Cementing (Monograph - Society of Petroleum Engineers; v. 4). Henry L.
Doherty Memorial Fund of AIME, Dallas, Texas.
8. US Environmental Protection Agency (US EPA). 2006. Terminology Services: Terms and
Acronyms. Available at http://iaspub.epa.gov/sor_internet/registry/termreg/home/
overview/home.do. Accessed December 1, 2014.
9. USEPA. 2011. Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water
Resources. Available at http://www2.epa.gov/sites/production/files/documents/
hf_study_plan_110211_final_508.pdf. Accessed December 1, 2014.
77
-------
Well Design and Construction
May 2015
[This page intentionally left blank.]
78
-------
Well Design and Construction
May 2015
[This page intentionally left blank.]
79
-------
EPA/601/R-14/002 I May 2015 I www.epa.gov/hfstudy
svEPA
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
------- |